I

The effects of 670nm light on retinal Müller cell gliosis following retinal

stress or injury: exploring the underlying cellular mechanisms using in vivo and in vitro models

Yen-Zhen Lu

December 2018

A thesis submitted for the degree of Doctor of Philosophy of The Australian National University

© Copyright by Yen-Zhen Lu 2018 All Rights Reserved

I

CANDIDATE STATEMENT This declaration certifies that the following work entitled ‘The effects of

670nm light on retinal Müller cell gliosis following retinal stress or

injury: exploring the underlying cellular mechanisms using in vivo and

in vitro models.’ is the authors own original work, complies with The

Australian National University Research Award Rules and has not been

previously accepted for award of a degree or diploma to any other

university or institution of higher learning. The Taiwan-Australian

National University Scholarship sponsored this thesis.

Word Counts: 53272 Signed: ___________________ Date: ___________________

II

ACKNOWLEDGMENTS First and foremost I would like to express my sincere gratitude to my supervisor, Associate Professor Krisztina Valter. I also thank Professor Jan Provis, Associate Professor Michele Madigan, and Dr. Riccardo Natoli for their continuous support of my Ph.D. Without their support, patient, and dedication, this thesis would not have been completed. I could not have imagined having better supervisor and advisors for my Ph.D. study. I am sincerely grateful to Krisztina for her passionate and support to my research and Ph.D. life. Because of her, I was able to have a beautiful Ph.D. journey to Australia in the past four years. Thanks to Jan for her insightful comments on my research projects, which helped me to develop my research with different perspectives. Thanks to my advisor Michele for her constructive research aspects and encouragement. I also would like to thank Riccardo. His guidance and brilliant ideas widened my research field. I also would like to thank my all colleagues and academic friends, including Matt Rutar, Nilisha Fernando, Helen Jiao, Tanja Racic, Joshua Chu-Tan, Riemke Aggio-Bruce, Yvette Wooff, Dina Satriawan, Nuan-Ting Huang, Zabrina Abdool, Peggy Chou, and Ryan Tseng for their scientific supports and sweet company. All of them lightened up my every day in JCSMR. Without them, I would not conquer all difficulties during my Ph.D. Thanks to my friends in Canberra, Sarah Cheng, Arthur Lee, Jofan Lin, Eric Wang, Alice Liao, and Regina Shen for their company and delicious comfort food. I also would like to give a huge thank to my overseas friends, Joy Lee, Chin-Ting Lin, Meng-Hsin Chen, Hsin-Lun Tsai, Chi-Chin Wu, Hsing-Hui Wang, Hang-Shiang Jiang, Chih Kuei Wang, and Abby Weng. Thanks for their support and unconditional friendship. And to my beloved fiancé, Kenny Lau, thank you for bringing me delight and support always. I would not have achieved any of this without your love. Finally, I would like to give sincere thanks to my family in Taiwan, my dear father (Ruey-Hsing), aunts, older brother (Chi-Chien), sister in law (Chia-Chin), as well as cutest niece (Pin-Yu, Lucy) and nephew (Bo-Chang). Their loves and smiles always warm up my heart throughout writing thesis and my life in general. Special thanks to my mother (Chia-Wen), This thesis is dedicated to her and her love.

III

‘‘The slowest person unless he does not lose the sight of the goal, is faster than the one who wanders aimlessly.’’

− Gotthold Lessing

IV

ABSTRACT Photobiomodulation (PBM) describes a process whereby light wavelengths of 600-

1000nm are used to initiate biological responses. PBM has been shown to attenuate

inflammation and accelerate wound healing in skin and mucosal tissues. In the nervous

system, it promotes recovery of injured spinal cord and optic nerve. Our laboratory has

found that irradiation with 670nm light, applied prior to retinal insult, reduced

photoreceptor death in retinal degeneration in vivo. However, very little attention was

paid to the non-neuronal component of the retina, the macroglia of the retina.

Müller cells (MCs), the principal macroglia of the retina, are involved in supporting

retinal structure and maintaining its homeostasis. MCs react to retinal stress or injuries,

described as gliosis, aiming to protect neurons. However when it enters into a

progressive state, it becomes detrimental to the retina. Activated MCs, if uncontrolled,

release a large amount of proinflammatory cytokines and chemokines, recruiting

microglias (MGs) and monocytes, which lead to further retinal inflammation. While

the retinal damage is extensive, MCs undergo mitosis and thickening of their processes,

which reach the subretinal space to form glial scars that inhibit nutrient delivery,

leading to further neuronal death. Thus, the aim of this thesis is to investigate the effects

of 670nm irradiation on activated MCs using in vivo and in vitro stress models,

exploring a new avenue that may prevent irreversible retinal degeneration.

In Chapter 3, the effects of 670nm light on activated MCs using in vitro and in vivo

stress models of retinal injury were investigated. Our results demonstrated that 670nm

modified MC activation, both its proinflammatory and proliferative processes. This

chapter additionally draws attention to the importance of appropriate timing of

treatment, as there is a finite therapeutic window to effectively mitigate gliotic changes.

In Chapter 4 investigated the effects of 670nm light on interaction between MCs and

photoreceptors, and on subsequent MC-derived MG activation in vitro. Results

confirmed that 670nm light mitigated MC gliosis induced by photo-oxidative damage

(PD), subsequently reducing MG activation. This protective mechanism of action of

670nm light in MCs was associated with increased mitochondrial activity. Chapter 5

explored the cell communication between MCs and MGs in vitro, focusing on the role

of exosomes. Exosomes have been discovered to carry microRNAs (miRNAs), which

V

allows the transfer of genetic information between cells. In this chapter, IL-1β

stimulation of MCs led to the release of exosomes, which stimulated MGs to upregulate

expression of proinflammatory cytokines. Several miRNAs implicated in regulating

inflammatory processes were identified in MC-derived exosomes in stressed MCs.

Treatment with 670nm significantly reduced the expression of some of the pro-

inflammatory genes.

The results from this thesis collectively indicate that following MC activation in retinal

damage, 670nm treatment post-damage can mitigate MC gliosis and subsequently

ameliorate retinal inflammation. Furthermore, this effect may be achieved through

down-regulation of proinflammatory cytokine production in the retina and modification

of exosome contents. Therefore, targeting activated MCs using 670nm light may be a

potential therapeutic strategy in mitigating inflammation associated with the

irreversible retinal degeneration.

VI

TABLE OF CONTENTS CANDIDATE STATEMENT ......................................................................................... IACKNOWLEDGMENTS .............................................................................................. IIABSTRACT.................................................................................................................... IVCONFERENCE ABSTRACTS .................................................................................. XIIABBREVIATION....................................................................................................... XIIICHAPTER ONEINTRODUCTION .......................................................................................................... 151.1 General Introduction ................................................................................................ 161.2 The function and structure of the retina .................................................................. 171.3 Photoreceptors .......................................................................................................... 201.4 Blood ocular barrier ................................................................................................. 231.5 Müller cells ............................................................................................................... 231.5.1 Origin and development ..................................................................................... 231.5.2 Basic properties of Müller cells (MCs) ............................................................. 241.5.3 The role of Müller cells (MCs) in vision formation ......................................... 241.5.4 The role of Müller cells (MCs) in neuroprotection ........................................... 271.5.5 Müller cells (MCs) activation............................................................................. 27

1.6 Extracellular vesicles (EVs) .................................................................................... 321.6.1 Microvesicles (MVs) .......................................................................................... 331.6.2 Exosomes ............................................................................................................. 331.6.3 Role of extracellular vesicles (EVs) in cell-cell communication ..................... 35

1.7 Age-related macular degeneration (AMD) ............................................................. 361.7.1 Inflammation in AMD ........................................................................................ 391.7.2 The role of Müller cells (MCs) in AMD ........................................................... 43

1.8 Light damage-induced retinal degeneration in rodent animals ............................. 431.9 Photobiomodulation (PBM) .................................................................................... 441.9.1. History ................................................................................................................ 451.9.2. Effective treatment: the optical window and dosage ....................................... 461.9.3. Mechanism of action .......................................................................................... 471.9.4. Treatment of eye disease by 670nm light ......................................................... 48

1.10 Aims ........................................................................................................................ 49 CHAPTER TWOMATERIALS AND METHODS.................................................................................. 512.1 Introduction .............................................................................................................. 522.2 Animal rearing and housing conditions .................................................................. 522.3 Photo-oxidative damage of animal models ............................................................ 522.3.1 Bright light-induced photo-oxidative damage of a rat model .......................... 522.3.2 Progressive photo-oxidative damage in mice .................................................... 53

2.4 670nm light treatment in vivo in rats ...................................................................... 532.5 Cell cultures .............................................................................................................. 542.5.1 661W Photoreceptor-like cells ........................................................................... 542.5.2 MIO-M1 cell line ................................................................................................ 542.5.3 Mouse microglia N11 cell line ........................................................................... 552.5.4 Preparation of rat primary MCs ......................................................................... 55

VII

2.5.5 Preparation of rat primary MGs/MΦ ................................................................. 572.5.6.The treatment with 670nm light in vitro ............................................................ 58

2.6 Tissue collection and fixation.................................................................................. 582.7 Histological Processing............................................................................................ 592.8 Histological Techniques .......................................................................................... 592.8.1 Measurement of retinal thickness....................................................................... 592.8.2 Counting of the number of rows of photoreceptor nuclei ................................ 602.8.3 Immunohistochemistry (IHC) of retinal sections .............................................. 61

2.9 The immunocytochemistry (ICC) staining ............................................................. 622.10 Gene expression ..................................................................................................... 652.10.1 RNA extraction from tissues ............................................................................ 652.10.2 RNA extraction from cells................................................................................ 652.10.3 Complementary DNA synthesis ....................................................................... 662.10.4 Quantitative real time polymerase chain reaction (qRT-PCR) ...................... 66

2.11 Western blotting ..................................................................................................... 682.12 Assessment of cell viability and cell toxicity ....................................................... 692.13 Statistical analysis .................................................................................................. 69

CHAPTER THREELIGHT TREATMENT WITH 670NM MITIGATES MÜLLER CELL GLIOSIS FOLLOWING RETINAL INJURY: EVIDENCE FROM IN VITRO AND IN VIVO MODELS OF STRESS....................................................................................... 703.1 Introduction .............................................................................................................. 713.2 Materials and methods ............................................................................................. 723.2.1 Experimental groups of animals ......................................................................... 723.2.2 Tissue collection.................................................................................................. 743.2.3 Immunohistochemistry (IHC) of retinal sections .............................................. 743.2.4 Photoreceptor cell death, and survival analysis ................................................ 753.2.5 RNA isolation and real-time quantitative polymerase chain reaction ............. 753.2.6 Maintenance of MIO-M1 Cells .......................................................................... 763.2.7 Estimating MC gliosis by using the in vitro uniform scratch model ............... 763.2.8 670nm red light treatment paradigm and experimental groups........................ 773.2.9 Immunocytochemistry staining (ICC) ............................................................... 783.2.10 Proliferation inhibition with mitomycin C (MMC) ........................................ 793.2.11 Cell spreading .................................................................................................... 793.2.12 Migratory capability assay ............................................................................... 793.2.13 Flow cytometry analysis for cell viability and cell cycle ............................... 803.2.14 Statistical analysis ............................................................................................. 80

3.3 Results ....................................................................................................................... 813.3.1 Early treatment with 670nm light reduced photoreceptor loss following photo-oxidative damage (PD) ................................................................................................. 813.3.2 Treatment with 670nm light reduced MC gliosis ............................................. 843.3.3 Retinal stress and inflammation-related gene expressions changed following PD and PBM ................................................................................................................. 853.3.4 Treatment with 670nm light modified the recruitment and mobilization of MG/MΦ post PD .......................................................................................................... 873.3.5 Direct effects of 670nm light on MC motility and division in vitro ................ 893.3.6 670nm light reduced injury-induced MC mitosis in vitro ................................ 893.3.7 The 670nm light treatment reduced cell death after injury in vitro ................. 913.3.8 Effects of 670nm light on cell spreading and migration................................... 92

VIII

3.3.9 In vitro assessment of cellular stress .................................................................. 923.4 Discussion ................................................................................................................. 953.4.1 670nm treatment post injuries suppressed MC activation ................................ 953.4.2 670nm light treatment reduced photoreceptor loss and slowed disease progression .................................................................................................................... 963.4.3 The effects of 670nm light on MC migration: implications for glial scarring 963.4.4 Early 670nm light treatment mitigated inflammatory changes ........................ 983.4.5 PBM may regulate retinal inflammatory response via inflammasomes .......... 983.4.6 Mechanism of action of 670nm light on MCs ................................................... 99

3.5 Conclusion .............................................................................................................. 100 CHAPTER FOUR PHOTOBIOMODULATION WITH 670NM LIGHT AMELIORATES MÜLLER CELL MEDIATED ACTIVATION OF MICROGLIA/MACROPHAGES IN RETINAL DEGENERATION ................ 1014.1. Introduction ........................................................................................................... 1024.2. Materials and methods .......................................................................................... 1044.2.1 Animals and light exposure .............................................................................. 1044.2.2 Maintenance of 661W photoreceptor-like cells .............................................. 1044.2.3 Preparation of rat primary MCs ....................................................................... 1044.2.4 Photo-oxidative damage in co-culture of 661W cell line with rat primary MCs ..................................................................................................................................... 1044.2.5 670nm light treatment of co-cultures ............................................................... 1054.2.6 Isolation and assessment of activation of rat primary MG/MΦ ..................... 1074.2.7 Assessment of cell viability and cell toxicity .................................................. 1074.2.8 Immunocytochemistry on rat primary MCs .................................................... 1074.2.9 Quantitative polymerase chain reaction........................................................... 1084.2.10 In situ hybridization ........................................................................................ 1094.2.11 Enzyme-Linked Immunosorbent Assay (ELISA) ......................................... 1094.2.12 Western blotting .............................................................................................. 1094.2.13 Mitochondrial membrane potential (ΔΨm) ................................................... 1094.2.14 Statistical analysis ........................................................................................... 110

4.3. Results .................................................................................................................... 1114.3.1 670nm light suppressed cytokine expression following PD in vivo .............. 1114.3.2 670nm light reduced stress in primary MCs exposed to damaged 661W cells ..................................................................................................................................... 1114.3.3 670nm light suppressed oxidative stress and inflammation in MCs following PD ................................................................................................................................ 1144.3.4 670nm light treatment has no effect on NLRP3 inflammasome activation in MCs ............................................................................................................................. 1144.3.5 Mitochondrial membrane potential was enriched by 670nm light following PD ................................................................................................................................ 1174.3.6 670nm light regulated MC-mediated activation of MG/MΦ following PD .. 119

4.4. Discussion .............................................................................................................. 1214.4.1 670nm light reduced MC-mediated activation of MG/MΦ ............................ 1214.4.2 670nm light modulated MC stress in response to damage ............................. 1224.4.3 The change in Δψm on regulating gene expression ........................................ 123

4.5. Conclusion ............................................................................................................. 124

IX

CHAPTER FIVE670NM RED LIGHT REGULATES RETINAL MÜLLER CELL-MEDIATED INFLAMMATION THROUGH EXTRACELLULAR VESICLES.................... 1255.1 Introduction ............................................................................................................ 1265.2 Material and methods ............................................................................................. 1285.2.1 MIO-M1 stimulation with IL-1β ...................................................................... 1285.2.2 670nm light treatment paradigm and experimental groups ............................ 1285.2.3 Morphological analysis of MIO-M1 cells ....................................................... 1295.2.4 Assessment of cell viability on MIO-M1 cells ................................................ 1295.2.5 Ultracentrifugation-isolated MVs .................................................................... 1295.2.6 Microbead-captured exosomes ......................................................................... 1305.2.7 MV isolation and purification from retinas ..................................................... 1305.2.8 Annexin V binding/staining ............................................................................. 1335.2.9 Detection of exosome surface markers ............................................................ 1335.2.10 Analysis of CD3 and CD81 expression on MVs .......................................... 1345.2.11 MV miRNA extraction ................................................................................... 1355.2.12 Profiling miRNAs ........................................................................................... 1355.2.13 Prediction of the target genes of candidate miRNAs .................................... 1365.2.14 Assessment of N11 activation ........................................................................ 1375.2.15 Assessment of N11 cell morphology ............................................................. 1375.2.16 Statistical analysis ........................................................................................... 137

5.3 Results ..................................................................................................................... 1385.3.1 Characterization of activated MIO-M1 following IL-1β challenge ............... 1385.3.2 Characterisation of activated MC-derived EVs .............................................. 1395.3.3 Characterization of MIO-M1-derived MVs .................................................... 1405.3.4 Composition of miRNAs in MVs .................................................................... 1435.3.5 Influences of 670nm light on the secretion of MVs and exosomes ............... 1485.3.6 Contribution of MIO-M1-derived MVs on proinflammatory and anti-inflammatory responses in microglia ........................................................................ 1505.3.7 The amount of MV changed in the mouse retina post PD.............................. 151

5.4 Discussion ............................................................................................................... 1535.4.1 IL-1β stimulated MC activation and their release of MVs and exosomes .... 1535.4.2 Function of EV surface markers on recipient cells ......................................... 1545.4.3 Target genes of MC-derived EVs carried miRNAs that modulate inflammation ............................................................................................................... 1545.4.4 The effects of IL-1β and 670nm light on MV release and its intraluminal content ......................................................................................................................... 1555.4.5 MVs and exosomes on retinal diseases ........................................................... 156

5.5 Conclusion .............................................................................................................. 157CHAPTER SIX ............................................................................................................. 158SUMMARY AND CONCLUSION ............................................................................ 158REFERENCES ............................................................................................................. 162

X

LIST OF FIGURES Figure 1.1 Structure of the human and mouse eye ------------------------------------------ 19Figure 1.2 Distribution and configuration of photoreceptors in the human eyes ----- 21Figure 1.3 Schematic drawing of cellular components of the retina --------------------- 25Figure 1.4 The schematic diagram displays the visual cycle in MCs -------------------- 26Figure 1.5 MC proliferative gliosis in injured retinas -------------------------------------- 31Figure 1.6 Release of microvesicles (MVs) and exosomes -------------------------------- 34Figure 1.7 Schematic transverse sections through the human eyeball: the classification

of AMD. ------------------------------------------------------------------------------ 38Figure 1.8 Schematic diagram presents the NLRP3 inflammasome complex formation

----------------------------------------------------------------------------------------- 41Figure 1.9 Optical window and hormetic effects of PBM --------------------------------- 47Figure 1.10 A schematic diagram presents proposed mitochondrial retrograde

signalling pathways and intracellular mechanism of action of 670nm light treatment in cells ----------------------------------------------------------- 50

Figure 2.1 Treatment SD rats with 670nm red light ---------------------------------------- 54Figure 2.2 Characteristics of rat primary MCs ----------------------------------------------- 56Figure 2.3 The WARP 10 appliance was applied to in vitro studies --------------------- 58Figure 2.4 Illustration of ONL thickness measurement ------------------------------------ 61Figure 3.1 Diagram of the rat PD model and the paradigms of the 670nm treatment. 74Figure 3.2 Diagram of the uniform scratch in MIO-M1 cells ----------------------------- 77Figure 3.3 Diagram of the measurement of the area of cell coverage ------------------- 77Figure 3.4 Timeline of the in vitro (scratch) model in MIO-M1 cells. ------------------ 78Figure 3.5 670nm light treatment on PR and MC gliosis following PD ---------------- 82Figure 3.6 Effects of 670nm light on PR survival post PD -------------------------------- 83Figure 3.7 Effects of 670nm light treatment on MC gliosis following PD ------------- 85Figure 3.8 The comparative fold changes of gene expression in the retina using

quantitative qRT-PCR ------------------------------------------------------------- 86Figure 3.9 Effects of 670nm light on MG/MΦ recruitment following PD ------------- 88Figure 3.10 Cell area coverage and proliferation following 670nm light treatment in

the in vitro scratch model --------------------------------------------------------- 90Figure 3.11 Effects of 670nm light treatment on MC cell cycle status and cell death

following the scratch --------------------------------------------------------------- 91Figure 3.12 Cell spreading of MIO-M1 cells was suppressed by 670nm light -------- 93Figure 3.13 Effects of 670nm light on MIO-M1 cell stress ------------------------------- 94Figure 4.1 Schematic diagram of the experimental design ------------------------------- 106Figure 4.2 Timeline of PD and 670nm light treatment ------------------------------------ 106Figure 4.3 In vivo expression of cytokines was mitigated by 670nm light treatment

following PD in rat retinas ------------------------------------------------------- 112Figure 4.4 670nm light treatment reduced primary MC stress when co-cultured with

661W cells exposed to PD ------------------------------------------------------- 113Figure 4.5 MC-derived inflammation was mitigated by exposure to 670nm light

following PD ----------------------------------------------------------------------- 115Figure 4.6 Effects of 670nm light on activation of the NLRP3 inflammasome in MCs

following PD ----------------------------------------------------------------------- 116Figure 4.7 The activation of cytochrome oxidase (COX5a) was enhanced in 670nm-

treated MCs after PD ------------------------------------------------------------- 118Figure 4.8 670nm light ameliorated MC-mediated activation of MG/MΦ following

PD ------------------------------------------------------------------------------------ 120Figure 5.1 Timeline of IL-1β stimulation and 670nm light treatment ------------------ 128Figure 5.2 Retinal MV isolation experimental flow chart -------------------------------- 132

XI

Figure 5.3 IL-1β affected cell morphology and metabolism in MIO-M1 cells ------- 138Figure 5.4 Release of MVs and exosomes from MIO-M1 in response to IL-1β

challenge ---------------------------------------------------------------------------- 140Figure 5.5 Profiles of exosome surface markers on MVs from ultracentrifugation

isolated-MVs and microbead captured-exosomes --------------------------- 141Figure 5.6 The composition of miRNAs was changed in MIO-M1-derived MVs in

response to IL-1β stimulation --------------------------------------------------- 143Figure 5.7 Effects of 670nm light on IL-1β stressed-MIO-M1 cells ------------------- 149Figure 5.8 670nm light triggered MIO-M1 cells to release MVs and exosomes ----- 149Figure 5.9 670nm light diminished MV associated-N11 activation and gene

expressions of proinflammatory cytokines ----------------------------------- 151Figure 5.10 Analysis of a percentage of Annexin V binding on MVs isolated from PD

retinas -------------------------------------------------------------------------------- 152

LIST OF TABLES Table 2.1 Primary antibodies for in vivo studies in rats ------------------------------------ 63Table 2.2 Primary antibodies for in vitro studies in rat primary MCs ------------------- 63Table 2.3 Primary antibodies were used in in vitro studies in MIO-M1 ---------------- 64Table 2.4 Details and information of secondary antibodies ------------------------------- 64Table 2.5 QRT-PCR program ------------------------------------------------------------------- 67Table 2.6 TaqMan® probes used in rat retinas ---------------------------------------------- 67Table 2.7 TaqMan® probes for rat primary MCs and MGs ------------------------------- 67Table 2.8 TaqMan® probes used in mouse MG cell line N11 ---------------------------- 68Table 3.1 Primary antibodies used for experiments in vivo and in vitro ---------------- 75Table 5.1 Exosome surface markers on MIO-M1-derived MVs ------------------------ 142Table 5.2 Expression of miRNAs in MVs from IL-1β-treated MCs -------------------- 144Table 5.3 Biological signalling pathways enrichment analysis for MV carrying

miRNA-targeting validated genes ------------------------------------------------ 145Table 5.4 Biological processes regulated by MV miRNA targeting predictive genes

------------------------------------------------------------------------------------------------- 146Table 5.5 Biological signalling pathways regulated by MV miRNA-targeting

validated and predictive genes ---------------------------------------------------- 147

PEER-REVIEWED JOURNAL ARTICLE Lu, Y.Z., Fernando, N., Natoli, R., Madigan, M., Valter, K., 2018. 670nm light treatment following retinal injury modulates Muller cell gliosis: Evidence from in vivo and in vitro stress models. Experimental eye research 169, 1-12. Lu, Y.Z., Natoli, R., Madigan, M., Fernando, N., Saxena, K., Aggio-Bruce, R., Jiao, H., Provis, J., Valter, K., 2017. Photobiomodulation with 670nm light ameliorates Muller cell-mediated activation of microglia and macrophages in retinal degeneration. Experimental eye research 165, 78-89. Natoli, R., Rutar, M., Lu, Y.Z., Chu-Tan, J.A., Chen, Y., Saxena, K., Madigan, M., Valter, K., Provis, J.M., 2016. The Role of Pyruvate in Protecting 661W Photoreceptor-Like Cells Against Light-Induced Cell Death. Current eye research 41, 1473-1481.

XII

CONFERENCE ABSTRACTS Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2017) The amelioration of Müller cell-derived proinflammation with 670nm red light under the in vitro photo-oxidative stress. The Australian Society of Medical Research - ACT New Investigators Forum in Canberra, Australia – oral presentation Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2016) The treatment of 670nm red light prevents activation of microglia in vitro: Moderation of Müller cells – photoreceptors interaction. XXII Biennial Meeting of the International Society for Eye Research in Tokyo, Japan – poster presentation Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2016) Post-injury treatment with 670nm red light modifies Müller cell gliosis and decelerates progressive retinal degeneration in the light-induce animal model of retinal degeneration. XVIIth International Symposium on Retinal Degeneration in Kyoto, Japan – poster presentation Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2016) Irradiation with 670nm red light modifies Müller cells – photoreceptor interactions in the in vitro model of light damage. Canberra Health Annual Research Meeting in Canberra, Australia – poster presentation Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2016) Irradiation with 670nm red light modifies Müller cells – photoreceptor interactions in the in vitro model of light damage. The Australian Society of Medical Research - ACT New Investigators Forum in Canberra, Australia – oral presentation Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2015) Irradiation with 670nm red light mitigates gliotic changes and slows progression of photoreceptor damage in the light-induced model of retinal degeneration in adult albino rats. European Molecular Biology Laboratory (EMBL) symposium in Melbourne, Australia – poster presentation-EMBL travel grant Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2015) Can 670nm light ameliorate proliferative gliosis in retinal Müller cells? : an in vitro study. Canberra Health Annual Research Meeting in Canberra, Australia - Best Laboratory Research of Radiation Oncology Private Practice Trust Fund Award Yen-Zhen Lu, Riccardo Natoli, Michele Madigan, Krisztina Valter (2015) Can 670nm light ameliorate proliferative gliosis in retinal Müller cells? : an in vitro study. The Australian Society of Medical Research - ACT new investigators Forum in Canberra, Australia– poster presentation Yen Zhen Lu, Krisztina Valter. (2014) Effects of 670nm red light on stressed Müller glia: Lessons learned from in vitro models. European Molecular Biology Laboratory (EMBL) symposium in Sydney, Australia - poster presentation

XIII

AMD age-related macular degeneration ERG electroretinography

ANOVA analysis of variance EV extracellular vesicle

APF Australian Phenomics Facility FBS fetal bovine serum

AQP4 aquaporin-4 FGF-2 fibroblast growth factor 2

ARVO Association for Research in Vision and Opthalmology

FACS fluorescence-activated cell sorter

ATP adenosine triphosphate GCL ganglion cell layer

BBZ Bisbenzimide GFAP glial fibrillary acidic protein

BRB blood-retinal barrier GS glutamine synthetase

BSA bovine serum albumin GAPDHglyceraldehyde 3-phosphate dehydrogenase

CCL2 chemokine (C-C motif) ligand 2 GM-CSFgranulocyte-macrophage colony-stimulating factor

cNDA complementary DNA GM growth cultured medium

CNS central nervous system HBSS Hank’s Balanced Salt Solution

COX cytochrome c oxidase ICC immunocytochemistry

CRALBP cellular retinaldehyde-binding protein IHC immunohistochemistry

ct cycle threshold iNOS inducible nitric oxide synthases

DIV days in vitro ILM inner limiting membrane

DNA deoxyribonucleic acid INL inner nuclear layer

dUTP deoxyribonucleotide triphosphate IPL inner plexiform layer

DR diabetic retinopathy IL-1β interleukin 1β

DMEM Dulbecco’s Modified Eagle Medium IL-6 interleukin 6

ABBREVIATION

XIV

IL-10 interleukin 10 OPL outer plexiform layer

IL-18 interleukin 18 OS outer segment

JCSMR John Curtin School of Medical Research

PFA paraformaldehyde

Kir inwardly rectifying potassium channels PBS phosphate buffer saline

LD light damage PD photo-oxidative damage

LED light emitting diode PBM photobiomodulation

LLLT low level light therapy P postnatal

MC Müller cell PYCARD PYD and CARD domain containing

MΦ macrophage qRTPCRquantitative real time polymerase chain reaction

mRNA messenger RNA RLBP1 retinaldehyde binding protein1

miRNA micro RNA ROS reactive oxygen species

MG microglia RPE retinal pigment epithelium

MΦ macrophage RP retinitis pigmentosa

MV microvesicle SD Sprague Dawley

MTT3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide

SN supernatant

NOX-1 NADPH oxidase 1 SOD2 superoxide dismutase 2, mitochondrial

NIR near infrared TUNEL terminal deoxynucleotidyl transferase nick end labelling

NLRP3 NLR family pyrin domain containing 3 TNF tumor necrosis factor

NLR NOD-like receptor VEGF vascular endothelium growth factor

NGS normal goat serum VS versus

OLM outer limiting membrane W weight

ONL outer nuclear layer W/V weight per volume

15

Chapter One

Introduction

16

1.1 General Introduction Age-related macular degeneration (AMD) is a degenerative disorder of the retina, and

the leading cause of irreversible vision loss in elderly people in developing countries

(Ambati and Fowler, 2012; Jager et al., 2008). It is a multifactorial illness, including

aging and environmental/epigenetic risk factors (Buschini et al., 2011). The disease

affects the central retina, the area responsible for high acuity and colour vision, the

macula. The pathology of AMD involves photoreceptor degeneration, retinal pigment

epithelium (RPE) atrophy, the infiltration and activation of microglia/macrophage

(MG/MΦ) and an increased deposition of debris in the macular sub-retinal region (Jager

et al., 2008).

Long-term and chronic abnormality of inflammation is considered as a crucial factor in

the development of AMD. The expression of several proinflammatory cytokines has

been found to be elevated either systemically in the serum or in the ocular of patients

with AMD, including interleukin-6 (IL-6) and tumour necrosis factor alpha (TNFα)

(Knickelbein et al., 2015). In the retinal degeneration animal model, elevated levels of

interleukin-1β (IL-1β), IL-6, and TNFα were observed in the damaged retina (Fernando

et al., 2016), which can be expressed by MG/MΦ (Fernando et al., 2016; Natoli et al.,

2017b). The aberration of MG/MΦ accumulation and activation within the retinal

tissues is present in all forms of AMD (Parmeggiani et al., 2012), where MG/MΦ

formed large aggregations in the photoreceptor layer, subretinal space, and RPE

(Ambati et al., 2013). Activated MG/MΦ released complement 3 (C3) to activate the

complement system (Natoli et al., 2017a), which is well-documented to contribute the

onset and progress of AMD. Chemokines within the retina are key regulators of

controlling MG/MΦ recruitment and activation in AMD, including C-C motif

chemokine ligand 2 (Ccl2) released by Müller cells (MCs) (Rutar et al., 2011b). These

chemokines are implicated in further exacerbating photoreceptor death in retinal

degeneration animal models (Ambati et al., 2013).

MCs, the principal macroglia of the retina, have been reported to participate in the

progression of retinal degeneration. In response to retinal diseases or degeneration,

MCs become activated, which is termed as reactive gliosis. This early activation is

aimed to protect the retina. In response to neuronal damage, MC’s release cytokines

and chemokines, such Ccl2, that recruit MG/MΦ in order to clean away damaged/dying

17

cells. At the same time, MCs express neuroprotective factors to guard neighbouring,

still healthy neurons. Despite having these beneficial properties, excessive

accumulation of these activated MG/MΦ, releasing complement into the retinal

environment, if the regulation of the complement system is compromised, can lead to

further retinal damage. In chronic or in more severe cases of retinal damage, MCs

undergo proliferative gliosis, when they migrate into the subretinal space forming glial

scars, which hinder nutrition and oxygen delivery from choroid to the underlying retina

causing further neuronal loss (Bringmann et al., 2006).

Photobiomodulation (PBM), is a low-energy photo-irradiation process using light

waves between 600nm to 1000nm wavelength, that has been shown to accelerate

wound healing in skin (Conlan et al., 1996), mucosa (Desmet et al., 2006), and soft

tissue (Herranz-Aparicio et al., 2013). In early AMD, the treatment with 670nm light

has been shown to reduce drusen volume and to improve visual acuity in human (Merry

et al., 2017). The mechanism of 670nm light on retinas is still being explored. Studies

in animal models of retinal degeneration have shown that 670nm light can mitigate

oxidative stress, inflammatory responses, and reduce neuronal cell death. However, the

effects of 670nm light on non-neuron cells, such as MCs, are still unclear.

This chapter provides an overview of the basic physiological properties of the retina in

human or rodents, the roles of MCs in healthy and diseased retinas, and the

pathogenesis of retinal degeneration, specifically AMD. The effects of 670nm light on

retinal degeneration and the related underlying mechanisms will also be presented.

1.2 The function and structure of the retina Vision is associated with intellectual development, growth promotion, and survival in

humankind (Swaroop et al., 2010). In humans, vision coordinates higher-order neuronal

functions including behaviour, memory, learning, and emotion (Swaroop et al., 2010).

Darkness and loss of sight drastically influence the quality of daily life in human beings.

The retina, the light-sensitive tissue at the back of the eyes, is the site where the process

of vision is initiated (Cuenca et al., 2014; Veleri et al., 2015). The fovea is a pit-like

depression region of the retina specialized for high acuity of the vision that requires a

18

high spatial density of cone photoreceptors. The fovea can be found within the retina

of many species, including fish, reptiles, birds, and mammals. However, amongst

mammals, the fovea appears to be present only in simian primates (Provis et al., 1998)

(Figure 1.1A&B). In human and non-human primates, the area, consisting of the fovea

centralis and containing the yellow macular pigments, carotenoids, zeaxanthin and

lutein, is described ophthalmoscopicly as the macula lutea (Volland et al., 2015)

(Figure1.1&1.2). In this region, there is a dramatic reduction in the density of rods,

accompanied by a dramatic increase in the density of cones (details are referred to

section 1.3) (Provis et al., 1998; Provis et al., 2013). Retinal thickness differs from

fovea to peripheral areas (Grover et al., 2010). An average of retinal thickness is

approximately is 249µm in human (Alamouti and Funk, 2003; Chan et al., 2006),

190µm in rat (Hariri et al., 2012; Lozano and Twa, 2012; Srinivasan et al., 2006), and

204µm in mice (Ferguson et al., 2013).

Vertebrate retina contains six types of neurons (rod and cone photoreceptors, horizontal,

bipolar, amacrine, and ganglion cells) (Swaroop et al., 2010) and four types of glial

cells (MCs, astrocyte, MG and oligodendrocyte) (Vecino et al., 2015). The major

characterization of various cells within the retina was firstly described by Santiago

Ramón Y Cajal (Ramón Y Cajal, 1988), a Spanish neuroscientist and pathologist

(Piccolino et al., 1989). He descried that the retina contains specific classes of nerve

cells connecting with other nerve cells, in order to convey the visual message, along

well-defined pathways to the cortical centre (Piccolino et al., 1989). His pioneering

work inspired intensive studies of the visual system over many decades. To date,

scientists have demonstrated that the retina contains three nuclear layers, which are the

outer nuclear layer (ONL), the INL, and the ganglion cell layer (GCL) (Madeira et al.,

2015; Veleri et al., 2015; Yue et al., 2016) (Figure 1.1C). The ONL contains cell bodies

of photoreceptors, cones and rods. The INL includes cell bodies of bipolar, horizontal,

and amacrine cells. The GCL contains cell bodies of ganglion cells, and displaces

amacrine cells. These neuronal cell layers are separated by plexiform layers, where

neuronal connections are formed.

19

Figure 1.1 Structure of the human and mouse eye (A-B) Schematic sections of the human and mouse eye. The fovea in the human eye is the central cone-only region and is charge of high resolution vision. The mouse retina lacks this specialized region (fovea). (C) The structure of a vertical section of the retina stained with hematoxylin and eosin. Three layers of cell nucleus (ONL, INL and, GCL) constitute structure of the retina. ONL contains rod and cone photoreceptor cell bodies. Amacrine, bipolar, horizontal neurons, and nucleus of MCs (not indicated in this image) reside in INL. GCL contains ganglion cells and axons of which from the optic nerve. OLM: outer limiting membrane; ONL: outer nuclear layer; OPL; outer plexiform layer; INL; inner nuclear layer; IPL; inner plexiform layer; GCL; ganglion cell layer; ILM: inner limiting membrane. Image A and B is reproduced with permission from (Veleri et al., 2015), Copyright © 2015 the Company of Biologists Ltd. Image C is reproduced by courtesy of the Webvision, http://webvision.med.utah.edu/. Images adapted from: A-B: (Veleri et al., 2015) ; C: (Madeira et al., 2015)

20

The neurons are supported by glial cells: two types of macroglia, MCs and astrocytes,

and resident MG. Nuclei of MCs are located in the INL and their processes span across

the retina (Goldman, 2014; Vecino et al., 2015) (details are referred to section1.5)

(Figure 1.3). MGs are found at variable densities in various retinal layers, including

nerve fibre layers (NFL), GCL, inner plexiform layer (IPL), INL, and outer plexiform

layer (OPL) in the healthy retina. Astrocytes, the other macroglia of the retina, inhabit

nearby NFLs (Figure 1.1). The forth glial type, oligodendrocyte is observed only

occasionally in the retina (Vecino et al., 2015).

1.3 Photoreceptors Mammalian photoreceptors are compartmentalized to perform their specialized

function of capturing light waves and transducing them into electrical signals in the

process of “phototransduction”. This process is performed by two major types of

photoreceptors of the retinas, cones and rods that are identified by their shape, type of

photopigments, retinal distribution, and pattern of synaptic connections (Mustafi et al.,

2009). Cone photoreceptors receive the signal from light to manage colour and high

resolution of visual images. Rod photoreceptors work under low light conditions

(Swaroop et al., 2010). In mice and humans, photoreceptors constitute over 70% of

retinal cells. The ratio of the cell number of rods to cones is 20:1 in human and 30:1 in

mice (Swaroop et al., 2010). The density of cones is much higher in the macula

compared to other regions of the retina. (Provis et al., 2013), and the fovea, found in

the centre of the macula, is contains only tightly-packed cones without rods (Provis et

al., 2013; Yue et al., 2016) (Figure1.1 &1.2). In the fovea centralis, an average

horizontal diameter of the rod-free zone is 0.035mm in human (Curcio et al., 1990). As

the macula has the higher concentration of the cone photoreceptors, and there is a one-

to-one relationship of cone to bipolar and ganglions cell connection of this region (Jager

et al., 2008; Mustafi et al., 2009; Yue et al., 2016), the human macula is responsible for

the high-resolution central visual acuity.

21

Figure 1.2 Distribution and configuration of photoreceptors in the human eyes (A) The fundoscopic image exhibits the location of macula, fovea, and the optic disc (blind spot) in the eyes. (B) A cross-sectional image of the retina, generated by optical coherence tomography, shows the fovea pit in the retina. (C) A pink-shaded area in the schematic representation stands for the area of macula, where a ratio of cones to rods is higher than the peripheral retina. (D) The compartmentalized morphology of rod and cone photoreceptors. The membranous discs in OS contains visual pigment (opsin). Organelles, including mitochondria (purple), are located in IS and responsible for metabolism in cells. Visual proteins, including opsins, are transported to OS form IS via a connecting cilium. Neurotransmitters are released from hyperpolarized cones or rods at presynaptic regions.

Image A and B are reproduced by courtesy of the Webvision, http://webvision.med.utah.edu/

Image C is reproduced with permission from the publication (Yue et al., 2016) (License number: 4310630923420 provided by Elsevier and Copyright Clearance Center), Copyright © 2016 Elsevier Ltd. All rights reserved.

Image D is adopted from (Veleri et al., 2015)

22

Photoreceptors consist of four distinct regions: the nucleus, the inner segment (IS), the

outer segment (OS), which continue to cell soma (perikaryon), and the synaptic region

(Veleri et al., 2015; Wolfrum and Schmitt, 2000) (Figure 1.2D). The OS and IS are

joined by a thin cellular bridge, the modified nonmotile connecting cilium (Wolfrum

and Schmitt, 2000). Phototransduction occurs in the OS that contains membrane discs

with photopigments, responsible for photon capture (details below) (Mustafi et al., 2009;

Swaroop et al., 2010). The IS contains all necessary organelles, including endoplasmic

reticulum, Golgi apparatus, and mitochondria, that is responsible to the replenishment

of spent lipoprotein membranes of the OS while providing the necessary high energy

support for the effective function of the cells (Gilliam et al., 2012; Khanna, 2015;

Wolfrum and Schmitt, 2000).

Opsin (also called visual pigment), including rhodopsin, is synthesised in IS (Nir et al.,

1987) and then transported into the disc membranes of OS (Khanna, 2015; Nir et al.,

1987) through the connecting cilium (Wolfrum and Schmitt, 2000). The type of opsin

present is unique to a photoreceptor subtype and defines its identity. In the human retina,

three subtypes of cones are identified based on a type of opsins they include and its

maximal spectral sensitivity; these are L- (long, 564 nm), M- (medium, 533nm), and S-

(short, 437 nm) wavelength cones. In the mouse retina, only two subtypes of cones (M-

and S- cones) are found. In addition, only one type of rods carrying rhodopsin is shown

in the vertebrate retina (Veleri et al., 2015). Rhodopsin (or opsin) is covalently bound

to a light-sensitive chromophore (11-cis retinal) in the dark (Shichida and Matsuyama,

2009; Veleri et al., 2015). Light absorption (photon) changes conformation of

chromophore isomerizes (rhodopsin and 11-cis retinal), which allows it to activate G

protein before decaying to yield free all-trans-retinal (Kaylor et al., 2013). The activated

G protein initiates an enzymatic signalling cascade and then eventually generates an

electrical response (hyperpolarization) in photoreceptor cells (Shichida and Matsuyama,

2009; Veleri et al., 2015). The hyperpolarized photoreceptors release neurotransmitters

at synaptic terminals, conveying signals to the second-order neurons (Veleri et al.,

2015). In addition, the other type of opsin, melanopsin, which is found in retinal

ganglion cells but not in cones or rods, might be a component of the photoreceptive

system for circadian rhythms of mammals (Ruby et al., 2002).

23

1.4 Blood ocular barrier The human retina has the highest oxygen consumption in the body. The retinal cells,

which have a high metabolic rate, need the specialized and distinct system to regulate

blood supply and to remove metabolic waste out of the retina (Campbell and Humphries,

2012). Blood-ocular barriers play a fundamental role in preserving and maintaining an

appropriate microenvironment of the retina (Cunha-Vaz, 2017). Two major types of

blood-ocular barriers are in the retina; they are the blood-aqueous barrier and blood-

retinal barrier (BRB) (Campbell and Humphries, 2012; Cunha-Vaz, 2017). The BRB,

including inner and outer BRB, restricts non-specific transport between the retinas and

blood circulation (Hosoya and Tachikawa, 2012) and acts as a physiological barrier to

regulates ion, protein, and water flux into or out of the retinas (Cunha-Vaz, 2017). The

inner BRB is composed of tight junctions of retinal endothelial cells and covered with

pericytes, MC processes, and astrocytes. The outer BRB, is formed by tight junctions

between neighbouring RPE cells, and is responsible for supporting photoreceptor

survival by controlling trafficking between the outer retina and the choroidal

vasculature that supply oxygen and glucose for the retina (Campbell and Humphries,

2012; Cunha-Vaz, 2017). In addition, RPE plays a role in the visual cycle, the removal

of metabolic waste, spent lipoprotein membranes and debris, in maintaining the proper

microenvironment with the retina (Campbell and Humphries, 2012).

1.5 Müller cells

1.5.1 Origin and development MCs, named after German anatomist, Heinrich Müller, who first described them as

radial fibre in the fish, frog, and pigeon retina in 1851(Bringmann et al., 2006; Hamon

et al., 2016; Müller, 1851). MCs are the major macroglia of the mammalian retina and

represent 4-5% of the retinal cells, in the human retina there are 8-10 million MCs

(Jadhav et al., 2009; Vecino et al., 2015). MC development starts in a relatively late

stage of retinal histogenesis, at embryonic day 18 (E18) and terminates at postnatal day

12 (P12) (Jadhav et al., 2009). As neurogenesis occurs from E10 to P12, the

development of neurons precedes gliogenesis (Aparecida Da Silva et al., 2013; Vecino

et al., 2015).

24

Most of the retinal cells is generated from common multipotent retinal progenitor cells

(Ohsawa and Kageyama, 2008), but their final cellular fate determined by differential

expression of selected genes (Vecino et al., 2015). During an early stage of the retinal

histogenesis, the expressions of glia-specific genes are suppressed with

deoxyribonucleic acid (DNA) methylation in the retinal progenitors. Approximately a

week after the commencement of neurogenesis, the by then abundant neuroblasts

express committed Notch ligands to effectively inhibit DNA methylation in their

surrounding progenitors, which in turn activate glia-specific genes. Around the same

time, some cytokines, such as ciliary neurotrophic factor (CNTF), secreted from

neurons, also aid the DNA hypomethylation. (Ahmad et al., 2011) that eventually

triggers the retinal progenitors to differentiate into MCs.

1.5.2 Basic properties of Müller cells (MCs) Cell bodies of MCs are located in INL (Goldman, 2014) and their processes are

elongated toward the nerve fibre layer (NFL) and outer limiting membrane (OLM),

thereby spanning across the retina (Bringmann et al., 2006; Bringmann and Wiedemann,

2012; Vecino et al., 2015) (Figure 1.3A). MCs connect to photoreceptors or other

retinal neurons by forming columnar structural units (Figure 1.3B), which support an

architectural stability of the retina. In addition to this anatomical support, MCs also

maintain homeostasis within the retina. MC processes constitutes inner BRB by

surrounding capillary endothelial cells and pericytes, which maintains a proper

microenvironment for neuronal activity in the retina (details are referred to section 1.4).

1.5.3 The role of Müller cells (MCs) in vision formation

In addition to their anatomical and metabolic support role, MCs are also involved in the

process of vision. Firstly, in an inverted retina of the vertebrate eyes, light has to pass

all retinal layers before it is captured by photoreceptors. Retinal cells, their processes,

and organelles are objects that cause light scattering. Light scattering within the retinal

layers reduces visual sensitivity and acuity. Due to their anatomical position and

structure, MCs have been suggested to function as fibre optics that help in guiding the

light and thereby increasing photon absorption in cones (Reichenbach and Bringmann,

2013).

25

(A) (B)

Figure 1.3 Schematic drawing of cellular components of the retina (A) The cell soma of MC (dark blue) is localised at INL and its processes (light blue) extend toward both surfaces of the retina. MC processes are enlarged to the endfoot (E) in NFL and GCL. MC processes, extending to ONL, form a honeycomb-like shape (microvilli). A: amacrine cells; AS: astrocytes in green; B: bipolar cells; BV: blood vessels; C: cones; Ch: choroid; E: endfoot; G: ganglion cells; H: horizontal cells; M: Müller cells; Mi: microglia in red; MIC: microvilli; NFL: nerve fibre layer; ON: optic nerve; PE: pigment epithelium; R: rods. (B) Artist’s view of a human MC (left black or right blue), enveloping distinct types of retinal neurons (green) and establishing contacts to retinal blood vessels (red). Cap: capillaries. Image A is cited from the publication (Vecino et al., 2015) (https://doi.org/10.1016/j.preteyeres.2015.06.003) and permitted to be reused in this thesis under the terms of Creative Commons Attribution-NonCommercial-No Derivatives License (CC BY NC ND). Image B is reproduced with permission from the publication (Bringmann et al., 2006) (License number: 4393500775476 provided by Elsevier and Copyright Clearance Center), Copyright © 2006 Elsevier Ltd. All rights reserved.

MCs, by acting as living optical fibres (light-guiding fibres) can minimise light

scattering helped by their funnel-shaped endfoot at the vitreal surface of the retina

(Franze et al., 2007). This configuration can assist light to transport from the inner

retinal surface to photoreceptors with minimal loss of light that results in increased

visual sensitivity and contrast (Franze et al., 2007; Reichenbach and Bringmann, 2013).

Secondly, MCs can support visual-chromophore regeneration in the retina (Kaylor et

al., 2013; Miyazono et al., 2008) (Figure 1.4). MCs take up all-trans-retinol released

26

from the cone OS where all-trans-retinal is reduced to all-trans-retinol after the

initiation of phototransduction (Kaylor et al., 2013; Miyazono et al., 2008) (details are

referred to section 1.3). And then all-trans-retinol is isomerized by dihydroceramide

desaturase-1 (DES1), a retinol isomerase, to 11-cis-retinol in MCs. Cellular

retinaldehyde-binding protein (CRALBP), which is identified as a carrier protein in

MCs (Boppana et al., 2012; Saari et al., 1982), interacts with DES1 and prevents 11-

cis-retinol from further isomerization (Kaylor et al., 2013). This complex interacts with

negatively charged phospholipid, where 11-cis-retinol is separated from CRALBP and

then released to the interphotoreceptor matrix (IPM) from MCs. The cone OS takes up

the released 11-cis-retinol and then oxidises it to 11-cis-retinal, a light sensitive

chromophore (Kaylor et al., 2013).

Figure 1.4 The schematic diagram displays the visual cycle in MCs The 11-cis-RAL, visual chromophore in cone opsins, is isomerized to all-trans-RAL by photos which are absorbed by cone opsins. All-trans-RAL is reduced to all-trans-ROL by retinol dehydrogenase 8 (RDH8) in the cone OS and then released to the interphotoreceptor matrix (IPM), where all-trans-ROL is taken up by MCs. Here, the all-trans-ROL is isomerized by dihydroceramide Δ4-desaturase-1 (DES1) to 11-cis-ROL, which is bound to CRALBP. Interaction with negatively charged phospholipids causes CRALBP to release 11-cis-ROL into the IPM, where is taken up by the cone OS. In the cone OS, 11-cis-ROL is oxidized by an unknown NADP+-dependent RDH to 11-cis-RAL , which combine with apo-opsin to form new opsin pigments. 11-cis-RAL: 11-cis-retinal; 11-cis-ROL: 11-cis-retinol; all-trans-RAL: all-trans-retinal; all-trans-ROL: all-trans-retinol; CRALBP: Cellular retinaldehyde-binding protein; DES1: dihydroceramide Δ4-desaturase-1; RDH8: retinol dehydrogenase type 8 Image is cited from the publication (Kaylor et al., 2013) (License number: 4480060927736 provided by Springer Nature and Copyright Clearance Center), Copyright © 2012 Springer Nature. All rights reserved.

27

1.5.4 The role of Müller cells (MCs) in neuroprotection

MCs maintain homeostasis and an appropriate metabolic microenvironment within the

retina to support neuronal function. MCs take up toxic neurotransmitters, and recycle

them as non-toxic precursors (glutamine) to be taken up by neurons. MCs are capable

of taking up extracellular glutamate via the electrogenic GLAST and convert it into

glutamine with the glial-specific enzyme glutamine synthetase (GS) in cells. Glutamine

is then released by MCs and then collected by neurons as precursors for

neurotransmitters (Reichenbach and Bringmann, 2013). Furthermore, MCs produce

antioxidants and glutathione that reduce retinal oxidative stress, which is a consequence

of a high level of oxygen consumption of photoreceptors in healthy retinas. (Bringmann

and Wiedemann, 2012; Pow and Crook, 1995). In addition, MCs release neurotrophins

and growth factors, such as CNTF, neurotropin-3 (NT-3), and basic fibroblast growth

factor-2 (FGF-2) to support survival and differentiation of retinal neurons (Vecino et

al., 2015). Without the support by MCs, the photoreceptor degeneration was apparent

and accompanied with BRB breakdown and intraretinal neovascularization in the

selective MC ablation mouse model (Shen et al., 2012).

1.5.5 Müller cells (MCs) activation MCs become activated (reactive) in response to pathological changes of the retina

(Bringmann et al., 2009). MC reaction is one component of a complex retinal response

to pathogenic stimuli to protect the retina from further damage and to preserve tissue

function (Bringmann et al., 2009). However, over-reactive MCs, termed as MC

proliferative gliosis, result in detrimental effects on the injured retina in retinal

degeneration and diseases (Bringmann et al., 2006; Bringmann and Wiedemann, 2012).

Therefore, MC activation is described as having a “Janus face”, which has both

cytoprotective and cytotoxic effects on the damaged retinas (de Hoz et al., 2016).

Characteristics of MC activation are described as follows.

1.5.5.1 Non-specific and specific Müller cells (MCs) response

MC reaction is characterised by unspecific and specific responses to retinal insults. The

former one is independent on stimulus of retinal stress or injuries. Three major and

crucial unspecific responses of MC reaction are: cellular hypertrophy, proliferation, and

up-regulation of intermediate filament proteins, including glial fibrillary acidic protein

28

(GFAP), nestin and vimentin, in MCs (Bringmann et al., 2009). Elevated GFAP

expression is a well-known and the most sensitive early stress marker of MC reaction

in response to retinal injury (Albarracin and Valter, 2012; Bringmann et al., 2009;

Bringmann et al., 2006; Bringmann and Wiedemann, 2012). Studies reported that

upregulation of GFAP may be associated with the initiation of MC reaction

(Bringmann et al., 2009; Verardo et al., 2008). In mice, deficient of GFAP, MCs display

thinner and shorter processes, as well as a reduced extent of glial scar formation

compared to the wild-type mice in a retinal detachment model (Nakazawa et al., 2007b;

Verardo et al., 2008). These changes also reduced monocyte infiltration into the injured

retinas and led to a decrease of photoreceptor death. In addition, GFAP has been

reported to play an important role in protecting the brain against glutamate-

excitotoxicity by assisting extracellular glutamate removal (Sullivan et al., 2007). The

study showed that GFAP had interaction with glutamate aspartate transporter (GLAST,

a glutamate transporter), a scaffolding protein (PDZ-binding protein), and a membrane-

cytoskeleton protein (Ezrin). COS7 cells, while expressing GFAP and GLAST,

displayed the higher capability of taking D-aspartate (a glutamate analogue). In an in

vivo model, the presence of GFAP was essential in retaining GLAST in the plasma

membranes of astrocytes in mammalian brains after hypoxic insult.

Specific MC gliotic response are dependent on the given pathogenic factors and

mechanisms (Bringmann et al., 2006). Altered expression of GS in MCs is a prominent

example of specific MC gliotic responses. GS, a specific enzyme in MCs, participates

in neurotransmitter recycling (details are referred to section 1.5.4). Its expression was

shown to decrease in MCs in the bright light-induced retinal degeneration in rats

(Grosche et al., 1995), and the retinal detachment model in cats (Lewis et al., 1989),

causing glutamate-induced photoreceptor death. But GS levels was shown to increase

in MCs, in an in vitro model of hepatic retinopathy, where GS is necessary to detoxify

cells from elevated levels of ammonia (Reichenbach et al., 1995). However, no

alternation of GS expression was shown in MCs of patients with diabetic retinopathy

or in rats following optic nerve crush, although injured retinas displayed up-regulation

of GFAP (Chen and Weber, 2002; Mizutani et al., 1998).

29

1.5.5.2 Protective effects of Müller cells (MCs) activation Shortly after retinal injury, or in less severe retinal damaged, MC activation is described

as “conservative” or “non-proliferative” reaction, whereby MC reaction is

neuroprotective and attempts to limit further retinal injuries by various mechanisms,

including the release of neurotrophic factors as well as clearance of cellular debris and

waste (Bringmann et al., 2009; de Hoz et al., 2016). Details are in following sections.

1.5.5.2.1 Retinal neuron survival Following retinal damage, MCs release neurotrophic factors to prevent neuronal death.

In the rat model of bright light-induced retinopathy, brain-derived neurotrophic factor

(BDNF), secreted by MCs, provides a feed-forward loop to stimulate neighbouring

MCs to release other neurotrophic factors, including CNTF and FGF-2 (Joly et al.,

2007), which enhance photoreceptor survival (Bringmann et al., 2009). It has been

described that elevated level of leukemia inhibitor factor (LIF) in MCs was essential to

prevent photoreceptor death (Burgi et al., 2009; Joly et al., 2008) and that it directly

increased FGF-2 gene expression of (Burgi et al., 2009), in bright light-induced, or

opsin mutant retinal degeneration animal models. This protective mechanism was

linked to intrinsic endothelin-2 (edn2) signalling cascade activation (Joly et al., 2008)

or Janus kinase/signal transducer and activator of transcription (Jak/STAT) signalling

pathway (Burgi et al., 2009).

1.5.5.2.2 Müller cells (MCs) phagocytosis Phagocytosis of exogenous particles, cell debris, and haemorrhage is crucial in

repairing injured retinas (Bringmann et al., 2009; Garcia and Vecino, 2003). In retinal

diseases or following RPE transplant, activated MCs were capable of phagocytosing

fragments of RPEs (Bringmann et al., 2009; Francke et al., 2001). However, the

association between the tissue regeneration/repair and MC phagocytosis remains

elusive.

1.5.5.3 Detrimental effects of Müller cells (MCs) activation Activated MCs, if uncontrolled, display dysfunctional metabolism, impair ion and

water homeostasis, release proinflammatory cytokines, and undergo proliferation

within the retina. The resulting glial scar inhibits tissue repair and causes further retinal

30

cell death. The following section will discuss the detrimental effects of MC activation

on injured retinas.

1.5.5.3.1 Disruption of glia-neuron interaction Declined expressions of MC-specific proteins or enzymes has been shown to be linked

to the functional uncoupling of neurons and glia in the diseased retina (Bringmann and

Wiedemann, 2012). For example, consistent with MC reaction (upregulation of GFAP),

MCs failed to convert glutamate into glutamine (Lieth et al., 1998), which causes an

accumulation of excitotoxic glutamate within the retina, leading to neuronal death

(Bringmann and Wiedemann, 2012). Downregulation of CRALBP in MCs (its function

is referred to section 1.5.3) has been reported to disrupt vision in degenerating retinas

(Lewis et al., 1994; Pfeiffer et al., 2016; Sheedlo et al., 1995). In addition, a

redistribution of Kir4.1 (inward rectifying potassium channel) and aquaporin-4 (water

channel) in damaged MCs causes the disruption of K+ and water transport, leading to

retinal oedema in the mouse model light-induced retinal degeneration (Iandiev et al.,

2008). Taken together, dysregulated protein expression of MCs disrupts glia-neuron

interactions, which increases the susceptibility to retinal insults and causes further

neuronal death in damaged retinas (Bringmann et al., 2009).

1.5.5.3.2 Inflammatory and related immune responses MC-related immune responses are initiated and accelerated with extravasated serum

component in the retina, such as immunoglobulin G (Chu et al., 1999), or blood-derived

immune cells, such as T lymphocytes (Richardson et al., 1996), following the

breakdown of BRB (Bringmann et al., 2009; Kumar et al., 2013). Activated MCs

express and release a wide variety of proinflammatory cytokines to trigger and amplify

inflammation responses in damaged retinas (Hollborn et al., 2008), for example Ccl2.

Ccl2 has been shown to facilitate MG/MΦ migration into the injured areas, in the bright

light-induced model of retinal degeneration (Rutar et al., 2012b). In addition, in vitro

research confirmed that activated MCs were capable of producing interleukin-1 beta

(IL-1β), interleukin-6 (IL-6), inducible nitric oxide synthase (iNOS) (Wang et al., 2011),

tumour necrosis factor alpha (TNFα) (Walker and Steinle, 2007), and reactive oxygen

species (ROS) (Deliyanti and Wilkinson-Berka, 2015; Singh et al., 2014). Intracellular

adhesion molecular (ICAM-1) protein, which triggers MG/MΦ infiltration (Wang et al.,

31

2011) and activates T cells (Richardson et al., 1996), was presented in MCs at injury

sites of the retina, in a retinal laser photocoagulation model (Richardson et al., 1996).

Taken together, evidence support that activated MCs play a role in amplifying

inflammation by promoting leucocyte adhesion and MΦ activation and migration.

1.5.5.3.3 Proliferation and glial scar formation

The most severe retinal insults induce another level of MC responses described as

“massive” or “proliferative” gliosis (Bringmann et al., 2009), where MCs re-enter the

cell cycle to undergo proliferation (Bringmann et al., 2006). Proliferative MCs migrate

toward the subretinal space and vitreal surface, to form subretinal or epiretinal

membranes (glial scars), respectively (Figure1.5) (de Hoz et al., 2016; Lewis et al.,

2010). These glial scars, fill in spaces left behind by dying or dead neurons (Bringmann

et al., 2006). However, the stiff and fibrotic glial scars impede the re-establishment of

synaptic connectivity and limit nutrient delivery from the choroid to the outer retina,

thereby causing the damage or death of neighbouring cells (Sethi et al., 2005).

Moreover, gliotic MCs increased the expression of extracellular matrix (Inatani et al.,

2000) and cell adhesion molecules, CD44 (Chaitin et al., 1996). The content of

extracellular matrix, such as small chondroitin/heparan sulphate proteoglycans, and

CD44 function as chemical inhibitors of axonal growth and neuronal regeneration

(Bringmann and Wiedemann, 2012; Canning et al., 1996; Ponta et al., 2003). In

addition, while glial scars that are formed on the vitreal surface cause vision distortion

and impairment by creating macular wrinkling, retinal folds, and retinal detachment

(Lewis et al., 2010).

Figure 1.5 MC proliferative gliosis in injured retinas Schematic drawing illustrates that MCs (orange) migrate to the subretinal space or vitreous body where they form glial membranes. Image is cited from the publication (de Hoz et al., 2016) (http://dx.doi.org/10.1155/2016/2954721) and permitted to be reused in this thesis under the terms of Creative Commons Attribution-NonCommercial-NoDerivatives License (CC BY NC ND).

32

1.6 Extracellular vesicles (EVs) Intercellular communication is a crucial characteristic in multi-cellular organisms. It

can occur through direct cell-cell contact (nanotubes), or by transfer of secreted soluble

factors binding to plasma receptors, to activate signalling cascades in target cells (Lee

et al., 2012; Minciacchi et al., 2015; Raposo and Stoorvogel, 2013). Cells are now

known to exchange information through a third mechanism, by the release of

membrane-enclosed particles called extracellular vesicles (EV) (Raposo and

Stoorvogel, 2013).

EVs are a heterogeneous population of membranous vesicles of various origins and

broadly organised into three categories depending on their biogenesis: exosomes,

microvesicles (MVs), and apoptotic bodies. A common character in these three EV

subtypes is a lipid bilayer membrane, that surrounds specific biomolecules, such as

DNA, ribonucleic acid (RNA), proteins, or cellular debris. However, their size and

buoyant densities are significantly different (Kalra et al., 2016). The description of the

size and buoyant density ranges for the diverse EV subtypes greatly varies in the

literature. Exosomes are described to be around 30-100 nm, while MVs are 100-1000

nm, and apoptotic bodies are 1000-5000 nm in diameter and have a buoyant density of

1.10 -1.14 g/ml (Gyorgy et al., 2011).

Recently, exosomes and MVs are proposed to be promising tools in gene therapy, due

to their natural characteristics of carrying and transferring genetic materials to target

cells. Micro RNAs (miRNAs), which are well-known for their ability to silence or

degrade target mRNAs (Bartel, 2004), has been identified in exosomes and MVs (Buzas

et al., 2014; Gyorgy et al., 2011; Lee et al., 2012; Ludwig and Giebel, 2012; Raposo

and Stoorvogel, 2013; Zhang et al., 2015). EV carrying miRNAs enter recipient cells,

where they can regulate gene expression (Ludwig and Giebel, 2012; Raposo and

Stoorvogel, 2013; Zhang et al., 2015). In addition, EVs act as lipid-based carriers that

can shield the enclosed content from degradation by plasma ribonucleases (RNase),

while EVs are secreted into the serum and transported to distant sites (tissues) (Lee et

al., 2012; Vickers and Remaley, 2012; Zhang et al., 2015). The biogenesis and

characteristics of exosomes and MVs are discussed in the following sections.

33

1.6.1 Microvesicles (MVs) MVs were first described by Chargaff and West in 1946 as a precipitable factor in

platelet-free plasma (Chargaff and West, 1946). MVs have been shown to be released

from endothelial cells and monocytes (Inal et al., 2012). Recent studies report that

neutrophils (Inal et al., 2012), progenitors of differentiated cells, and tumour cells

(Muralidharan-Chari et al., 2010) are also capable of secreting MVs. MVs are created

by the formation of outward buds (budding/blebbing) at specific sites of the cell

membrane followed by fission and subsequent release of the vesicle into extracellular

space (Gyorgy et al., 2011; Minciacchi et al., 2015) (Figure 1.6). During MV biogenesis,

the plasma membrane undertakes several molecular rearrangements at the sites of MV

formation, including alterations of Ca2+ level in cells as well as the lipid and protein

composition at the cell membrane (Minciacchi et al., 2015). Upregulated Ca2+ levels

trigger the recruitment and activation of calcium-dependent enzymes, such as

scramblase and floppase, modifying the plasma membrane composition.

Polymerization of actin cytoskeleton, such as lamellipodia and filopodia formation, are

present during MV shedding from human blood platelets via Cdc42/Rac1/p21-activated

kinase (PAK)-dependent pathway (Crespin et al., 2009; Minciacchi et al., 2015).

1.6.2 Exosomes Exosomes were discovered in the 1980s (Harding et al., 2013). Stahl (Harding and Stahl,

1983) and Johnstone (Pan and Johnstone, 1983) observed the release of multi-vesicular

mature endosomes from differentiated reticulocytes into the extracellular environment

(Lee et al., 2012; Raposo and Stoorvogel, 2013). Few years later, the name “exosomes”

was coined by Johnstone and colleagues to define these multivesicular mature

endosomes (Buzas et al., 2014; Johnstone et al., 1987; Zhang et al., 2015), although the

term had in fact been used a few years earlier, when referring to other membrane

fragments isolated from biological fluids (Trams et al., 1981). Presently, exosomes

identify small extracellular vesicles ranging from 30 to 100 nm in size and are derived

from an endocytic recycling pathway (details below) (Minciacchi et al., 2015). Most

cell types, including T cells, B cells, mast cells, neurons, endothelium, epithelial cells,

embryonic stem cells, and tumour cells, are capable of releasing exosomes (Evans-

Osses et al., 2015; Raposo and Stoorvogel, 2013).

34

Exosomes are created via an endocytic recycling pathway (Figure 1.6). The endosomal

system manipulates the uptake and processing of various types of macromolecules from

the extracellular milieu and plasma membrane inside the cells (Ludwig and Giebel,

2012). The endosomal system is composed of various interconnected vesicular

organelles, including primary endocytic vesicles, early endosomes, recycling

endosomes, late endosomes, and lysosomes. Endocytic vesicles form at the plasma

membrane and then fuse to form early endosomes (Lee et al., 2012). Early endosomes

mature and then become the late endosomes (mature endosomes), where intraluminal

vesicles bud off parts of membranes of late endosomes to form multivesicular

endosomes (MVEs) (Buschow et al., 2009). MVEs either fuse with lysosomes to

degrade their intraluminal content, such as messenger RNAs (mRNAs) or miRNAs, or

with the plasma membrane to release exosomes into the extracellular space (Lee et al.,

2012; Ludwig and Giebel, 2012).

Figure 1.6 Release of microvesicles (MVs) and exosomes MVs bud directly at sites of the plasma membrane, whereas exosomes are derived from an endocytic recycling pathway. Early endosomes are formed from the plasma membrane and then become late endosomes (or called MVE), where intraluminal vesicles bud off into an intra-cytoplasmic lumen. MVEs either fuse with plasma and then release exosomes into extracellular space or incorporate with lysosomes where the contents are degraded. CCV: clathrin-coated vesicles; ER: endoplasmic reticulum; MVE: multivesicular endosome. Image is reproduced with permission from (Raposo and Stoorvogel, 2013) (License number: 4310690862821 provided by Copyright Clearance Center), Copyright © Rockefeller University Press

35

Exosomes released from secreting cells are delivered to, and their contents are released

into recipient (target) cells by binding to molecules at the cell membrane of their target.

Studies have shown that the protein content of exosomes or those on their membranes

may be involved in manipulating exosome uptake by and function of recipient cells.

Exosomes are enriched by mannose-and sialic acid-containing glycoproteins, that play

a role in the exosomal uptake by recipient cells (Escrevente et al., 2011). Tspan8, a

member of the tetraspanin family, is well-known to form a complex with integrin. The

tspan8-CD49d complex, on the rat aortic endothelial cell-derived exosome is required

for exosome internalization (Nazarenko et al., 2010). Further, other members of the

tetraspanin family, CD9 and CD81, were also identified to control exosome targeting

by recipient cells (Morelli et al., 2004).

1.6.3 Role of extracellular vesicles (EVs) in cell-cell communication

EVs have emerged as important vehicles and tools for the transfer of genetic

information in physiological and pathological conditions, regulating immune activities,

cancer formation, tumour invasion and metastasis, as well as the spread of viruses (Lee

et al., 2012; Ludwig and Giebel, 2012). In the immune systems, exosomes released

from different immune cells can harbour a broad range of immunoregulation-related

molecules, such as major histocompatibility complex (MHC) class II, ICAM-1, cluster

of differentiation 86 (CD86) that manipulate a variety of immunological functions

including T cell activation, immune tolerance induction, and dendritic cell maturation

(Karlsson et al., 2001; Lee et al., 2012; Skokos et al., 2003; Sprent, 2005). In response

to endotoxin stimulation, dendritic cell-released exosomes, carrying miR-155 and miR-

146a that were subsequently taken up by neighbouring dendritic cells, regulated the

expression of pro-inflammatory genes in recipient cells (Alexander et al., 2015). In

addition, EVs re-shape the local tumour environment into a more favourable niche for

tumour growth, invasion, and metastasis (Lee et al., 2012). In cancer environment, EV-

carrying miRNAs let-7, miR15b, and oncogenic receptors were transferred to other

non-cancer cells, resulting in further tumour growth and metastasis (Skog et al., 2008).

EVs secreted by lung cancer cells has been shown to deliver pro-angiopoietic factors

to surrounding non-cancer cells and stromal cells, which accelerates cancer metastasis

(Wysoczynski and Ratajczak, 2009). Moreover, EVs result in disease spreading by

transferring vesicle-mediated receptors to non-infected cells. EVs derived from human

36

peripheral blood platelets and megakaryocytes were described to transfer C-X-C

chemokine receptor type 4 (CXCR4) to CD4+/CXCR4-null cells, increasing their

susceptibility to infection by human immunodeficiency virus (HIV) (Rozmyslowicz et

al., 2003).

Roles of EVs in eye diseases gained focus in the last decade. To date, most studies

concentrated on exosomes derived from RPE, in response to insults in eyes (Klingeborn

et al., 2017). In response to oxidative stress, RPE-derived exosomes carried vascular

endothelium growth factor (VEGF) receptors, which promoted angiogenesis on

endothelial cells (Atienzar-Aroca et al., 2016). In addition, a larger amount of MVs was

found in the vitreous of diabetic retinopathy (DR) patients compared to samples

collected from non-diabetic patients. These MVs have been proven to originate from

endothelium, platelets, and retinal cells, and their content induced endothelial cell

proliferation and new vessel formation in vitro (Chahed et al., 2010). In contrast,

exosomes derived from retinal glial cells inhibited angiogenesis. Astroglial cell-derived

exosomes, containing endostatin, blocked retinal vessel leakage, inhibited angiogenesis,

and suppressed MΦ migration in a laser-induced CNV mouse model (Hajrasouliha et

al., 2013).

1.7 Age-related macular degeneration (AMD) Age-related macular degeneration (AMD), a progressive condition, is the principal

cause of irreversible blindness in people aged 50 years or older in developed countries

(Ambati and Fowler, 2012; Jager et al., 2008). AMD is a multifactorial disease and its

aetiology remains largely unknown. To date, aging, race, heredity, and history of

smoking have been confirmed to be risk factors in the development of AMD (Jager et

al., 2008).

There are two types of AMD, dry and wet forms. The more common dry form (90% of

cases), is characterised by relentless, gradual vision loss and the formation of drusen in

the macula (Ambati and Fowler, 2012; Buschini et al., 2011). Wet AMD, is

accompanied by neovascularization from choroid into sub-RPE and subretinal regions.

This form affects 10-15% of AMD patients (Ambati and Fowler, 2012). In contrast to

dry AMD, wet AMD emerges abruptly and rapidly progresses into blindness, if left

37

untreated. Although wet AMD represents a small proportion of total AMD cases, it

contributes to the majority of blindness associated with AMD. In clinics, AMD patients

generally develop dry AMD first, and wet AMD occurs following dry AMD. Therefore,

dry AMD is regarded as a risk factor or even a precursor state for wet AMD (Ambati

and Fowler, 2012).

A shared clinical symptom of these two forms at early stages of AMD is a focal

deposition of cellular polymorphous debris, drusen, between the RPE and Bruch’s

membrane in macular regions (Jager et al., 2008; Kim, 2015) (Figure 1.7).

Histologically, drusen deposits can be divided into hard (small with defined margins)

and soft drusen (larger and without defined margins). Hard drusen deposits are few and

distributed at peripheral retinas with aging. They may coalesce into soft ones, causing

RPE abnormalities. Along with the development of atrophy of RPE cells, the damaged

area turns into a late stage dry AMD, also known as geographic atrophy (GA) (Ambati

and Fowler, 2012; Buschini et al., 2011). As survivals and functions of photoreceptors

are fundamentally supported by RPE cells, GA is associated with degenerated

photoreceptors and vision loss (Ambati and Fowler, 2012; Buschini et al., 2011). In wet

AMD, choroidal neovascularisation (CNV), newly-formed blood vessels grow into the

outer retina from the underlying choroid (Ambati and Fowler, 2012). This abnormal

blood vessel proliferation results in bleeding and fluid exudation below the retina,

causing significant macular oedema. Neovascular membranes then form a big scar in

the macula, eventually leading to a dramatic decrease in central vision (Ambati and

Fowler, 2012). Currently, anti-VEGF treatment is the main therapy for wet AMD.

However, there is still no effective treatment for GA (dry AMD) (Knickelbein et al.,

2015).

Pathogenesis of AMD is undoubtedly multifactorial and complicated. Its pathogenesis

still need to be further verified. Currently, RPE cell dysfunction and atrophy, precede a

later stage of AMD, may be the fulcrum of AMD pathogenesis (Ambati and Fowler,

2012). Subsequent events, oxidative stress (Hernandez-Zimbron et al., 2018), immune

dysfunction (Ambati et al., 2013; Ambati and Fowler, 2012), chronic local

inflammation (Knickelbein et al., 2015), complement activation (Buschini et al., 2011;

Knickelbein et al., 2015), and inflammasome activation (Gao et al., 2015) facilitate

38

AMD pathogenesis and cause further damage in the retinas. Following sections

discusses contribution of immune responses and retinal cells to AMD pathogenesis.

Figure 1.7 Schematic transverse sections through the human eyeball: the classification of AMD. (A&B) A healthy human retina. (C&D) Retina of an individual with dry AMD. (E&F) Retina of a patient with wet AMD. (C) A fundus photograph displayed presence of numerous yellow deposits, drusen, in central macula. (D) A cross-section of macula shows three drusen under RPE cells. (E) Severe macula oedema and foveal haemorrhage observed in a fundus photograph, whereas only small sparse deposits, drusen, were presented. (F) An optical coherence tomography scan through a fovea area showed the formation of intraretinal fluid cysts. Oedema resulted in a foveal pit disappeared and disorganised retinal layers. Images are cited from the publication (Kauppinen et al., 2016)(doi: 10.1007/s00018-016-2147-8) and permitted to be reused in this thesis under the terms of Creative Commons Attribution-NonCommercial-NoDerivatives License (CC BY NC ND).

39

1.7.1 Inflammation in AMD In AMD, intense inflammatory reactions are not present in patients’ eyes. Instead, a

chronic, low-grade inflammation, described as parainflammation, has been suggested

as a factor causing subsequent AMD pathology. Immune cells, MΦ/MG,

proinflammatory cytokines, chemokines, and activated inflammasomes participate in

the regulation of parainflammation in degenerating retinas (Knickelbein et al., 2015).

1.7.1.1 The roles of microglias and macrophages (MGs/MΦ) in AMD Growing evidence show that residential MGs and choroidal MΦ become activated and

migrate into the area of injury in the course of AMD, having a crucial role in initiating

and propagating the inflammation process (Madeira et al., 2015). In AMD patients,

CX3C chemokine receptor 1 (CX3CR1)-dependent MG accumulation has been

described in outer retina and subretinal spaces at affected sites, which may be a driving

force in AMD pathogenesis (Combadiere et al., 2007). Activated MGs have been shown

to express several chemokines, including C-C motif chemokine ligand 3 (Ccl3), Ccl4

and Ccl7, mediating photoreceptor degeneration in the rat model of retinal degeneration

(Rutar et al., 2015). Aggregation of MΦ was observed in or around drusen in the

subretinal space in patients affected by AMD. These MΦ generated pro-angiogenic

factors, VEGF (Knickelbein et al., 2015), which has a well-defined role in triggering

proliferations of endothelial cells (Oh et al., 1999) and recruiting MΦ (Cursiefen et al.,

2004). In an animal model of retinal degeneration, a large number of MΦ was present

in the photoreceptor layer and around PRE, expressing Ccl3, Ccl4, and IL-6 (Fernando

et al., 2016). Polarization of MΦ may potentially have an effect on AMD progression.

A higher ratio of M1: M2 chemokine transcripts was presented in MΦ in maculae of

dry AMD patients, compared to normal autopsied eyes (Cao et al., 2011), implying that

MΦ in dry AMD patients’ eyes tends to be pro-inflammatory. However, whether

MG/MΦ activation is involved in disease progression or is an accumulative

consequence of the AMD pathogenesis is still not fully elucidated.

1.7.1.2 Proinflammatory cytokines and chemokines

Several proinflammatory cytokines/chemokines were found in serum and aqueous

humour of AMD patients (Buschini et al., 2011; Jonas et al., 2012; Klein et al., 2014;

Knickelbein et al., 2015; Miao et al., 2012). The systemic IL-6 level in serum is

40

associated with a 20-year cumulative incidence of early AMD, independent of ages,

smoking status, and other factors (Klein et al., 2014). Patients with diabetic macular

oedema had higher concentration of IL-6 and IL-8 in aqueous humour samples

compared to patients undergoing cataract surgery (Jonas et al., 2012). The level of IL-

6, presented in aqueous fluid in wet AMD patients, is strongly correlated with volumes

of macular oedema in patients with CNV compared to patients with cataract, idiopathic

epiretinal membrane, or macular hole (Knickelbein et al., 2015; Miao et al., 2012).

However, it is still not clear whether these cytokines are involved in the primary

pathogenesis of AMD.

Chemokines are a large family of molecules with potent chemoattractant properties in

the CNS (Bajetto et al., 2002). The higher protein and gene levels of Ccl2, a

chemoattractant for MΦ (Yoshimura et al., 1989) and MG (Nakazawa et al., 2007a),

were found in serum in AMD patients (Anand et al., 2012) and in degenerated retinas

in rats (Rutar et al., 2011b). Laboratory studies have proven that the presence of Ccl2

is strongly linked to activation (Ambati et al., 2003) and recruitment of MG (Rutar et

al., 2012b) at sites of injury in rats. Chemokine receptors, such as chemokine

(fractalkine) receptor chemokine (C–X3–C motif) receptor 1 (CX3CR1) and C-C-

chemokine receptor 2 (CCR2), has been shown to regulate immune cell trafficking into

the degenerated retinas (Combadiere et al., 2007; Rangasamy et al., 2014). Aged mice

deficient with CX3CR1 genes displayed aberrant immune cell migration, causing MG

accumulation in the affected retinas and subretinal space. These effects may correlate

to thinning of photoreceptor layers (Combadiere et al., 2007). CCR2-or Ccl2 deficient

aged mice, accompanied by impaired in recruiting MΦ, exhibited overexpression of

complement component 5a (C5a) and IgG in RPE and choroid, causing RPE and/or

choroidal endothelial cells to generate VEGF (Ambati et al., 2003), well-known to

mediate CNV development in wet AMD (Buschini et al., 2011).

1.7.1.3 Inflammasome activation in AMD Inflammasome refers to multi-protein intracellular complexes in response to

stimulation of pattern recognition receptors, including NOD-like receptor (NLR), by

various cellular stress (Knickelbein et al., 2015). Following cellular stress, such as

pathogen-associated molecular pattern (PAMPs) from microorganisms or damage-

41

associated molecular patterns (DAMPs) from endogenous signals, NLR family pyrin

domain-containing 3 (NLRP3) oligomerises through homotypic interactions between

NAIP, CIITA, HET-E and TP1 (NACHT) domains. PYD of NLRP3 then interacts with

PYD of apoptosis-associated speck-like protein containing a caspase-recruitment

domain CARD (ASC). In turn, CARD of ASC recruits pro-caspase 1 via CARD-CARD

interactions (Kauppinen et al., 2016; Tschopp and Schroder, 2010). These protein

complexes lead to production of active caspase-1 that cleaves and activates process and

secretion of IL-1β and interleukin -18 (IL-18) (Knickelbein et al., 2015; Tschopp and

Schroder, 2010) (Figure 1.8).

Figure 1.8 Schematic diagram presents the NLRP3 inflammasome complex formation In the presence of PAMPs or DAMPs from endogenous signals, NLRP3 oligomers are assembled at NACHT domain, recruiting ASC and pro-caspase 1. Following the formation of this complex, caspase 1 is activated to trigger the generation and release of IL-1β and IL-18. ASC: Apoptosis-associated speck-like protein containing a caspase-recruitment domain; CARD: caspase-recruitment domain; DAMP: damage-associated molecular patterns; PAMP: pathogen-associated molecular pattern; NACHT: NAIP, CIITA, HET-E and TP1 (NACHT); NLRP3: NLR family, pyrin domain-containing 3; LRRs: leucine-rich repeats; PYD: pyrin domain. Image is reproduced with permission from (Tschopp and Schroder, 2010)(License number: 4396971510306 provided by Springer Nature and Copyright Clearance Center), Copyright © 2010, Springer Nature

42

Inflammasome activation has been implicated in AMD pathogenesis (Ambati et al.,

2013). Emerging evidence show that RPE is one source of inflammasome activation in

diseased retinas. RPE cells exhibited NLRP3 staining at lesion sites in retinas of dry

or wet AMD patients (Tseng et al., 2013). Drusen component of AMD patients,

amyloid-beta 1-40, triggered NLRP3 activation in RPE cells and rat retinas in vivo

(Kauppinen et al., 2016; Liu et al., 2013), which was accompanied by upregulation of

proinflammatory cytokines (IL-6, TNF α, IL-1β), IL-18, and caspase-1. In an in vitro

study, IL-1α-stimulated RPE cells expressed pro-1β and pro-caspases-1 via activation

of NLRP3 inflammasome (Tseng et al., 2013).

Inflammasome activation is also presented in immune cells within the retinas. Under

oxidative stress, MGs polarised toward M1 state via NLRP3 inflammasome activation

and expressed a upregulated level of angiogenic factors. Transplantation of these MGs

into subretinal space facilitated CNV development (Indaram et al., 2015). Besides,

drusen extraction from AMD patients’ eyes activated human peripheral blood

mononuclear cells and bone marrow-derived MΦ, generating IL-1β via NLRP3

inflammasome activation (Doyle et al., 2012b). Complement component 1q (C1q) and

oxidative-stress–related protein-modification carboxyethylpyrrole (CEP), identified as

components in the drusen, have been link to activate or prime inflammasome complexes

in these leukocytes (Doyle et al., 2012b).

IL-1β and IL-18 are effectors of inflammasome activation, as mentioned above (Figure

1.8). The discovery of NLRP3 activation in AMD has prompted investigations into

roles of IL-1β and IL-18 in AMD pathogenesis. IL-1β has been linked to the

acceleration of CNV development in a laser-induced retinal model of neovascular AMD

in mice by increasing choroidal endothelial cell outgrowth (Lavalette et al., 2011).

Conversely, IL-18 may show a protective effect against CNV in mice (Doyle et al.,

2012b). Same group demonstrated that human IL-18, when injected in non-human

primates showed a moderate improvement in laser-induced CNV in these animals

(Doyle et al., 2015). However, the therapeutic effects of IL-18 on AMD are still

questionable, since another report showed no effect of IL-18 on laser-induced CNV in

mice (Hirano et al., 2014).

43

1.7.2 The role of Müller cells (MCs) in AMD The role of MCs in AMD has not been fully elucidated, but emerging clinical and

laboratory studies support that MC gliosis may be involved in AMD pathogenesis. MC

processes shown to express high levels of GFAP in retinas in dry AMD patients around

drusen deposits along the vitreal surface and in the outer retina, compared to those, in

age-matched human eyes (Edwards et al., 2016; Wu et al., 2003). These MC processes,

showed an upregulation of GFAP that may be linked to BRB disruption in patients with

GA (Wu et al., 2003). Laboratory studies have also reported that MCs displayed

upregulation of GFAP and vimentin across degenerated retinas (Albarracin and Valter,

2012; Begum et al., 2013; Marco et al., 2013) and an increased level of Ccl2, which

mediates MG/MΦ recruitment into degenerated retinas in vivo (Rutar et al., 2011b).

1.8 Light damage-induced retinal degeneration in rodent animals In 1966, Noell found that high intensity of light (only 2-3 times above normal room

light) cause progressive photoreceptor death and RPE atrophy in rats and ultimately

resulted in vision loss (Noell et al., 1966). Its molecular basis involves toxic

photoproduct from vitamin A, metabolic dysfunction, and light-induced oxidative

stress (Noell et al., 1966). Because of remarkable similarities between retinal cell

remodelling in retinas of a late-stage light-damage rodent model and the anatomical

changes found in patients with dry AMD (Marc et al., 2008), the rodent light-induced

model of retinal damage have served as an experimental model for investigating human

degenerative diseases, such as AMD (Organisciak and Vaughan, 2010). These

remarkable hallmarks include,

1. After exposure to bright light, focal photoreceptor degeneration appears at an early

stage, which is followed by RPE dysfunction and an increased level of retinal oxidative

stress. These symptoms are also present in dry AMD patients.

2. Bright light-induced damages occur in a “sensitive” region of the retina, in the

superio-temporal retina, the area centralis of rodents, where the choroidal vasculature

is sparser compared to the inferior retina. Similarly, human AMD is biased toward the

macula, where the vasculature appears to have a lower perfusion rate. The area

centralis is the functional equivalent of the human macula.

44

3. Light-induced retinal injuries are regional, with a very sharp transition between the

affected and surviving retina, similar to AMD in humans, but not in retinitis pigmentosa

(RP).

4. Bright light produces the spatially central RPE-choriocapillaris ablation and BRB

breakdown in the rat eyes. Severe choriocapillaris degeneration emerges in affected

regions of the retinas at a late stage of dry AMD in patients.

In the light-induced retinal degeneration rodent model, bright light is an effective

trigger for a large-scale tissue remodelling to ultimately decimate the entire retina

(Marc et al., 2008). In addition, a distinctive advantage of the light damage model over

animal models expressing hereditary retinal degenerations is the synchronous

involvement of practically all photoreceptor death during the damage processes. This

acute retinal degeneration model therefore offers an opportunity to investigate the

temporal sequence of retinal pathology, such as cell death (Organisciak and Vaughan,

2010). The oxidative stress within RPE cells is a well-accepted component of AMD

pathogenesis (Knickelbein et al., 2015). Anti-oxidants only show protection in light-

induced retinal degeneration, but not in inherited rhodopsin mutation-induced

degeneration of P23H or S334ter rats (Kaldi et al., 2003; Organisciak and Vaughan,

2010), suggesting that the mechanism of light-induced retinal degeneration and the

pathogenesis of AMD have parallel in associating with oxidative stress. In summary,

the light-induced retinal degeneration in rodents is a useful model to mimic many of

the hallmark pathologies of AMD: focal photoreceptor cell death, RPE atrophy,

choroidal ablation, and retinal tissue remodelling (Marc et al., 2008).

1.9 Photobiomodulation (PBM) Photobiomodulation (PBM), also referred to as low-level light therapy (LLLT), is an

approach of regulating biological function by using light., wavelengths between 600

and 1000 nm (far-red to near-infrared, NIR) (Desmet et al., 2006; Hamblin, 2016). PBM

is being used in clinical applications for over 40 years (Desmet et al., 2006). Beneficial

effects of PBM have been documented on the board scope of diseases, pathological

responses, and biological dysfunction including oedema, inflammation, chronic joint

disorders, wound healing, as well as neurological disorder and pain (Chung et al., 2012).

45

Low-power lasers and light-emitting diodes (LEDs) are two well-accepted light devices

for PBM (Anders et al., 2015). LED array offers a relatively larger and flatter array of

lights to tissues without generating heat compared to low-power laser (Whelan et al.,

2001). Notably, LED has deemed as a non-significant risk and obtained for light therapy

in humans by the Food and Drug Administration (FDA) (Whelan et al., 2001; Wong-

Riley et al., 2005), while efficacy of PBM to protect against cell death is dependent on

the severity of tissue damage, such as retinal degeneration (Chu-Tan. et al., 2016).

1.9.1. History

PBM had documented that ancient Greeks and Romans maintained general health by

using sunbath as therapeutic processes (McDonagh, 2001). Between the 17th to 19th

century, previous studies showed that actinic, ultraviolet component of sunlight, is

lethal to bacteria and other microorganisms. At the end of the 18th century, Danish

physician, Niels Finsen, treated small skin blisters with red light and lupus vulgaris with

UV light. These achievements won him the Nobel Prize in Medicine and Physiology in

1903 (Geneva, 2016).

In 1968, a few years after the first working laser was invented, Hungarian scientist,

Endre Mester applied a power ruby laser (694 nm) to the shaved dorsal skin of mice to

investigate whether the light with wavelength of 694 nm caused skin cancer (Mester et

al., 1968). Unexpectedly, he found that the low-powered ruby laser accelerated hair

growth instead of causing skin cancer (Chung et al., 2012). Based on these findings,

subsequently, he started using light in the treatment of non-healing skin ulcers in

patients (Conlan et al., 1996; Geneva, 2016; McDonagh, 2001). To date, the

development of PBM into therapeutic procedures is applied in three main ways: 1) to

reduce inflammation, oedema and chronic joint disorders; 2) to promote wound healing

in deeper tissues and nerves, and 3) to treat neurological disorders and pain (Chung et

al., 2012).

In recent decades, the use of PBM has been extended to diverse clinical fields. PBM is

now used to mitigate side effects of cancer therapy by reducing chemotherapy-induced

oral mucositis (Desmet et al., 2006), in diabetic patients to reduce blood glucose levels,

and insulin resistance (Fukuoka et al., 2016), as well as to reduce peripheral neuropathy

46

(Argenta et al., 2017). Interestingly, athletes’ muscle performance is enhanced with

PBM by increasing muscle mass, decreasing inflammation, and mitigating oxidative

stress in muscle biopsies (Ferraresi et al., 2016).

1.9.2. Effective treatment: the optical window and dosage Effective tissue penetration and consequently, the best biological responses fall

between 600 and 1000nm. Outside this range, light is absorbed by molecules abundant

in the body, like water, haemoglobin, oxyhaemoglobin, melanin, etc. (Chung et al.,

2012). The principal tissue chromophores have high absorption bands at wavelength

shorter than 600nm. Water begins to absorb significantly at wavelength greater than

1150nm. For these reasons, there is an “optical window” of effective treatment in tissue

covering the wavelengths appropriately from 600 to 1000nm that maximise the

absorbance of light by alternative photo-acceptors (Figure 1.9A). PBM with

wavelengths in the range of 600–700 nm is utilised to treat superficial tissues. The lights

with longer wavelengths, in the range of 780–950 nm, are used to treat deeper tissues.

However, lights with wavelengths between 700–770 nm have limited biochemical

activity. For example, there is no beneficial effect of 728 and 770 nm light on enhanced

mitochondria activity and ATP generation in functionally inactivated neurons by toxins

(Wong-Riley et al., 2005). Therefore, lights with wavelength within 700-700 nm are

not used as PBM (Chung et al., 2012).

Biological effects of PBM are characterised by hormetic response-dependent effects,

but not classical linear dose-response pharmacological effects (Figure 1.9B). A

hormetic response, also known as U-shape or bell-shaped dose-response, is

characterized by stimulation of a biological process at low doses and inhibition of that

process at high doses (Rojas and Gonzalez-Lima, 2011). PBM with energy density from

0.01 to 10J/cm2 on tissues displays photo-stimulatory responses, while PBM with

energy density higher than 10J/cm2 shows significant photo-inhibitory biological

effects (Rojas and Gonzalez-Lima, 2011). Hormetic effects are superior to linear

threshold models in their capacity to accurately predict low-dose responses below a

pharmacological threshold and are essential for the development of PMB application in

clinical trials (Calabrese et al., 2008).

47

(A) (B)

Figure 1.9 Optical window and hormetic effects of PBM (A) An optical window of PBM is defined based on the principal tissue chromophores (haemoglobin and melanin) and water. (B) The effects of PBM on cells or tissues does not meet a classical linear dose-dependent response curve. Maximal stimulatory effects are observed at the intermediate doses (0.1-10 J/cm2). However, the linear relationship does not hold at high does. At high doses (>10 J/cm2), PBM shows inhibitory biological effects instead. Image A is reproduced with permission from (Chung et al., 2012)(License number: 4325030170530 provided by Springer Nature and Copyright Clearance Center), Copyright © Biomedical Engineering Society 2011 Image B is cited from the publication (Rojas and Gonzalez-Lima, 2011) (doi: 10.2147/EB.S21391). This is an Open Access article which permits unrestricted noncommercial use. Copyright © 2011 Rojas and Gonzalez-Lima, publisher and licensee Dove Medical Press Ltd.

1.9.3. Mechanism of action The mechanism of action of 670nm light is still being elucidated. A number of studies

supports that cells, treated with PBM, increase ATP synthesis, mitochondrial

membrane potential (Δψm), and RNA synthesis in mitochondria (Karu, 1999).

Experiments using HeLA cells examined the action spectra and absorption spectrum of

light and found that the photoacceptor molecules, that respond to PBM have different

types of Cu (I-III) centres. Such structure is present in cytochrome c oxidase (COX) in

mitochondrial transmembrane, which is a rate-limiting enzyme in terminal

phosphorylation and therefore play an important role in cellular metabolism and

survival (Figure 1.10). Therefore, a widely supported hypothesis is that COX, the most

likely primary photoacceptor to 670nm light (Giacci et al., 2014; Karu, 1999). During

mitochondrial oxidative phosphorylation, protons are transferred across the inner

48

mitochondrial membrane into the intermembrane space coupled to electron transfer at

complex I, III and IV. This is known as a proton-motive force (an electrochemical

proton gradient), which generates the mitochondrial membrane potential (Δψm) and

ATP (Kang and Pervaiz, 2012). COX is the terminal complex (complex IV) of an

electron transfer chain in mitochondria. It catalyses the final step in mitochondrial

electron transfer chain (Li et al., 2006), receiving electrons from cytochrome c

molecules and then transferring electrons to an oxygen molecule to form moles of water

(Kim, 2014) (Figure 1.10)

The upregulation of COX activity by PBM are found in vivo and in vitro models (Gkotsi

et al., 2014; Kaynezhad et al., 2016), which may be a key step to subsequently elevate

Δψm in mitochondria (Kokkinopoulos et al., 2013) and increase ATP production in

cells or tissues after PBM treatment (Gkotsi et al., 2014). In addition, the alternation of

Δψm by PBM may trigger the mitochondrial retrograde signalling pathway in cells

(Karu, 2008), regulating gene expression in nucleus (Gavish et al., 2004), which

subsequently modify cellular responses and metabolism (Schroeder et al., 2007) (Figure

1.10). The mechanism of the mitochondrial retrograde signalling pathway is still being

investigated. Emerging evidence show that retrograde signalling pathway transduces

signals, redox sensitive transcription factors, activator protein-1 (AP-1) and nuclear

factor (NF)-kB, to cell nuclei, which in turn regulate gene expressions in cells (Karu,

2008; Kim, 2014).

1.9.4. Treatment of eye disease by 670nm light PBM has been demonstrated to improve the visual acuity in patients with AMD (Merry

et al., 2017), DR, amblyopia, and RP (Geneva, 2016). Although parameters and

delivery protocols are not consistent between clinical trials, PBM offers a hope to

patients suffering from eye diseases. Laboratory studies are exploring and revealing the

underlying mechanism of 670nm light on eye diseases. 670nm light (rather than 830

nm light) effectively decreased photoreceptor death in the light-induced retinal

degeneration rat model (Albarracin and Valter, 2012; Giacci et al., 2014) or reduced

neuron death caused by toxin, methanol, or hyperglycaemia in vitro (Eells et al., 2003;

Liang et al., 2008; Tang et al., 2013; Wong-Riley et al., 2005). Furthermore, the

treatment with 670nm light mitigates inflammatory responses, known to mediate AMD

49

pathogenesis (Kowluru and Odenbach, 2004; Whitcup et al., 2013). Additionally, the

treatment with 670nm light showed the capability of mitigating the recruitment of MG/

MΦ into aged retinas (Kokkinopoulos, 2013a; Kokkinopoulos et al., 2013), inhibiting

leukostasis in retinal blood vessels in a DR rat model (Tang et al., 2013), and decreasing

pro-inflammatory cytokines and complement generation in degenerated retinas

(Kokkinopoulos et al., 2013; Rutar et al., 2012a).

According to our current knowledge, the effects of PBM on non-neuron cells, MCs

which participate in the development of AMD and other eye diseases, are still being

explored. Our laboratory previously demonstrated that treatment with 670nm light prior

to retinal degeneration can reduce MC activation and photoreceptor death in the light-

induced retinal degeneration rat model (Albarracin et al., 2011; Albarracin and Valter,

2012). Other research also reported that PBM reduced MC activation in aged retinas

(Begum et al., 2013). However, it is still unclear whether the reduction of MC activation

is directly conducted by PBM or is the consequence of the decreased photoreceptor

death.

1.10 Aims The aim of this thesis is to explore whether PBM could modulate MC gliosis, which

subsequently mitigate photoreceptor degeneration and mitigate inflammation in injured

retinas. To reach this aim, I investigated 1) the influence of 670nm light on activated

MCs; 2) the direct effects of 670nm light on MC activation; 3) the underlying cellular

mechanism of 670nm light in MCs under stress. To verify these three objectives, I

performed experiments by using in vivo and in vitro models. The results are reported in

the following three chapters. In Chapter 3, I applied 670nm light on activated MCs by

using the light-induced retinal degeneration model in rats and the in vitro uniform

scratch model. In Chapter 4, using a co-culture model of rat primary MCs, and 661W

photoreceptor cell line, I examined the direct interaction between neuron and glia in

vitro. In addition, I examined the effects of this interaction on MG activation, using rat

primary MGs. Studies in Chapter 5 explored the influence of 670nm light on cell

communication between MCs and MGs in the in vitro and in vivo pro-inflammatory

environment and, specifically, its effect on EV-mediated communication.

50

Figure 1.10 A schematic diagram presents proposed mitochondrial retrograde signalling pathways and intracellular mechanism of action of 670nm light treatment in cells Cells that absorbs energy from 670nm light show increased Δψm, which initiates retrograde signalling responses, whereby signals, such as transcription factors, are transferred back to the nucleus to regulate gene expressions. In addition, upregulation of ATP synthesis in cells with 670nm light facilitates cAMP formation and Ca2+ release, which act as second messengers to activate different metabolic pathways.

51

Chapter Two

Materials and Methods

52

2.1 Introduction The in vivo and in vitro experiments were performed to understand the effects of 670nm

red light on activated MCs and investigate the potential cellular mechanism of 670nm

light in MCs in Chapter 3 to 5. This chapter provides detailed descriptions of materials

and methods used in this thesis. Brief description of materials and methods, presented

in this chapter, and additional techniques are respectively mentioned in Chapters 3 to 5.

2.2 Animal rearing and housing conditions All procedures were performed in accordance with the Association for Research in

Vision and Ophthalmology (ARVO) statement for the Use of Animals in Ophthalmic

and Vision Research and with ethics approval from the Australian National University

(ANU) Animal Experimentation Ethics Committee (Ethics ID: A2014/56). Sprague

Dawley (SD) albino rats (aged 8-12 days, 50-60 days, or 90 to 120 days) and C57BL/6J

mice (aged 60 to 80 days) were obtained from the Australian Phenomics Facility (APF)

and housed at the John Curtin School of Medical Research (JCSMR) animal holding

facility. Animals were born and raised in dim (5 lux illumination) cyclic light conditions

with a 12h light, 12h Dark cycle. Food and water were provided in constant supply and

cages changed on a weekly basis.

2.3 Photo-oxidative damage of animal models

2.3.1 Bright light-induced photo-oxidative damage of a rat model The photo-oxidative stress (PD) caused photoreceptor apoptosis via excess oxidative

stress in the outer segment (Noell et al., 1966). The atrophy of RPE and choriocapillaris,

as well as the recruitment of MG/MΦ in damaged retinas, were followed. To establish

the PD model in rats, SD albino rats (postnatal (P) 90 – 120 days) were used. PD was

conducted according to previously described methodology (Albarracin et al., 2011;

Albarracin and Valter, 2012; Jager et al., 2008; Rutar et al., 2010). Procedures of

establishing the PD model are described as follows. Rats were transferred to the

individual transparent chambers with food placed on the cage floor and water was

provided in transparent bottles to avoid shading of the light entering the cage.

Fluorescent light tubes (COLDF2, 2x36W, IHF; Thorn Lighting, Spennymoor, UK)

were placed 200 mm above the bottom of cages, so the light intensity reached 1000 lux

53

on the cage floor (The total delivery energy: 2.35 Watts; Irradiance: 17W/m2;

Luminance: 138.4 lumens). Before exposure to light, animals were dark adapted

overnight. Exposure of bright light started at 9 AM in all groups. To encourage

increased animal activity during the light exposure period, packaging rice puffs were

dispersed. Animals were exposed to bright light for 24 hours and then returned to a low

light environment (5 lux) to recover for 1 month.

2.3.2 Progressive photo-oxidative damage in mice

Procedures of inducing the progressive PD-induced retinal degeneration in a mouse

model were followed as described in previous publications (Natoli et al., 2017a; Natoli

et al., 2016a). C57BL/6J were housed in Perspex boxes coated with a reflective interior,

and exposed to 100K lux of white light from an LED source for 5 days, with free access

to food and water. Each animal was administered with pupil dilator eye-drops twice

daily during the 5-day PD (Atropine Sulphate 1% w/v eye-drops; Bausch and Lomb,

NSW, Australia). On the 5th day of PD, mice were euthanized with CO2 prior to tissue

collection. Fresh retinas were collected for isolation of MVs ( referred to section 5.2.7).

2.4 670nm light treatment in vivo in rats To treat animals with 670nm light, animals were wrapped in a cloth, to aid manual

handling, and were placed under the near-infrared 670-nm light emitting diode array

(WARP 75, Quantum devices, Barneveld, WI) (Figure 2.1). Animals were positioned,

so that eye level was approximately 2.5 cm away from the 670nm light source for 3

minutes at 50 mW/cm2 output at eye level. This treatment protocol produces the overall

energy of ~9 J/cm2 delivered to the both eyes. Animals were treated with 670nm light

once daily for five consecutive days to account for one complete treatment. Sham

treated animals were handled identically, but the LED array was not switched on

and regarded as the control group.

54

(A) (B) Figure 2.1 Treatment SD rats with 670nm red light (A) The WARP75 appliance to conduct the 670nm light treatment. (B) A individual rat was wrapped in a cloth and eyes were position approximately 2-3 cm from the WARP75 appliance for three minutes. The overall energy output is ~9J/cm2 at the eye level during the 3-minute exposure.

2.5 Cell cultures

2.5.1 661W Photoreceptor-like cells

Murine photoreceptor-derived 661W cells were kindly gifted by Dr. Muayyad R. Al-

Ubaidi (Department of Cell Biology, University of Oklahoma Health Sciences Centre,

Oklahoma City, OK, USA). Cells for experimental purposes were used within five

passages of authentication, and validation of authenticity was performed using gene

expression of green cone pigments and cone arrestin. Cells were further validated for

species authenticity (Cellbank, Sydney, Australia). Cells were cultured in growth

cultured medium (GM), containing Dulbecco’s Modified Eagle Medium (DMEM;

Sigma-Aldrich, MO, USA) supplemented with 10% fetal bovine serum (FBS; Sigma-

Aldrich), 6mM L-glutamine (ThermoFisher Scientific, MA, USA), and 1% antibiotic-

antimycotic (100U/ml of penicillin, 100µg/ml of streptomycin and 0.25µg/ml of

Fungizone; ThermoFisher Scientific), in the dark in a humidified atmosphere of 5%

CO2 at 37°C.

2.5.2 MIO-M1 cell line MIO-M1, spontaneously immortalised human Müller cell line (Limb et al., 2002b), was

used in Chapter 3 and 5, was validated for species authenticity (Cellbank, Sydney,

Australia). MIO-M1 is identified by its characteristic morphology under phase-contrast

microscopy and by its expression of GS, GFAP, α-smooth muscle actin, vimentin,

55

CRALBP, and epidermal growth factor receptor (EGF-R) (Limb et al., 2002b). Similar

to non-immortal human MCs, MIO-M1 is capable of having the strong expression of

cytokines and growth factors, such as Ccl-2 and G-CSF (Eastlake et al., 2016). Notably,

the presence of blue opsin and melanopsin in MIO-M1 has been confirmed by using

RT-PCR and immunocytochemistry, that shown to have slow and fast transient calcium

responses to 480 nm light, but not 600 nm light (Hollborn et al., 2011). Therefore,

opsins in MIO-M1 should not be activated by 670nm light used in the current studies.

Cells were maintained in GM in a humidified atmosphere of 5% CO2 at 37°C and only

used with ten passages.

2.5.3 Mouse microglia N11 cell line Mouse microglia (N11 immortalised cell line) is an immortalized microglia cell line

derived from mouse embryonic brain cultures (Murata et al., 1997). Cells were

maintained in GM in humidified atmosphere of 5% CO2 at 37°C and passaged every 4

to 5 days before cells aggregated and the colour of GM became yellow. Cells were used

within 10 passages in all experiments. Prior to incubation with MC-derived MVs

(Chapter 5), N11 cells were placed in 6-well plates at the density of 200,000 cells per

well in GS and incubated for 24 hours at 37°C with 5% CO2. Then cells were washed

once with fresh 1X phosphate buffered saline (PBS) and then incubated in DMEM

supplemented with 1% FBS, 3% L-glutamine and 1% antibiotic-antimycotic for further

24 hours at 37°C and 5% CO2.

2.5.4 Preparation of rat primary MCs All procedures were in accordance with ARVO statement for the Use of Animals in

Ophthalmic and Vision Research and with the requirements of the Australian National

University Animal Experiments Ethics Committee. Rat retinal MCs were prepared as

described previously and with modification (Hicks and Courtois, 1990; Wang et al.,

2003). Albino SD rats (P8-12 days) were euthanized by cervical dislocation. Retinas

were collected from both eyes and soaked in DMEM containing 6mM glutamine and

1% antibiotic-antimycotic for 2 hours on the ice. The retinas were chopped into small

fragments and then incubated with DMEM containing 0.1 trypsin, 70U/ml collagenase

type 4 and 1% antibiotic-antimycotic (Life Technologies, ThermoFisher) for an hour at

37°C. Following incubation, small pieces of tissues were collected by centrifugation at

56

1200 rpm (250xg) at 4 °C for 10 minutes and cells were dissociated by triturating with

plastic serological pipettes over 20 times. Cell suspension was seeded onto T25 flask

(Nunc, ThermoFisher Scientific) in GM and then incubated in a humidified incubator

with 5% CO2 at 37°C (cells from 2~3 retinas were placed in one T25 flask). GM was

left unchanged for first 4 days and then was replenished with fresh GM of 5 ml on 5th

day. After 7-8 days, aggregation of neuronal cells, loosely attached to the bottom of the

T25 flask, was removed by forcing medium onto the flask twice. And then fresh GM

was replenished. Until cells reached 90% of confluency, cells were detached with

0.25% trypsin/EDTA (Life Technologies, ThermoFisher) and then transferred into T75

flasks (Nunc, ThermoFisher Scientific). It is regarded as the 1st passage. After reaching

confluence in the T75 flask, the cells were passaged into desired plates and then

performed for various analyses. These cells are considered as the 2nd passage. All

experiments in this thesis used cells at the 2nd passage. Phenotypic characteristics were

investigated using antibodies against vimentin (1:1200, Sigma-Aldrich), S100 β (1:400,

Sigma-Aldrich), and GS (1:100, Millipore) as MC markers. In addition, antibodies,

GFAP (1: 200, ThermoFisher) or RPE65 (1:100, Santa Cruz Biotech, Dallas, Texas,

USA) were used to exclude out the contamination of astrocytes and RPE cells (Figure

2.2).

Figure 2.2 Characteristics of rat primary MCs The purity of primary rat MC culture was assessed by using antibodies against vimentin (A), S100β (B), GS (C), RPE65 (D), and GFAP (E). Almost all cells in MC culture showed strong expression of typical MC markers, vimentin and S100β, but only a few cells manifested weak RPE65 and GFAP staining, indicating that the high purity of primary MC culture. Moreover, primary MCs expressed GS, an enzyme to convert the toxic glutamate to non-toxic glutamine.

57

2.5.5 Preparation of rat primary MGs/MΦ Adult rats (P50-60 days) were euthanized. Retinas immediately were captured from

whole eye globes and then transferred into the cold Hank’s balanced salt solution

(HBSS) containing 5mM MgCl2. The solution was removed and retinas were cut into

small fragments. Tissues were digested with the digestion buffer containing the papain

suspension (2mg/ml; Worthington; Lakewood, NJ, USA), L-Cystine (5.5mM; Sigma-

Aldrich), Gentamycin (10µg/ml; Sigma-Aldrich), DNase I (200U/ml; Roche, Basel,

Switzerland), superoxide dismutase (SOD) (5ug/ml; Worthington), Catalase (5ug/ml;

Sigma-Aldrich) and MgCl2 (5mM; Sigma-Aldrich) in HBSS (no calcium, no

magnesium and no phenol red) (Life Technologies) and incubated at 7°C for 45 minutes

and then at 37°C for 7 minutes. The digested tissues were dissociated with vigorous

trituration 10 times with pipettes and then were centrifuged at 1350 rpm (150xg) for 5

minutes at 4°C. The supernatant was removed and the cell pellet was resuspended in

the neutralization buffer containing antipain (50ug/ml; Roche), DNase I (200U/ml;

Roche), SOD (5ug/ml; Worthington), Gentamycin (10ug/ml; Sigma-Aldrich), Catalase

(5ug/ml; Sigma-Aldrich), MgCl2 (5mM; Sigma-Aldrich) and 4% Bovine Serum

Albumin (BSA, Sigma-Aldrich) in HBSS for 10 minutes at 4°C. Cells were spun down

at 1350 rpm for 5 minutes at 4°C and then washed with the staining buffer containing

1% BSA in HBSS. Cells were incubated in the staining buffer containing an anti-rat

CD11b antibody conjugated to PE (1:500, BioLegend, San Deigo, CA, USA) for 30

minutes at 4°C in the dark and then were centrifuged at 1350 rpm for 5 minutes at 4°C.

Cell pellets were washed with the washing buffer containing 1%BSA and

ethylenediaminetetraacetic acid (EDTA) (2mM; Sigma-Aldrich) in HBSS once and

then centrifuged at 1350 rpm at 4°c for 5 minutes. Cell pellets were resuspended in the

staining buffer and run through a fluorescence-activated cell sorter (FACS) (BD

FACSAria II; BD Biosciences, Franklin Lake, NJ, USA) at 4°C at the NEI Flow

Cytometry Core Facility. Approximately 4000-8000 CD11b+ MGs were obtained from

both retinas of each experimental rat in a well of a 24-well-plate. Cells were cultured

in DMEM culture medium containing mouse granulocyte-macrophage colony-

stimulating factor (GM-CSF) (1ng/ml, Vancouver, BC, Canada) in dark and humidified

atmosphere of 5% CO2 at 37°C. Fresh medium was replaced every 3-4 days until cells

reached 80% confluency.

58

2.5.6.The treatment with 670nm light in vitro 670-nm LED array (WARP 10, Quantum devices, Barneveld, WI) (Figure 2.3) was

applied to MIO-M1(Chapter 3&5) and rat primary MCs (Chapter 4). A firm cardboard

box sustains the device WARP 10. The small opening was made on the topside of the

cardboard box where 670nm light emission passed (Figure 2.3). The 670-nm LED array

is approximately placed 4.5 cm above the bottom of cell culture plates where cells were

placed. The 670nm light irradiated cell culture plates containing cells for 3 minutes in

one session. The energy delivered to cells is 60mW/cm2 (~9J/cm2) in each session. Cells

received three sessions one day and the interval of each session is 4 to 5 hours. The

sham treated control group was placed underneath the WARP 10 device, but light

source of the device was switched off. The detailed protocol and timelines are presented

in Chapter 3, 4 or 5.

(A) (B) Figure 2.3 The WARP 10 appliance was applied to in vitro studies (A) WARP 10 appliance. (B) Cell culture plates were placed underneath a cardboard box positioning a WARP 10 device 4.5 cm above the culture plates.

2.6 Tissue collection and fixation Animals were euthanized by administering an intraperitoneal injection of barbiturate

(pentobarbitone sodium) overdose ( >60 mg/kg body weight, Lethabarb Euthanasia

Injection,Valabarb; Virbac, Carros, France). The retina, lens, and a part of vitreous

body of the right eye of each animal were harvest by forceps through a corneal incision

and then placed in 200µl of RNAlater solution (Ambion Biosystems, Austin, TX, USA).

Later, lens and vitreous body were gradually separated from the retina while the tissues

were stored in the RNAlater solution. Samples were stored at 4°C overnight then

59

transferred to -80°C. The left eye of each animal was collected for histological

assessments. For each left eye, the superior aspect of eyeballs was marked with a black

or blue permanent marker prior to enucleation for the orientation. The eyeball was

enucleated by using forceps and iris scissors and then was injected with fresh 4%

paraformaldehyde using a 1ml syringe with 27gauge needle through the limbus. The

eyeball was then immersed in 4% paraformaldehyde for 3 hours at 4°C. Then the

eyeball was washed with cold 0.1M PBS and then stored in 15% sucrose made up in

0.1M PBS at 4°C over 24 hours.

2.7 Histological Processing

- Cryopreservation, Embedding and Sectioning After 24 hours of immersing eyeballs in 15% sucrose, the eyeballs were placed into

plastic moulds filled with Tissue Tek Optimal cutting temperature (OCT) mounting

medium (Tissue-Tek®, Sakura Finetechnical, Japan). For orientating, the superior

marked side on the eyeball was placed to parallel to a wall of a mould, which is opposite

to tags on the mould. A mould was then snap frozen in a solution of Acetone (Merk

Millipore, Bayswater, VIC, Australia), which was cooled by dry ice. Embedded blocks

containing eyeballs were stored at -20°C until required. Embedded eyes were

cryosectioned at the 16µm thickness in a sagittal plane, to allow a dorsal to ventral

observation of the retinas using Leica CM1850 Cryostat (Leica Microsystems,

Nossloch, Germany). Eye sections were mounted directly onto SuperFrost®- Plus

slides (Menzel-Glaser) coated in gelatin (Sigma-Aldrich) and poly-L-lysine (Sigma-

Aldrich). To enable compare changes of the histological structure across different

animals and groups, only tissue sections containing the optic nerve (ON) head were

used for all experiments in this thesis. CryoSections were stored at -20°C until required.

2.8 Histological Techniques

2.8.1 Measurement of retinal thickness To quantify photoreceptors survival, DNA-specific dye bisbenzimide (BBZ) (Sigma-

Aldrich) was used to counterstain the nuclei. There are two measurements to quantify

the viability of photoreceptors in this thesis. The first approach is the measurement of

ONL thickness. To remove OCT mounting medium, the retinal cryosections were

60

thawed at room temperature (RT) for 15 minutes and then sequentially washed with

70% ethanol for 10 minutes. The sections were then rehydrated by sequentially

immersing in distilled water for 10 minutes and 0.1M PBS for 10 minutes. A

hydrophobic barrier (PAP pen) (ProSciTech, Kirwan, QLD, Australia) was drawn

along the edge of tissue sections to avoid leakage of reagents. The tissue sections were

incubated with BBZ in a 1:10000 dilution for 10 minutes, followed by a final wash in

0.1M PBS for 5 minutes and covered with coverslips by using Aqua-Poly/Mount

aqueous mounting medium. The retinal ONL thickness measurement was performed

on digital images of the stained tissue sections. Each animal had two representative

retinal sections. Digitised images of each stained retinal section were captured with a

NIKON A1 Confocal Microscope (Zeiss, NSW, Australia) at 10X magnification

(Figure 2.4A). Ten specific fields were measured at increments of 1 mm along the full

length of an retinal section (Figure 2.4A) using AxioVision software (Version 4.6; Zeiss,

NSW, Australia). The relative location of each field along the length of superior or

inferior retina is expressed as a distance (mm) eccentricity from the ON head. The

hotspot of the rat retina is located at 2mm eccentricity from the ON head in the superior

retina. In each field of one retinal section, 5 measurements of the ONL thickness and

corresponding the outer member thickness (GCL-OLM) were obtained (Figure 2.4B).

The ratio of ONL length (red line) to the GCL-OLM length (green line) was used, to

account for obliquely cut sections. Five measurements in each field are averaged for

each animal.

2.8.2 Counting of the number of rows of photoreceptor nuclei The second method to assess photoreceptor survivability is to count the number of rows

of photoreceptor cell bodies in ONL. Retinal sections were stained with BBZ as the

mention in section 2.8.1. Each animal has two representative retinal sections. The ten

fields along the superior and inferior retina of each retinal section were captured (Figure

2.4A). Details are referred to section 2.8.1.

61

A) (B)

Figure 2.4 Illustration of ONL thickness measurement

(A) The ONL thickness was measured at increments of 1 mm along the full length of the retina. The five fields were measured along the superior or inferior retina. The numbers (green) denote the distance (mm) away from the ON head along the superior or inferior retina. (B) The ratio of the ONL length (red line) to the GCL-OLM length (green line) is an indication of photoreceptor survivability. ON: optic nerve GCL: ganglion cell layer; INL: inner nuclear layer; OLM: outer limiting membrane; ONL: outer nuclear layer. Six counts were made for one field along the retina and then were averaged. Each

experimental animal has two representative retinal sections. In total, 12 counts were

made in each field along the retina for one experimental animal. The data is presented

as an average of the number of rows of photoreceptor nuclei in each field along the

retina in one experimental group.

2.8.3 Immunohistochemistry (IHC) of retinal sections The steps of performing retinal section rehydration was mentioned above. To prevent

the leakage of reagents, a hydrophobic barrier (PAP pen) was drawn along the edge of

the retinal sections. The 0.2% triton X-100 (Sigma-Aldrich) was pipetted on sections

for 5 minutes at RT and then washed out with 0.1M PBS twice. Tissue sections were

immersed in 10% normal goat serum (NGS) for 2 hours at RT and then incubated with

diverse antibodies diluted in 1% NGS overnight at 4°C in a humidified chamber (the

62

information of primary antibodies and optimal dilution were shown in (Table 2.1). Next

day, tissue sections were washed in 0.1M PBS twice and then incubated with the

appropriate secondary antibody in 1% NGS at RT for 90 minutes in the dark (Table

2.4). Sections were then incubated with DNA-specific binding BBZ (1:10000) for 10

minutes and covered with coverslips in Aqua-Poly/Mount aqueous mounting medium.

Immunofluorescence was visualized using A1 Nikon Confocal Microscope (Nikon,

NSW, Australia), and acquired using NIS-Element AR software (Nikon). The

brightness and contrast of images were adjusted for publication using Photoshop CS6

software (Adobe Systems, CA, USA), which was standardised among images.

2.9 The immunocytochemistry (ICC) staining

MIO-M1 and rat primary MCs were plated in desired plates or chamber slides at the

optimal density (details are referred to section 3.2.9 and 4.2.8). After different

treatments, cells were fixed with 2% paraformaldehyde for 30 minutes at RT and then

washed with 0.1 M PBS twice. Cells were permeabilised with 0.02 % triton X-100 for

5 minutes and then washed with 0.1 M PBS three times. 10% NGS (Sigma-Aldrich)

was used to cover non-specific binding for 2 hours at RT. Cells were incubated with

the primary antibodies diluted in 1% NGS at 4°C in a humidified environment

overnight. The information of primary antibodies applied on rat primary MCs was

shown in or performed in MIO-M1 was presented in Table 2.2 & 2.3. The secondary

antibodies conjugated with Alexa-flour®-488 or 594 (Table 2.4) diluted in 1% NGS

were added onto cells and then incubated at RT in the dark for 90 minutes. The

secondary antibodies were then washed away with 0.1M PBS three times. In some

conditions, following the incubation of the primary antibody, cells were incubated with

the secondary antibody, anti-mouse IgG (H+L) conjugated with biotin (SAB3701153;

Sigma-Aldrich) (Table 2.4) in 1% NGS at RT in the dark for 90 minutes, followed by

the incubation with the streptavidin conjugated with Alexa-flour®-488 (S32453;

ThermoFisher) diluted in 1% NGS at RT in the dark environment for further 90

minutes. For visualisation of cell nuclei, cell nuclei were counterstained with BBZ

(Sigma-Aldrich) for 10 minutes. Cells were covered using the aqueous mounting

medium. Images were visualized using confocal microscopy (A1 Nikon Confocal

Microscope, Nikon, NSW, Australia) and captured by NIS-Element AR software. The

brightness and contrast of images were adjusted for publication using Photoshop CS6

software (Adobe Systems, CA, USA), which was standardised among images.

63

Table 2.1 Primary antibodies for in vivo studies in rats

Table 2.2 Primary antibodies for in vitro studies in rat primary MCs

Primary antibody Source Catalogue

number Host

species

Isotype Epitope Positive control

Dilution factor

ED-1 Merc-Millipore MAB1435 Mouse IgG1

Macrophage, Monocytes, Lymphoid

organs

1:200

GFAP Dako Z0334 Rabbit Astrocytes in the brain 1:400

Glutamine Synthetase

Merc-Millipore MAB302 Mouse IgG2a Rat brain

tissue 1:500

Iba1 Wako 019-19741 Rabbit C-terminus 1:500

Vimentin Sigma-Aldrich V6630 Mouse IgG1 1:150

Primary antibody Source

Catalogue

number

Host species Isotype Epitope Positive

control Dilution factor

COX5a Abcam Ab110262 Mouse IgG2a

Rat hearts, Mouse hearts,

Hela cells

1:500-1:1000

GFAP ThermoFisher Scientific MS1376 Mouse IgG1

IMR-5 cells, Brain.

Astrocytoma 1:150

Glutamine Synthetase

Merc-Millipore MAB302 Mouse IgG2a Rat brain

tissue 1:100

RPE65 Santa Cruz SC-32893 Rabbit IgG Amino acids

256-340

Mouse eye extract

1:100

S100 β Sigma-Aldrich S2532 Mouse IgG1 β-chain 1:200

Vimentin Sigma-Aldrich V6630 Mouse IgG1 1:1200

64

Table 2.3 Primary antibodies were used in in vitro studies in MIO-M1

Table 2.4 Details and information of secondary antibodies

Primary antibody Source Catalogue

number Host

species Isotype Epitope Positive control

Dilution factor

FGF-2 Merc-Millipore 05-118 Mouse IgG1κ Fetal liver or

kidney tissue 1:200

GFAP ThermoFisher Scientific MS1376 Mouse IgG1

IMR-5 cells, Brain.

Astrocytoma 1:200

Ki67 Merc-Millipore AB9260 Rabbit

Breast carcinoma, lymph node

tissue, normal tonsil

1:300

Nestin ThermoFisher Scientific MA1-110 Mouse IgG1κ 1:400

Vimentin Sigma-Aldrich V6630 Mouse IgG1 1:500

Secondary antibody Source Catalogue number Host species Dilution

factor

Anti-mouse IgG (H+L) Alex-488

ThermoFisher Scientific A11029 Goat 1:500

Anti-rabbit IgG (H+L) Alex-488

ThermoFisher Scientific A11034 Goat 1:500

Anti-mouse IgG (H+L) Alex-594

ThermoFisher Scientific A11032 Goat 1:500

Anti-rabbit IgG (H+L) Alex-594

ThermoFisher Scientific A11037 Goat 1:500

Anti-mouse IgG (H+L) with biotin Sigma-Aldrich SAB3701153 Donkey 1:500

65

2.10 Gene expression

2.10.1 RNA extraction from tissues The RNA extraction form the rat retinas was conducted using TRIzol (TRIzol reagent;

ThermoFisher Scientific) and RNAqueous® kit (Ambion, ThermoFisher Scientific).

The retinas were collected and stored as mentioned in section 2.6. Retinas then were

thawed and then RNAlater solution was removed from the retinas. TRIzol was

conducted in RNA extraction. Retinas were homogenised with 200µl of TRIzol on the

ice and then added with additional 800µl of TRIzol. After gently vortexing for 20

seconds, the homogenized retinas in TRIzol reagent were placed at RT for further 10-

15 minutes. To separate RNAs from DNAs, proteins, and cell debris, the homogenized

retinas were mixed well with 200µl of chloroform (Chem-Supply, Adelaide, Australia)

(1/4 volume of TRIzol) and then the mixture was centrifuged at 13,000rpm for 10

minutes at 4°C. An aqueous layer was transferred into a clean 1.5 ml tube and then

added with an equal volume of 64% ethanol. The reagent was mixed well by vortex and

then passed through the filter. The purified and DNase-treated RNA was collected using

50µl of elution solution (heated at 75 °C) and then stored in -80°C until the beginning

of cDNA synthesis. All procedure was followed the manufacture’s protocol

(RNAqueous® total RNA isolation kit protocol, Ambion™, ThermoFisher). Purified

RNA was quantified using the ND-1000 (Nanodrop Technologies, Wilmington, DE)

spectrophotometer. To ensure the RNA purity and quality, only RNA samples with an

A260/A280 ratio above 1.8 were used for the following study. The RNA samples were

stored at -80 °C until the further analyses.

2.10.2 RNA extraction from cells Cells were washed with cold 0.1M PBS once, and 1ml of the TRIzol (TRIzol reagent;

Invitrogen; ThermoFisher Scientific) was added onto cells. The 200µl of Chloroform

was mixed with 1 ml of the TRIzol solution containing cells and then shook by hands

for 30 seconds. The mixed reagent was placed at RT for 15 minutes and then centrifuged

at 13,000xg for 10 minutes at 4°C. The aqueous layer was recovered and then

transferred into a clean 1.5 ml tube. The recovered solution was added with half volume

of 100% ethanol and then mixed well by the vortex. The solution was then passed

through the filter. After that, the filter was washed sequentially with the Wash Solution

66

1 once and the Wash Solution 2/3 twice. The purified and DNase-treated RNA was

collected using 20-25µl of elution solution (heated at 75 °C) and then stored in -80°C.

All procedure was followed the manufacture’s protocol. (RNAqueous®- Micro total

RNA isolation kit protocol; Ambion-Applied Biosystem, ThermoFisher). To ensure the

RNA purity and quality, only RNA samples with an A260/A280 ratio above 1.8 were

used for the following study.

2.10.3 Complementary DNA synthesis The synthesis of complementary DNA (cDNA) was achieved by revere transcribing the

500ng of the total RNA, according to the prescribed manufacturer’s protocol (Tetro

cDNA Synthesis Kit, Bioline, London, UK). The manufacturer’s direction is described

as follows: Purified RNA was primed in a sterile 0.2ml tube by mixing 500 ng RNA,

made up to 12µl with UltraPureTM DNase/RNase-Free Distilled water (ThermoFisher).

Each sample was mixed with 1µl Oligo (dT)18 primers (500µg/ml), and 1µl dNTPs

(10mM). The final volume of the mixture is 14µl. The mixture was incubated at 70°C

for 5 minutes and placed on ice for 5 minutes. Then samples were centrifuged and added

with 6µl of reverse transcription mix containing 200U Tetro Reverse Transcriptase, 4µl

5x RT buffer and 10U Ribosafe RNase Inhibitor. After that, the tubes were placed into

the VERITI® PCR machine (Applied Biosystems). The cycling program was run as

follows: 30 minutes at 45°C for the reaction to occur, followed by 5 minutes at 85°C to

terminate the reaction and hold at 4°C.The final product of cDNA was stored at -20°C.

2.10.4 Quantitative real time polymerase chain reaction (qRT-PCR)

For measuring the fold change in gene expression, gene expressions were determined

by qRT-PCR using gene analysis probes. The program for qRT-PCR is listed in Table

2.5. Taqman probes are shown in Table 2.6, 2.7, and 2.8.

The fold change in the gene expression was compared to samples in the control group

and determined by the comparative the critical cycle threshold (Ct) method (∆∆Ct). The

calculation and equation are as follows:

Normalised Target Gene expression level = 2(-ΔΔCt)

Where, ΔCt = Ct (Target Gene) – Ct (Endogenous control Gene)

ΔΔCt = ΔCt (Experimental sample) – ΔCt (Control sample)

67

Table 2.5 QRT-PCR program

* For establishing the Ct cycle, the fluorescence level was automatically measured at the completion of each cycle.

Table 2.6 TaqMan® probes used in rat retinas

Table 2.7 TaqMan® probes for rat primary MCs and MGs

Step Temperature(°C) Duration Cycles UDG incubation 50 2 minutes 1X AmpliTaq Gold, UP Enzyme activation

95 10 minutes 1X

Denature 95 15 seconds 40X Anneal and Extend* 60 1 minutes

68

Table 2.8 TaqMan® probes used in mouse MG cell line N11

2.11 Western blotting Protein extraction was performed using the CellyticTM buffer (Sigma-Aldrich)

containing protease inhibitor cocktail (Sigma-Aldrich). After the different treatments,

cells were washed with chilled 0.1M PBS once and then 60-80µl of the CellyticTM

buffer was added onto cells. Cells were scraped by the wide opening end of a 200µl

pipette and then the cell lytic buffer containing protein and cellular debris was

transferred to a 1.5-ml vial. Vials were placed on ice for 20-30 minutes and then

centrifuged at 14,000 rpm (16400xg) for 15 minutes at 4°C to remove cellular debris.

The protein-containing supernatant was transferred to a clean vial and then stored in -

80°C until the onset of Western blotting. The concentration of protein in each sample

was determined by Bradford assay (Bio-Rad, California, USA). Protein (20µg) in each

sample was mixed with an equal volume of 2X loading dye (Laemmili Sample Buffer,

# 161-0737, Bio-Rad) and then denatured at 95 °C for 5 minutes. Denatured samples

were loaded onto 4% to 20% Mini-Protean TGX Precast Protein gel (#4561093, Bio-

Rad) in chilled 1X TGS (Tris/ Glycine/ SDS, # 161-0732, Bio-Rad) buffer to separate

proteins.

Protein in the gel was transferred to nitrocellulose (NC) membrane in chilled transfer

buffer (1X TGS buffer containing 20% Methanol). After transferring, NC membrane

was blocked with 5 % non-fat milk for 1 hour at RT (or 3% BSA for 3 hours) prior to

incubating with the primary antibody (COX5a) (1: 500-1:1000, #ab110262, Abcam,)

or GAPDH (1:2000, #9549, Sigma-Aldrich) in PBS overnight. The membrane was

washed three times with PBST, containing 0.1% tween 20 and then soaked in the

secondary antibody (1:3000, goat anti-rabbit conjugated with HRP [# 170-6515, Bio-

Rad] or goat anti-mouse conjugated with HRP [# 170-6516, Bio-Rad]) in PBS for 2

hours at RT. Then the membrane was washed three times with PBST and then soaked

69

in the chemiluminescence using Clarity Western ECL kit (# 1705060, Bio-Rad) for 1

minute. The protein was visualised with the detection of chemiluminescence and

images were captured and analysis by using Chemidoc MP (Bio-Rad). The expression

of COX5a in each sample was respectively normalised to the expression of matching

loading control (GAPDH) (Begum et al., 2013; Walker and Steinle, 2007).

2.12 Assessment of cell viability and cell toxicity For measuring cell viability, an ATPlite 1 step assay (PerkinElmer, MA, USA) and an

MTT assay (3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide; Sigma-

Aldrich) were used according to the manufacturer’s instructions. In the ATPlite 1 step

assay, the luminescence of samples was detected using the TECAN Infinite 200 PRO

plate reader (TECAN, Männedorf, Switzerland), in which the value of luminescence is

proportional to ATP levels. For the MTT assay, all procedures were performed in the

dark environment. Cells were incubated with MTT reagent for 4 hours in a humidified

atmosphere of 5% CO2 at 37°C. After 4 hours, the stop solution (MTT solubilisation

solution) was added onto cells, incubating for a further 24 hours. The optical density of

supernatant was measured at 570nm using a TECAN plate reader. A CellTox assay

(Promega, Madison, WI, USA) was performed to measure cell death according to

previously described methods (Natoli et al., 2016b). Briefly, culture medium were

removed from each well and detection solution was added onto cells for 20 minutes at

room temperature. Fluorescent cells (dead/dying cells) were counted using

488nmEX/525nmEM under an Axiovert 200 microscope (Zeiss, West Germany). The

number of fluorescent cells in 2-3 separate fields of each well was quantified and

averaged. The view of each field is 2mm in diameter.

2.13 Statistical analysis Data were analysed with one-way ANOVA and post-hoc comparisons Tukey’s

Multiple Comparisons test. Data were reported as mean ± SEM. Values of P< 0.05

were considered to be statistically significant. All statistical analyses were performed

using GraphPad Prism 6.0 (GraphPad software, San Diego, CA, USA).

70

Chapter Three

Light treatment with 670nm mitigates Müller cell gliosis following retinal injury: evidence from in vitro

and in vivo models of stress The results in this chapter is accepted in January 2018 and publish in Experimental Eye Research (Volume 169; Page 1-12; 2018).

71

3.1 Introduction MCs are the principal macroglia of the vertebrate retina and as such, they have a key

role in maintaining retinal structure and effective function. Cell bodies of MCs are

located in the INL (Bringmann et al., 2006; Bringmann and Wiedemann, 2012) and

their processes traverse the entire retina. End-feet of MC processes forms the inner and

outer limiting membrane therefore contributes to BRB (Albarracin and Valter, 2012;

Omri et al., 2010; Tout et al., 1993). Through their intimate relationship with retinal

neurons, MCs provide essential support and maintenance of homeostasis and

metabolism by buffering potassium, recycling of excess glutamate, and regulating

water levels within the retina (Bringmann et al., 2006; Bringmann and Wiedemann,

2012). Moreover, MCs indirectly participate in vision formation by acting as guidance

to minimise light scattering within the retinal layers, increasing visual sensitivity and

contrast (details are referred to section 1.5.3) (Reichenbach and Bringmann, 2013).

MCs become reactive upon retinal injury, or during degeneration (Bringmann et al.,

2009). Initially, this gliosis aims to protect the retina, however severe injury or chronic

stress leads to progressive MC gliosis, exacerbating retinal injuries (Bringmann et al.,

2009). The early response of MCs in retinal stress involves the release of

neuroprotective factors such as CNTF and FGF-2 (Bringmann and Wiedemann, 2012;

Shen et al., 2012) that have been shown to rescue photoreceptors in the light-induced

retinal degeneration (Albarracin and Valter, 2012). Activated MCs express Ccl2, a

known chemoattractant and activator for monocytes and MGs in vitro (Matsushima et

al., 1989; Nakazawa et al., 2007a; Yoshimura et al., 1989). In in vivo models, exposure

to damaging levels of white light induces Ccl2 expression in MCs, leading to MΦ

recruitment to the areas of severe damage (Rutar et al., 2012b). In chronic or severe

acute retinal damage when a large number of neurons are lost, MCs enter into

proliferative gliosis (Bringmann et al., 2009), forming glial scars that hinder nutrient

delivery to surviving retinal neurons, causing further cell death thereby leading to a

progression of the disease (Albarracin and Valter, 2012).

PBM is a low-energy photo irradiation that has been shown to accelerate wound healing

on skin (Conlan et al., 1996), mucosa (Desmet et al., 2006), and soft tissue wound

healing (Herranz-Aparicio et al., 2013), and had recently gained FDA approval for

72

clinical use (Desmet et al., 2006; Whelan et al., 2001). Beneficial effects have also been

reported in central nervous tissue injury models, such as spinal cord injury, retina, and

optic nerve damages caused by partial optic nerve transection (Albarracin et al., 2011;

Fitzgerald et al., 2010; Hu et al., 2016). Preclinical studies demonstrated that 670nm

light treatment is neuroprotective, preventing or slowing photoreceptor or ganglion cell

death (Albarracin et al., 2011; Albarracin et al., 2013; Eells et al., 2003; Giacci et al.,

2014; Natoli et al., 2013; Tang et al., 2013), and mitigating the secretion of

inflammatory cytokines, the recruitment of MΦs, and the activation of the complement

system in retinal degeneration models (Calaza et al., 2015; Kokkinopoulos et al., 2013).

Clinical studies have conducted in the past decade and substantiated earlier claims of

its benefits in the retina (Ivandic and Ivandic, 2008, 2012; Tang et al., 2014). Most

notably, PBM has been shown to reduce retinal oedema clinically, suggesting its effect

on mitigating inflammation (Ivandic and Ivandic, 2008). The underlying mechanism of

670nm light is thought to be related to the mitochondrial enzyme COX, which is a rate-

limiting enzyme in terminal phosphorylation in the mitochondrial respiratory chain and

has been posited as the most likely primary photoacceptor for 670nm light (Karu, 1999;

Karu, 2008).

Although a growing body of research has investigated the effects of 670nm light on

retinal neurons, its influence on non-neuronal cells, including MCs, have not been

studied extensively. Previously, our publication demonstrated that pre-injury treatment

with 670nm light can ameliorate the activation of MCs following exposure to damaging

bright white light in albino rats (Albarracin et al., 2011; Albarracin and Valter, 2012).

The aim of this study is to explore the direct effect of 670nm light on activated MCs

using in vitro and in vivo models of stress.

3.2 Materials and methods

3.2.1 Experimental groups of animals All procedures and conditions of animal rearing and housing are mentioned in section

2.2. Adult albino SD rats (P100-120) (n=39) were used in this chapter. PD was

performed as described in section 2.3.1. Some animals were exposed to bright light

(1000 lux) for 24 hours and then returned to the dim environment (5 lux) to recover for

1 month. During this recovery period, some of these animals were treated with 670nm

73

light for 3 minutes daily for 5 consecutive days, at different time points after PD (details

are described below). The procedures and details of the treatments of 670nm light refer

to section 2.4. Sham treated animals were handled identically, but the LED array was

not switched on. Animals were divided into 5 groups. Information and details are as

follows,

SD rats were

Control group (N=6):

In this group, animals were raised in the dim environment (5 lux) for 30 days and not

exposed to either PD or 670nm red light.

PD only group (N=9):

The animals were exposed to PD and then returned to the dim environment (5 lux) for

30 days. The rat was individually wrapped in a towel and kept under an unswitched the

670nm red light LED array box for 3 minutes. This sham treatment was conducted once

per day for 5 consecutive days.

PD+R0 group (N=7):

The rats of this group received PD and then were treated with 670nm light immediately

after PD for 3 minutes daily for 5 consecutive days. Aside from the treatment with

670nm light, animals were raised in the low light circumstance (5 lux) for 30 days post

PD.

PD+R3 group (N=10):

Animals underwent PD and then stayed in the dim environment (5 lux) for 30 days.

Animal received the treatment with 670nm light for 3 minutes daily for continuous 5

days commencing at 3 days post PD.

PD+R14 group (N=7):

Rats were exposed to PD for 24 hours and then returned to the dim environment (5 lux)

for 30 days. The treatment with 670nm light was performed on PD rats for 3 minutes

every day for 5 days starting at 14 days post PD.

74

All samples were collected at 30 days after PD. Samples were accessed with the

following methods/techniques. The paradigm of the treatment of 670nm light illustrates

in Figure 3.1.

Figure 3.1 Diagram of the rat PD model and the paradigms of the 670nm treatment.

3.2.2 Tissue collection One eye of each animal was enucleated, immersion fixed in 4% paraformaldehyde, and

prepared for histological assessment as described in section 2.6. The retina of the other

eye was collected and immersed in RNALater, stored at 4 oC overnight, then transferred

to -80oC for storage for molecular analysis as described in section 2.10.

3.2.3 Immunohistochemistry (IHC) of retinal sections The preparation of cryosections and the procedures of IHC were conducted as described

in the section 2.8. The detailed information of antibodies refers to Table 2.1, Table 2.3,

and Table 2.4. The brief information of primary antibodies used in this chapter is listed

in Table 3.1. The secondary antibodies (anti-mouse IgG-AlexaFluor or anti-rabbit IgG-

AlexaFluor) were diluted in 1% NGS to incubated with retinal sections following

incubation of primary antibodies. Immunofuorescence was visualized using A1 Nikon

Confocal Microscope (Nikon, NSW, Australia), and acquired using NIS-Element AR

75

software (Nikon). The brightness and contrast of images were adjusted for publication

using Photoshop CS6 software (Adobe Systems, CA, USA), which was standardised

among images.

Table 3.1 Primary antibodies used for experiments in vivo and in vitro

Primary

antibody Source

Catalogue

number

Host

species

Dilution

factor

FGF-2 Millipore 05-118 Mouse 1:200

GFAP (retinas) Dako Z0334 Rabbit 1:400

GFAP (cells) ThermoFisher

Scientific MS1376 Mouse 1:200

Iba1 Wako 019-19741 Rabbit 1:500

Ki67 Millipore AB9260 Rabbit 1:300

Vimentin Sigma-Aldrich V6630 Mouse 1:150

3.2.4 Photoreceptor cell death, and survival analysis ONL thickness and the number of rows of photoreceptor nuclei were quantified. The

detailed protocols have been published previously (Albarracin and Valter, 2012; Natoli

et al., 2017a) and described in section 2.7 and section 2.8 in this thesis.

3.2.5 RNA isolation and real-time quantitative polymerase chain reaction Retinas were collected and stored in RNA stabiliser (RNAlater: Ambion-Applied

Biosystems, Foster City, CA) overnight at 4 °C and then stored in -80 °C until the onset

of RNA extraction. The protocols of RNA extraction is described in detail in section

2.10. For validating of the fold change in gene expression, target genes were determined

by qRT-PCR using gene analysis probes. Probes were shown in Table 2.6. The

procedures of qRT-PCR are mentioned in section 2.10. The fold change in the gene

expression was compared to samples in the control group (mentioned in section 3.2.1)

76

and determined by the comparative cycle threshold method (∆∆Ct). To account for the

variability, the amplification of each sample was performed in duplicate.

3.2.6 Maintenance of MIO-M1 Cells MIO-M1, spontaneously immortalized human Müller cell line (Limb et al., 2002b) was

used within 10 passages in this chapter. Details are mentioned in the section 2.5.2.

3.2.7 Estimating MC gliosis by using the in vitro uniform scratch model To assess MC activation in vitro, we used a well-established scratch model using the

MIO-M1 cell line (Romo et al., 2011). An in vitro scratch assay has been reported as

a convenient and inexpensive method for analysis of cell migration (Liang et al., 2007).

Cells were seeded at a density of 13X104 per well in a 6-well plate at 37 °C in a

humidified CO2 incubator for two days to reach 90% confluence. A uniform scratch

(cross configuration) was then made using a 1 ml pipette tip (Figure 3.2). A grid beneath

each well was used to assure a uniform scratch-wound for all wells. After scratching,

floating cells and debris were removed by washing wells with 0.1M PBS twice. Then

wells were replenished with fresh GM and incubated at 37 °C in a humidified CO2

incubator for another 72 hours. The wound edges were traced by hand (dashed lines in

Figure 3.3), and digital images of the wound edges were captured in the same location

at T=0, 24, 48 and 72 hours after the scratch using a phase-contrast light microscope

Zeiss Axiovert 3 (Zeiss). The experiment was repeated three times on separate

occasions (N=9 in total). To analyse cell migration, the area of cell coverage (the area

with black oblique lines in Figure 3.3) was calculated by determining the number of

pixels with using Image J software (Rasband, W.S., ImageJ, U. S. National Institutes

of Health, Bethesda, Maryland, USA). Data is presented as change in pixels from T=0

at each time point.

77

Figure 3.2 Diagram of the uniform scratch in MIO-M1 cells (A-B) MIO-M1 cells reached 90% confluence in a well. (C) A cross configuration of a scratch was made on the confluency cell layer. (D) The distribution of cells on the wound edges.

Figure 3.3 Diagram of the measurement of the area of cell coverage (A) Wound edges of the scratch at T = 0 (black dashed line). (B) Image was taken at the identical regions and the new edges of the scratch were demarcated at T=72 hours (blue dashed line). (C) The edges of the scratch at T=0 and T=72 hours were merged by using ImageJ software. (D) Areas between edges (oblique black lines) were measures in pixels as the value of cell coverage area.

3.2.8 670nm red light treatment paradigm and experimental groups For the irradiation of cells, the 670-nm LED array (WARP 10; Quantum Devices)

(Figure 2.3 ) was used. Cells were irradiated 3 times (sessions) with 670nm light within

first 10 hours after the scratch-injury. The energy of 670nm light delivered is

60mW/cm2 (~9J/cm2) every session (details are referred to section 2.3). The first

treatment was delivered immediately after the scratch followed by two further

78

irradiations in four to five hour-intervals (Figure 3.4). Four experimental groups were

created and three wells were included in each group. In the NS Non scratch (NS) group

cells were used as control, as neither scratch injury nor 670nm light treatment was used

in this group. In the Scratch group (S), no 670nm irradiation was delivered at any time,

though the cells underwent identical treatment to the 670nm light groups (NS+670nm

and S+670nm), with the LED array remaining unswitched. NS+670nm and S+670nm

groups both received 670nm light treatment. Baseline measurements were taken

immediately after the first intervention relevant for each group (scratch +/- 670nm light),

and was expressed as T=0. Further cell harvest was performed at T=24, 48, and 72

hours

Figure 3.4 Timeline of the in vitro (scratch) model in MIO-M1 cells. Cells received three session of 670nm light within 10 hours after the scratch.

3.2.9 Immunocytochemistry staining (ICC) Cells were placed at 1.3X104 each well into 8-well chamber slides (Millipore, Millicell

EZ SLIDE) in GM for two days at 37 °C in a humidified CO2 incubator. A uniform

scratching wound was created using 1 ml pipette tip. In each group, cells were

respectively collected at T=0, 24, 48 and 72 hours and fixed with 2% paraformaldehyde

for 30 minutes at RT. Procedures of ICC are described in section 2.9. The primary

antibodies primary antibodies, 1:200 GFAP (ThermoFisher, MS1376), 1:50 Ki67

(Millipore, AB9260), vimentin (Sigma-Aldrich, V6630),or 1:200 FGF-2 (Millipore,

05-118) diluted in 1% NGS, were used. Alexa-488 anti-mouse or anti-rabbit secondary

antibodies were added onto cells following the incubation with primary antibodies.

Images were visualised using confocal microscopy (A1 Nikon Confocal Microscope,

Nikon, NSW, Australia) and captured by using NIS-Element AR software (Nikon).

Each well was randomly captured for 7-8 images along wound sites. The brightness

79

and contrast of images were adjusted for publication using Photoshop CS6 software

(Adobe Systems, CA, USA), which was standardised among images. The intensity

fluorescence of FGF-2 per cell in images was analysed with NIS-Element AR software.

Positive staining of cells (GFAP+ or Ki67+) and the total cell number were counted in

each image. A ratio of positive staining cells to the total cell number in each image is

obtained and an average of ratios of 7-8 images is represented as a quantification in one

well. Data is presented as an fold change in percentage compared to a ratio average in

samples of the S group (or NS group) at 0 hour.

3.2.10 Proliferation inhibition with mitomycin C (MMC) Cells were grown to reach 90% of confluence in 6-well plates in GM for 2 days and

then treated with 10µg/ml of mitomycin C (Sigma-Aldrich) in GM for 2 hours prior to

the scratch injury (Arranz-Valsero et al., 2014). Then, medium was removed and cells

were washed twice with PBS. Fresh GM was added into each well and a scratch wound

was created as described above. The same paradigm of the treatment of 670nm light

was performed as described above.

3.2.11 Cell spreading Cellular cytoskeleton was observed for cells immuno-labelled with vimentin (Sigma-

Aldrich, V6630). Steps of ICC are referred to section 2.9. The extent of cell spreading

of individual cell was accessed by calculating the ratio of the area of cell cytoplasm in

relation to the area of the nucleus. Cell spreading was quantified by analysing 20-30

cells along the wound edges in each well (Lu et al., 2013).

3.2.12 Migratory capability assay The protocol previously published was modified for this study (Lu et al., 2013). Cells

(3000 cells per well) were placed onto inserts of transwells (5.0 um pore-size, Millipore)

and cultured in GM in a humidified 5% CO2 at 37°C. After 48 hours, serum-free

medium was added to the inserts, while fresh GM (10% FBS) was placed in the lower

chamber. Cell migration was assessed over 72 hours. Cell migration was based on the

number of cells in the insert membrane that migrated across the membrane toward the

lower surface. Cells located on the surface of insert membrane were removed using

80

cotton swabs and then rinsed with PBS, as these were categorised as non-migrating

cells. Migrating cells were quantified by immersing the insert membrane in 4%

paraformaldehyde for 10 minutes at RT, then counterstained with BBZ (Sigma-Aldrich)

for 5 minutes. Cell nuclei were visualised under a fluorescent microscope Zeiss

Axiovert 3 (Zeiss) and the cell number on inserts was quantified as an indication of a

migratory capability. Seven to eight fields were captured for each well. The dimensions

of each field observed is 2mm in diameter.

3.2.13 Flow cytometry analysis for cell viability and cell cycle Flow cytometry was used to assess cell cycle status and cell death. Vibrant ®

DyeCycleTM Violet Stain (Invitrogen, Life Technologies, USA) is able to stain for DNA.

DNA content can discriminate the phases of the cell cycle. 7-Aminoactinomycin D (7-

AAD) (ThermoFisher, USA) is a membrane impermeant dye and displays a high

affinity to DNA. Therefore, 7-AAD only can enter into non-living cells and bind its

DNA to detect dead cells. Cells were grown at a density of 10X104 per well in a 6-well-

plate in GM for 2 days. Following various treatments, cells were trypsinized for 5

minutes and centrifuged at 1000rpm for 10 minutes. Cells were resuspended in 500µl

of HBSS (Life Technologies, USA) and then incubated with Vybrant ® DyeCycleTM

Violet Stain (1µl in 1ml of cell suspension) at 37°C for 2 hours in the dark. Following

2 hours of incubation, 7-AAD was added (5µl in 200µl of cell suspension) at 37°C for

another 10 minutes in the dark. Cells were analysed using Flow cytometry (LSRII, BD,

CA, USA). The population of different cell cycle phases and the percentage of dead

cells in each sample were analysed by using a software, FlowJo (FLOWJO, LLC,

OR,USA). The fraction of cell cycle phase in G0/G1, S, or G2/M phase was determined

based on the DNA content. A percentage of dead cells in one sample was defined by

the higher fluorescent intensity of 7-AAD in cells compared to the background.

3.2.14 Statistical analysis Data were analysed with one-way ANOVA, with post-hoc comparisons Tukey’s

Multiple Comparisons test. Data were reported as mean ± SEM. Values of P < 0.05

were considered to be statistically significant. All statistical analyses were performed

using GraphPad Prism 6.0 (GraphPad software, San Diego, CA, USA).

81

3.3 Results

3.3.1 Early treatment with 670nm light reduced photoreceptor loss following photo-oxidative damage (PD) Recoverin labelling showed a disruption of photoreceptor (PR) structures following PD.

At the lesion site, the ONL was largely obliterated centrally, while on the periphery of

the lesion ONL was present but severely distorted, the remaining PRs were scattered

and their inner and outer segments were absent (Figure 3.5E and F left). At the

penumbra of the lesion, the ONL was more compact, and inner and outer segments

were observed (Figure 3.5E and F right). The presence or absence of IS/OS and the

organisation of the ONL were used to define the edge of the lesion (dotted lines in

Figure 3.5F). A similar pattern of photoreceptor damage was found in the lesion and

penumbra of PD+R0 (Figure 3.5H&I), PD+R3 (Figure 3.5K&L), and PD+R14 (Figure

3.5N&O) retinas, suggesting that the 670nm light cannot restore photoreceptor

structural damage.

We have previously shown that following PD, there is a focal loss of photoreceptors in

the superior retina, which becomes the centre of a progressive lesion, which may engulf

the entire retina over time (Rutar et al., 2010). To assess whether 670nm light had an

effect on lesion expansion, we measured the size of the lesion along the retina at 30

days following PD, using the criteria to define the edge of lesion, as described above.

In control retinas, no lesion was detected (Figure 3.5D). In the PD group, approximately

80% of the superior retina displayed photoreceptor loss indicating an expansive field of

lesion (Figure 3.5G&Q). Early treatment with 670nm light (PD+R0 and PD+R3),

resulted in a significant reduction in lesion size (P<0.05), affecting approximately half

of the superior retina only (Figure 3.5J, M, and Q). Treatment at 14 days following PD

(PD+R14) resulted in no significant change in lesion size compared to PD only retinas

(Figure 3.5P&Q).

82

Figure 3.5 670nm light treatment on PR and MC gliosis following PD

(A) Protocols of PD and 670nm light treatment. (B, C, E, F, H, I, K, L, N, O) Sections immunolabelled with vimentin (red), recoverin (green), and counterstained with BBZ (blue). Increased presence of vimentin in MC processes indicated gliosis at the lesion edges. White squares denote the area shown in C, F, I, L, O at higher magnification: White arrowheads indicate PR IS/OS. White dashed lines denote the border between lesion (on left) and penumbra (on right). (D, G, J, M, P) ONL lesion size in the superior retina of the control (D), PD (G), PD+R0 (J), PD+R3 (M) and PD+R14 (P). D, G, J, M, P were captured at 10X magnification. Blue arrows denote the lesion edge. (Q) Lesion size increases following PD, and was significantly decreased after 670nm light treatment at 0 and 3 days. ONL: outer nuclear layer. INL: inner nuclear layer. All data is presented as the mean ± SEM. * denotes P<0.05. N=6-10 for all groups.

83

PD caused a significant reduction in the number of rows of photoreceptors and

consequently a significant thinning of the ONL, which was partially reversed by 670nm

light (Figure 3.6). In non-treated PD retinas (Figure 3.6, blue line), there was a

significant loss of photoreceptors along the retina, but the loss was especially severe in

the superior retina where, in a circumscribed area, approximately 2-4 mm supero-

temporal from the optic disc (grey highlighted rectangles), the ONL was obliterated.

Three paradigms of 670nm light treated-groups were represented by the red lines in

each graph (Figure 3.6A-F). Compared to non-treated injured eyes (PD), retinas with

the early (PD + R0) and middle treatment (PD + R3) of 670nm light, had significantly

more photoreceptor cells along the retina ( Figure 3.6 A-B, D-E). Notably, the most

severe loss was confined to an approximately 1-2mm area in the superior retina in the

PD+R0 and PD+R3 groups. When 670nm light treatment started later (PD+R14), there

was no significant difference in the ONL thickness and the number of rows of

photoreceptor cell nuclei compared to non-treated PD retinas (Figure 3.6 C&F).

Figure 3.6 Effects of 670nm light on PR survival post PD (A-C) The ONL thickness in retinal sections in different groups. (D-F) The number of rows of photoreceptor nuclei in ONL. Lines in graph A-F respectively denote the control group (black lines), the PD group (blue lines), and the 670nm light-treated groups (red lines). ONL thickness significantly reduced in all PD retinas, however, they remained significantly thicker at the majority of length of the retina of PD+R0 (A&D) and PD+ R3 group (B&E) animals compared to the PD retinas. Grey highlighted rectangles in the superior retina denote the retinal injured area (lesion) in the PD group. All data is presented as mean ± SEM. * denotes P< 0.05 compared to the value at the corresponding time point in the PD group. N=6-10 for all groups.

84

3.3.2 Treatment with 670nm light reduced MC gliosis To assess MC structural changes, vimentin immunolabelling was used, a cytoskeletal

protein marker. It showed thickened MC process that surrounded the surviving PRs in

the area of the lesion (yellow colour in Figure 3.5F), indicating glial scar formation in

the ONL. Co-localisation of recoverin and vimentin fluorescence was barely present in

the 670nm light-treated retinas (Figure 3.5I, L, and O).

To measure retinal stress, GFAP protein expression was assessed via

immunohistochemistry. In control retinas, GFAP was only expressed by astrocytes

(Figure 3.7A). Following PD, GFAP labelling was present in MCs, expanding the entire

length of the inner and outer processes of the cells in the area of lesion (Figure 3.7B).

Treatment with 670nm light modified GFAP expression, as evidenced by the reduced

intensity of fluorescence (Figure 3.7D, F, and H left ). At the penumbra, GFAP labelling

was still present in MCs, however it did not extend beyond the OPL (Figure 3.7D, F,

and H right). To quantify GFAP expression in the penumbra, NIS-Element AR software

was used to measure the intensity of fluorescence along the MCs from GCL to the inner

edge of the OPL (Figure 3.7C, E, G, and I). GFAP labelling was significantly increased

in PD retinas compared to controls for almost the entire length of the MCs (Figure 3.7C,

P<0.05). In PD+R0 and PD+R3 retinas, the level of GFAP staining in the GCL was not

different from those of non-treated PD retinas, indicating that 670nm irradiation does

not have an effect on the GFAP expression in the astrocytes. In early-treatment retinas

(PD+R0 and PD+R3), GFAP immunostaining of MCs was significantly reduced

between the IPL and OPL (Figure 3.7E&G). Interestingly, the entire length of the

measured area from GCL to ONL had a significantly reduced immunolabelling

compared to PD retinas in the PD+R14 retinas, (Figure 3.7I, P<0.05).

Thickened GFAP+ MC processes were evident in the outer retina, within and outside

of the remnants of ONL at the lesion site following PD (white star, Figure 3.7B),

suggesting glial scar formation. Such glial scars were not observed outside the lesion

area. In retinas irradiated with 670nm light, GFAP+ MC processes structures were

present in the outer side of the ONL in the area of the lesion however, their staining

was much weaker (white starts, Figure 3.7D, F, and H) than non-treated PD retinas

(Figure 3.7B).

85

Figure 3.7 Effects of 670nm light treatment on MC gliosis following PD (A, B, D, F, H) Sections immunolabelled with GFAP (red) showed that in control retinas only astrocytes were labelled, while MCs became GFAP+ after PD indicating tissue stress . White * denotes glial scars in and outside of the ONL in the lesion area. (C, E, G, I) GFAP intensity profiles across MC processes in the penumbra adjoining the lesion (white dash line boxes). GCL: ganglion cell layer; IPL: inner plexiform layer; INL: inner nuclear layer; ONL: outer nuclear layer; OPL: outer plexiform layer All data is presented as the mean + SEM. N=6-10 for all groups.

3.3.3 Retinal stress and inflammation-related gene expressions changed following PD and PBM Retinal stress status was assessed by determining Gfap and Fgf-2 gene expression. Gene

expression of Gfap increased approximately 20-fold following PD (Figure 3.8A)

compared to the control group. In the PD+R0, PD+R3 groups, the early treatment with

670nm light significantly moderated the stress response, showing only 10-fold increase

compared to control retinas (Figure 3.8A). Late treatment (PD+R14) did not reduce

stress response significantly. Gene expression of Fgf-2 showed a similar pattern

(Figure 3.8B). Photo-oxidative stress led to a significant, 10-fold increase of Fgf-2

expression in non-treated retinas. In retinas that received treatment early (PD+R0) only

an approximately 5-fold increase was observed, which represented a significant

reduction compared to non-treated injured retinas (Figure 3.8B). Retinas where

treatment started later, at 3 (PD+R3) or 14 (PD+R14) days, did not show a significant

reduction in Fgf-2 gene expression (Figure 3.8B). In addition, the expression of genes

related to the innate immune response including cytokines (Ccl2, Il-1β) and

inflammasome components (Nlrp3, Casp8) was investigated. All inflammatory genes

86

showed a significant increase in expression following PD. Ccl2 increased by

approximately 20-fold (Figure 3.8D). Ccl2 expression was not diminished by treatment

with 670nm light (Figure 3.8D). The expression of Il-1β, Nlrp3, and Casp8 was

increased in the PD retinas (Figure 3.8 C, E and F). Only early treatment (PD+R0

group) reduced inflammasome-related genes (Il-1β, Nlrp3, and Casp8) compared to the

PD group (Figure 3.8 C, E and F).

Figure 3.8 The comparative fold changes of gene expression in the retina using quantitative qRT-PCR Effects of 670nm light treatment on gene expressions in the PD retinas. Expression levels were compared to dim-reared control retinas. (A) GFAP expression in the retinas was significantly decreased after 670nm light treatment at 0 (PD + R0) and 3 days (PD+R3) following PD, compared to non-treated PD controls. (B) Fgf-2 expression significantly decreased in the PD + R0 670nm light-treated group compared to PD controls. (C-F) Inflammatory gene expression of Il-1β (C), Nlrp3 (E), and Casp8 (F) decreased significantly in the PD + R0 670nm light-treated group compared to PD controls, whereas Ccl2 expression did not significantly change compared to controls (D). All data is presented as the mean ± SEM. * denotes P< 0.05 and N = 6–10 for all groups.

87

3.3.4 Treatment with 670nm light modified the recruitment and mobilization of MG/MΦ post PD

IHC with antibodies against Iba1 and ED-l was used to characterise inflammatory

responses in the superior retinas following PD. In the non-injured control group, Iba1+

cells were mainly found in the inner retina (from GCL to IPL), but no cells were present

in the ONL (Figure 3.9A top row). Iba1+ cells were ramified indicating resting MGs.

Following PD, numerous Iba1+ cells were found across the retina (Figure 3.9A top

row). Some amoeboid Iba1+ cells, typical of active MGs, were seen in retinas of the

PD group (white arrows, Figure 3.9A top row). In the PD+R0 and PD+R3 groups, a

number of Iba1+ cells was significantly reduced both in the outer retina and across the

retinal layers (whole retina) compared to the PD group (Figure 3.9B), however those

that displayed an amoeboid appearance were found primarily in the outer retina (white

arrows, Figure 3.9A top row). In the PD+R14 group, a large number of amoeboid Iba1+

cells were observed across the retinal layers, similar to samples in the PD group (Figure

3.9A top row & B).

In the control group, only a few ED-1+ cells were seen scattered across the retinal layers

(Figure 3.9A bottom row). In the PD group, a large number of ED-1+ cells were

distributed across the retinal layers, however we noted a strong accumulation of these

cells in the outer retina (Figure 3.9A bottom row). In the PD+R0 and PD+R3 groups,

significantly less ED-1+ cells were observed in the retinas (Fig. 3.9A bottom row & C),

however these cells were mainly located in the outer retina (Figure 3.9 A bottom row).

In the PD+R14 group, the overall numbers of ED-1+ cells were significantly reduced

however, the late treatment did not significantly suppress the number of ED-1+ cells

in the outer retina following PD (Figure 3.9 A bottom row & C).

88

Figure 3.9 Effects of 670nm light on MG/MΦ recruitment following PD (A) Representative images of the superior retinas showed that Iba1+ cells (green) and ED+1 cells (red) were recruited into the outer retina following PD. They were reduced following 670nm light treatment at 0 day (PD+R0), 3 days (PD+R3), and 14 days (PD+R14). Cell nuclei were counterstained with BBZ (blue). (B&C) The quantification of the number of Iba1+ (B) or ED-1+ cells (C) in the superior retina or along the ONL layer of the superior retina. Iba1+ or ED-1+ cell counts showed that the reduction was significant after the 670nm light treatment compared to the control group. GCL: ganglion cell layer. IPL: inner plexiform layer. INL: inner nuclear layer. OPL: outer plexiform layer. ONL: outer nuclear layer. N= 7 (control), 9 (PD), 7 (PD+R0), 10 (PD+R3), and 7 (PD+R14). The error bar represents as SEM. * denotes P< 0.05.

89

3.3.5 Direct effects of 670nm light on MC motility and division in vitro To assess MCs reactivity in vitro, the well-established scratch model in MIO-M1 cell

line was used in this thesis. The rate of area coverage was compared between 670nm

treated and non-treated cells (Figure 3.10B-D) and it showed that PBM reduced the rate

of area coverage significantly at each time point (Figure 3.10E). Within 24 hours, the

area covered in the non-treated wells was approximately double (2.26 ± 0.16 pixels per

hour) that of the treated wells (1.04 ± 0.13 px/hr), representing a significant slowing of

motility by PBM (p<0.05). After the first 24 hours, when 670nm light treatment ceased,

the rate of area coverage was almost identical in the two groups (1.62 ± 0.08 px/h in

the S group and 1.67± 0.09 px/h in the S+670nm group, p>0.05). These data suggest

that the reduction of area coverage occurred within the duration of 670nm light

treatment. However, the difference in area coverage remained significantly different

between the two groups up to 72 hours (Figure 3.10E)

3.3.6 670nm light reduced injury-induced MC mitosis in vitro To investigate whether MIO-M1 displayed proliferative gliosis in this model, Ki67, a

known nucleic proliferation marker, was used. The number of Ki67+ cells between the

Non-scratch (NS) and the NS+670nm groups did not differ over 72 hours (Figure

3.10N). Ki67 labelling was present in MIO-M1 cellular nuclei at 24 hours post-insult

(the S group)(white arrows, Figure 3.10G), and this represented a significant increase

when compared to T=0 (Figure 3.10O). The number of mitotic cells presented in the

wells stayed high thereafter (no significant difference between T=24 and T48 or

between T=24 and T=72) (Figure 3.10 H, I, and O). In the S+670nm group, 670nm

light treatment did not alter the number of mitotic cells in the early stages (T=24&48)

following the scratch compared to samples of the S group (Figure 3.10K, L, and O).

However, 670nm light treated-cells showed significantly less mitosis at 72 hours post-

injury (the S+670nm group) compared to cells of the S group (Figure 3.10M&O).

90

Figure 3.10 Cell area coverage and proliferation following 670nm light treatment in the in vitro scratch model (A) Experimental paradigm. (B-D) Representative images showed control cells (B). Untreated cells (C) and 670nm light-treated cells were at 72 h following the scratch (E) Quantification of cell area coverage showed that 670nm light significantly reduces the rate of area coverage at each time point. (F-M) Ki67 ICC showed that mitosis increases in MIO-M1 cells following the scratch injury, and this rate persists for up to 72 h (F–I), whereas cells treated with 670nm light had a reduced expression of Ki67 at 72 hours (M). (N) Proliferation in non-scratched groups (NS, NS+670nm) showed no difference over the time course. (O) Quantification of Ki67 + cells showed an upregulation of proliferation following the scratch injury, and a reduction in cells treated with 670nm light at 72 hours. All data is presented as the mean ± SEM. * denotes P < .05 and experiments were performed with N = 3 in technical triplicate.

91

3.3.7 The 670nm light treatment reduced cell death after injury in vitro

Flow cytometry was used to confirm whether the 670nm light-induced reduction in

mitotic cell numbers was related to cell cycle status or cell death. A percentage of cell

population in G0/G1, G2/M and, S phase was calculated (Figure 3.11A-C) and found

no significant differences between the two groups at any of the time points examined.

The reduction in the G2 phase population observed at T=24 in the treated group did not

reach significance, when compared to non-treated cells at the same time (Figure 3.11B).

MMC significantly slowed the rate of cell area coverage in non-treated cells (dashed

black line, Scratch+MMC, Figure 3.11D), however had no effect on 670nm light-

treated cells (Scratch+670nm+MMC, red lines in Figure 3.11D). The percentage of

dead cells in all groups was low (~2-3%), during the time period observed (Figure

3.11E). Notably, the percentage of dead cells, at T=24 and T=48 were significantly

lower in the Scratch+670nm group compared to the Scratch group at the same time

point (Figure 3.11E).

Figure 3.11 Effects of 670nm light treatment on MC cell cycle status and cell death following the scratch (A-C) The population of MIO-M1 cells at each phase of the cell cycle was analysed by flow cytometry, and it was found that there were no differences between 670nm-treated and non-treated groups across the time course for the G0/G1 phase (A), G2/M phase (B) and, the S phase (C) post the scratch. (D) MMC pre-treatment of cells prior to the scratch significantly reduced the rate of area coverage in non-670nm light-treated cells (dashed black line) but did not have an effect on 670nm light-treated cells (red lines). (E) 670nm light treatment reduced the percentage of dead cells at 24 and 48 h compared to scratch controls post the scratch. All data is presented as the mean ± SEM. * denotes P<0.05. Experiments were performed with N=3 experiments in technical triplicate.

92

3.3.8 Effects of 670nm light on cell spreading and migration

To assess differences in cell sizes, vimentin and nucleic acid labelling to visualise the

cytoplasm and nucleus of cells near the scratch were used (Figure 3.12 A). Cytoplasm

area was significantly increased at T=24 and maintained at 48 and 72 hours following

mechanical scratch, consistent with cell spreading (top row, Figure 3.12A). Cells

treated with 670nm light showed no significant change in cell area across all time points

(bottom row, in Figure 3.12A). Following the scratch, the ratio of cytoplasm to nuclear

area was significantly higher in the non-treated group at all time points compared to the

670nm light-treated group, suggesting that 670nm light suppressed cell spreading

(Figure 3.12B). To assess the effect of 670nm light on cell motility, transwell chambers

were used. Only a few migrating cells were noted in the first 24 hours, but a significant

increase in migration was observed at 48 hours and a further increase in migration was

present at 72 hours in both groups. There was no significant difference found in the

number of migrating cells between the two groups, suggesting that 670nm light did not

have an impact on the migratory ability (motility) in MIO-M1 cells (Figure 3.12C).

3.3.9 In vitro assessment of cellular stress To assess stress status in MIO-M1 cells, GFAP and FGF-2 immunolabelling were

performed (Figure 3.13 A & B). In the Scratch group, the percentage of GFAP+ cells

at the scratch site increased significantly following injury (Figure 3.13A, top row). The

increased GFAP labelling was maintained thereafter, but by 72 hours incubation, there

was a significant reduction in GFAP presence (Figure 3.13 C). Treatment with 670nm

light reduced stress response at all time points (Figure 3.13 A bottom row). Data suggest

that mechanical injury causes an early mild stress response that 670nm light can prevent

(Figure 3.13 C).

Another stress-inducible protein, FGF-2 was also examined. MCs expressed FGF-2 in

the cellular cytoplasm immediately after injury (Figure 3.13 B, top row). A significant

increase in FGF-2 labelling was apparent in the cytoplasm in the Scratch group at T=24

and T=48 hours (Figure 3.13 D). Notably, more intense fluorescence was observed

around the nucleus. By T= 72 hours, FGF-2 labelling subsided almost reaching the T=0

level, but was still higher than the T= 0 level (P<0.05) (Figure 3.13 D). In the

Scratch+670nm group, FGF-2 labelling increased following injury, however this

93

increase was significantly smaller than in their non-treated counterparts and remained

at this significantly lower level all through the time points observed (Figure 3.13B,

bottom row and D).

Figure 3.12 Cell spreading of MIO-M1 cells was suppressed by 670nm light (A) Vimentin immunocytochemistry (ICC, green) and cell nuclei staining (BBZ, blue) allowed the measurement of the area of the cytoplasm and the area of nucleus (yellow dashed lines). (B) Quantification of the ratio of the cytoplasmic area to the nuclear area showed that there was a significant decrease in cell spreading following 670nm light treatment. (C) Cell migratory capability across a membrane of the transwell was measured as the number of cells per field (2mm in diameter). Cell migration increased across the time course. However, 670nm light did not have an effect on the migratory capacity of MIO-M1 cells. All data is presented as the mean ± SEM. * denotes P< .05 and experiments were performed with N=3 experiments in technical triplicate.

94

Figure 3.13 Effects of 670nm light on MIO-M1 cell stress (A) The number of MIO-M1 cells, which were labelled with GFAP (ICC, green), was increased following the scratch injury. Treatment with 670nm light reduced the number of GFAP+ cells on the wound edges. (B) MIO-M1 cells were labelled for FGF-2 (green), which also increased in the scratch model. Treatment with 670nm light reduced FGF-2 expression. (C&D). Quantification of GFAP (C) and FGF-2 (D) labelling showed that 670nm light treatment significantly reduced the number of GFAP+ cells and FGF-2 expression in MIO-M1 cells across the time course following the scratch injury. All data is presented as the mean ± SEM. * denotes P<0.05 and experiments were performed with N=3 experiments in technical triplicate.

95

3.4 Discussion

This chapter demonstrated the effects of 670nm light on the gliotic changes of MCs in

in vivo and in vitro stress models. Current data have shown that 670nm light modulated

gliotic changes that includes the reduction of tissue and cellular stress, the extent of

glial scar formation, the migration and proliferation of MCs. In addition, 670nm light

decreased the recruitment of MG/MΦ and mitigated the expression of proinflammatory

genes, which are associated with MC activation. Through the stabilisation of MCs,

PBM can increase the survivability of retinal neurons during injury, and thereby to

minimise irreversible and progressive degenerative changes.

3.4.1 670nm treatment post injuries suppressed MC activation To assess the effects of 670nm red light on activated MCs, following injury, cellular

changes in MCs were observed. In vivo, time points for PBM treatments were chosen

based on MC stress (GFAP expression) pattern in this animal model in our previous

publication (Rutar et al., 2010). Our previous findings demonstrated that the expression

of GFAP first increased at 12 hours, reached its peak at 24 hours post PD (Rutar et al.,

2011b; Rutar et al., 2010). Therefore, in this study, the treatment with 670nm light

commenced after activation of MCs in vivo. Current in vivo results are consistent with

the previous finding in the age-related AMD mouse model (16 month old CFH-/- mice),

where 670nm light reduced GFAP expression and the length of MC processes in the

aging retinas while MCs already displayed up-regulated GFAP (Begum et al., 2013).

Current in vitro and in vivo data demonstrated that 670nm light treatment reduced gene

and protein expression of FGF-2 and GFAP, MC stress markers, in MCs. PBM

treatment of the retina significantly reduced MC stress, their expression of

proinflammatory agents and consequently, the recruitment of MG/MΦ in the PD-

retinas, when treatment commenced shortly after injury (0-3 days). Although, treatment

commencing late (2 weeks post-injury) did not mitigate MCs reaction overall, but there

was an indication of a mild effect on MΦ recruitment. Up-regulation of GFAP in the

late AMD may be associated with the disruption of RPE and blood-retinal barrier (Wu

et al., 2003), which is related to the recruitment of MG/MΦ (Ambati et al., 2013; Omri

et al., 2011).

96

3.4.2 670nm light treatment reduced photoreceptor loss and slowed disease progression

Our previous publication showed that treatment with 670nm light prior to retinal injury

mitigated photoreceptor loss and reduced MC activation. However, it remained unclear

if this was the consequence of a direct effect of PBM on the Müller glia, or was

subsequent to the survival of photoreceptors. In the current study, the PBM treatment

commenced after the activation of MCs, to assess theeffects of 670nm light on gliosis.

Studies in this chapter demonstrated that, while even the earliest treatment regime (R0)

could not prevent the development of a focal severe disruption of the ONL. However,

treatment commenced early after injury (0-3 days) limited the size of this ‘hot spot’ to

a 2 mm region around the area centralis of the rat retina. In contrast, retinas where

treatment started with a longer delay (R14), the ONL thickness was reduced across the

entire length of retina and was not significantly different from that of the non-treated

eyes. Our previous findings indicated that the majority of neuronal damage occur in the

first 2 weeks following PD in this model, and only slow progression was present

thereafter for up to 2 months (Rutar et al., 2010), suggesting that there is a limited time

when PBM treatment can effectively mitigate the effects of PD.

3.4.3 The effects of 670nm light on MC migration: implications for glial scarring

The underlying mechanism of the effects of 670nm light on the progression of

degeneration was investigated in the in vitro scratch model using MCs. This system

allowed the assessment of the direct impact of PBM on MCs in the hope to provide

some mechanistic insight. This study demonstrated that 670nm light reduced the rate

by which MCs were able to cover physical space, which was related to the lower level

of cell spreading and proliferation.

Previous studies reported that MC nuclei migrated into the damaged outer retina,

undergoing mitosis thereby forming subretinal glial scars following retinal detachment

(Lewis et al., 1999; Lewis et al., 2010), suggesting that MC nuclei movement and MC

proliferation are relevant to glial scar formation. In our study, 670nm irradiation

reduced spreading of and mitosis in MCs. In addition, our data showed the reduction of

FGF-2 and GFAP protein expression in MCs irradiated with 670nm light. FGF-2 has

97

been shown to induce increased expression of GFAP in MCs, and that have been linked

to glial scar formation in the rabbit retinal detachment model (Lewis et al., 1992).

Our findings suggested that PBM has a mitigating effect on MC proliferative gliosis

(Figure 3.10). MC migration has been shown to be regulated by matrix metalloprotease

(MMP) and protease inhibitor α2 –macroglobulin (α2M) (Barcelona et al., 2013;

Barcelona et al., 2011; Limb et al., 2002a). α2M activity has been implicated in the

induction of GFAP expression in MCs in vivo and in vitro (Barcelona et al., 2013;

Barcelona et al., 2011). In addition, 670nm light has been shown to reduce MMP

expression in skin wound healing of diabetic rats (Aparecida Da Silva et al., 2013). It

is possible therefore, that the decreased GFAP expression observed in our models may

be related to changes in MMP or α2M activity in activated MCs, subsequently

influencing MC migration (proliferative gliosis). However further studies exploring the

association between MMP activity in MCs and 670nm irradiation are required.

However, 670nm light-ameliorated cell migration is not strongly associated with down-

regulation of cell proliferation in MCs in the uniform scratch model in the current study.

Although 670nm light mitigated the number of Ki67+ cells at wound site at 72 hours

post the scratch, the effects of 670nm light on the reduction of the wound coverage

rate was only shown at the first 24 hours after the scratch. In addition, 670nm light also

has no impact on the cell cycle in MCs over 72 hours after the scratch. These evidences

implicated that 670nm light-mediated the reduction of cell proliferation (fewer Ki67+

cells) at 72 hours might be involved in other mechanism. Extensive studies in recent

years have implicated that MCs are potential stem cells in the retina (Hamon et al.,

2016). Adult MCs in zebrafish are capable of re-entering cell cycle and then dividing

asymmetrically near the ONL to generation transient amplifying progenitors, which

exit from cell cycle and then differentiate into photoreceptors (Goldman, 2014).

Although it was believed that mammalian MCs could only undergo reactive gliosis and

not neurogenesis, Takahashi’s group showed that some of rat MCs can dedifferentiate,

re-enter the cell cycle, and generate new bipolar and rod photoreceptors in response to

a neurotoxicity-induced retinal damage model (Ooto et al., 2004). Although the

mechanism of MC de-differentiation are still being explored, some reports has

demonstrated that MCs might exit from cell cycle prior to dedifferentiate. Zhou’s group

demonstrated that the expression of cyclin D1 and cyclin D3 was rapidly up-regulated

98

and then down-regulated in MCs in a N-methyl-N-nitrosourea-induced retinal

degeneration rat model, which was accompanied with that some of MC-derived cells

expressed rhodopsin (MC-dedifferentiation) (Wan et al., 2008). In the current study,

670nm light-caused the reduction of cell proliferation at the late stage (72 hours post

the scratch) in the in vitro model may be linked to accelerating MCs to exit from the

cell cycle, which subsequently may initiate de-dedifferentiation. However, further

investigation is needed.

3.4.4 Early 670nm light treatment mitigated inflammatory changes Our previous study reported that in the PD model of retinal degeneration, there is an

early and rapid increase in the expression of the monocyte chemo-attractant molecule

(Ccl2) in activated MCs shortly after retinal injury, leading to the recruitment of

MG/MΦ into the damaged areas of the retina (Rutar et al., 2012b; Rutar et al., 2011b;

Rutar et al., 2010). Other study have shown that PBM reduced Ccl2 (MCP1) expression

in LPS-activated macrophage cell line (Fernandes et al., 2015). The present study

demonstrated that PBM-treated retinas had a fewer number of MG/MΦ following PD,

which is consistent with previous findings in rodent retinal degeneration models

(Albarracin et al., 2011; Begum et al., 2013; Kokkinopoulos et al., 2013). As MCs have

been shown to express Ccl2 (Rutar et al., 2012b), the reduction of MG/MΦ in our model

of PD following 670nm irradiation may be linked to the reduced stress and activation

of MCs and the reduction of Ccl2. Although the current data shows no difference in

Ccl2 gene expression among the treatment groups and PD control retinas, the reduction

of MΦ numbers in the damaged area of retinas suggests that PBM has an effect on the

reduction of immune cell recruitment into the retina. It is possible that the effect of

670nm light on Ccl2 gene expression is transient, therefore it is not observable at this

late stage post-injury.

3.4.5 PBM may regulate retinal inflammatory response via inflammasomes

This study demonstrates that 670nm light reduced inflammation-related gene

expressions, including IL-1β, in retinas after PD. It has been reported that, when

incubated with activated MGs, or grown in hyperglycaemic condition, MCs express IL-

1β (Liu et al., 2012; Wang et al., 2011). This activation of the MCs, in turn, activated

99

MG/MΦ that produce proinflammatory cytokines and complement proteins (Fernando

et al., 2016; Natoli et al., 2017b). Therefore, the reduction of inflammation-related

genes observed in our study in 670nm treated PD retinas, may be the consequence of

the decreased activation of MCs.

Inflammasomes have been implicated in a number of retinal diseases. Caspase-1,

involved in inflammasome activation, was present in rat Müller cell line in

hyperglycaemic condition (Trueblood et al., 2011). There is a growing body of evidence

showing that inflammasome activation was also present in RPE and MG/MΦ (Astray

et al., 2016). Caspase-8 has been indicated in the activation of NLRP3, the well

characterised, common inflammasome complex (Kauppinen et al., 2016; Tschopp and

Schroder, 2010). Present data indicates that 670nm light can reduce the expression of

inflammasome-related genes, including NLRP-3 and CASP8, in the retina. The effect

of 670nm red light on inflammasomes is a novel finding, however, due to their

ubiquitous presence, further investigations into the role of MCs on inflammasome

activation are necessary. In our previous study, using a photoreceptor and MCs co-

culture system, MCs displayed ability of producing NLRP3 inflammasomes, when

incubated with degenerated photoreceptors, however 670nm irradiation in that system

did not result in a reduction of NLRP3 inflammasome activation (Lu et al., 2017). Thus

the present finding of the reduction in the expression of NLRP3 inflammasome and

related genes may not be through the direct effect of 670nm light on MCs, hence further

research is required to clarify the mechanisms behind this anti-inflammatory effect.

3.4.6 Mechanism of action of 670nm light on MCs The exact mechanism of 670nm light is still debated, but there is good evidence for its

role in mitochondrial function. It has been proposed that 670nm light is absorbed by

COX, a component of the electron transport chain and a rate-limiting enzyme in the

intermembrane space of mitochondria (Karu, 1999). 670nm light has been shown to

increase COX expression (Begum et al., 2013) and elevate mitochondrial membrane

potential (Kokkinopoulos et al., 2013), which is proportional to ATP production

(Wong-Riley et al., 2005). ATP can be released from MCs into extracellular space upon

the osmotic stress and glutamate stimulation (Voigt et al., 2015). Extracellular ATP and

its downstream signalling cascade have been reported to inhibit MC swelling, an

indication of MC gliosis, under hypo-osmotic conditions (Uckermann et al., 2006).

100

Consequently, 670nm light may trigger MCs to release ATP into the extracellular space

under stress, which subsequently help maintain retinal homeostasis.

Our study in the in vivo model demonstrated that 670nm red light has an effect on retinal

gene expressions. PBM has been indicated to facilitate the synthesis of DNA and RNA

in HeLa cells (Karu and Kolyakov, 2005). This agrees with our previous findings in

vivo, demonstrating that 670nm light modulates the expression of a variety of genes

including complements (Rutar et al., 2012a), neuroprotection (Albarracin and Valter,

2012), and oxidative stress (Albarracin et al., 2013). However, our current data show

that only early treatment of 670nm light (R0) can effectively reduce the expression of

MC activation-related genes (GFAP, FGF-2, IL-1β, CASP8, and NLRP3) after the PD.

Our previous studies showed that IL-1 β, GFAP, and FGF-2 reached a peak level with

24 hours after the retinal injury in the same model (Rutar et al., 2015; Rutar et al.,

2011b). These gene expressions showed a significant decrease (IL-1β and FGF-2) or a

downward trend (GFAP) at 3 or 4 days after lesioning (Rutar et al., 2015; Rutar et al.,

2011b; Valter et al., 2005), although its expression was still higher than the non-injured

retinas. These evidences imply that 670nm light may effectively suppress rather than

mitigate gene expression. Transcription factors, AP-1 and NF-kB, which has been

shown to reduce inflammation in periodontal ligament cells at gene levels by PBM,

(Lee et al., 2018). In addition, our previous study showed that 670nm light changes the

gene expression of non-coding RNAs in normal healthy and PD rodent retina (Natoli

et al., 2010). Our unpublished data reported that miRNAs, as non-coding RNAs, were

modulated by 670nm light in PD rat retinas. These extensive studies suggest that 670nm

light may act on functional groups of genes by regulating their transcription.

3.5 Conclusion Our findings showed that the 670nm red light, commenced shortly after retinal injuries,

can modulate gliotic activity in MCs both in vitro and in vivo. Our data suggested that

670nm light may be beneficial to mitigate retinal damage, and slow the progression of

retinal degeneration even after the onset of disease. Studies in this chapter also show

that there is a therapeutic window, when 670nm light can be effectively used. Our data

demonstrate a potential use of 670nm light in conditions where MC activation is a

concern in retinal degenerative diseases, such as AMD, DR, and retinal detachment.

101

Chapter Four

Photobiomodulation with 670nm light ameliorates Müller cell mediated activation of microglia/macrophages in retinal degeneration

The results in this chapter is accepted in September 2017 and publish in Experimental Eye Research (Volume 165; Page 78-89; 2017).

102

4.1. Introduction Irradiation with low energy light wavelengths from far red to the near infrared spectrum

(600nm-1000nm), termed as PBM, has been shown to display beneficial effects on

various tissue injuries in animals models and clinical trials, including accelerated

wound healing of skin, decreased pain perception in joint disorders (Herranz-Aparicio

et al., 2013), reduced inflammation in autoimmune diseases (Brosseau et al., 2005), and

brain and spinal cord injury in animal models (Giacci et al., 2014; Hu et al., 2016).

PBM has been shown to be beneficial in human retinal diseases such as AMD, DR, and

RP (Abrahan et al., 2009; Albarracin et al., 2011; Geneva, 2016). Irradiation with

670nm light has been shown to protect photoreceptors from death in animal models of

retinal degeneration caused by hyperoxia, continuous bright light, or toxic substances

(Abrahan et al., 2009; Albarracin et al., 2011; Chu-Tan. et al., 2016; Nelson et al., 2011;

Wong-Riley et al., 2005) and to reduce neuronal cell death caused by toxin or

hyperglycaemia in vitro (Liang et al., 2008; Tang et al., 2013; Wong-Riley et al., 2005)

Functional analysis showed that 670nm light improved retinal response as determined

using electroretinography (ERG) in rat retinal degeneration (Albarracin et al., 2011;

Tang et al., 2013). Effectiveness of 670nm light to protect against photoreceptor death

has been shown to be dependent on the severity of retinal injuries (Chu-Tan. et al.,

2016). Regarding potential cellular signal pathways and mechanism of 670nm light

treatment, the widely supported hypothesis is that COX, which is the rate-limiting

enzyme in terminal phosphorylation in the mitochondria, is the most likely primary

photoacceptor of 670nm light (Desmet et al., 2006; Karu, 1999). Exposure to 670nm

light has been shown to enhance COX activity in retinas (Begum et al., 2013;

Kaynezhad et al., 2016) and primary neurons (Desmet et al., 2006; Wong-Riley et al.,

2005), mediating the increase of redox states in mitochondria (Kaynezhad et al., 2016),

increasing ATP production (Calaza et al., 2015; Gkotsi et al., 2014; Wong-Riley et al.,

2005), and upregulating mitochondrial membrane potential (Δψm) (Kokkinopoulos et

al., 2013).

The treatment with 670nm light has been linked to a reduction in retinal inflammation,

a key feature of several retinal diseases including AMD and DR (Whitcup et al., 2013).

670nm light suppressed the infiltration of MG/MΦ into the retinas in CFH-/- or aged

103

mice (Kokkinopoulos, 2013a; Kokkinopoulos et al., 2013) as well as inhibited

leukostasis in retinal blood vessels in DR (Tang et al., 2013). The morphology of

MG/MΦ remained ramified, indicating a resting state, in 670nm light-treated retinas

following retinal degeneration or aging (Albarracin and Valter, 2012; Begum et al.,

2013). Proinflammatory cytokines and complement components including TNFα and

C3 by MG/MΦ displayed lower expression by 670nm light in retinal degenerations

(Kokkinopoulos et al., 2013; Rutar et al., 2012a).

However, due to the lack of understanding of the precise cellular signalling events

during 670nm irradiation, there are still roadblocks in the translation of 670nm light

therapy to the clinic (Hamblin, 2016). Previously, our studies have demonstrated that

MCs are the key source of chemokine Ccl2 in retinal photo-oxidative damage (Rutar et

al., 2012b; Rutar et al., 2011b). Subsequently, it was found that activated MCs

expressed other chemokines and proinflammatory cytokines, which are responsible for

the recruitment and activation of MG/MΦ in retinas (Bamforth et al., 1997; Fischer et

al., 2014; Grigsby et al., 2014; Krady et al., 2005; Natoli et al., 2017b; Rivera et al.,

2013; Rutar et al., 2015). As interactions between MCs and MG/MΦ are correlated with

activation of a broad inflammatory response (Natoli et al., 2017b; Rutar et al., 2015;

Wang et al., 2011), we hypothesised that 670nm light may reduce MC-mediated

inflammation in retinal degeneration.

Our previous in vivo study suggested that pre-treatment with 670nm light mitigated

photo-oxidative stress-induced structural changes in MCs (Albarracin and Valter, 2012).

The aim of this study is to investigate the direct effects of 670nm light on MC activation

and its mediated proinflammation response following PD in vivo and in vitro. Our data

showed that 670nm light reduced the expression of proinflammatory cytokines in rat

primary MCs as well as in MC-mediated activation of rat primary MG/MΦ. In addition,

the underlying mechanism of 670nm light in rat primary MCs may be associated with

higher levels of COX5a and Δψm. Our data suggests that 670nm light may be beneficial

in reducing MC-mediated inflammatory changes in the retina, and could have

therapeutic potential in slowing the progression of retinal degenerations.

104

4.2. Materials and methods

4.2.1 Animals and light exposure Details of animal rearing and housing conditions are described in section 2.2. PD and

the treatment of animals with 670nm light were conducted according to previously

described methodology (section 2.3 and section 2.4). Briefly, prior to PD, animals were

held 2.5cm away from the 670nm LED array (Quantum Devices, WI, USA) to receive

~9 J/cm2 energy to both eyes once per day for 5 consecutive days. After the 5th 670 nm

light treatment, animals were then transferred to individual transparent cages and

exposed to PD. Following the 24-hour PD, whole eyes and retinas were collected for

histological analysis and RNA extraction. Age-matched SD rats, held in a dim

environment (5 lux), were used as dim-reared controls.

4.2.2 Maintenance of 661W photoreceptor-like cells Murine photoreceptor-derived 661W cells were used within five passages. The details

of cell culture are mentioned in section 2.5.1.

4.2.3 Preparation of rat primary MCs Rat primary MCs were isolated from retinae of SD rats aged P8-10 according to

previously described methodology (section 2.5.4). Rat primary MCs were used at the

second passage for all experiments in this chapter.

4.2.4 Photo-oxidative damage in co-culture of 661W cell line with rat primary MCs

For mimicking in vivo interactions between photoreceptors and MCs, a transwell co-

culture system was used. Primary MCs were seeded into 24-well plates at a density of

2.5x104 cells per well or 6-well plates at density of 2x105 cells per well in GM, and

plates incubated in 5% CO2 at 37°C for 48 hours to reach 90% confluence. In separate

plates, 661W cells were seeded onto the membranes of transwell inserts (pore size

0.4µm; Corning, NY, USA) at a density of 4x103 cells per insert (24-well transwell), or

at a density of 5x104 cells per insert (6-well transwell) in GM, and were incubated with

5% CO2 at 37°C for 24 hours to reach 80% confluence. Following incubations, both

primary MCs and 661W cells were washed with PBS and incubated in reduced-serum

105

DMEM (supplemented with 1% FBS, L-glutamine and antibiotic-antimycotic). The

inserts containing 661W cells were then transferred into the plates with primary MCs;

as such, 661W cells were co-cultured above the MCs. This co-culture system was

exposed to 15,000 lux light (1.5mW/cm2; irradiance measured with PM100D optical

power meter, THORLABS, NJ, USA) from two white fluorescent lamps (LFD10;

Crompton, NSW, Australia) for 4 hours in a humidified atmosphere with 5% CO2 at

37°C. The intensity of light was measured and monitored using a hand-held light meter

(HD450; Extech, MA, USA). Control plates were placed in the same incubator, but

shielded with aluminium foil to avoid light exposure. For air/gas exchange, small

incisions were made on the aluminium foil. Following 4 hours of photo-oxidative

damage (PD), co-cultures were incubated under dim light conditions with 5% CO2 at

37°C for 24 hours of recovery. After that, cells or supernatant were collected for

following analysis. The setting was presented in Figure 4.1.

4.2.5 670nm light treatment of co-cultures The 670nm LED array (Quantum Devices, details are referred to section 2.5.6) was

applied to different co-culture groups (with or without PD stress) as follows. (A) PD +

670nm group - inserts containing 661W cells were removed from the co-culture during

670nm light exposure. Only MCs were exposed to 670nm light and were treated three

times over the first 12 hours of recovery following PD. Energy of 670nm light delivered

was 60mW/cm2 (60mW/cm2 [~9J/cm2]) each time. (B) 670nm group - MCs received

treatment with 670nm light using the same paradigm, but with no PD. For performing

the sham treatment, (C) PD only group and (D) control group (control), the inserts

were removed from plates and only MCs were exposed to the 670nm LED array, but

with the light source switched off (Figure 4.1&4.2).

106

Figure 4.1 Schematic diagram of the experimental design MCs and 661W cells grew and reached confluency separately before the onset of the PD. 661W and MCs were co-incubated and then immediately exposed to the PD (yellow arrow). Then Co-cultured system of MCs and 661W was placed in the dim environment with 5% CO2 at 37°C (grey arrow) for the 24-hour recovery. MCs were moved and exposed to 670nm light for 3 minutes every session and then replaced back to the co-cultured system (red trapezoid). DIV: days in vitro.

Figure 4.2 Timeline of PD and 670nm light treatment Cells in each group were incubated with 5 % CO2 at 37°C. (A) PD + 670nm light group. (B) 670nm light group. (C) PD only. (D) Control group without PD or 670nm light.

107

4.2.6 Isolation and assessment of activation of rat primary MG/MΦ To isolate rat retinal MG/MΦ, a protocol previously published was modified for this

study (Fernando et al., 2016; Ma et al., 2013; Rutar et al., 2015) and described in section

2.5.5 in detail. Supernatant (SN) was collected from the co-cultured system of

MCs/661W for the following groups: 1. Control SN, 2. 670nm SN, 3. PD SN and 4.

PD+670nm SN. MG/MΦ cells were incubated with SN of these groups for 24 hours

with 5% CO2 at 37°C. MG/MΦ were then collected for RNA extraction, or

immunostained with CD11b/CD86 for flow cytometry. Following incubation with SN,

MG/MΦ were collected and fixed in 2% paraformaldehyde for 10 minutes on ice. Cells

were blocked with 1% BSA and then incubated with biotin CD86 (1:100; Biolegend),

or anti-rat CD11b PE antibody in 1% BSA for 30 minutes on ice. Cells were washed

once in PBS containing 0.2% Tween-20 and resuspended in the secondary antibody

with streptavidin conjugated with Alexa-Fluor-488 (S32453; ThermoFisher Scientific)

for 30 minutes on ice. Cells were washed once and resuspended in cold 0.1M PBS. The

expression of CD11b or CD86 was measured using FACSort (LSRII; BD, CA, USA)

and data was analysed using FlowJo (FLOWJO, OR, USA).

4.2.7 Assessment of cell viability and cell toxicity For measuring cell viability, an ATPlite 1 step assay (PerkinElmer, MA, USA) and an

MTT assay were used according to the manufacturer’s instructions (details are referred

to section 2.12). A CellTox assay (Promega, Madison, WI, USA) was performed to

measure cell death according to previously described methods (Natoli et al., 2016b).

Briefly, media was removed from each well and detection solution was added for 20

minutes at room temperature. Fluorescent cells (dead/dying cells) were counted using

488nmEX/525nmEM under an Axiovert 200 microscope (Zeiss, West Germany). The

number of fluorescent cells in 2-3 separate fields of each well was quantified and

averaged. The view of each field is 2mm in diameter.

4.2.8 Immunocytochemistry on rat primary MCs Coverslips (12mm diameter) were sterilised with 75% ethanol for 5 minutes, then

washed twice in sterilised water. Each coverslip was placed in a well of a 24-well-plate

and then incubated with the poly-L-lysine solution (100µg/ml; Sigma-Aldrich)

overnight at room temperature. Next day, coverslips were washed twice with sterile

108

water and then dried for at least 2 hours under UV light. MCs were then seeded onto

poly-L-lysine-coated glass coverslips at a density of 2.5X104 cells/mm2 for 48 hours

and then co-cultured with 661W cells. Cells either received PD, 670nm light treatment,

or both as described above. Coverslips containing MCs were fixed with 2%

paraformaldehyde for 30 minutes and then washed twice with cold 0.1M PBS.

Immunocytochemistry was performed as described previously (Albarracin et al., 2011;

Albarracin and Valter, 2012) and in section 2.9. The information of antibodies refer to

Table 2.2 and Table 2.4. Each well was incubated with primary antibody for COX5a

(1:500, #ab110262, Abcam) in 1% NGS at 4°C overnight in a humid environment and

then incubated in the secondary antibody, anti-mouse IgG (H+L) conjugated with biotin

(SAB3701153; Sigma-Aldrich) in 1% NGS for 90 minutes at RT in dim conditions.

Following incubation, cells were incubated with streptavidin conjugated with Alexa-

Fluor-488 (S32453; ThermoFisher Scientific) in 1% NGS in the dim light condition for

further 90 minutes at RT. Cell nuclei were counterstained with BBZ (Sigma-Aldrich).

Fluorescence was visualised using a laser-scanning A1+ confocal microscope (Nikon,

Tokyo, Japan) and captured with NIS-Element AR software (Nikon). The brightness

and contrast of images were adjusted for publication using Photoshop CS6 software

(Adobe Systems, CA, USA), which was standardised among images.

4.2.9 Quantitative polymerase chain reaction The preparation of RNA and the synthesis of cDNA from retinas and cells are described

in section 2.10. Gene expression was determined by RT-qPCR, using Taqman

hydrolysis probes (Table 2.7; ThermoFisher Scientific) and Taqman Gene Expression

Master Mix (ThermoFisher Scientific), which were applied according to the

manufacturer’s instructions by using a QuantStudio Flex 12K instrument

(ThermoFisher Scientific). Data analysis was performed using the comparative cycle

threshold method (ΔΔCt), which was normalised to the expression of the Gapdh

reference gene, which does not change in expression following retinal light damage

(Chen et al., 2004; Rohrer et al., 2007). Data is presented as fold change of the gene

expression compared to samples in the dim-reared control group in Figure 4.3,

compared to MCs of the control group (referred to section 4.2.5) in Figure 4.4, 4.5 and

4.6, or compared to MGs cocultured in the control SN (referred to section 4.2.6) in

Figure 4.8 .

109

4.2.10 In situ hybridization To localise Ccl2 mRNA expression in retinas, Ccl2 was cloned from PCR products

(550-bp amplicon) using cDNA synthesis from retinal RNA (as described above). A

digoxigenin (DIG)-labelled riboprobe for Ccl2 mRNA was designed and synthesised

by members of our lab (Rutar et al., 2011b). In situ hybridisation was used on retinal

cryosections as described previously (Cornish et al., 2005). Briefly, the Ccl2 riboprobe

was hybridized overnight at 55°C and then washed in a series of saline sodium citrate

solutions (pH 7.4) at 60°C. The bound probe was visualised using nitro-blue tetrazolium

chloride (NBT)/ 5-bromo-4-chloro-3'-indolyphosphate p-toluidine salt (BCIP). The

labeling was counted as Ccl2 positive cells across the retina. Four retinal sections were

counted for each animal. Each experimental group has 5 animals. The brightness and

contrast of images were adjusted for publication using Photoshop CS6 software (Adobe

Systems, CA, USA), which was standardised among images.

4.2.11 Enzyme-Linked Immunosorbent Assay (ELISA) Cell culture media were assayed for IL-6 (#R6000B, R&D Systems, MN, USA) and

Ccl2 (#ab100777, Abcam, Cambridge, UK) using a sandwich enzyme-linked

immunosorbent assay (Cornish et al., 2005). Supernatant was collected from the co-

cultured plates (MCs with or without PD) and stored at -80°C. Samples were analysed

in duplicate as per the manufacturer’s instructions.

4.2.12 Western blotting Whole cell protein lysates were extracted from MCs. The expression of COX5a in the

whole cell protein lysates were detected by using Western blotting. The expression of

COX5a was normalised to GAPDH (Begum et al., 2013; Walker and Steinle, 2007).

All details are referred to section 2.11.

4.2.13 Mitochondrial membrane potential (ΔΨm) JC-1 dye (Sigma-Aldrich) was used to assess changes in the mitochondrial membrane

potential (ΔΨm) of living cells. In cells with a high and polarized ΔΨm, JC-1 monomers

retained in the mitochondrial matrix formed aggregates that emitted red fluorescence.

With a low and depolarized ΔΨm, the JC-1 was unable to accumulate in mitochondria

and remained as monomers in the cytoplasm, and emitted green fluorescence. The ratio

110

of red to green fluorescence intensity was used to indicate the level of ΔΨm. MCs were

trypsinised, collected by centrifugation, then resuspended in warm HBSS containing

JC-1 staining dye (2µM) for 40 minutes in the dark in a humidified atmosphere of 5%

CO2 at 37°C. The excess of JC-1 dye was washed away with warm HBSS (37°C) once

and then resuspended in HBSS for analysis of flow cytometry (FACSort, LSRII). Data

were collected at 525nm emission for green fluorescence and 575nm for red

fluorescence and then analysed by using the software FlowJo. The level of ΔΨm is

presented as a ratio of red to green fluorescence intensity.

4.2.14 Statistical analysis Statistical analysis was performed using Prism 6 (GraphPad Software, San Diego, CA)

were analysed using one-way ANOVA with a Tukey's multiple comparisons post-test,

with P<0.05 considered to represent a statistically significant difference. All data is

represented as the mean ± SE

111

4.3. Results

4.3.1 670nm light suppressed cytokine expression following PD in vivo The expression of Il-1β and Ccl2 expression in rat retinas following 670nm treatment

and PD was examined. The expression of Il-1β and Ccl2 increased significantly in PD

retinas compared to dim-reared controls (P<0.05, Figure 4.3A, B). 670nm light-treated

animals had significantly reduced levels of Il-1β and Ccl2 expression in PD-retinas

compared to untreated PD retinas (P<0.05). Il-1β and Ccl2 expression was comparable

between 670nm-treated dim-reared animals and dim-reared controls (without 670nm

treatment or PD, Figure 4.3A, B). For localisation of Ccl2 in the PD retinas, Ccl2

expression was detected by using in situ hybridisation (Figure 4.3 C-F). A large number

of Ccl2-positive cells were apparent in INL of PD retinas (Figure 4.3E), where the

nuclei of MCs reside. Conversely, 670-treated retinas had fewer Ccl2-positive cells in

the INL in response to PD compared to untreated PD retinas (Figure 4.3F).

Quantification of the number of Ccl2-positive cells in the INL demonstrated

significantly fewer cells in the INL of 670nm light-treated PD-retinas compared to

untreated PD retinas (P<0.05, Figure 4.3 G).

4.3.2 670nm light reduced stress in primary MCs exposed to damaged 661W cells The effects of 670nm light treatment on primary MCs in response to 661W cells that

had undergone PD were assessed. The level of ATP, MTT and cell death in MCs were

unchanged following PD or 670nm light treatment (Figure 4.4B-D). Compared to MC

alone, the Gfap expression, a marker of gliosis, was upregulated in MCs co-cultured

with healthy 661W cells (no PD, data not shown). However, 670nm light did not

suppress this increase in MCs co-cultured with healthy 661W (no PD, data not shown).

Notably, 670nm-treated MCs exhibited significant down-regulation of Gfap, (P<0.05,

Figure 4.4 E) and an elevated expression of Rlbp1 compared to MCs under PD (with

degenerated 661W cells) (P<0.05, Figure 4.4 F). As Rlbp1 is required for maintaining

normal metabolic homeostasis in the retina, it is suggested that physiological function

of MCs is corrected by 670nm light in response to PD.

112

Figure 4.3 In vivo expression of cytokines was mitigated by 670nm light treatment following PD in rat retinas (A-B) PD significantly increased levels of Il-1β (A) and Ccl2 (B) in retinas compared to dim-reared controls (P<0.05). However, Il-1β and Ccl2 were significantly reduced after 670nm light treatment compared to PD retinas (P<0.05). (C-F) Ccl2 in situ hybridization revealed that numerous Ccl2-positive cells (arrows) were present in the INL of PD retinas (E) compared to dim-reared controls (C). After 670nm light treatment, a lower number of Ccl2-positive cells was detected in PD retinas (F). (G) 670nm light significantly reduced the number of Ccl2-positive MCs in the INL following PD (P<0.05). Scale bar stands for 25µm (A). GCL, ganglion cell layer; INL, inner nuclear layer; ONL, outer nuclear layer. N=5 was used for all experimental comparisons, except for (B) (N=10), and was performed in biological duplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

To further confirm whether 670nm light influences the function of MCs in PD, the

neuroprotective effect of MCs on 661W photoreceptor cells was measured. For 661W

cells (in the absence of MCs) with PD, the levels of ATP and MTT were decreased and

cell death was increased in 661W cells compared to control 661W cells (without PD or

670nm treatment; P<0.05, Figure 4.4 G-I). When co-cultured with MCs, 661W cells

had elevated levels of ATP and MTT and a reduced level of cell death following PD,

compared to 661W cells with PD only (P<0.05). In the control conditions (no PD),

treated MCs with 670nm light had no effect on MTT, ATP and cell death of 661W cells

(Figure 4.4 G-I).

25 μm

113

Figure 4.4 670nm light treatment reduced primary MC stress when co-cultured with 661W cells exposed to PD

(A) Timeline of experimental paradigm. (B-D) Neither PD nor 670nm light influenced ATP (B), MTT (C) and cell death (D) in MCs. (E) 670nm light reduced Gfap expression in MCs following PD (P<0.05). (F) Expression of Rlbp1 was elevated in 670nm-treated MCs compared to untreated MCs (P<0.05). (G-I) Following PD, co-culturing MCs (M) with 661W cells increased ATP (G) and MTT (H) and diminished cell death (I) in 661W cells compared to 661W cells cultured alone (P<0.05). ATP, MTT and cell death in healthy 661W cells were not affected by 670nm light-treated MCs. DIV; days in vitro. N=6 was used for all experimental comparisons and was performed in biological triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

114

4.3.3 670nm light suppressed oxidative stress and inflammation in MCs following PD

As MC-mediated inflammation is a key factor in the progression of retinal

degenerations, gene expressions of MC-related inflammation cytokines were examined

in MCs when co-cultured with 661W cells exposed to PD. Following damage, 670nm-

treated MCs had reduced Nox1 (NAPHD oxidase 1) and Nox4 (NAPHD oxidase 4)

gene expression (involved in the generation of reactive oxygen species, ROS),

compared to untreated MCs (P<0.05, Figure 4.5B, C). Following PD, Ccl2 and Il-6

gene expression reduced in 670nm-treated MCs compared to untreated MCs (P<0.05,

Figure 4.5 D, E). Similarly, secreted CCL-2 and IL-6 proteins in co-cultured

supernatants were significantly reduced in the PD+670nm group compared to the PD

group (P<0.05, Figure 4.5 F, G), suggesting that 670nm light regulates inflammatory

molecules in MCs.

4.3.4 670nm light treatment has no effect on NLRP3 inflammasome activation in MCs NLRP3 inflammasome activation has been implicated in retinal degenerations

including AMD (Doyle et al., 2012a; Tarallo et al., 2012). Our previous data have

shown that IL-1β production by MG/MΦ was promoted by chemokine expression by

retinal cells including MCs, and this was associated with inflammasome activation

(Natoli et al., 2017b). In this chapter, inflammasome-related gene expressions were

examined in 670nm light-treated MCs. It was found that 670nm light did not affect the

expression of Casp8, Nlrp3, Pycard, Casp1 and Il-18 in MCs following PD (Figure

4.6 B-F), although Il-1β was significantly reduced in 670nm-treated MCs compared to

untreated MCs following PD (P<0.05, Figure 4.6 G), correlating with the in vivo data

presented in Figure 4.1.

115

Figure 4.5 MC-derived inflammation was mitigated by exposure to 670nm light following PD (A) Timeline of experimental paradigm. (B-E) MCs exposed to 670nm light had significantly reduced Nox1 (B), Nox4 (C), Ccl2 (D) and Il-6 (E) expression compared to untreated MCs (P<0.05). (F&G) Ccl2 and IL-6 in co-culture supernatant was significantly decreased by 670nm light post-PD (P<0.05). DIV; days in vitro. N=6 was used for all experimental comparisons and was performed in biological triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

116

Figure 4.6 Effects of 670nm light on activation of the NLRP3 inflammasome in MCs following PD (A) Timeline of experimental paradigm. (B-F) Expression of inflammasome genes including Casp8 (B), Nlrp3 (C), Pycard (D), Casp1 (E), and Il-18 (F) were comparable between 670nm-treated and untreated MCs following PD (P>0.05). (G) Il-1β expression was significantly lowered in 670nm-treated MCs compared to untreated MCs (P<0.05). DIV; days in vitro. N=6 was used for all experimental comparisons and was performed in biological triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

117

4.3.5 Mitochondrial membrane potential was enriched by 670nm light

following PD Cytochrome c oxidase (COX) is the terminal component of the electron transport chain

in mitochondria and is the most likely photoacceptor of 670nm light (Desmet et al.,

2006; Diaz, 2010; Karu, 1999). COX5a is one of 13 subunits of the COX enzyme in

mammals. To understand whether COX participates in the effects of 670nm light on

MCs, the expression of COX5a and Δψm were measured in MCs. Low expression of

COX5a was detected in the cytoplasm of the soma in MCs co-cultured with healthy

661W cells (Figure 4.7 B). Upon treatment with 670nm light or the exposure of PD,

the expression of COX5a was enriched and scattered in the cytoplasm of MCs.

Following PD, COX5a staining was robustly increased in 670 nm-treated MCs

cytoplasm. For quantification of COX5a expression, Western blotting for COX5a was

performed in MCs. In the MCs w/wo PD, 670nm-treated cells had a higher expression

of COX5a than untreated MCs (without PD or 670nm light) (p<0.05, Figure 4.7 C, D),

confirming that the mechanistic effects of PBM. However, the finding that exposure to

PD caused a higher expression of COX5a in MCs compared to untreated MCs (without

PD or 670nm light) (P<0.05) was unexpected. Its underlying mechanism needs further

investigation.

The mitochondrial membrane potential (Δψm) was assessed by JC-1 polarisation assay

in the matrix of mitochondria. The Δψm significantly dropped in MCs after PD

compared to control MCs (no PD or 670nm light; P<0.05, Figure 4.7 E). MCs treated

with 670nm light had a higher Δψm in response to PD than untreated MCs with PD

(P<0.05, Figure 4.7 E). This confirmed that 670nm light affects the activation of

mitochondria in MCs following PD.

118

Figure 4.7 The activation of cytochrome oxidase (COX5a) was enhanced in 670nm-treated MCs after PD (A) Timeline of experimental paradigm. (B) MCs were immunolabelled with COX5a antibody (green). Low levels of COX5a were detected in MCs co-cultured with normal 661W cells. Following PD or 670 nm light, a higher intensity of COX5a was detected in MCs compared to controls. The intensity of COX5a was further elevated by 670nm light in MCs compared to untreated MCs after PD. (C-D) Western blotting quantification relative to GAPDH expression confirmed the higher expression of COX5a by 670nm light-treated MCs with/wo PD (P<0.05). (E) The JC-1 ratio in mitochondria to cytoplasm was used to measure Δψm. Following PD, 670nm light triggered a higher level of Δψm in 670nm light-treated MCs compared to untreated MCs (P<0.05). DIV; days in vitro. N=6 was used for all experimental comparisons and was performed in biological triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

119

4.3.6 670nm light regulated MC-mediated activation of MG/MΦ following PD After incubation with supernatant ( SN ) from the co-cultured MCs and normal 661W

cells (control SN, Figure 4.8 B), or 670nm-treated MCs and normal 661W cells (670nm

SN, Figure 4.8C), primary MG/MΦ displayed a ramified morphology with defined

processes, indicating a resting state. However, the shape of MG/MΦ became more

amoeboid with shorter processes following incubation with SN from MCs co-cultured

with PD 661W cells (PD SN, Figure 4.8D), indicating activation of these cells. After

incubation with SN from 670nm light-treated MCs, co-cultured with PD 661W cells

(PD+670nm SN, Figure 4.8E), MG/MΦ were more ramified and processes extending

from cell soma were apparent, indicating that 670nm light reduced activation of

MG/MΦ.

For quantification, flow cytometry was used to detect the fluorescence intensity of

CD11b (Figure 4.8 F-G) and CD86 expression (Figure 4.8 H-I) in MG/MΦ stimulated

with differing SN. The histograms demonstrated that MG/MΦ tended to have a higher

expression of CD11b and CD86 after incubation with PD SN compared to cells

incubated with PD+670nm SN (P<0.05). A higher mean intensity of fluorescence (MIF)

of CD11b and CD86 (indicators of activation) was detected in MG/MΦ incubated with

PD SN compared with cells cultured with the PD+670nm SN (P<0.05, Figure 4.8 F-I).

The expression of MG/MΦ activation-related genes including Il-1β, Ccl2, Tnfα, Il-6,

Il-10 and Sod2 (superoxide dismutase 2), were analysed in MG/MΦ (Figure 4.8J-O).

The expression of these genes was significantly lowered in MG/MΦ incubated with

PD+670nm SN compared to PD SN (P<0.05). This indicates that 670nm light

ameliorated MC-derived inflammation following PD, influencing the activation state

of MG/MΦ.

120

Figure 4.8 670nm light ameliorated MC-mediated activation of MG/MΦ following PD (A) Timeline of experimental paradigm. (B-E) After incubation with PD SN, MG/MΦ cells displayed an amoeboid morphology. However, a ramified shape was observed in MG/MΦ incubated with SN of the PD+670nm group. (F-I) A higher mean of fluorescence intensity (MFI) of CD11b (F, G) and CD86 (H, I) in MG/MΦ. (J-O) Gene expression of Ccl2 (J), Il-1β (K), Tnfα (L), Il-6 (M), Il-10 (N), and Sod2 (O) in MG/MΦ incubated with SN of different groups. DIV; days in vitro. N=6 was used for all experimental comparisons and was performed in biological triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

121

4.4. Discussion Treatment with 670nm red light has been shown to reduce inflammation in retinal

diseases (Geneva, 2016). However, there is still a lack of understanding of the precise

cellular signalling pathways that are activated upon exposure to 670nm light. The

current study offers insight into the cellular signalling pathways influenced by 670nm

light during retinal degeneration, and demonstrates that treatment with 670nm light

reduced MC-mediated inflammation in PD, which in vivo induces oxidative stress that

leads to focal retinal degeneration and inflammation (Natoli et al., 2016a; Rutar et al.,

2011a; Rutar et al., 2011b; Rutar et al., 2010). Firstly, 670nm light suppressed

inflammation by reducing the expression of proinflammatory cytokines, chemokines

and, oxidative stress components in MCs exposed to damaged photoreceptors.

Secondly, 670nm light reduced activation of MG/MΦ by influencing the expression of

inflammatory activators by MCs. Thirdly, our data demonstrate that a potential

mechanism of action of 670nm light that MCs involved enhancement of Δψm, an

increase in the expression of COX5a, and the augmentation of its metabolic function

(Rlbp1) following PD.

4.4.1 670nm light reduced MC-mediated activation of MG/MΦ Our current data showed that 670nm light directly mitigated the expression of IL-1β

and CCL2 by primary MCs in retinal damage. These results are consistent with previous

investigations that MCs are a key source of Ccl2 and IL-1β (Liu et al., 2012; Ni et al.,

2008; Walker and Steinle, 2007; Wang et al., 2011) in models of photo-oxidative retinal

degeneration (Ni et al., 2008; Rutar et al., 2012b; Rutar et al., 2011b), DR (Kowluru

and Odenbach, 2004), and retinal ischemia (Yoneda et al., 2001). This indicates that

670nm light targets the expression of proinflammatory cytokines and chemokines by

MCs.

Higher levels of IL-1β and Ccl2 have been linked to proliferative DR as well as AMD

(Anand et al., 2012; Kowluru and Odenbach, 2004; Spandau et al., 2003; Xu et al.,

2015; Zhao et al., 2015b). These proinflammatory factors, as well as IL-6 are known to

mediate the activation and recruitment of retinal MGs and blood-borne MΦ (Bamforth

et al., 1997; Fischer et al., 2014; Grigsby et al., 2014; Krady et al., 2005; Rivera et al.,

2013; Rutar et al., 2015; Rutar et al., 2012b; Wang et al., 2011). Our previous

122

publication have demonstrated that inhibition of Ccl2 and IL-1β reduces the infiltration

MG/MΦ into the damaged outer retina in photo-oxidative retinal degeneration (Natoli

et al., 2017b; Rutar et al., 2012b). The activation and accumulation of MG/MΦ in the

subretinal space is a well-established feature of retinal degenerations including AMD

(Fernando et al., 2016; Knickelbein et al., 2015), which is associated with photoreceptor

cell death (Natoli et al., 2017a; Ni et al., 2008; Rivera et al., 2013; Zhao et al., 2015a).

670nm light mitigated MC expression of Ccl2, IL-6 and IL-1β, as well as several other

inflammatory factors were observed in retinal degeneration, and that this led to a

reduced activation of MG/MΦ (Kobayashi et al., 2013; Roy et al., 2006). It agrees with

our previous studies that 670nm light resulted in reduced activation and recruitment of

MG/MΦ in retinal damage (Albarracin et al., 2011; Begum et al., 2013; Kokkinopoulos,

2013a; Kokkinopoulos et al., 2013). Other studies have demonstrated the interplay

between MCs and microglia in retinal degeneration, namely that activated MCs were

capable of inducing MG activation and further production of proinflammatory

cytokines including IL-1β (Natoli et al., 2017b; Rutar et al., 2015; Wang et al., 2011).

Inflammasome has been linked to the progression of retinal degenerations such as AMD

(Doyle et al., 2012a; Tarallo et al., 2012), and generation of IL-1β is one of the direct

consequences of NLRP3 inflammasome activation (Ambati et al., 2013; Schroder and

Tschopp, 2010). In the current study, the down-regulation of IL-1β by 670nm light in

MCs in response to PD was observed. However, 670nm light did not influence the

expression of components of the NLRP3 inflammasome in MCs, indicating that the

production of IL-1β may be inflammasome-independent in some retinal cell types

(Netea et al., 2015). Moreover, it is thought that the RPE and MG/MΦ are the primary

retinal cells contributing to inflammasome activation in AMD, not MCs (Doyle et al.,

2012a; Kataoka et al., 2015; Kauppinen et al., 2016).

4.4.2 670nm light modulated MC stress in response to damage In the current study, 670nm light reduced the expression of Gfap in MCs following PD,

which was consistent with previous in vivo studies of PD (Albarracin et al., 2011;

Marco et al., 2013) and ageing (Begum et al., 2013). These effects were accompanied

with the reduction of photoreceptor cell death (Albarracin et al., 2011; Marco et al.,

123

2013). However what is unclear is whether 670nm light directly regulates MC stress,

or indirectly influences the activation of MCs caused by the damaged photoreceptors

(Abrahan et al., 2009). In this study, only MCs were exposed to 670nm light, not 661W

photoreceptors. And the down-regulated Gfap expression in MCs, a well-known

marker of MC gliosis in retinal diseases, was observed. Mice, deficient in GFAP

exhibited less glial scars after retinal detachment, followed by a lower level of MG/MΦ

recruitment and MCP-1 expression (Nakazawa et al., 2007b; Verardo et al., 2008). This

reduction may lead to a consequence decrease of Ccl2 and monocyte infiltration into

retinas following damage (Nakazawa et al., 2007b). Conversely, an increase in Rlbp1

in MCs following 670nm light treatment was observed, compared to PD damaged MCs.

RLBP1 is required for the maintenance of normal metabolic function (Taylor et al.,

2015), so the increase towards control levels suggests that treatment with 670nm light

supported MCs in maintaining their metabolic, and possibly other cellular functions, to

withstand the tissue stressed caused by PD and thereby to protect photoreceptors.

Moreover, the loss of Rlbp-1 has been proved to be consistent with proliferative MCs

in vitro (Abrahan et al., 2009; Guidry, 1996), implying that MCs initiated

dedifferentiation and re-entered the into cell cycle. Taken together, the increase or

maintenance of expression in Rlbp-1 might correlate with the less extent of MC gliosis

in our model in this chapter.

4.4.3 The change in Δψm on regulating gene expression This study showed that the treatment of 670nm light increased the Δψm and expression

of COX5a in MCs under PD. Cytochrome c oxidase (COX) is complex IV of the

mitochondrial respiration chain as known to absorb energy from far red to near infrared

light (Schroeder et al., 2007). This can trigger a higher Δψm which generates ATP in

cells (Verardo et al., 2008). 670nm light has been shown to increase the level of

oxidized COX, where the up-regulated activity of COX lasted 1-2 hours after 5 minutes

of exposure to 670nm light (Kaynezhad et al., 2016). Elevated production of ATP and

higher protein expression of COX were also observed in retinas in other models of

retinal degeneration (Begum et al., 2013; Gkotsi et al., 2014). The Δψm was reported

to be increased in RPE cells in response to 670nm light in aged mice (Begum et al.,

2013; Kokkinopoulos et al., 2013). The current study is consistent with these results,

and although we did not detect a higher level of ATP production in 670nm light-treated

124

MCs, the higher levels of Δψm and COX5a were observed in this study. Therefore, we

speculate that ATP production is a consequence of up-regulated Δψm and COX5a.

PBM is known to potentially influence cellular DNA and RNA synthesis (Karu, 1999).

As photoacceptors are located in the mitochondria, Karu’s laboratory proposed that

mitochondrial retrograde signalling can be triggered by PBM in cells (Karu, 2008). The

mitochondrial retrograde signal is broadly defined as cellular responses to changes in

the functional state of mitochondria (Schroeder et al., 2007). The change of Δψm, ROS,

and calcium mobilisation may be involved in mitochondrial retrograde signalling in

cells by PBM, subsequently regulating DNA and RNA expression in the cell nucleus.

Indeed, growing evidence suggests that increased Δψm is related to changes in gene

expression. Δψm was increased in human keratinocytes immediately after 780 nm light

exposure, but reduced gene expression of IL-6 was measured at 2 hours following 780

nm light exposure (Gavish et al., 2004), suggesting the change of Δψm occurred prior

to the alternation of gene expressions. Treatment with an inhibitor of the electron

transport chain abolished gene expression changes in fibroblasts caused by irradiation

with 760 nm or 1140nm light (Schroeder et al., 2007). Therefore, in the current study,

the change in Δψm may participate in the regulation of gene expression in MCs exposed

to 670nm light following PD. However, the full effects of 670nm light on mitochondrial

retrograde signalling in MCs remains to be further explored.

4.5. Conclusion

Our findings suggest that 670nm light can directly affect MCs by mitigating their stress-

induced inflammatory reaction, which subsequently lead to the reduction of MG/MΦ

recruitment and activation. Further, the maintenance of mitochondrial function in MCs

may enhance their tissue support function, and thereby lead to the neuro-protective

effect of 670nm light. These data confirm our previous in vivo data and suggest that

670nm light plays a key role in controlling retinal inflammation during stress. As a non-

invasive and relatively non-expensive treatment, 670nm light has adjuvant therapeutic

potential for retinal degenerations where inflammation, MG/MΦ recruitment, and

photoreceptor loss play a key role.

125

Chapter Five

670nm red light regulates retinal Müller cell-mediated inflammation through

extracellular vesicles

126

5.1 Introduction The interaction between MCs and MGs is becoming more apparent in vision diseases.

The activation and recruitment of MG/MΦ and the upregulation of proinflammatory,

that are all hallmarks of AMD, glaucoma, and DR, can be linked to uncontrolled MC

activation (Albarracin and Valter, 2012; Bringmann et al., 2009; Bringmann et al., 2006;

Rutar et al., 2012b). On the other hand, activated MGs present a positive feedback loop,

which further activate MCs (Wang et al., 2011). Our previous research demonstrated

that MG-derived IL-1β activated MCs in vivo (Natoli et al., 2017b).

EVs are regarded as the third intercellular communication pathway (Skog et al.), in

addition to cell-to-cell contacts and secreted soluble factors (Minciacchi et al., 2015;

Raposo and Stoorvogel, 2013). EVs are traditionally categorised by their size; vesicles

over 1000nm are described as macrovesicles (Raposo and Stoorvogel, 2013), while

smaller vesicles are further distinguished to two major sub-populations, microvesicles

(MVs) (100-1000nm) and exosomes (30-100nm) (Buzas et al., 2014; Gyorgy et al.,

2011; Lee et al., 2012). The latter two groups have gained more attention due to their

ability to transport genetic material, including miRNAs and mRNAs between cells

(Hajrasouliha et al., 2013; Lee et al., 2012; Ludwig and Giebel, 2012; Raposo and

Stoorvogel, 2013; Zhang et al., 2015). EVs have been shown to be released by a number

of cell types and are found in body fluids (serum, urine, breast milk, and saliva) in both

physiological and pathological conditions (Evans-Osses et al., 2015; Raposo and

Stoorvogel, 2013), including in AMD. CD81 and LAMP1, both specific EV markers,

have been identified in drusen in AMD eyes in humans (Wang et al., 2009). In this

chapter we investigated whether MC communicate via EVs to regulate immune

responses in MGs and whether miRNAs are the key agents in this process.

MiRNAs are a class of small, non-coding RNAs [17-24 nucleotides (nt) long], which

silence or degrade target mRNAs by binding to the 3’ untranslated-region (UTR) or the

open reading frame of target mRNAs (Bartel, 2004). The paring of mature miRNA-

mRNA is mediated with a partial sequence complementarity by a 6-8nt long region at

the 5’ end of miRNA called the ‘seed’ (Nigita et al., 2016). A single mRNA can be

targeted by several miRNAs, and a single miRNA might target multiple mRNAs (Wang

et al., 2012), which may modulate the translation of a large part of the transcriptome to

127

manipulate biological function (Nigita et al., 2016). MiRNAs have been shown to play

crucial roles in pathological processes, including angiogenesis, oxidative stress, and

inflammation (Natoli et al., 2010; Saxena et al., 2015; Wang et al., 2012). These

features are all important hallmarks in the AMD pathogenesis (Ambati and Fowler,

2012).

Recent report showed that a group of miRNAs is highly expressed in MCs in normal

mouse retina. However they did not report on the function of these miRNAs, and neither

did they investigate whether this expression pattern changes in activated or injured MCs

(Wohl and Reh, 2016). Thus, the roles of miRNAs in regulating the function of MCs

are still unclear. MiRNA-124 has been linked to the expression of MCP-1 (Ccl2)

(Nakamachi et al., 2009), which was mainly released by MCs (Rutar et al., 2012b; Rutar

et al., 2011b). Our previous study found that miRNAs were presented in the rodent

retina and its expression was significantly alternated in PD (Saxena et al., 2015). These

data implicated that miRNAs may play a role in regulating MC activation and the

interaction between MCs and other retinal cell types in retinal diseases. In this chapter,

IL-1β was used to stimulate the human Müller cell line (MIO-M1) for investigating 1)

the potential of EV secretion by both resting and activated MIO-M1 cells, 2) whether

these EVs contain miRNAs and 3) the effects of MC-derived EVs on the activation of

MGs. Furthermore, the examined whether treatment with 670nm light has effects on

these responses.

128

5.2 Material and methods

5.2.1 MIO-M1 stimulation with IL-1β Characteristics and maintenance of MIO-M1 cell line are referred to section 2.5.2.

MIO-M1 cells was seeded in 6-well plates at a density of 200, 000 cells per well in GM

for 24 hours. Cells were washed once with fresh 0.1M PBS and were incubated in

DMEM supplemented with 1% FBS, 3% L-glutamine and 1% antibiotic-antimycotic

for a further 24 hours. To activate MIO-M1 cells, culture medium was replaced with

10ng/ml IL-1β (Miltenyi Biotec, MACS, Bergisch Gladbach, Germany) in 1% FBS

DMEM and cells were incubated in the dark at 37°C with 5% CO2 for further 24 hours.

The culture medium was then used for EV collection.

5.2.2 670nm light treatment paradigm and experimental groups MIO-M1 cells were irradiated with 670nm red light (670nm) for 3 minutes (~9J /cm2)

three times within the first ten hours of IL-1β stimulation, performed as follows. The

first treatment of 670nm commenced immediately after the onset of the IL-1β

stimulation. The second irradiation of 670nm light was delivered 4 to 5 hours later.

Cells were treated the third time after 10 hours of IL-1β stimulation. To compare the

effect of IL-1β and 670nm light on the secretion of EVs, four experimental groups were

defined as follows: 1) MIO-M1 cells without IL-1β or 670nm treatment (control group);

2) MIO-M1 cells treated with IL-1β (IL-1β group); 3) 670nm treated MIO-M1 cells

without IL-1β (670nm group); 4) MIO-M1 cells incubated with IL-1β and then

exposed to 670nm irradiation (670nm + IL-1β group). In the first and second groups,

cells were receive an identical treatment to the 670 groups (group 3 and 4); however

with the LED array remaining switched off (Figure 5.1).

Figure 5.1 Timeline of IL-1β stimulation and 670nm light treatment

Cells were treated with 670nm light after the commence of IL-1β challenge within 10 hours. Samples (culture medium) were harvested after the 24-hour IL-1β stimulation.

129

5.2.3 Morphological analysis of MIO-M1 cells The morphology of MIO-M1 cells was observed under a phase-contrast light

microscope Zeiss Axiovert 3 (Zeiss), and images were captured by using the ProgRes®

CapturePro microscope camera software (JENOPTIK, Jena, Germany). To characterise

and quantify the change in morphology, the size and shape of individual cells was

analysed using ImageJ (Rasband, W.S., ImageJ, U. S. National Institutes of Health,

Bethesda, Maryland, USA) (Wang et al., 2011). Images were analysed using “Analyze

Particle” function, which automatically measures the area and a perimeter of an

individual cell. The particle size setting was defined as between 500-2000 (pixel2) to

exclude cell debris and cell aggregates. The circularity of MIO-M1 was defined and

calculated by using the equation (4π(area of a cell)/(cell perimeter2)), where a higher

value indicates a circular shape. Alternatively, a smaller value of circularity in a cell

indicates an elongated shape. The data of circularity of MIO-M1 in each group is

collected and analysed from 10 to 12 images, captured from three respective wells using

a phase-contrast light microscope Zeiss Axiovert 3 (Zeiss). Data is presented as the

mean ± SEM.

5.2.4 Assessment of cell viability on MIO-M1 cells To measure MIO-M1 cell viability and cellular metabolism, cells were seeded into 96-

well plates at the density of 5000 cells per well in GM. IL-1β (10 ng/ml) stimulation

was performed as described above. Cell survivability was measured by using an

ATPlite 1 step assay (PerkinElmer, MA, USA) and an MTT assay (mentioned in section

2.12)

5.2.5 Ultracentrifugation-isolated MVs MVs were isolated from the cell supernatant of each group by using sequential

centrifugation procedures which are modified based on previous publications

(Hajrasouliha et al., 2013; Koliha et al., 2016; Pospichalova et al., 2015; Wang et al.,

2009). The supernatant from MIO-M1 cell culture was collected and then purified by

removing cell pellets, cellular debris and apoptotic cells after centrifugation for 10

minutes at 300xg and then for 30 minutes at 2000xg. To remove larger EVs, the

supernatant was centrifuged for further 70 minutes at 10,000xg. Then the supernatant

130

was harvested and filtered through a 0.22µm membrane to remove remaining large

vesicles. The filtered supernatant was then centrifuged for 4 hours at 100,000xg at 4°C.

The pellet, containing MVs, was resuspended in chilled PBS and stored at -80 °C until

analysis. For MV miRNA extraction, the pellet was resuspended in Lysis/Binding

buffer (mirVana miRNA isolation kit, Ambion Biosystems) and then MV miRNA

extraction was performed according to manufacturer’s instruction. To avoid the

differential efficiency of recovering MVs by ultracentrifugation, supernatant collected

from different groups were placed in the same centrifuge rotor and then spun down at

the same time.

5.2.6 Microbead-captured exosomes Cell culture medium was collected from four experimental groups. The isolation of

exosomes from the supernatant was carried out using the Exosome Isolation Kit Pan

(MACS, Germany) by following the manufacturer’s instructions. First, the supernatant

was centrifuged with serial centrifugation at 300xg for 10 minutes, 2000xg for 30

minutes, and 10,000xg for 45 minutes to remove cell debris and large vesicles,. The

supernatant was aspirated after each centrifugation step and then transferred to a new

tube. After the last centrifugation, the supernatant (4ml) was mixed with 50 µL of

Exosome Isolation Microbeads recognizing CD9, CD63, and CD81 antigens. The

mixture was placed in the roller mixer and then incubated at RT for 2 hours in the dark.

Then the mixture passed through the equilibrated µ column, which was placed in a

µMACS Separator (Miltenyi Biotec), to collect CD9, CD63, or CD81+ MVs. Steps are

as follows; 1) the µ column was equilibrated with 100µL of equilibration, then 3 times

100µL of isolation buffer. 2) the mixture was slowly passed through the equilibrated µ

column. 3) the µ column was washed with 200µL of isolation buffer four times. 4) The

µ column was removed from the µMACS Separator and then placed onto 1.5-ml tubes.

The microbeads conjugated with MVs (CD9, CD63, or CD81+) were eluted with

100µL of isolation buffer by firmly pushing the plunger into the µ column. 5) The

eluted MVs (CD9, CD63, or CD81+) were stored in -80°C until analysis.

5.2.7 MV isolation and purification from retinas Procedures are modified with previous publications (Perez-Gonzalez et al., 2012;

Pospichalova et al., 2015) and a flow chart of the protocol is presented in Figure 5. 2.

131

Retinas were collected from either dim-reared animals (control) or post-PD (details

refer to section 2.3.2). Fresh mouse retinas were harvested and immersed in cold HBSS.

The retinas were roughly fragmented into small pieces by scissors and then mildly

treated with 20 units/ml papain (Worthington; Lakewood, NJ, USA) in 500µl of HBSS

at 37 °C for 20 minutes. Then tissue fragments were gently triturated with a pipette.

Notably, the papain treatment and gentle trituration are needed to prevent

contamination from intracellular organelles vesicles released from ruptured cells. To

stop the reaction, the tissue suspension was diluted in 10 volumes of cold HBSS (5ml).

The diluted tissue suspension was filtered sequentially through a 40-µm mesh (BD,

Biosciences, NJ, USA), and a 0.2-µm syringe filter (Minisart, Sigma-Aldrich) to

remove cell debris and large vesicles. The filtrate was sequentially centrifuged at 300xg

for 10 minutes, 2000xg for 10 minutes, and 10, 000xg for 45 minutes at 4°C to remove

remaining cell debris and large vesicles. After this, the supernatant was collected and

then centrifuged at 100,000xg for 4 hours at 4°C to collect MVs. To further purify MVs,

pellets containing MVs were resuspended in 0.95M sucrose solution (1.8ml) in 20mM

Tris buffer (pH 7.6). MV suspension in 0.95M sucrose solution was inserted inside a

sucrose step gradient column (six 1.8-ml steps starting from 2M sucrose up to (bottom)

to 0.25M (top) in 0.35M increments). The sucrose step gradient was then centrifuged

at 200, 000xg for 16 hours at 4°C. The fractions (b, c, and d) (1.8 ml X 3) (Figure 5.2)

were pooled together into a clean vial. This mixture was then diluted up to 35 ml with

cold 0.1M PBS and then centrifuged at 100,000 g for 4 hours at 4°C to recover purified

MVs. MVs, collected from sucrose gradient fraction pellets, were resuspended in cold

Annexin V binding buffer for Annexin V staining (referred to section 5.2.8.1).

132

Figure 5.2 Retinal MV isolation experimental flow chart The steps of the experimental procedures were designed to isolate and purify mouse retinal MVs.

133

5.2.8 Annexin V binding/staining

5.2.8.1 MVs isolated by ultracentrifugation MVs were isolated from cell culture medium or retinas with the ultracentrifugation

approach and then suspended in Annexin V binding buffer (10mM HEPES, 140mM

NaCl, and 2.5mM CaCl2 in 2dH2O, pH 7.4). The suspension was incubated with 10µL

of Annexin V-conjugated to PE (MACS) for 20 minutes in the dark at RT. To stop the

reaction, an equal volume of Annexin V binding buffer was added. The intensity of

Annexin V binding on MVs was measured by flow cytometry using a BD Fortessa (BD,

CA, USA). MV size was determined by comparing to nanofluorescent polystyrene

particle beads (450 nm and 880 nm). Data was analysed using FlowJo (FLOWJO, OR,

USA) and presented as the mean of the intensity of Annexin V per particle in each

sample.

5.2.8.2 Exosomes isolated by MACS Exosome Isolation Kit Pan

Microbead-conjugated MVs were centrifuged at 10,000xg for 10 minutes at 4°C. The

supernatant was discarded, and the pellet was re-suspended in 100µL of sterilized

Annexin V binding buffer containing 10µL of Annexin V-conjugated to PE. The

mixture was incubated for 20 minutes in the dark at RT. Then 1 ml of Annexin V

binding buffer was added to stop the reaction. The mixture was centrifuged at 10,000xg

for 10 minutes at 4°C to wash out excessive Annexin V-conjugated to PE. The pellet

was resuspended in 150µL of Annexin V binding buffer and then was analysed by a

flow cytometry (BD LSRII, BD). Data was analysed using FlowJo (FLOWJO) and

presented as the mean of the fluorescence intensity of Annexin V per particle in each

sample.

5.2.9 Detection of exosome surface markers

The expression of surface epitopes was determined by a flow cytometry using the

MACSPlex Exosome Kit (Miltenyi Biotec, Bergish Gladbach, Germany). All

procedures are referred to the manufacturer's instruction and the previous publication

(Koliha et al., 2016). First of all, 120µL of exosomes isolated from MACS Microbeads

or 100µL of exosomes isolated by ultracentrifugation was mixed with 15µL of

MACSPlex Exosome Capture Beads. The mixture was incubated in the dark overnight

on an orbital shaker at RT. The mixture containing Capture Beads conjugated to

134

exosomes was washed with 1 ml of MACSPlex Buffer and then centrifuged at 3000xg

for 5 minutes. The supernatant was discarded, 135µL was left in the tube to incubate

with 15µL of MACSPlex Exosome Detection Reagent cocktail (CD9, CD63, and

CD81) at RT for 1 hour in the dark. To stop the reaction, 1 ml of MACSPlex buffer

was added into the tube and then incubated at RT for another 15 minutes. The mixture

was centrifuged at 3000xg for 5 minutes, and all but 150µL of the supernatant was left.

The Exosome Capture Beads were resuspended by pipetting and then analysed by a

flow cytometry using a flow cytometry (BD LSRII). Each population of Capture Beads

is coated with a specific antibody binding the respective exosome surface epitopes

(surface markers). Thirty-nine populations of Capture Beads, including two types of

isotypes, can be distinguished by different fluorescence intensities in FITC and PE

channels. Exosomes bound to Capture Beads were detected with a cocktail of

antibodies (CD9, CD63 and CD81) conjugated with APC. Thirty-nine signal Capture

Bead populations were gated to allow the determination of the APC signal intensity in

the respective Capture Beads population. The data was analysed by using the software

FlowJo (FLOWJO) and was presented as the APC medians in each Capture Beads

population.

5.2.10 Analysis of CD3 and CD81 expression on MVs MVs were isolated by ultracentrifugation and then re-suspended in chilled PBS

containing 0.5% BSA for one hour on ice to block non-specific binding.

Immunostaining was used to detect the expression of CD3 on the MVs. The procedures

are as follows. MVs were incubated with mouse anti-human CD3 conjugated with FITC

(1:200, #300405, BioLegend, USA) for one hour on ice in the dark. An equal volume

of chilled PBS was mixed with the suspension to stop the reaction. The CD81

expression on MVs was measured by indirect immunostaining. Primary antibody

mouse anti-human CD81 (1:100, sc-166029, Santa Cruz, USA) was incubated with the

MVs for one hour on ice. The secondary antibody-APC/Cy7 goat anti-mouse IgG

(minimal x-reactivity) (1:500, #405316, BioLegend, USA) was added to incubate for

another hour on ice in the dark. To stop the reaction, an equal volume of chilled PBS

was mixed with each sample. The intensity of CD3 or CD81 in cells was measured by

a flow cytometry (BD Fortessa, BD) and the data were analysed using FlowJo

135

(FLOWJO). The data is presented as the percentage of CD3+ or CD81+ of MV

population in each sample.

5.2.11 MV miRNA extraction MiRNA was extracted from ultracentrifugation-isolated MVs using the mirVana

miRNA isolation kit (Ambion Biosystems) following the manufacturer’s instruction.

Steps are as follows. MVs were lysed with 300µL of Lysis/Binding Buffer and 30µL

of miRNA Homogenate Additive (1/10 volume). This solution was incubated on ice for

10 minutes and then mixed with 300µL of Acid-Phenol: Chloroform (AmbionTM,

ThermoFisher Scientific) (an equal volume to initial lysate volume). The mixture was

centrifuged at 13,000xg for 10 minutes at 20°C. An aqueous layer was transferred into

a clean 1.5mL tube and then combined with 1.25 volume of 100% ethanol with well

mixture. To purify miRNAs, the mixture was passed through the filter, sequentially

washed with Wash Solution 1 once and Wash Solution 2/3 twice. To collect purified

miRNAs, warm elution solution (95 °C) was added on the filter and then passed though

the filter by centrifugation. Purified miRNAs was stored at -80°C until cDNA synthesis.

5.2.12 Profiling miRNAs The composition of miRNA in MVs was analysed with semi-quantitative analysis,

using MACSPlex miRNA Kit (Cancer) (Miltenyi Biotec, Bergish Gladbach, Germany).

MACSPlex miRNA Kits are designed for detection of miRNAs based on MACSPlex

miRNA cancer beads, which display defined fluorescence properties and can be

identified using a flow cytometry. MACSPlex miRNA cancer beads contain a cocktail

of various fluorescently labelled bead populations. Each population of the labelled

cancer beads is coupled with a specific oligonucleotide complementary to one of the

respective enzymatic labelling of miRNAs, labelled with Vio® 670, in samples. The

hybridized beads, where labelled miRNAs were hybridized to the specific

oligonucleotide, is analysed based on fluorescence characteristics of both MACSplex

miRNA cancer beads and the hybridized labelled miRNA. Steps are described as

follows. Purified miRNA was quantified on an ND-1000 (Nanodrop Technologies,

Wilmington, DE) spectrophotometer. The concentration of miRNA was adjusted to

0.25µg/µL with RNAse-free water. Then miRNAs in samples were labeled with Vio®

670 fluorescence. To label miRNAs with Vio® 670 fluorescence, 4µL of miRNA

136

sample was added into MACSPlex Enzyme 1 tube and then incubated at 37°C for 30

minutes. The enzyme 1 tube was incubated with the addition of 2.1µL of DMSO at

90°C for 5 minutes and then the reaction was stopped by cooling in an ice-water bath

for 10 minutes. The enzyme 1 tube was added with 3.9µL of MACSPlex Reconstitution

Buffer to a final volume of 10µL. This reaction buffer was transferred to the enzyme 2

tube and then incubated at 16°C for 2 hours. Samples were stored at -20°C or

immediately used in a hybridization assay as follows. Hybridization assays were

conducted in the dark. The labelled miRNA was heated to 70°C for 2 minutes and then

cooled in an ice-water bath for 10 minutes. 4µL of MACSPlex miRNA Cancer Beads

and 7µL of MACSPlex miRNA Hybridization Buffer were mixed well with the labelled

sample. The reaction was performed at 48°C for 3 hours (the sample was shaken every

30 minutes). The hybridized sample was washed with 1 ml of MACSPlex miRNA

Buffer and then centrifuged at 6000xg for 10 minutes. The sample was resuspended in

100µL of MACSPlex miRNA Buffer. The measurement of the hybridized sample was

performed by using flow cytometry (BD LSRII, BD) and data was analysed with the

software FlowJo (FLOWJO). The 39 populations of beads were defined based on the

fluorescent intensity of FITC and PE. The intensity of Vio® 670 fluorescence in APC

channel in each beads population was detected and its intensity is proportional to the

miRNA expression. The data is presented as the relative miRNA expression in the

treated group (IL-1β) compared to the control group (without IL-1β).

5.2.13 Prediction of the target genes of candidate miRNAs To predict potential target genes of miRNAs existing in the isolated MVs in this study,

the experimental validated miRNA-target interactions (MTIs) database, miRTarBase,

was used. The MTIs database in miRTarbase is validated experimentally by reporter

assay, western blot, and microarray experiments with overexpression or knockdown of

miRNA (Hsu et al., 2011). The potential target genes of miRNA, predicted with MTIs

database in MiRTarbase, are termed as validated genes in this chapter. Another open-

source library, the miRmap library, was employed to rank potential target genes of

miRNA with a comprehensive features and information by using thermodynamic,

probabilistic, evolutionary, and sequence-based approaches beyond the seed match

(Vejnar and Zdobnov, 2012). The potential target genes of miRNAs within the first-

hundred rankings by the predictive power, determined with the miRmap library, were

137

included as predictive genes in this chapter. To discover biological meaning of these

validated or predictive genes, the functional gene-annotation enrichment analysis is

used with DAVID bioinformatics resources, consists of an intergraded biological

knowledgebase and analytic tool (Huang da et al., 2009a, b). The biological function of

target genes is interpreted as biological signalling pathways based on the gene-

enrichment in annotation terms ( p<0.05 with modified Fisher Exact test).

5.2.14 Assessment of N11 activation The maintenance of N11 cells and other related details are referred to section 2.5.3.

N11 cells were incubated with ultracentrifugation-isolated MV (referred to section

5.2.5) from the four groups of MCs (mentioned in section 5.2.2) for 24 hours. N11 cells

were collected for RNA extraction to assess gene expression of proinflammatory

cytokines (refer to section 2.10). Ultracentrifugation-isolated MVs were grouped as

follows: 1. Control (con MV), 2. IL-1β only (IL-1β MV), 3. 670nm light only (670nm

MV), 4. IL-1β+670nm light (670nm + IL-1β MV).

5.2.15 Assessment of N11 cell morphology The morphology of N11 cells was observed under a phase-contrast light microscope

Zeiss Axiovert 3 (Zeiss), and images were captured by using the ProgRes® CapturePro

microscope camera software (JENOPTIK, Jena, Germany). To describe and compare

ramified versus amoeboid cell shapes, the grid cross analysis method was used (Luckoff

et al., 2017). Images were analysed using a grid system (Image J) to count the number

of grid crossing points per cell (N = 60-70 cells from 5 to 7 images per experimental

group). The image size and resolution were kept consistent through all images. The data

is presented as an average of grid crossing points per cell in each experimental group.

5.2.16 Statistical analysis Statistical analysis was performed using Prism 6 (GraphPad Software, San Diego, CA)

or DAVID bioinformatics Resources. Data were analysed using one-way ANOVA with

a Tukey's multiple comparisons post-test, an unpaired Student t test, or a modified

Fisher exact test, with P<0.05 considered to represent a statistically significant

difference. All data is presented as the mean ± SEM.

138

5.3 Results

5.3.1 Characterization of activated MIO-M1 following IL-1β challenge In response to IL-1β challenge, MIO-M1 displayed elongated processes (Figure 5.3A),

and a lower circularity score compared to the control cells (Figure 5.3B), indicating

MC activation. MIO-M1 viability was measured with MTT assay and showed no cell

death under IL-1β challenge (Figure 5.3C), However, ATP production of activated cells

was reduced compared to control cells (Figure 5.3D). This data shows that IL-1β

stimulation activated MIO-M1 cells and affects their metabolism, but did not influence

cell viability.

Figure 5.3 IL-1β affected cell morphology and metabolism in MIO-M1 cells IL-1β stimulation had impact on morphology and metabolism in cells. (A) The morphology of MIO-M1. (B) The scores of circulatory in MIO-M1. The lower number (scores) of circularity indicates an increasingly elongated shape. (C&D) Cell viability (C) and ATP production (D) in MIO-M1. N=3 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

139

5.3.2 Characterisation of activated MC-derived EVs EVs were isolated from the culture media of MIO-M1 cells in the control and the IL-

1β groups by using sequential ultracentrifugation. To identify EVs, a size and the

expression of Annexin V labeling (a marker for both exosomes and MVs) were

determined using flow cytometry (Figure 5.4 A&B). Due to the limits of the technique,

the black dots in the flow cytometry plot showed noise that include Annexin V-

particles (Figure 5.4 A). Annexin V+ labeling indicates EVs, which appear as purple

dots in Figure 5.4A and B. Only a few Annexin V+ particles were distributed between

450 nm and 880 nm. The size of most of the Annexin V+ particles was smaller than

450 nm, indicating that the majority of isolated vesicles were MVs. Thereby, we

successfully demonstrated that MIO-M1 are able to release MVs, however due to the

limitations of conventional flow cytometry we were not able to prove the presence of

the smallest EV group, exosomes through this method.

To quantify the amount of released MVs from control and IL-1β-challenged MIO-M1

cells, Annexin V expression was analysed by flow cytometry. There was a significant,

4.15-fold increase (p<0.05) of Annexin V binding in MVs isolated from the IL-1β

group compared to the control group, suggesting that IL-1β stimulate MIO-M1 cells to

release MVs into the supernatant (Figure 5.4 C). As Annexin V labels both MVs and

exosomes, to specifically identify exosomes in MIO-M1 culture medium, microbeads

that recognize CD9, CD63 or CD81, exosome-specific markers, were used. Annexin

V+ vesicles bound to these microbeads are considered exosomes. Our current data

showed that a significantly higher amount of exosomes was in the culture medium of

the IL-1β-activated MIO-M1 compared to the control group (1.79-fold change, p<0.05,

Figure 5.4 D). Notably, we isolated a larger number of Annexin V+ vesicles by the

ultracentrifugation than by the microbead-capture technique from the IL-1β-incubated

MIO-M1, indicating that EVs released by activated MIO-M1 include both MVs and

exosomes.

140

Figure 5.4 Release of MVs and exosomes from MIO-M1 in response to IL-1β challenge

(A) Flow cytometry plot. The staining of Annexin V on ultracentrifugation-isolated MVs (purple). Black dots include noise signals and Annexin V- particles. (B) The major Annexin V+ population in a FSC/SSC plot. A green square indicates the distribution of 450nm set-up beads. A red rectangle accounts for the 880nm set-up beads. The size of most EVs isolated from MIO-M1 culture medium is smaller than 450 nm. (C-D) The expression of Annexin V on the ultracentrifugation-isolated EVs (C) and microbeads captured exosomes (D). N=3 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

5.3.3 Characterization of MIO-M1-derived MVs First of all, the MACSPlex kit to profile exosome surface markers was used. MVs were

isolated from MIO-M1 culture medium by using the ultracentrifugation method, then

were tested for labelling by a number of exosome-specific surface markers, listed in

Table 5.1. In response to the IL-1β stimulation, the expression of ten out of thirty-seven

exosomal epitopes was increased on MVs compared to the control group (Figure 5.5A).

The typical exosome markers, CD9, CD63, or CD81, are included in these ten epitopes

(Figure 5.5A), indicating these MVs contain a higher level of exosome markers. Next,

141

exosome markers in microbeads-captured exosomes were characterized (Table 5.1).

Following the IL-1β stimulation, the microbeads-captured exosomes displayed a higher

expression of the CD3, CD14 and CD81 epitopes and lower expressions of the CD44,

CD209 and SEA-4 epitopes compared to the control group (Figure 5.5B). The

upregulation of CD81 on the microbeads-captured exosomes coincides with the higher

level of CD81 labelling on the ultracentrifugation-isolated EVs, confirming that MVs

isolated via ultracentrifugation included exosomes. Based on results from these two

approaches, CD3 and CD81 may be used to identify MIO-M1 - derived exosomes with

high fidelity.

Figure 5.5 Profiles of exosome surface markers on MVs from ultracentrifugation isolated-MVs and microbead captured-exosomes Compared to the control group, the expression of several exosome markers on MVs isolated by using ultracentrifugation (A) or on microbeads captured-exosomes (B) significantly changed. N=3 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change compared to the respective control (P<0.05).

142

Table 5.1 Exosome surface markers on MIO-M1-derived MVs

* denotes a significant change compared to the respective control group (P<0.05).

MarkerCD3 49.84 ± 18.10 111.17 ± 16.17 * 12.62 ± 2.22 31.02 ± 5.96 *CD4 ND ND ND ND

CD19 ND ND ND 5.35 ± 2.81

CD8 ND ND ND ND

HLA-DRDPDQ 2.83 ± 6.66 19.83 ± 9.02 ND 3.03 ± 13.08

CD56 51.83 ± 4.51 107.50 ± 34.31 28.29 ± 1.47 24.43 ± 1.30

CD105 46.50 ± 7.22 48.83 ± 11.67 13.92 ± 6.70 21.58 ± 4.10

CD2 ND ND 0.18 ± 7.72 9.18 ± 8.68

CD1c ND ND ND 0.62 ± 2.13

CD25 18.17 ± 6.25 19.83 ± 4.70 19.48 ± 6.72 26.32 ± 12.97

CD49e 1418.50 ± 245.90 2193.17 ± 44.46 * 78.38 ± 7.93 60.75 ± 9.95

ROR1 12.83 ± 8.41 48.17 ± 23.59 1.48 ± 4.37 2.35 ± 3.41

CD209 ND 2.83 ± 2.96 6.62 ± 1.56 ND *CD9 1755.50 ± 167.70 2956.67 ± 127.30 * 10.45 ± 2.13 3.62 ± 1.27

SSEA-4 496.50 ± 48.77 934.33 ± 14.15 * 90.59 ± 10.40 47.99 ± 3.00 *HLA-ABC 400.17 ± 9.68 499.56 ± 34.42 * 37.33 ± 9.03 35.59 ± 8.57

CD63 3796.50 ± 231.80 4786.83 ± 356.80 * ND ND

CD40 0.83 ± 6.57 8.50 ± 8.74 ND ND

CD62P ND ND ND ND

CD11c ND ND 5.35 ± 4.33 ND

CD81 2357.50 ± 139.70 3056.17 ± 147.30 * ND 24.86 ± 5.27 *MCSP 1927.17 ± 377.00 4010.67 ± 255.50 * 59.38 ± 3.67 91.32 ± 20.28

CD146 385.83 ± 71.37 525.83 ± 118.40 28.05 ± 13.37 38.35 ± 10.20

CD41b 50.17 ± 8.67 140.83 ± 24.25 3.43 ± 3.82 10.69 ± 5.05

CD42a ND ND ND ND

CD24 21.50 ± 11.05 0.17 ± 10.15 6.18 ± 5.61 7.48 ± 2.94

CD86 4.50 ± 6.57 1.83 ± 3.48 5.35 ± 9.77 5.75 ± 5.68

CD44 3584.83 ± 608.60 4665.83 ± 779.40 106.35 ± 11.26 67.02 ± 6.83 *CD326 10.17 ± 4.04 11.50 ± 8.09 22.02 ± 8.17 11.35 ± 9.10

CD131 15.17 ± 6.81 11.50 ± 14.62 ND 1.48 ± 3.67

CD29 2368.83 ± 329.80 3354.67 ± 216.20 * 92.98 ± 17.85 57.72 ± 8.56

CD69 ND ND 0.65 ± 8.52 ND

CD142 162.50 ± 20.80 369.17 ± 46.77 * 20.35 ± 12.88 11.75 ± 6.68

CD45 0.83 ± 1.45 9.83 ± 11.32 4.88 ± 11.21 26.75 ± 7.17

CD31 1.50 ± 4.81 8.50 ± 1.33 6.65 ± 6.23 13.05 ± 7.03

CD20 ND 23.50 ± 11.02 0.65 ± 9.27 ND

CD14 2.83 ± 8.09 ND 2.78 ± 7.98 37.48 ± 6.66 *

Ultracentrifugation Microbeads

Control

Intensity of fluorescence

IL-1β Control IL-1β

143

5.3.4 Composition of miRNAs in MVs To demonstrate that MIO-M1-derived MVs are capable of carrying miRNAs and

investigate whether the composition of miRNAs in MVs is influenced by IL-1β

challenge, MACSPlex miRNA-kit to detect the expression of thirty-five miRNAs in

isolated MVs was used. The relative miRNA expressions in IL-1β-treated MIO-M1-

derived MVs compared to the control group are shown in Table 5.2. The expression

level of nine miRNAs was significantly elevated in MVs from the IL-1β-treated MIO-

M1 cells compared to the control group (P<0.05). These nine genes are let-7f-5p, miR-

9-5p, miR-15a-5p, miR-23b-3p, miR-101-3p, miR-103a-3p, miR-141-3p, miR378b,

and miR-200b-3p (Figure 5.6). The expression of other twenty-six miRNAs was

unchanged (Table 5.2). This data indicates that MIO-M1-derived MVs carried specific

miRNAs in response to IL-1β stimulation. Based on results of the identified validated

predictive target genes of these nine miRNAs, these nine miRNA associated-signalling

pathways are listed in Table 5.3 and 5.4. Their primary target pathways are related to

stem cell pluripotency regulation, MAPK, and TGFB signalling. In addition, they were

also has been linked to cancer (Table 5.5).

Figure 5.6 The composition of miRNAs was changed in MIO-M1-derived MVs in response to IL-1β stimulation MVs released from IL-1β-stimulated MIO-M1 cells were enriched with miRNAs. The relative expression of miRNAs in MVs from the IL-1β-treated MIO-M1 cells was normalised to the miRNA in MVs from the control MIO-M1 cells. N=5-6 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change compared to the control (P<0.05).

144

Table 5.2 Expression of miRNAs in MVs from IL-1β-treated MCs

miRNAID Relativeintensitytocontrol SEM p-value Sigificantlydifferent

(p<0.05)

miR378b 1.150 0.060 0.0365 *miR-141-3p 1.685 0.189 0.0068 *miR-23a-3P 0.213 0.654 0.2632

miR-17-5p 0.837 0.215 0.4706

miR-20a-5p 1.153 0.258 0.5706

miR-200a-3p 0.964 0.413 0.9335

miR-29a-3p 0.955 0.222 0.8451

miR-429 1.609 0.432 0.1963

miR-143-3p 1.290 0.164 0.1154

miR-21-5p 1.646 0.410 0.1538

miR-19b-3p 1.266 0.257 0.3305

miR-31-5p 1.128 0.300 0.6808

let-7f-5p 1.597 0.181 0.0108 *miR-335-5p 1.269 0.189 0.1912

miR-200b-3p 1.979 0.275 0.0073 *miR-34a-5p 1.046 0.089 0.6204

let-7a-5p 1.333 0.253 0.2247

miR-101-3p 1.918 0.393 0.0478 *miR-373-3p 1.227 0.224 0.3417

miR-122-5p 1.078 0.240 0.7552

miR-107 1.691 0.310 0.0563

hsa.miR-30b-5p 1.309 0.354 0.4086

miR-15b-5p 0.663 0.469 0.493

miR-520c-3p 0.992 0.364 0.984

miR-220c-3p 1.030 0.097 0.7655

miR-9-5p 1.389 0.133 0.0194 *miR-23b-3p 1.703 0.199 0.0078 *let-7g-5p 1.491 0.336 0.1817

miR-145-5p 1.215 0.181 0.2669

let-7b-5p 0.996 0.061 0.9499

miR-103a-3p 1.355 0.142 0.037 *miR-10b-5p 0.917 0.068 0.2573

miR-16-5p 1.015 0.290 0.959

miR-126-3p 1.307 0.292 0.324

miR-15a-5p 1.402 0.148 0.0264 *

Table II Intensities for specific microRNAs extraceted from the ultracentrifugation-isolated Evs

145

Table 5.3 Biological signalling pathways enrichment analysis for MV carrying miRNA-targeting validated genes

Pathway Number of genes %

Modified Fisher Exact P-Value

(enrichment)

MicroRNAs in cancer 14 23.3 2.900E-10

Signaling pathways regulating pluripotency of stem cells 9 15 2.600E-07

Pathways in cancer 12 20 2.000E-06

Colorectal cancer 6 10 1.400E-05

Pancreatic cancer 6 10 1.500E-05

Chronic myeloid leukemia 6 10 2.400E-05

FoxO signaling pathway 7 11.7 4.300E-05

Small cell lung cancer 6 10 5.100E-05

Hepatitis B 7 11.7 6.900E-05

HTLV-I infection 8 13.3 3.500E-04

Melanoma 5 8.3 3.800E-04

Cell Cycle 5 8.3 7.200E-04

Hippo signaling pathway 6 10 8.000E-04

Prostate cancer 5 8.3 8.700E-04

Transcriptional misregulation in cancer 6 10 1.200E-03

p53 Signaling Pathway 4 6.7 1.800E-03

Non-small cell lung cancer 4 6.7 2.500E-03

Proteoglycans in cancer 6 10 3.000E-03

Cell cycle 5 8.3 3.100E-03

Glioma 4 6.7 3.900E-03

NFkB activation by Nontypeable Hemophilus influenzae 4 6.7 6.800E-03

Influence of Ras and Rho proteins on G1 to S Transition 4 6.7 7.600E-03

Bladder cancer 3 5 1.700E-02

Focal adhesion 5 8.3 1.800E-02

Epstein-Barr virus infection 5 8.3 2.100E-02

Endometrial cancer 3 5 2.700E-02

PI3K-Akt signaling pathway 6 10 2.800E-02

Human Cytomegalovirus and Map Kinase Pathways 3 5 3.000E-02

Acute myeloid leukemia 3 5 3.100E-02

Wnt signaling pathway 4 6.7 3.200E-02

MAPK signaling pathway 5 8.3 3.500E-02

VEGF signaling pathway 3 5 3.500E-02

Adrenergic signaling in cardiomyocytes 4 6.7 3.700E-02

Telomeres, Telomerase, Cellular Aging, and Immortality 3 5 3.700E-02

Adherens junction 3 5 4.900E-02

Prolactin signaling pathway 3 5 5.000E-02

Cyclins and Cell Cycle Regulation 3 5 5.600E-02

TGF-beta signaling pathway 3 5 6.500E-02

NF-kappa B signaling pathway 3 5 8.200E-02

Rap1 signaling pathway 4 6.7 8.700E-02

Toll-like receptor signaling pathway 3 5 8.800E-02

Chagas disease (American trypanosomiasis) 3 5 9.100E-02

HIF-1 signaling pathway 3 5 9.200E-02

T cell receptor signaling pathway 3 5 9.200E-02

Validated genes (60 genes)

146

Table 5.4 Biological processes regulated by MV miRNA targeting predictive genes

Pathway Number of genes % Modified Fisher Exact

P-Value

CTCF 5 0.8 1.80E-02

MAPKsignalingpathway 17 2.6 2.30E-03

TGF-betasignalingpathway 9 1.4 2.80E-03

Oxytocinsignalingpathway 12 1.9 5.40E-03

Purinemetabolism 12 1.9 1.30E-02

Hypertrophiccardiomyopathy(HCM) 7 1.1 2.50E-02

Dilatedcardiomyopathy 7 1.1 3.10E-02

Signalingpathwaysregulatingpluripotencyofstemcells 9 1.4 4.30E-02

Sphingolipidmetabolism 5 0.8 4.70E-02

BLymphocyteCellSurfaceMolecules 3 0.5 4.80E-02

ActivationofCskbycAMP-dependentProteinKinaseInhibitsSignalingthroughtheTCellReceptor4 0.6 4.90E-02

Proteinprocessinginendoplasmicreticulum 10 1.5 5.00E-02

Endocytosis 14 2.2 5.40E-02

Stathminandbreastcancerresistancetoantimicrotubuleagents 4 0.6 5.50E-02

Circadianrhythm 4 0.6 5.70E-02

Hipposignalingpathway 9 1.4 6.60E-02

Axonguidance 8 1.2 7.50E-02

Ovariansteroidogenesis 5 0.8 7.90E-02

Regulationoflipolysisinadipocytes 5 0.8 7.90E-02

RNAdegradation 6 0.9 8.40E-02

mTORsignalingpathway 5 0.8 8.70E-02

ChREBPregulationbycarbohydratesandcAMP 3 0.5 9.40E-02

Oocytemeiosis 7 1.1 9.40E-02

MicroRNAsincancer 13 2 9.50E-02

Gapjunction 6 0.9 9.80E-02

Predictive genes (646 genes)

147

Table 5.5 Biological signalling pathways regulated by MV miRNA-targeting validated and predictive genes

148

5.3.5 Influences of 670nm light on the secretion of MVs and exosomes

The IL-1β-induced activation of MIO-M1 cells was reduced by 670nm treatment as

evidence by the reduction of elongation of treated MIO-M1 cells. (Figure 5.7A&B).

Furthermore, MTT and ATP assay were used to access MIO-M1 cell viability and

showed that 670nm treatment did not cause cell death, but increased ATP production

compared to non-treated cells (Figure 5.7B&C). MVs in the 670nm light-treated MIO-

M1 cell culture medium (Figure 5.7) was quantified. Our data showed that 670nm

treatment of non-activated control MIO-M1 cells (670nm group) led to an increased

release of Annexin V+ MVs compared to the control group (Figure 5.8A). 670nm

treatment of IL-1β-activated MIO-M1 cells (670nm+IL-1β group) resulted in the higher

release of MVs compared to non-treated activated MIO-M1 cells (IL-1β group) (Figure

5.8A). These data suggest that 670nm light increased the secretion of MVs from healthy

and activated MIO-M1 cells. In another experiment, CD81 and CD3 antibodies were

used to quantify MVs isolated via ultracentrifugation (Figure 5.8C&D). In this method,

670nm light increased MVs release from both healthy and activated MIO-M1 cells.

Furthermore, the effects of 670nm light on exosome secretion were also assessed using

microbeads captured exosomes (Figure 5.8B). 670nm light treatment increased the

exosome release in the healthy MIO-M1 compared to non-treated cells (Figure 5.8 B),

however did not alter the amount of exosome release by activated MIO-M1. The effects

of 670nm light on EV secretion from the healthy MIO-M1 cells and activated MIO-M1

cells were not consistent between two approaches. Due to the high variation in the

microbeads isolation, the further experiments are required.

149

Figure 5.7 Effects of 670nm light on IL-1β stressed-MIO-M1 cells IL-1β caused cell morphology changed and increased ATP production in MIO-M1 cells, but did not affect cell viability. (A) Morphology of MIO-M1 in different groups. (B) The indication of circularity of MIO-M1. The lower number of circularity indicates an increasingly elongated shape. (C&D) The cell survival (C) and ATP production (D) in MIO-M1 cells. N=3 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

Figure 5.8 670nm light triggered MIO-M1 cells to release MVs and exosomes (A &B) The Annexin V binding on ultracentrifugation isolated MVs (A) or on microbeads-captured exosomes (B). (C&D) The percentage of CD3 (C) or CD81+ MVs (D) on ultracentrifugation isolated-MVs. N=3 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. * denotes a significant change (P<0.05).

150

5.3.6 Contribution of MIO-M1-derived MVs on proinflammatory and anti-inflammatory responses in microglia To investigate the effects of MIO-M1-derived MVs on inflammatory responses, MVs

was isolated from MIO-M1 cell culture medium, and then incubated N11 microglia cell

line with MVs for 24 hours. The morphology of N11 cells, incubated with control MVs

displayed a ramified shape with elongated processes (Figure 5.9A). In contrast, the IL-

1β MVs transformed the morphology of N11 to an amoeboid shape (Fig. 5.9A). N11

cells cultured with 670nm MVs displayed a ramified shape, but with shorter processes

(Figure 5.9A). To determine and quantify the morphological changes in MV-stimulated

N11 cells, a grid-cross analysis method was used. Our data showed that a number of

grid-cross points was significantly lower in the N11 cells, which had been co-cultured

with the IL-1β MVs, compared to N11 with control MVs (Figure 5.9B). The 670nm

MVs did not significantly change the grid-cross points in N11 cells compared to the

control MVs (p>0.05), however the grid-cross points in N11, treated with 670nm + IL-

1β MVs, were higher than N11 co-cultured with the IL-1β MVs (P<0.05). These data

suggest that 670nm treatment can ameliorate MC-derived MVs on activating N11 cells.

To further confirm this, the expression level of proinflammatory cytokines in N11 cells,

after incubation with MVs, was analysed (Figure 5.9C-F). IL-6 and IL-1β expression

was significantly reduced in N11 cells incubated with MVs from 670nm-treated

activated MIO-M1 cells (670nm+IL-1β MVs) compared to MVs from non-treated

activated MIO-M1 cells (IL-1β MV) (Figure 5.9C-F). However, Ccl2 was not

significantly different between three experimental groups (Figure 5.9D). Surprisingly,

TNFα expression was significantly reduced in N11 cells incubated with MVs from

activated MIO-M1 cells (IL-1β MV) and 670nm light treated MIO-M1 cells (670nm

MV). But 670nm light diminished this effect showing TNFα expression was similar to

control in N11 cells co-cultured with MVs from IL-1β-stimulated MIO-M1 cells (IL-

1β MVs vs. 670nm+IL-1β MVs) (Figure 5.9E).

151

Figure 5.9 670nm light diminished MV associated-N11 activation and gene expressions of proinflammatory cytokines (A) The morphology of N11 cells with/wo MVs from MIO-M1 culture medium in different groups. (B) The quantification of N11 morphology by using grid-cross points. (C-F) The expressions of proinflammation genes in N11 cells. N=5 was used for all experimental comparisons and was performed in technical triplicate. The data is presented as the mean ± SEM. # denotes a significant change compared to the control group (P<0.05). * denotes a significant change (P<0.05). 5.3.7 The amount of MV changed in the mouse retina post PD To investigate whether MVs play a role in retinal degeneration, the in vivo PD model

in mice was used. MV isolated from mouse retinas were purified by using sucrose

gradient steps (Figure 5.10A-D). A decreased tendency of Annexin V+ particles

percentage was shown in the PD group in comparison of the dim-reared control group

152

(Figure 5.10E), but the sample size (N=3) is too small to reach statistically significantly

difference. These data offers an insight that a lower percentage of MVs existed in the

intercellular spaces in the retinas following 5-day PD. However, the further

experiments are needed to investigate the underlying mechanism.

Figure 5.10 Analysis of a percentage of Annexin V binding on MVs isolated from PD retinas (A-D) Annexin V+ vesicles (purple) in a dot plot and a FSC/SSC plot in the sample of MV isolated without a sucrose step gradient. (A-B) or the sample of MVs collected with a sucrose step gradient (C-D). The 450-nm or 880-nm square account for the distribution of the set-up beads. (E) The quantification of Annexin V biding on MVs with a sucrose step gradient. N=3 per group, data is presented as the mean ± SEM.

153

5.4 Discussion This chapter highlights a role of MIO-M1 cells-derived MVs in regulating the

activation of microglia. Our data have shown that MVs released by control healthy

MIO-M1 cells and activated MIO-M1 cells contain miRNAs that may explain the

interaction between MIO-M1 cells and MGs. This interaction can be manipulated by

670nm red light. This chapter has four main findings. Firstly, we have developed robust

isolation techniques allowing for the investigation of MVs from both cells and the

retinas. Secondly, our data demonstrated that IL-1β-activated MIO-M1 cells increased

their release of MVs and exosomes. Thirdly, target genes of the selective miRNAs in

MVs from IL-1β-treated MIO-M1 cells are associated with MAPK pathway, which

regulates inflammatory responses in MGs. Lastly extracted MIO-M1 cell derived-MVs

regulated the activity and inflammation profile of MGs. This work lays the foundation

for future exploration for understating the roles that miRNAs and exosomes play in the

communication of MGs and MCs.

5.4.1 IL-1β stimulated MC activation and their release of MVs and exosomes In this study, our data demonstrated that IL-1β triggered MIO-M1 cells to secrete MVs

into their culture media. To our knowledge, this is the first study to show that MCs are

a source of MVs. The size is a widely accepted methodology to define MVs (100-1000

nm) (Lee et al., 2012) and exosomes (30-100 nm) (Atienzar-Aroca et al., 2016; Lee et

al., 2012; Minciacchi et al., 2015). However, we are unable to accurately measure the

size of MIO-M1cell-derived MVs due to the limit of detection with conventional flow

cytometry (< 300 nm) (Buzas et al., 2014; Dragovic et al., 2011; Raposo and Stoorvogel,

2013). Therefore, we need to use other devices, such as fluorescence nanoparticle

tracking analysis (NTA), to determine the population of exosomes and MVs by the size.

However, by using exosome specific markers, we were able to confirm that MIO-M1

cell-derived MVs include exosomes.

The quantification and characteristics of MIO-M1 cell-derived exosomes (microbead-

captured) are not fully consistent with those of MIO-M1 cell-secreted MVs isolated by

using the ultracentrifugation under the IL-1β stimulation w/wo 670nm light in this

154

chapter. While using the ultracentrifugation approach, all subpopulations of exosomes

were harvested and separated from other EVs based on the difference in their sizes and

buoyant densities. However, microbead-captured exosomes were selectively collected

by recognizing exosome membrane proteins, CD3, CD63, and CD81. Several studies

have shown that classical markers of exosomes, such as CD63 and CD81, are also

enriched in vesicles which originate through budding from the plasma membrane and

could not be successfully distinguished from other EVs (Andreu and Yanez-Mo, 2014).

Therefore, it might explain that the variation of exosome quantification (Annexin V

binding) and exosome markers characteristics between ultracentrifugation-isolated

MVs and microbead-captured exosomes from MIO-M1 cell culture supernatant in this

chapter.

5.4.2 Function of EV surface markers on recipient cells The interaction between EVs and recipient cells are still not fully explored. Some

reports showed that exosome surface markers and lipid raft-associated proteins on

exosomes play a role in regulating EV trafficking to, and the activation of signalling

pathway in recipient cells (Andreu and Yanez-Mo, 2014; Blanchard et al., 2002; Tetta

et al., 2013). Annexins, which are related to cell adhesion and growth, play a crucial

role in regulating uptake of EVs in breast carcinoma cell lines (Koumangoye et al.,

2011). CD81-positive exosomes promoted breast cancer cell motility and metastasis

via a Wnt-PCP signalling pathway (Luga et al., 2012). Exosomes secreted from Jurkat

T cell line harboured the CD3/TCR complex, which may be a specific delivery signal

to cells carrying the right combination of peptide/MHC complexes (Blanchard et al.,

2002). In the current study, an elevated expression of CD3, CD81, and Annexin V was

found on the membranes of MIO-M1 cell-derived MVs, implicating that these

membrane markers might be associated with the uptake by, and the initiation of the

down-steam signalling pathway in target cells, such as MGs.

5.4.3 Target genes of MC-derived EVs carried miRNAs that modulate inflammation Our current studies showed that several miRNAs existed in MIO-M1 cell-derived MVs.

MiR-101, found in MC-derived MVs in this chapter, has been reported to regulate the

155

production of proinflammatory cytokines IL-6 in MGs during LPS stimulation (Saika

et al., 2017). Moreover, target genes of miR-101 are predicted to participate in MAPK

signalling pathway by which the superoxide production in diabetic retinopathy (Du et

al., 2010). Therefore, MV containing miR-101-3p may play a crucial role in modulating

inflammation in retinal degeneration via the regulation of the MAPK signalling

pathway. Further experiments are needed to verify this hypothesis.

5.4.4 The effects of IL-1β and 670nm light on MV release and its intraluminal content

Our current study found that 670nm light treatment led to increased secretion of MVs

from normal MIO-M1 cells and IL-1β-treated MIO-M1 cells. Furthermore, 670nm light

treatment of activated MIO-M1 cell-derived MVs drove the reduction of IL-6 and IL-

1β expression in MGs. Both of which are the part of MAPK pathway (Yang et al., 2013),

while no change of CCL-2 and a reverse effect on TNFα expression were observed in

the current study, which might be activated by other signalling pathways. These suggest

that the anti-inflammatory effects of MVs are through the regulation of the MAPK

pathway in the target cells, which might be involved in the mechanism of 670nm light

treatment in the in vitro model in this chapter. This agrees with the previous study that

670nm light reduced the phosphorylation of p38 MAPK in retinal ganglion cells in the

hyperglycaemic condition (Tang et al., 2013).

670nm light increased secretion of MVs. However, the mechanism or cellular

signalling pathways of this response is still not clear. It is possible to be associated with

the ATP production in cells, which is a significant effect of 670nm light treatment on

cells or tissues (Gkotsi et al., 2014; Wong-Riley et al., 2005). The previous study shows

that astrocyte-derived ATP as the endogenous factor is responsible for MVs shedding

and IL-1β release in primary hippocampal MGs or N9 murine microglial cell line

(Bianco et al., 2005). The ATP-induced MV release is abrogated by the ATP-degrading

enzyme apyrase or by P2X(7) receptor (purinergic receptor) antagonist. In Chapter 4,

our data showed that 670nm light increased COX5a expression and mitochondrial

membrane potential (Δψm) in MCs post PD, which can be expected to cause the

subsequence increase in ATP production. Taken together, 670nm light may trigger

156

MCs to generate and then release ATP to extracellular space, which may prompt

neighbouring MCs to release EVs in the in vitro model in this chapter.

On the other hand, 670nm light-induced MV release might be not purely quantitative.

It could include changes of the composition of MVs, thereby delivered changed

miRNAs to MGs. This hypothesis is supported by Hajrasouliha, who reported that

exosomes derived from mast cells in the oxidative state had different RNA content from

samples grown under normal conditions (Lee et al., 2012). Therefore, the content in

MVs could be a highly selective and may be dependent on the surrounding environment

or the activation state of the host cell. However, we also need to take the amount of

MVs into consideration in our model of this study and need further experiments to

support this aspect. Our current study demonstrated that 670nm light is capable of

facilitating the secretion of EVs from MCs and also influencing the packing of MVs in

the normal or inflammation condition.

5.4.5 MVs and exosomes on retinal diseases Research focusing on the role of MVs and exosomes in eye diseases only began to

emerge in the past decade (Klingeborn et al., 2017). Exosome markers, CD81 and

LAMP2, were localized in the retina of patients with dry AMD (Wang et al., 2009),

potentially implicating that CD81-positive exosomes may participate in drusen

formation. Retinal exosomes were indicated to regulate immune responses, including

T cell activation and proliferation as well as monocyte death (Knickelbein et al., 2016),

which have been known to be involved in the AMD pathogenesis (Ambati et al., 2013;

Ambati and Fowler, 2012). Our in vivo data of this chapter showed a reduced amount

of MVs in the PD retinas, suggesting that MVs play a role in retinal degeneration in

vivo. The reduced amount of MVs in the extracellular space in PD retinas may indicate

a more rapid uptake of these vesicles and the utilisation of their content by target cells,

such as MGs. The route and mechanism of EV uptake by MGs have been partially

discovered. Clathrin-mediated endocytosis, which is involved in cellular internalization,

and micropinocytosis, an endocytic uptake pathway, both participate in controlling EV

uptake by MGs (Mulcahy et al., 2014). In our study, PD may trigger these two

pathways-related protein expressions in or on MGs in the injured retinas, subsequently

157

prompting EV uptake by MGs. However, further experiments are needed to confirm

these theories.

5.5 Conclusion Our results indicate that 670nm light can regulate MIO-M1 cell-mediated MG

activation at a gene level possibly via the transportation of miRNAs by MVs and

exosomes. Using MV and exosome markers it is possible to manipulate their target.

PBM with 670nm light may be used to alter MC gene expression patterns, which

mitigate MC gliosis, and in addition by the increased release of MV and exosomes.

These EVs allow the delivery of modified genetic material, especially miRNAs that can

directly regulate inflammatory responses in target cells. The understanding the contents

and surface markers of MVs will lead to a greater understanding of diseases where

retinal inflammation is a key feature, and also lead to the development of retinal cell-

targeted gene-based therapies in retinal degenerations, such as AMD.

158

Chapter Six

Summary and Conclusion

159

This thesis demonstrates the therapeutic potential of PBM in targeting retinal MCs

following retinal injury. Short, 3-minute long 670nm irradiation of the retinas over 5

consecutive days, or isolated primary MCs within a 24-hour period were effective in

moderating MCs activation following injury. This work shows that PBM treatment can

be effective on an already activated MCs, offering a new and exciting therapeutic target

in chronic degenerative conditions of the retina. The reduction of inflammatory

activation, and proliferative gliosis, the hallmarks of many such conditions, can lead to

the mitigation of neuronal loss and the slowing of the progression of the disease.

This work also provides insight into some of the mechanisms through which PBM

exerts its protective effects on the retina. The novel findings of this work are

1. 670nm light can mitigate proliferative gliosis of MCs

Experiments using the in vitro scratch model showed that PBM with 670nm light

slowed the rate of area coverage by the human Müller cell line (MIO-M1) (Chapter 3).

This slow ‘wound’ coverage was the consequence of the reduced spreading, and

proliferation of the treated MCs. Both in the in vitro and in vivo stress models (Chapter

3), we noted the upregulation of GFAP, an intermediate filament protein, is responsible

for changing the cytoskeletal structure of MCs in retinal degeneration. The upregulation

of this protein has been linked to tissue stress as well as cellular hypertrophy in MCs

and been suggested to play a role in glial scar formation (Lewis et al., 2010; Lewis and

Fisher, 2003; Lu et al., 2011). The reduction in GFAP expression in the 670nm-

irradiated cells and retinas suggests that PBM may have a mitigating effect on glial scar

formation. We noted glial scar formation in the severely affected areas of all PD-injured

retinas. Although we were not able to quantify the volume of glial scars, the reduction

of their overall spread was presented in retinas treated with 670nm light shortly after

PD injury. In addition, our previous data demonstrated that 670nm light reversed the

altered expression vimentin and S100B, other cytoskeletal proteins, supporting the

complex architecture of MCs, after the exposure of PD (Albarracin and Valter, 2012).

These two proteins have been linked to MC nuclear migration, cell division, and

subretinal glial scars formation in a rabbit retinal detachment model (Lewis et al.,

2010), implying other cytoskeletal proteins may also participate in the change of MC

160

cytoskeletal structure in the PD model and the protective mechanism of the 670nm

light treatment in this thesis.

2. 670nm light affects MCs undergoing reactive gliosis and alleviate

inflammatory changes in the injured retina

Following retinal stress or injury, MCs respond by the expression of chemokines, which

are strong chemo-attractants that lead the recruitment of immune cells to the damaged

regions of the retina (Chapter 3). Activated MCs, together with the recruited MG/MΦ

express proinflammatory cytokines thereby further exacerbating inflammation in the

injured retina. The present work has shown that 670nm light treatment can reduce the

recruitment of MG/MΦ and the production of proinflammatory cytokines, as well as

the activation of inflammasomes in the injured retina in vivo (Chapter 3&4).

3. MCs treated with 670nm light display a reduced production and secretion of

proinflammatory cytokines

In the MC-photoreceptor co-culture system (Chapter 4), MCs had an increased

expression of proinflammatory cytokines following PD, indicating that they were

undergoing reactive gliosis. Cells treated with 670nm light subsequent to the PD

showed a significant reduction in proinflammatory gene expression. In addition, the

supernatant from the wells of 670nm light-treated MCs contained significantly lower

amount of proinflammatory cytokine proteins. As these gene and protein expression

changes occurred when cell were treated in isolation, our data successfully

demonstrated that 670nm light irradiation has a direct effect on activated MCs.

4. 670nm irradiated MCs suppress MG activation Results from both of our in vitro models showed that the downstream effect of 670nm

treatment of MCs in the mitigation of MG activation (Chapter 4&5). When MG were

incubated in the supernatant extracted from wells of irradiated MCs, their activation

became muted, as evidenced by the reduction of the production of proinflammatory

cytokines (IL-1β, IL-6, Ccl2) and the down-regulation of surface markers indicating

MG activation (CD11b and CD86). As these effects were not the consequence of direct

contact between MCs and MGs, we further investigated to try to identify the specific

161

product released by irradiated MCs that can mediate these changes in MG activation.

One way MCs may communicate with MGs is through EVs. One population of these

vesicles is the exosomes. In the present work, our results have shown that MCs release

exosomes upon stress (IL-1β) (Chapter 5). Isolated exosomes from the MC culture

medium induced activation of N11 microglia cells. 670nm treatment did not reduce the

amount of the released exosomes, however, when N11 cells were incubated with these

exosomes, their activation was significantly reduced. Exosomes have been described to

carry genetic materials, including miRNAs. In this work, our results demonstrated that

exosomes released by activated MCs contain several miRNAs. MiRNAs can modulate

gene expression through the regulation of mRNA translation. I identified a number of

miRNAs that are implicated in the regulation of inflammatory pathways. Therefore,

this data suggest that the 670nm light has both direct and secondary effect on immune

cell activation following retinal injury.

One of the most novel aspects of this research is the link/finding that PBM treatment

may influence the composition of (genetic) materials, miRNAs, transported from MCs

to surrounding retinal cells, including MGs. Our data suggest that influence do occur

however we have not yet confirmed this hypothesis in vivo. Based on previous research,

miRNAs have shown to regulate choroidal neovascularization (Wang et al., 2012) and

immune responses (Saxena et al., 2015) in dry AMD. Therefore, regulating miRNA

composition in EVs by using 670nm light might be a future direction for studying the

application of 670nm light in retinal degeneration.

This thesis showed that 670nm light commenced after the onset of retinal injuries and

MC activation have several beneficial effects on slowing progressive retinal

degeneration via some potential mechanism-of-action on MCs. These findings

supported the safety and efficacy of 670nm light treatment on retinal injuries, and may

assist in standardising and optimising treatment paradigms for future clinical studies,

based on the extent of MC gliosis.

162

References • Abrahan, C.E., Insua, M.F., Politi, L.E., German, O.L., Rotstein, N.P., 2009.Oxidative stress promotes proliferation and dedifferentiation of retina glialcellsinvitro.JNeurosciRes87,964-977.

• Ahmad, I., DelDebbio, C.B., Das, A.V., Parameswaran, S., 2011.Mullerglia: apromising target for therapeutic regeneration. InvestOphthalmolVisSci52,5758-5764.

• Alamouti,B.,Funk,J.,2003.Retinalthicknessdecreaseswithage:anOCTstudy.BrJOphthalmol87,899-901.

• Albarracin,R.,Eells,J.,Valter,K.,2011.Photobiomodulationprotectstheretinafromlight-inducedphotoreceptordegeneration.InvestOphthalmolVisSci52,3582-3592.

• Albarracin, R., Natoli, R., Rutar, M., Valter, K., Provis, J., 2013. 670 nm lightmitigates oxygen-induced degeneration in C57BL/6J mouse retina. BMCNeurosci14,125.

• Albarracin, R., Valter, K., 2012. 670 nm red light preconditioning supportsMullercellfunction:evidencefromthewhitelight-induceddamagemodelintheratretina.PhotochemPhotobiol88,1418-1427.

• Alexander,M.,Hu,R.,Runtsch,M.C.,Kagele,D.A.,Mosbruger,T.L.,Tolmachova,T., Seabra, M.C., Round, J.L., Ward, D.M., O'Connell, R.M., 2015. Exosome-deliveredmicroRNAsmodulatetheinflammatoryresponsetoendotoxin.NatCommun6,7321.

• Ambati,J.,Anand,A.,Fernandez,S.,Sakurai,E.,Lynn,B.C.,Kuziel,W.A.,Rollins,B.J.,Ambati,B.K.,2003.Ananimalmodelofage-relatedmaculardegenerationinsenescentCcl-2-orCcr-2-deficientmice.NatMed9,1390-1397.

• Ambati, J., Atkinson, J.P., Gelfand, B.D., 2013. Immunology of age-relatedmaculardegeneration.NatRevImmunol13,438-451.

• Ambati,J.,Fowler,B.J.,2012.Mechanismsofage-relatedmaculardegeneration.Neuron75,26-39.

• Anand,A.,Sharma,N.K.,Gupta,A.,Prabhakar,S.,Sharma,S.K.,Singh,R.,Gupta,P.K., 2012. Single nucleotide polymorphisms in MCP-1 and its receptor areassociated with the risk of age related macular degeneration. PLoS One 7,e49905.

• Anders, J.J., Lanzafame, R.J., Arany, P.R., 2015. Low-level light/laser therapyversusphotobiomodulationtherapy.PhotomedLaserSurg33,183-184.

• Andreu,Z.,Yanez-Mo,M.,2014.Tetraspaninsinextracellularvesicleformationandfunction.FrontImmunol5,442.

• AparecidaDaSilva,A.,Leal-Junior,E.C.,Alves,A.C.,Rambo,C.S.,DosSantos,S.A.,Vieira,R.P.,DeCarvalhoPde,T.,2013.Wound-healingeffectsoflow-levellasertherapyindiabeticratsinvolvethemodulationofMMP-2andMMP-9andtheredistributionofcollagentypesIandIII.JCosmetLaserTher15,210-216.

• Argenta,P.A.,Ballman,K.V.,Geller,M.A.,Carson,L.F.,Ghebre,R.,Mullany,S.A.,Teoh, D.G., Winterhoff, B.J., Rivard, C.L., Erickson, B.K., 2017. The effect ofphotobiomodulation on chemotherapy-induced peripheral neuropathy: Arandomized,sham-controlledclinicaltrial.GynecolOncol144,159-166.

• Arranz-Valsero, I., Soriano-Romani, L., Garcia-Posadas, L., Lopez-Garcia, A.,Diebold, Y., 2014. IL-6 as a corneal wound healing mediator in an in vitroscratchassay.ExpEyeRes125,183-192.

163

• Astray,G., Soto,B., Lopez,D., Iglesias,M.A.,Mejuto, J.C., 2016.ApplicationoftransitdataanalysisandartificialneuralnetworkinthepredictionofdischargeofLorRiver,NWSpain.WaterSciTechnol73,1756-1767.

• Atienzar-Aroca,S.,Flores-Bellver,M.,Serrano-Heras,G.,Martinez-Gil,N.,Barcia,J.M., Aparicio, S., Perez-Cremades, D., Garcia-Verdugo, J.M., Diaz-Llopis, M.,Romero, F.J., Sancho-Pelluz, J., 2016. Oxidative stress in retinal pigmentepithelium cells increases exosome secretion and promotes angiogenesis inendothelialcells.JCellMolMed20,1457-1466.

• Bajetto, A., Bonavia, R., Barbero, S., Schettini, G., 2002. Characterization ofchemokines and their receptors in the central nervous system:physiopathologicalimplications.JNeurochem82,1311-1329.

• Bamforth,S.D.,Lightman,S.L.,Greenwood,J.,1997.Ultrastructuralanalysisofinterleukin-1 beta-induced leukocyte recruitment to the rat retina.InvestigativeOphthalmology&VisualScience38,25-35.

• Barcelona, P.F., Jaldin-Fincati, J.R., Sanchez, M.C., Chiabrando, G.A., 2013.Activated alpha2-macroglobulin induces Muller glial cell migration byregulatingMT1-MMPactivitythroughLRP1.FASEBJ27,3181-3197.

• Barcelona, P.F., Ortiz, S.G., Chiabrando, G.A., Sanchez, M.C., 2011. alpha2-Macroglobulin induces glial fibrillary acidic protein expressionmediated bylow-density lipoprotein receptor-related protein 1 in Muller cells. InvestOphthalmolVisSci52,778-786.

• Bartel,D.P.,2004.MicroRNAs:genomics,biogenesis,mechanism,andfunction.Cell116,281-297.

• Begum,R.,Powner,M.B.,Hudson,N.,Hogg,C.,Jeffery,G.,2013.Treatmentwith670 nm light up regulates cytochrome C oxidase expression and reducesinflammation in an age-related macular degeneration model. PLoS One 8,e57828.

• Bianco, F., Pravettoni, E., Colombo, A., Schenk, U., Moller, T., Matteoli, M.,Verderio,C., 2005.Astrocyte-derivedATP inducesvesicle sheddingand IL-1betareleasefrommicroglia.JImmunol174,7268-7277.

• Blanchard,N.,Lankar,D.,Faure,F.,Regnault,A.,Dumont,C.,Raposo,G.,Hivroz,C.,2002.TCRactivationofhumanTcellsinducestheproductionofexosomesbearingtheTCR/CD3/zetacomplex.JImmunol168,3235-3241.

• Boppana, S., Scheglov, A., Geffers, R., Tarabykin, V., 2012. Cellularretinaldehyde-binding protein (CRALBP) is a direct downstream target oftranscriptionfactorPax6.BiochimBiophysActa1820,151-156.

• Bringmann,A.,Iandiev,I.,Pannicke,T.,Wurm,A.,Hollborn,M.,Wiedemann,P.,Osborne,N.N.,Reichenbach,A.,2009.CellularsignalingandfactorsinvolvedinMullercellgliosis:neuroprotectiveanddetrimentaleffects.ProgRetinEyeRes28,423-451.

• Bringmann,A.,Pannicke,T.,Grosche,J.,Francke,M.,Wiedemann,P.,Skatchkov,S.N., Osborne, N.N., Reichenbach, A., 2006. Muller cells in the healthy anddiseasedretina.ProgRetinEyeRes25,397-424.

• Bringmann, A., Wiedemann, P., 2012. Muller glial cells in retinal disease.Ophthalmologica227,1-19.

• Brosseau,L.,Robinson,V.,Wells,G.,Debie,R.,Gam,A.,Harman,K.,Morin,M.,Shea, B., Tugwell, P., 2005.Low level laser therapy (Classes I, II and III) fortreatingrheumatoidarthritis.CochraneDatabaseSystRev,CD002049.

• Burgi, S., Samardzija, M., Grimm, C., 2009. Endogenous leukemia inhibitoryfactorprotectsphotoreceptorcellsagainstlight-induceddegeneration.MolVis15,1631-1637.

164

• Buschini, E., Piras, A., Nuzzi, R., Vercelli, A., 2011. Age related maculardegenerationanddrusen:neuroinflammationintheretina.ProgNeurobiol95,14-25.

• Buschow,S.I.,Nolte-'tHoen,E.N.,vanNiel,G.,Pols,M.S.,tenBroeke,T.,Lauwen,M., Ossendorp, F., Melief, C.J., Raposo, G., Wubbolts, R., Wauben, M.H.,Stoorvogel,W.,2009.MHCIIindendriticcellsistargetedtolysosomesorTcell-inducedexosomesviadistinctmultivesicularbodypathways.Traffic10,1528-1542.

• Buzas, E.I., Gyorgy, B., Nagy, G., Falus, A., Gay, S., 2014. Emerging role ofextracellularvesicles in inflammatorydiseases.NatRevRheumatol10,356-364.

• Calabrese,E.J.,Stanek,E.J.,3rd,Nascarella,M.A.,Hoffmann,G.R.,2008.Hormesispredicts low-dose responses better than thresholdmodels. Int J Toxicol 27,369-378.

• Calaza,K.C.,Kam,J.H.,Hogg,C.,Jeffery,G.,2015.Mitochondrialdeclineprecedesphenotypedevelopment in the complement factorHmousemodelof retinaldegenerationbutcanbecorrectedbynearinfraredlight.NeurobiolAging36,2869-2876.

• Campbell,M.,Humphries,P.,2012.Theblood-retinabarrier:tightjunctionsandbarriermodulation.AdvExpMedBiol763,70-84.

• Canning, D.R., Hoke, A., Malemud, C.J., Silver, J., 1996. A potent inhibitor ofneurite outgrowth that predominates in the extracellularmatrix of reactiveastrocytes.IntJDevNeurosci14,153-175.

• Cao,X.,Shen,D.,Patel,M.M.,Tuo,J.,Johnson,T.M.,Olsen,T.W.,Chan,C.C.,2011.Macrophagepolarizationinthemaculaeofage-relatedmaculardegeneration:apilotstudy.PatholInt61,528-535.

• Chahed,S.,Leroyer,A.S.,Benzerroug,M.,Gaucher,D.,Georgescu,A.,Picaud,S.,Silvestre,J.S.,Gaudric,A.,Tedgui,A.,Massin,P.,Boulanger,C.M.,2010.IncreasedVitreous Shedding of Microparticles in Proliferative Diabetic RetinopathyStimulatesEndothelialProliferation.Diabetes59,694-701.

• Chaitin,M.H.,Ankrum,M.T.,Wortham,H.S.,1996.DistributionofCD44intheretinaduringdevelopmentandtherdsdegeneration.BrainResDevBrainRes94,92-98.

• Chan,A.,Duker,J.S.,Ko,T.H.,Fujimoto,J.G.,Schuman,J.S.,2006.Normalmacularthickness measurements in healthy eyes using Stratus optical coherencetomography.ArchOphthalmol124,193-198.

• Chargaff,E.,West,R.,1946.Thebiologicalsignificanceof thethromboplasticproteinofblood.JBiolChem166,189-197.

• Chen, H., Weber, A.J., 2002. Expression of glial fibrillary acidic protein andglutaminesynthetasebyMullercellsafteropticnervedamageandintravitrealapplicationofbrain-derivedneurotrophicfactor.Glia38,115-125.

• Chen,L.,Wu,W.,Dentchev,T.,Zeng,Y.,Wang,J.,Tsui,I.,Tobias,J.W.,Bennett,J.,Baldwin,D.,Dunaief,J.L.,2004.Lightdamageinducedchangesinmouseretinalgeneexpression.ExperimentalEyeResearch79,239-247.

• Chu,Y.,Alder,V.A.,Humphrey,M.F.,Constable,I.J.,1999.LocalizationofIgGinthenormalanddystrophicratretinaafterlaserlesions.AustNZJOphthalmol27,117-125.

• Chu-Tan., J.A.,Rutar.,M.,Saxena.,K.,Wu.,Y.,Howitt.,L.,Valter.,K.,Provis., J.,Natoli, R., 2016. Efficacy of 670 nm Light Therapy to Protect againstPhotoreceptor Cell Death Is Dependent on the Severity of Damage.InternationalJournalofPhotoenergy2016.

165

• Chung,H.,Dai,T.,Sharma,S.K.,Huang,Y.Y.,Carroll, J.D.,Hamblin,M.R.,2012.Thenutsandboltsoflow-levellaser(light)therapy.AnnBiomedEng40,516-533.

• Combadiere,C.,Feumi,C.,Raoul,W.,Keller,N.,Rodero,M.,Pezard,A.,Lavalette,S.,Houssier,M.,Jonet,L.,Picard,E.,Debre,P.,Sirinyan,M.,Deterre,P.,Ferroukhi,T.,Cohen,S.Y.,Chauvaud,D.,Jeanny,J.C.,Chemtob,S.,Behar-Cohen,F.,Sennlaub,F., 2007. CX3CR1-dependent subretinal microglia cell accumulation isassociatedwith cardinal featuresof age-relatedmaculardegeneration. JClinInvest117,2920-2928.

• Conlan,M.J.,Rapley,J.W.,Cobb,C.M.,1996.Biostimulationofwoundhealingbylow-energylaserirradiation.Areview.JClinPeriodontol23,492-496.

• Cornish,E.E.,Madigan,M.C.,Natoli,R.,Hales,A.,Hendrickson,A.E.,Provis,J.M.,2005. Gradients of cone differentiation and FGF expression duringdevelopmentofthefovealdepressioninmacaqueretina.VisualNeuroscience22,447-459.

• Crespin,M.,Vidal,C.,Picard,F.,Lacombe,C.,Fontenay,M.,2009.ActivationofPAK1/2 during the shedding of platelet microvesicles. Blood CoagulFibrinolysis20,63-70.

• Cuenca,N.,Fernandez-Sanchez,L.,Campello,L.,Maneu,V.,DelaVilla,P.,Lax,P.,Pinilla, I., 2014.Cellular responses following retinal injuriesand therapeuticapproachesforneurodegenerativediseases.ProgRetinEyeRes43,17-75.

• Cunha-Vaz, J.,2017.TheBlood-RetinalBarrier in theManagementofRetinalDisease:EURETINAAwardLecture.Ophthalmologica237,1-10.

• Curcio, C.A., Sloan, K.R., Kalina, R.E., Hendrickson, A.E., 1990. Humanphotoreceptortopography.JCompNeurol292,497-523.

• Cursiefen,C.,Chen,L.,Borges,L.P.,Jackson,D.,Cao,J.,Radziejewski,C.,D'Amore,P.A., Dana, M.R., Wiegand, S.J., Streilein, J.W., 2004. VEGF-A stimulateslymphangiogenesisandhemangiogenesisininflammatoryneovascularizationviamacrophagerecruitment.JClinInvest113,1040-1050.

• deHoz,R.,Rojas,B.,Ramirez,A.I.,Salazar,J.J.,Gallego,B.I.,Trivino,A.,Ramirez,J.M.,2016.RetinalMacroglialResponsesinHealthandDisease.BiomedResInt2016,2954721.

• Deliyanti, D., Wilkinson-Berka, J.L., 2015. Inhibition of NOX1/4 withGKT137831: a potential novel treatment to attenuate neuroglial cellinflammationintheretina.JournalofNeuroinflammation12.

• Desmet, K.D., Paz, D.A., Corry, J.J., Eells, J.T., Wong-Riley, M.T., Henry, M.M.,Buchmann,E.V.,Connelly,M.P.,Dovi,J.V.,Liang,H.L.,Henshel,D.S.,Yeager,R.L.,Millsap, D.S., Lim, J., Gould, L.J., Das, R., Jett, M., Hodgson, B.D., Margolis, D.,Whelan, H.T., 2006. Clinical and experimental applications of NIR-LEDphotobiomodulation.PhotomedLaserSurg24,121-128.

• Diaz,F.,2010.Cytochromecoxidasedeficiency:patientsandanimalmodels.BiochimBiophysActa1802,100-110.

• Doyle,S.L.,Campbell,M.,Ozaki,E.,Salomon,R.G.,Mori,A.,Kenna,P.F.,Farrar,G.J., Kiang, A.-S., Humphries,M.M., Lavelle, E.C., O'Neill, L.A.J., Hollyfield, J.G.,Humphries, P., 2012a. NLRP3 has a protective role in age-related maculardegenerationthroughtheinductionofIL-18bydrusencomponents.NatMed18,791-798.

• Doyle,S.L.,Campbell,M.,Ozaki,E.,Salomon,R.G.,Mori,A.,Kenna,P.F.,Farrar,G.J., Kiang, A.S., Humphries, M.M., Lavelle, E.C., O'Neill, L.A., Hollyfield, J.G.,Humphries, P., 2012b. NLRP3 has a protective role in age-related macular

166

degenerationthroughtheinductionofIL-18bydrusencomponents.NatMed18,791-798.

• Doyle,S.L.,Lopez,F.J.,Celkova,L.,Brennan,K.,Mulfaul,K.,Ozaki,E.,Kenna,P.F.,Kurali,E.,Hudson,N.,Doggett,T.,Ferguson,T.A.,Humphries,P.,Adamson,P.,Campbell,M.,2015.IL-18ImmunotherapyforNeovascularAMD:TolerabilityandEfficacyinNonhumanPrimates.InvestOphthalmolVisSci56,5424-5430.

• Dragovic,R.A.,Gardiner,C.,Brooks,A.S.,Tannetta,D.S.,Ferguson,D.J.,Hole,P.,Carr,B.,Redman,C.W.,Harris,A.L.,Dobson,P.J.,Harrison,P.,Sargent,I.L.,2011.Sizing and phenotyping of cellular vesicles using Nanoparticle TrackingAnalysis.Nanomedicine7,780-788.

• Du, Y., Tang, J., Li, G., Berti-Mattera, L., Lee, C.A., Bartkowski, D., Gale, D.,Monahan, J., Niesman,M.R., Alton, G., Kern, T.S., 2010. Effects of p38MAPKInhibitiononEarlyStagesofDiabeticRetinopathyandSensoryNerveFunction.InvestOphthalmolVisSci51,2158-2164.

• Eastlake,K.,Banerjee,P.J.,Angbohang,A.,Charteris,D.G.,Khaw,P.T.,Limb,G.A.,2016.Mullergliaasanimportantsourceofcytokinesandinflammatoryfactorspresentintheglioticretinaduringproliferativevitreoretinopathy.Glia64,495-506.

• Edwards,M.M.,McLeod,D.S.,Bhutto,I.A.,Villalonga,M.B.,Seddon,J.M.,Lutty,G.A., 2016. Idiopathic preretinal glia in aging and age-related maculardegeneration.ExpEyeRes150,44-61.

• Eells,J.T.,Henry,M.M.,Summerfelt,P.,Wong-Riley,M.T.,Buchmann,E.V.,Kane,M., Whelan, N.T., Whelan, H.T., 2003. Therapeutic photobiomodulation formethanol-inducedretinaltoxicity.ProcNatlAcadSciUSA100,3439-3444.

• Escrevente,C.,Keller,S.,Altevogt,P.,Costa,J.,2011.Interactionanduptakeofexosomesbyovariancancercells.BMCCancer11,108.

• Evans-Osses, I., Reichembach, L.H., Ramirez, M.I., 2015. Exosomes ormicrovesicles? Two kinds of extracellular vesicles with different routes tomodifyprotozoan-hostcellinteraction.ParasitolRes114,3567-3575.

• Ferguson,L.R.,Dominguez,J.M.,2nd,Balaiya,S.,Grover,S.,Chalam,K.V.,2013.Retinal Thickness Normative Data in Wild-Type Mice Using CustomizedMiniatureSD-OCT.PLoSOne8,e67265.

• Fernandes,K.P., Souza,N.H.,Mesquita-Ferrari,R.A., SilvaDde,F.,Rocha,L.A.,Alves, A.N., Sousa Kde, B., Bussadori, S.K., Hamblin,M.R., Nunes, F.D., 2015.Photobiomodulation with 660-nm and 780-nm laser on activated J774macrophage-like cells: Effect on M1 inflammatory markers. J PhotochemPhotobiolB153,344-351.

• Fernando, N., Natoli, R., Valter, K., Provis, J., Rutar, M., 2016. The broad-spectrumchemokineinhibitorNR58-3.14.3modulatesmacrophage-mediatedinflammationinthediseasedretina.JNeuroinflammation13,47.

• Ferraresi,C.,Huang,Y.Y.,Hamblin,M.R.,2016.Photobiomodulationinhumanmuscle tissue: anadvantage in sportsperformance? JBiophotonics9,1273-1299.

• Fischer,A.J.,Zelinka,C.,Gallina,D.,Scott,M.A.,Todd,L.,2014.ReactiveMicrogliaand Macrophage Facilitate the Formation of Muller Glia-Derived RetinalProgenitors.Glia62,1608-1628.

• Fitzgerald, M., Bartlett, C.A., Payne, S.C., Hart, N.S., Rodger, J., Harvey, A.R.,Dunlop,S.A.,2010.NearinfraredlightreducesoxidativestressandpreservesfunctioninCNStissuevulnerabletosecondarydegenerationfollowingpartialtransectionoftheopticnerve.JNeurotrauma27,2107-2119.

167

• Francke,M.,Makarov,F.,Kacza, J.,Seeger, J.,Wendt,S.,Gartner,U.,Faude,F.,Wiedemann, P., Reichenbach, A., 2001. Retinal pigment epitheliummelaningranules are phagocytozed by Muller glial cells in experimental retinaldetachment.JNeurocytol30,131-136.

• Franze, K., Grosche, J., Skatchkov, S.N., Schinkinger, S., Foja, C., Schild, D.,Uckermann,O.,Travis,K.,Reichenbach,A.,Guck,J.,2007.Mullercellsarelivingopticalfibersinthevertebrateretina.ProcNatlAcadSciUSA104,8287-8292.

• Fukuoka,C.Y.,TorresSchroter,G.,Nicolau,J.,Simoes,A.,2016.Low-powerlaserirradiation in salivary glands reduces glycemia in streptozotocin-induceddiabeticfemalerats.JBiophotonics9,1246-1254.

• Gao,J.,Liu,R.T.,Cao,S.,Cui,J.Z.,Wang,A.,To,E.,Matsubara,J.A.,2015.NLRP3inflammasome:activationandregulationinage-relatedmaculardegeneration.MediatorsInflamm2015,690243.

• Garcia, M., Vecino, E., 2003. Role of Muller glia in neuroprotection andregenerationintheretina.HistolHistopathol18,1205-1218.

• Gavish,L.,Asher,Y.,Becker,Y.,Kleinman,Y.,2004.Lowlevellaserirradiationstimulates mitochondrial membrane potential and disperses subnuclearpromyelocyticleukemiaprotein.LasersSurgMed35,369-376.

• Geneva, II,2016.Photobiomodulationforthetreatmentofretinaldiseases:areview.IntJOphthalmol9,145-152.

• Giacci, M.K., Wheeler, L., Lovett, S., Dishington, E., Majda, B., Bartlett, C.A.,Thornton,E.,Harford-Wright,E.,Leonard,A.,Vink,R.,Harvey,A.R.,Provis, J.,Dunlop,S.A.,Hart,N.S.,Hodgetts,S.,Natoli,R.,VanDenHeuvel,C.,Fitzgerald,M.,2014. Differential effects of 670 and 830 nm red near infrared irradiationtherapy: a comparative study of optic nerve injury, retinal degeneration,traumaticbrainandspinalcordinjury.PLoSOne9,e104565.

• Gilliam, J.C., Chang, J.T., Sandoval, I.M., Zhang, Y., Li, T., Pittler, S.J., Chiu,W.,Wensel,T.G.,2012.Three-dimensionalarchitectureoftherodsensoryciliumanditsdisruptioninretinalneurodegeneration.Cell151,1029-1041.

• Gkotsi,D.,Begum,R.,Salt,T.,Lascaratos,G.,Hogg,C.,Chau,K.Y.,Schapira,A.H.,Jeffery,G.,2014.Rechargingmitochondrialbatteriesinoldeyes.Nearinfra-redincreasesATP.ExpEyeRes122,50-53.

• Goldman,D.,2014.Mullerglial cell reprogrammingand retina regeneration.NatRevNeurosci15,431-442.

• Grigsby, J.G., Cardona, S.M., Pouw,C.E.,Muniz, A.,Mendiola,A.S.,Tsin,A.T.C.,Allen,D.M.,Cardona,A.E.,2014.TheRoleofMicrogliainDiabeticRetinopathy.JournalofOphthalmology.

• Grosche,J.,Hartig,W.,Reichenbach,A.,1995.Expressionofglialfibrillaryacidicprotein(GFAP),glutaminesynthetase(GS),andBcl-2protooncogeneproteinbyMuller(glial)cells inretinallightdamageofrats.NeurosciLett185,119-122.

• Grover, S.,Murthy,R.K.,Brar,V.S.,Chalam,K.V., 2010.Comparisonof retinalthickness in normal eyes using Stratus and Spectralis optical coherencetomography.InvestOphthalmolVisSci51,2644-2647.

• Guidry, C., 1996. Isolation and characterization of porcine Muller cells.Myofibroblasticdedifferentiationinculture.InvestOphthalmolVisSci37,740-752.

• Gyorgy, B., Szabo, T.G., Pasztoi, M., Pal, Z., Misjak, P., Aradi, B., Laszlo, V.,Pallinger,E.,Pap,E.,Kittel,A.,Nagy,G.,Falus,A.,Buzas,E.I.,2011.Membranevesicles, current state-of-the-art: emerging roleof extracellularvesicles.CellMolLifeSci68,2667-2688.

168

• Hajrasouliha,A.R.,Jiang,G.,Lu,Q.,Lu,H.,Kaplan,H.J.,Zhang,H.G.,Shao,H.,2013.Exosomes from retinal astrocytes contain antiangiogenic components thatinhibit laser-induced choroidal neovascularization. J Biol Chem 288, 28058-28067.

• Hamblin, M.R., 2016. Photobiomodulation or low-level laser therapy. JBiophotonics9,1122-1124.

• Hamon,A.,Roger,J.E.,Yang,X.J.,Perron,M.,2016.Mullerglialcell-dependentregeneration of the neural retina: An overview across vertebrate modelsystems.DevDyn245,727-738.

• Harding,C.,Stahl,P.,1983.Transferrinrecyclinginreticulocytes:pHandironareimportantdeterminantsofligandbindingandprocessing.BiochemBiophysResCommun113,650-658.

• Harding, C.V., Heuser, J.E., Stahl, P.D., 2013. Exosomes: looking back threedecadesandintothefuture.JCellBiol200,367-371.

• Hariri, S., Moayed, A.A., Choh, V., Bizheva, K., 2012. In vivo assessment ofthickness and reflectivity in a rat outer retinal degeneration model withultrahighresolutionopticalcoherencetomography.InvestOphthalmolVisSci53,1982-1989.

• Hernandez-Zimbron, L.F., Zamora-Alvarado, R., Ochoa-De la Paz, L., Velez-Montoya, R., Zenteno, E., Gulias-Canizo, R., Quiroz-Mercado, H., Gonzalez-Salinas, R., 2018. Age-Related Macular Degeneration: New Paradigms forTreatmentandManagementofAMD.OxidMedCellLongev2018,8374647.

• Herranz-Aparicio,J.,Vazquez-Delgado,E.,Arnabat-Dominguez,J.,Espana-Tost,A.,Gay-Escoda,C.,2013.Theuseoflowlevellasertherapyinthetreatmentoftemporomandibular jointdisorders.Reviewof the literature.MedOralPatolOralCirBucal18,e603-612.

• Hicks,D.,Courtois,Y.,1990.ThegrowthandbehaviourofratretinalMullercellsinvitro.1.Animprovedmethodforisolationandculture.ExpEyeRes51,119-129.

• Hirano,Y.,Yasuma,T.,Mizutani,T.,Fowler,B.J.,Tarallo,V.,Yasuma,R.,Kim,Y.,Bastos-Carvalho,A.,Kerur,N.,Gelfand,B.D.,Bogdanovich,S.,He,S.,Zhang,X.,Nozaki,M.,Ijima,R.,Kaneko,H.,Ogura,Y.,Terasaki,H.,Nagai,H.,Haro,I.,Nunez,G., Ambati, B.K., Hinton, D.R., Ambati, J., 2014. IL-18 is not therapeutic forneovascularage-relatedmaculardegeneration.NatMed20,1372-1375.

• Hollborn,M.,Francke,M.,Iandiev,I.,Buhner,E.,Foja,C.,Kohen,L.,Reichenbach,A., Wiedemann, P., Bringmann, A., Uhlmann, S., 2008. Early activation ofinflammation- and immune response-related genes after experimentaldetachmentoftheporcineretina.InvestOphthalmolVisSci49,1262-1273.

• Hollborn,M.,Ulbricht,E.,Rillich,K.,Dukic-Stefanovic,S.,Wurm,A.,Wagner,L.,Reichenbach,A.,Wiedemann,P.,Limb,G.A.,Bringmann,A.,Kohen,L.,2011.ThehumanMullercelllineMIO-M1expressesopsins.MolVis17,2738-2750.

• Hosoya, K., Tachikawa,M., 2012. The inner blood-retinal barrier:molecularstructureandtransportbiology.AdvExpMedBiol763,85-104.

• Hsu,S.D.,Lin,F.M.,Wu,W.Y.,Liang,C.,Huang,W.C.,Chan,W.L.,Tsai,W.T.,Chen,G.Z.,Lee,C.J.,Chiu,C.M.,Chien,C.H.,Wu,M.C.,Huang,C.Y.,Tsou,A.P.,Huang,H.D.,2011. miRTarBase: a database curates experimentally validated microRNA-targetinteractions.NucleicAcidsRes39,D163-169.

• Hu, D., Zhu, S., Potas, J.R., 2016. Red LED photobiomodulation reduces painhypersensitivity and improves sensorimotor function following mild T10hemicontusionspinalcordinjury.JNeuroinflammation13,200.

169

• Huangda,W.,Sherman,B.T.,Lempicki,R.A.,2009a.Bioinformaticsenrichmenttools:pathstowardthecomprehensivefunctionalanalysisoflargegenelists.NucleicAcidsRes37,1-13.

• Huangda,W.,Sherman,B.T.,Lempicki,R.A.,2009b.SystematicandintegrativeanalysisoflargegenelistsusingDAVIDbioinformaticsresources.NatProtoc4,44-57.

• Iandiev,I.,Pannicke,T.,Hollborn,M.,Wiedemann,P.,Reichenbach,A.,Grimm,C.,Reme,C.E.,Bringmann,A.,2008.Localizationofglialaquaporin-4andKir4.1inthelight-injuredmurineretina.NeurosciLett434,317-321.

• Inal, J.M.,Ansa-Addo,E.A.,Stratton,D.,Kholia,S.,Antwi-Baffour,S.S., Jorfi,S.,Lange,S.,2012.Microvesicles inhealthanddisease.ArchImmunolTherExp(Warsz)60,107-121.

• Inatani, M., Tanihara, H., Oohira, A., Honjo, M., Kido, N., Honda, Y., 2000.Upregulatedexpressionofneurocan,anervoustissuespecificproteoglycan,intransientretinalischemia.InvestOphthalmolVisSci41,2748-2754.

• Indaram,M.,Ma,W.,Zhao,L.,Fariss,R.N.,Rodriguez,I.R.,Wong,W.T.,2015.7-Ketocholesterol increases retinal microglial migration, activation, andangiogenicity: a potential pathogenic mechanism underlying age-relatedmaculardegeneration.SciRep5,9144.

• Ivandic, B.T., Ivandic, T., 2008. Low-level laser therapy improves vision inpatientswithage-relatedmaculardegeneration.PhotomedLaserSurg26,241-245.

• Ivandic,B.T.,Ivandic,T.,2012.Low-levellasertherapyimprovesvisualacuityinadolescentandadultpatientswithamblyopia.PhotomedLaserSurg30,167-171.

• Jadhav, A.P., Roesch, K., Cepko, C.L., 2009. Development and neurogenicpotentialofMullerglialcellsinthevertebrateretina.ProgRetinEyeRes28,249-262.

• Jager,R.D.,Mieler,W.F.,Miller,J.W.,2008.Age-relatedmaculardegeneration.NEnglJMed358,2606-2617.

• Johnstone, R.M., Adam,M.,Hammond, J.R., Orr, L., Turbide, C., 1987. Vesicleformation during reticulocytematuration. Association of plasmamembraneactivitieswithreleasedvesicles(exosomes).JBiolChem262,9412-9420.

• Joly, S., Lange, C., Thiersch, M., Samardzija, M., Grimm, C., 2008. Leukemiainhibitory factor extends the lifespan of injured photoreceptors in vivo. JNeurosci28,13765-13774.

• Joly,S.,Pernet,V.,Chemtob,S.,DiPolo,A.,Lachapelle,P.,2007.Neuroprotectioninthe juvenileratmodeloflight-inducedretinopathy:evidencesuggestingaroleforFGF-2andCNTF.InvestOphthalmolVisSci48,2311-2320.

• Jonas,J.B.,Jonas,R.A.,Neumaier,M.,Findeisen,P.,2012.Cytokineconcentrationinaqueoushumorofeyeswithdiabeticmacularedema.Retina32,2150-2157.

• Kaldi,I.,Dittmar,M.,Pierce,P.,Anderson,R.E.,2003.L-NAMEprotectsagainstacutelightdamageinalbinorats,butnotagainstretinaldegenerationinP23HandS334tertransgenicrats.ExpEyeRes76,453-461.

• Kalra,H.,Drummen,G.P.,Mathivanan,S.,2016.FocusonExtracellularVesicles:IntroducingtheNextSmallBigThing.IntJMolSci17,170.

• Kang, J., Pervaiz, S., 2012.Mitochondria: redoxmetabolism and dysfunction.BiochemResInt2012,896751.

170

• Karlsson,M.,Lundin,S.,Dahlgren,U.,Kahu,H.,Pettersson,I.,Telemo,E.,2001."Tolerosomes"areproducedby intestinal epithelial cells.Eur J Immunol31,2892-2900.

• Karu,T.,1999.Primaryandsecondarymechanismsofactionofvisibletonear-IRradiationoncells.JPhotochemPhotobiolB49,1-17.

• Karu,T.I.,2008.Mitochondrialsignalinginmammaliancellsactivatedbyredandnear-IRradiation.PhotochemistryandPhotobiology84,1091-1099.

• Karu, T.I., Kolyakov, S.F., 2005. Exact action spectra for cellular responsesrelevanttophototherapy.PhotomedLaserSurg23,355-361.

• Kataoka,K.,Matsumoto,H.,Kaneko,H.,Notomi,S.,Takeuchi,K.,Sweigard,J.H.,Atik,A.,Murakami,Y.,Connor,K.M.,Terasaki,H.,Miller,J.W.,Vavvas,D.G.,2015.Macrophage- and RIP3-dependent inflammasome activation exacerbatesretinaldetachment-inducedphotoreceptorcelldeath.CellDeathDis6,e1731.

• Kauppinen, A., Paterno, J.J., Blasiak, J., Salminen, A., Kaarniranta, K., 2016.Inflammationanditsroleinage-relatedmaculardegeneration.CellMolLifeSci73,1765-1786.

• Kaylor,J.J.,Yuan,Q.,Cook,J.,Sarfare,S.,Makshanoff,J.,Miu,A.,Kim,A.,Kim,P.,Habib,S.,Roybal,C.N.,Xu,T.,Nusinowitz,S.,Travis,G.H.,2013.IdentificationofDES1asavitaminAisomeraseinMullerglialcellsoftheretina.NatChemBiol9,30-36.

• Kaynezhad, P., Tachtsidis, I., Jeffery, G., 2016. Optical monitoring of retinalrespirationinrealtime:670nmlightincreasestheredoxstateofmitochondria.ExpEyeRes152,88-93.

• Khanna,H.,2015.PhotoreceptorSensoryCilium:TraversingtheCiliaryGate.Cells4,674-686.

• Kim, H.P., 2014. Lightening up Light Therapy: Activation of RetrogradeSignalingPathwaybyPhotobiomodulation.BiomolTher(Seoul)22,491-496.

• Kim, S.Y., 2015. Retinal phagocytes in age-related macular degeneration.Macrophage(Houst)2,e698.

• Klein, R., Myers, C.E., Cruickshanks, K.J., Gangnon, R.E., Danforth, L.G.,Sivakumaran, T.A., Iyengar, S.K., Tsai, M.Y., Klein, B.E., 2014. Markers ofinflammation, oxidative stress, and endothelial dysfunction and the 20-yearcumulative incidence of early age-relatedmacular degeneration: theBeaverDamEyeStudy.JAMAOphthalmol132,446-455.

• Klingeborn, M., Dismuke, W.M., Rickman, C.B., Stamer, W.D., 2017. Roles ofexosomesinthenormalanddiseasedeye.ProgRetinEyeRes59,158-177.

• Knickelbein, J.E., Chan, C.C., Sen, H.N., Ferris, F.L., Nussenblatt, R.B., 2015.Inflammatory Mechanisms of Age-related Macular Degeneration. IntOphthalmolClin55,63-78.

• Knickelbein,J.E.,Liu,B.,Arakelyan,A.,Zicari,S.,Hannes,S.,Chen,P.,Li,Z.,Grivel,J.C., Chaigne-Delalande, B., Sen, H.N., Margolis, L., Nussenblatt, R.B., 2016.Modulation of Immune Responses by Extracellular Vesicles From RetinalPigmentEpithelium.InvestOphthalmolVisSci57,4101-4107.

• Kobayashi,K.,Imagama,S.,Ohgomori,T.,Hirano,K.,Uchimura,K.,Sakamoto,K.,Hirakawa, A., Takeuchi, H., Suzumura, A., Ishiguro, N., Kadomatsu, K., 2013.MinocyclineselectivelyinhibitsM1polarizationofmicroglia.CellDeathDis4,e525.

• Kokkinopoulos,I.,2013a.670nmLEDamelioratesinflammationintheCFH(-/-)mouseneuralretina.JPhotochemPhotobiolB122,24-31.

171

• Kokkinopoulos,I.,Colman,A.,Hogg,C.,Heckenlively,J., Jeffery,G.,2013.Age-related retinal inflammation is reduced by 670 nm light via increasedmitochondrialmembranepotential.NeurobiolAging34,602-609.

• Koliha,N.,Wiencek,Y.,Heider,U.,Jungst,C.,Kladt,N.,Krauthauser,S.,Johnston,I.C.,Bosio,A.,Schauss,A.,Wild,S.,2016.Anovelmultiplexbead-basedplatformhighlightsthediversityofextracellularvesicles.JExtracellVesicles5,29975.

• Koumangoye, R.B., Sakwe, A.M., Goodwin, J.S., Patel, T., Ochieng, J., 2011.Detachmentofbreasttumorcellsinducesrapidsecretionofexosomeswhichsubsequentlymediatecellularadhesionandspreading.PLoSOne6,e24234.

• Kowluru,R.A.,Odenbach,S.,2004.Roleofinterleukin-1betainthedevelopmentof retinopathy in rats: effect of antioxidants. Invest Ophthalmol Vis Sci 45,4161-4166.

• Krady,J.K.,Basu,A.,Allen,C.M.,Xu,Y.P.,LaNoue,K.F.,Gardner,T.W.,Levison,S.W., 2005. Minocycline reduces proinflammatory cytokine expression,microglial activation, andcaspase-3activation ina rodentmodelofdiabeticretinopathy.Diabetes54,1559-1565.

• Kumar,A.,Pandey,R.K.,Miller,L.J.,Singh,P.K.,Kanwar,M.,2013.Mullergliainretinalinnateimmunity:aperspectiveontheirrolesinendophthalmitis.CritRevImmunol33,119-135.

• Lavalette,S.,Raoul,W.,Houssier,M.,Camelo,S.,Levy,O.,Calippe,B.,Jonet,L.,Behar-Cohen, F., Chemtob, S., Guillonneau, X., Combadiere, C., Sennlaub, F.,2011.Interleukin-1betainhibitionpreventschoroidalneovascularizationanddoesnotexacerbatephotoreceptordegeneration.AmJPathol178,2416-2423.

• Lee, J.H.,Chiang,M.H.,Chen, P.H.,Ho,M.L., Lee,H.E.,Wang,Y.H.,2018.Anti-inflammatoryeffectsoflow-levellasertherapyonhumanperiodontalligamentcells:invitrostudy.LasersMedSci33,469-477.

• Lee, Y., El Andaloussi, S., Wood, M.J., 2012. Exosomes and microvesicles:extracellularvesiclesforgeneticinformationtransferandgenetherapy.HumMolGenet21,R125-134.

• Lewis,G.,Mervin,K.,Valter,K.,Maslim,J.,Kappel,P.J.,Stone,J.,Fisher,S.,1999.Limiting the proliferation and reactivity of retinal Muller cells duringexperimentalretinaldetachment:thevalueofoxygensupplementation.AmJOphthalmol128,165-172.

• Lewis,G.P.,Chapin,E.A.,Luna,G.,Linberg,K.A.,Fisher,S.K.,2010.ThefateofMuller'sgliafollowingexperimentalretinaldetachment:nuclearmigration,celldivision,andsubretinalglialscarformation.MolVis16,1361-1372.

• Lewis, G.P., Erickson, P.A., Guerin, C.J., Anderson, D.H., Fisher, S.K., 1989.Changes in the expression of specific Muller cell proteins during long-termretinaldetachment.ExpEyeRes49,93-111.

• Lewis,G.P.,Erickson,P.A.,Guerin,C.J.,Anderson,D.H.,Fisher,S.K.,1992.Basicfibroblastgrowthfactor:apotentialregulatorofproliferationandintermediatefilamentexpressionintheretina.JNeurosci12,3968-3978.

• Lewis,G.P.,Fisher,S.K.,2003.Up-regulationofglialfibrillaryacidicproteininresponse to retinal injury: its potential role in glial remodeling and acomparisontovimentinexpression.IntRevCytol230,263-290.

• Lewis,G.P.,Guerin,C.J.,Anderson,D.H.,Matsumoto,B.,Fisher,S.K.,1994.Rapidchangesintheexpressionofglialcellproteinscausedbyexperimentalretinaldetachment.AmJOphthalmol118,368-376.

• Li, Y., Park, J.S., Deng, J.H., Bai, Y., 2006. Cytochrome c oxidase subunit IV isessential for assembly and respiratory function of the enzyme complex. JBioenergBiomembr38,283-291.

172

• Liang,C.C.,Park,A.Y.,Guan,J.L.,2007.Invitroscratchassay:aconvenientandinexpensivemethodforanalysisofcellmigrationinvitro.NatProtoc2,329-333.

• Liang,H.L.,Whelan,H.T.,Eells,J.T.,Wong-Riley,M.T.,2008.Near-infraredlightvia light-emitting diode treatment is therapeutic against rotenone- and 1-methyl-4-phenylpyridinium ion-induced neurotoxicity. Neuroscience 153,963-974.

• Lieth,E.,Barber,A.J.,Xu,B.,Dice,C.,Ratz,M.J.,Tanase,D.,Strother,J.M.,1998.Glialreactivityandimpairedglutamatemetabolisminshort-termexperimentaldiabeticretinopathy.PennStateRetinaResearchGroup.Diabetes47,815-820.

• Limb,G.A.,Daniels,J.T.,Pleass,R.,Charteris,D.G.,Luthert,P.J.,Khaw,P.T.,2002a.Differential expression ofmatrixmetalloproteinases 2 and 9 by glialMullercells: response to soluble and extracellular matrix-bound tumor necrosisfactor-alpha.AmJPathol160,1847-1855.

• Limb, G.A., Salt, T.E., Munro, P.M., Moss, S.E., Khaw, P.T., 2002b. In vitrocharacterizationofaspontaneouslyimmortalizedhumanMullercellline(MIO-M1).InvestOphthalmolVisSci43,864-869.

• Liu,R.T.,Gao,J.,Cao,S.,Sandhu,N.,Cui,J.Z.,Chou,C.L.,Fang,E.,Matsubara,J.A.,2013.Inflammatorymediatorsinducedbyamyloid-betaintheretinaandRPEin vivo: implications for inflammasome activation in age-related maculardegeneration.InvestOphthalmolVisSci54,2225-2237.

• Liu,Y.,BiarnesCosta,M.,Gerhardinger,C.,2012.IL-1betaisupregulatedinthediabeticretinaandretinal vessels:cell-specificeffectofhighglucoseandIL-1betaautostimulation.PLoSOne7,e36949.

• Lozano,D.C.,Twa,M.D.,2012.QuantitativeevaluationoffactorsinfluencingtherepeatabilityofSD-OCTthicknessmeasurementsintherat.InvestOphthalmolVisSci53,8378-8385.

• Lu, Q., Jiang, Y.R., Qian, J., Tao, Y., 2013. Apelin-13 regulates proliferation,migrationandsurvivalofretinalMullercellsunderhypoxia.DiabetesResClinPract99,158-167.

• Lu, Y.B., Iandiev, I., Hollborn, M., Korber, N., Ulbricht, E., Hirrlinger, P.G.,Pannicke,T.,Wei,E.Q.,Bringmann,A.,Wolburg,H.,Wilhelmsson,U.,Pekny,M.,Wiedemann,P.,Reichenbach,A.,Kas,J.A.,2011.Reactiveglialcells:increasedstiffnesscorrelateswithincreasedintermediatefilamentexpression.FASEBJ25,624-631.

• Lu,Y.Z.,Natoli,R.,Madigan,M.,Fernando,N.,Saxena,K.,Aggio-Bruce,R.,Jiao,H.,Provis,J.,Valter,K.,2017.Photobiomodulationwith670nmlightamelioratesMuller cell-mediated activation of microglia and macrophages in retinaldegeneration.ExpEyeRes165,78-89.

• Luckoff,A.,Scholz,R.,Sennlaub,F.,Xu,H.,Langmann,T.,2017.Comprehensiveanalysisofmouseretinalmononuclearphagocytes.NatProtoc12,1136-1150.

• Ludwig, A.K., Giebel, B., 2012. Exosomes: small vesicles participating inintercellularcommunication.IntJBiochemCellBiol44,11-15.

• Luga, V., Zhang, L., Viloria-Petit, A.M., Ogunjimi, A.A., Inanlou,M.R., Chiu, E.,Buchanan, M., Hosein, A.N., Basik, M., Wrana, J.L., 2012. Exosomes mediatestromal mobilization of autocrine Wnt-PCP signaling in breast cancer cellmigration.Cell151,1542-1556.

• Ma,W.,Cojocaru,R.,Gotoh,N.,Gieser,L.,Villasmil,R.,Cogliati,T.,Swaroop,A.,Wong, W.T., 2013. Gene expression changes in aging retinal microglia:relationship to microglial support functions and regulation of activation.NeurobiolAging34,2310-2321.

173

• Madeira, M.H., Boia, R., Santos, P.F., Ambrosio, A.F., Santiago, A.R., 2015.Contributionofmicroglia-mediatedneuroinflammationtoretinaldegenerativediseases.MediatorsInflamm2015,673090.

• Marc, R.E., Jones, B.W., Watt, C.B., Vazquez-Chona, F., Vaughan, D.K.,Organisciak,D.T.,2008.Extremeretinalremodelingtriggeredbylightdamage:implicationsforagerelatedmaculardegeneration.MolVis14,782-806.

• Marco,F.D.,Romeo,S.,Nandasena,C.,Purushothuman,S.,Adams,C.,Bisti,S.,Stone, J., 2013. The time course of action of two neuroprotectants, dietarysaffronandphotobiomodulation,assessedintheratretina.AmJNeurodegenerDis2,208-220.

• Matsushima,K., Larsen, C.G., DuBois, G.C., Oppenheim, J.J., 1989. Purificationand characterization of a novelmonocyte chemotactic and activating factorproducedbyahumanmyelomonocyticcellline.JExpMed169,1485-1490.

• McDonagh, A.F., 2001. Phototherapy: from ancient Egypt to the newmillennium.JPerinatol21Suppl1,S7-S12.

• Merry, G.F., Munk, M.R., Dotson, R.S., Walker, M.G., Devenyi, R.G., 2017.Photobiomodulation reducesdrusenvolumeand improvesvisual acuityandcontrastsensitivityindryage-relatedmaculardegeneration.ActaOphthalmol95,e270-e277.

• Mester,E.,Szende,B.,Gartner,P.,1968.[Theeffectoflaserbeamsonthegrowthofhairinmice].RadiobiolRadiother(Berl)9,621-626.

• Miao,H.,Tao,Y.,Li,X.X.,2012. Inflammatorycytokines inaqueoushumorofpatientswithchoroidalneovascularization.MolVis18,574-580.

• Minciacchi, V.R., Freeman, M.R., Di Vizio, D., 2015. Extracellular vesicles incancer: exosomes,microvesicles and the emerging role of large oncosomes.SeminCellDevBiol40,41-51.

• Miyazono, S., Shimauchi-Matsukawa,Y.,Tachibanaki, S.,Kawamura, S., 2008.Highly efficient retinalmetabolism in cones. Proc Natl Acad Sci U S A 105,16051-16056.

• Mizutani,M.,Gerhardinger,C.,Lorenzi,M.,1998.Mullercellchangesinhumandiabeticretinopathy.Diabetes47,445-449.

• Morelli,A.E.,Larregina,A.T.,Shufesky,W.J.,Sullivan,M.L.,Stolz,D.B.,Papworth,G.D.,Zahorchak,A.F.,Logar,A.J.,Wang,Z.,Watkins,S.C.,Falo,L.D.,Jr.,Thomson,A.W.,2004.Endocytosis,intracellularsorting,andprocessingofexosomesbydendriticcells.Blood104,3257-3266.

• Mulcahy, L.A., Pink, R.C., Carter, D.R., 2014. Routes and mechanisms ofextracellularvesicleuptake.JExtracellVesicles3.

• Müller,H.,1851.ZurHistologiederNetzhaut.ZeitschriftfürWissZool3,234-237.

• Muralidharan-Chari,V.,Clancy, J.W., Sedgwick,A.,D'Souza-Schorey,C., 2010.Microvesicles: mediators of extracellular communication during cancerprogression.JCellSci123,1603-1611.

• Murata, J., Ricciardi-Castagnoli, P., Dessous L'Eglise Mange, P., Martin, F.,Juillerat-Jeanneret, L., 1997. Microglial cells induce cytotoxic effects towardcolon carcinoma cells: measurement of tumor cytotoxicity with a gamma-glutamyltranspeptidaseassay.IntJCancer70,169-174.

• Mustafi,D.,Engel,A.H.,Palczewski,K.,2009.Structureofconephotoreceptors.ProgRetinEyeRes28,289-302.

• Nakamachi,Y.,Kawano,S.,Takenokuchi,M.,Nishimura,K.,Sakai,Y.,Chin,T.,Saura,R.,Kurosaka,M.,Kumagai,S.,2009.MicroRNA-124aisakeyregulatorof

174

proliferationandmonocytechemoattractantprotein1secretioninfibroblast-likesynoviocytesfrompatientswithrheumatoidarthritis.ArthritisRheum60,1294-1304.

• Nakazawa, T., Hisatomi, T., Nakazawa, C., Noda, K., Maruyama, K., She, H.,Matsubara,A.,Miyahara,S.,Nakao,S.,Yin,Y.,Benowitz,L.,Hafezi-Moghadam,A.,Miller, J.W., 2007a.Monocyte chemoattractant protein1mediates retinaldetachment-inducedphotoreceptorapoptosis.ProcNatlAcadSciUSA104,2425-2430.

• Nakazawa,T.,Takeda,M.,Lewis,G.P.,Cho,K.S.,Jiao,J.,Wilhelmsson,U.,Fisher,S.K.,Pekny,M.,Chen,D.F.,Miller, J.W.,2007b.Attenuatedglial reactionsandphotoreceptordegenerationafterretinaldetachmentinmicedeficientinglialfibrillaryacidicproteinandvimentin.InvestOphthalmolVisSci48,2760-2768.

• Natoli,R.,Fernando,N.,Jiao,H.,Racic,T.,Madigan,M.,Barnett,N.L.,Chu-Tan,J.A.,Valter,K.,Provis,J.,Rutar,M.,2017a.RetinalMacrophagesSynthesizeC3andActivateComplementinAMDandinModelsofFocalRetinalDegeneration.InvestOphthalmolVisSci58,2977-2990.

• Natoli,R.,Fernando,N.,Madigan,M.,Chu-Tan,J.A.,Valter,K.,Provis,J.,Rutar,M.,2017b.Microglia-derivedIL-1βpromoteschemokineexpressionbyMüllercellsandRPEinfocalretinaldegeneration.MolNeurodegener12,31.

• Natoli,R.,Jiao,H.,Barnett,N.L.,Fernando,N.,Valter,K.,Provis,J.M.,Rutar,M.,2016a.Amodelofprogressivephoto-oxidativedegenerationandinflammationinthepigmentedC57BL/6Jmouseretina.ExperimentalEyeResearch.

• Natoli, R., Rutar,M., Lu, Y.Z., Chu-Tan, J.A., Chen, Y., Saxena, K.,Madigan,M.,Valter, K., Provis, J.M., 2016b. The Role of Pyruvate in Protecting 661WPhotoreceptor-Like Cells Against Light-Induced Cell Death. Curr EyeRes41,1473-1481.

• Natoli, R., Valter, K., Barbosa,M., Dahlstrom, J., Rutar,M., Kent, A., Provis, J.,2013.670nmphotobiomodulationasanovelprotectionagainstretinopathyofprematurity:evidencefromoxygeninducedretinopathymodels.PLoSOne8,e72135.

• Natoli,R.,Zhu,Y.,Valter,K.,Bisti,S.,Eells,J.,Stone,J.,2010.GeneandnoncodingRNA regulation underlying photoreceptor protection: microarray study ofdietaryantioxidantsaffronandphotobiomodulationinratretina.MolVis16,1801-1822.

• Nazarenko,I.,Rana,S.,Baumann,A.,McAlear,J.,Hellwig,A.,Trendelenburg,M.,Lochnit, G., Preissner, K.T., Zoller,M., 2010. Cell surface tetraspaninTspan8contributes to molecular pathways of exosome-induced endothelial cellactivation.CancerRes70,1668-1678.

• Nelson,B.R.,Ueki,Y.,Reardon,S.,Karl,M.O.,Georgi,S.,Hartman,B.H.,Lamba,D.A., Reh, T.A., 2011. Genome-wide analysis of Muller glial differentiationrevealsarequirementforNotchsignalinginpostmitoticcellstomaintaintheglialfate.PLoSOne6,e22817.

• Netea,M.G.,vandeVeerdonk,F.L.,vanderMeer,J.W.,Dinarello,C.A.,Joosten,L.A., 2015. Inflammasome-independent regulation of IL-1-family cytokines.AnnuRevImmunol33,49-77.

• Ni, Y.Q., Xu, G.Z., Hu,W.Z., Shi, L., Qin, Y.W., Da, C.D., 2008. Neuroprotectiveeffectsofnaloxoneagainstlight-inducedphotoreceptordegenerationthroughinhibiting retinalmicroglial activation. Invest Ophthalmol Vis Sci 49, 2589-2598.

• Nigita, G., Acunzo, M., Romano, G., Veneziano, D., Laganà, A., Vitiello, M.,Wernicke, D., Ferro, A., Croce, C.M., 2016. microRNA editing in seed region

175

alignswithcellularchangesinhypoxicconditions.NucleicAcidsRes44,6298-6308.

• Nir,I.,Sagie,G.,Papermaster,D.S.,1987.OpsinaccumulationinphotoreceptorinnersegmentplasmamembranesofdystrophicRCSrats.InvestOphthalmolVisSci28,62-69.

• Noell,W.K.,Walker,V.S.,Kang,B.S.,Berman,S.,1966.Retinaldamagebylightinrats.InvestOphthalmol5,450-473.

• Oh,H.,Takagi,H.,Takagi,C., Suzuma,K.,Otani,A., Ishida,K.,Matsumura,M.,Ogura,Y.,Honda,Y.,1999.Thepotentialangiogenicroleofmacrophagesintheformationofchoroidalneovascularmembranes.InvestOphthalmolVisSci40,1891-1898.

• Ohsawa,R.,Kageyama,R.,2008.Regulationofretinalcellfatespecificationbymultipletranscriptionfactors.BrainRes1192,90-98.

• Omri,S.,Behar-Cohen,F.,deKozak,Y.,Sennlaub,F.,Verissimo,L.M.,Jonet,L.,Savoldelli, M., Omri, B., Crisanti, P., 2011. Microglia/macrophages migratethrough retinal epithelium barrier by a transcellular route in diabeticretinopathy:roleofPKCzetaintheGotoKakizakiratmodel.AmJPathol179,942-953.

• Omri,S.,Omri,B., Savoldelli,M., Jonet,L.,Thillaye-Goldenberg,B.,Thuret,G.,Gain, P., Jeanny, J.C., Crisanti, P., Behar-Cohen, F., 2010. The outer limitingmembrane(OLM)revisited:clinicalimplications.ClinOphthalmol4,183-195.

• Ooto,S.,Akagi,T.,Kageyama,R.,Akita,J.,Mandai,M.,Honda,Y.,Takahashi,M.,2004. Potential for neural regeneration after neurotoxic injury in the adultmammalianretina.ProcNatlAcadSciUSA101,13654-13659.

• Organisciak,D.T.,Vaughan,D.K.,2010.Retinallightdamage:mechanismsandprotection.ProgRetinEyeRes29,113-134.

• Pan, B.T., Johnstone, R.M., 1983. Fate of the transferrin receptor duringmaturation of sheep reticulocytes in vitro: selective externalization of thereceptor.Cell33,967-978.

• Parmeggiani, F., Romano, M.R., Costagliola, C., Semeraro, F., Incorvaia, C.,D'Angelo,S.,Perri,P.,DePalma,P.,DeNadai,K.,Sebastiani,A.,2012.Mechanismofinflammationinage-relatedmaculardegeneration.MediatorsInflamm2012,546786.

• Perez-Gonzalez, R., Gauthier, S.A., Kumar, A., Levy, E., 2012. The exosomesecretory pathway transports amyloid precursor protein carboxyl-terminalfragments from the cell into the brain extracellular space. J Biol Chem 287,43108-43115.

• Pfeiffer,R.L.,Marc,R.E.,Kondo,M.,Terasaki,H.,Jones,B.W.,2016.Mullercellmetabolicchaosduringretinaldegeneration.ExpEyeRes150,62-70.

• Piccolino,M.,Strettoi,E.,Laurenzi,E.,1989.SantiagoRamonyCajal,theretinaandtheneurontheory.DocOphthalmol71,123-141.

• Ponta,H.,Sherman,L.,Herrlich,P.A.,2003.CD44:fromadhesionmoleculestosignallingregulators.NatRevMolCellBiol4,33-45.

• Pospichalova, V., Svoboda, J., Dave, Z., Kotrbova, A., Kaiser, K., Klemova, D.,Ilkovics,L.,Hampl,A.,Crha,I.,Jandakova,E.,Minar,L.,Weinberger,V.,Bryja,V.,2015.Simplifiedprotocolforflowcytometryanalysisoffluorescentlylabeledexosomes and microvesicles using dedicated flow cytometer. J ExtracellVesicles4,25530.

• Pow,D.V.,Crook,D.K.,1995.Immunocytochemicalevidenceforthepresenceofhighlevelsofreducedglutathioneinradialglialcellsandhorizontalcellsintherabbitretina.NeurosciLett193,25-28.

176

• Provis,J.M.,Diaz,C.M.,Dreher,B.,1998.Ontogenyoftheprimatefovea:acentralissueinretinaldevelopment.ProgNeurobiol54,549-580.

• Provis,J.M.,Dubis,A.M.,Maddess,T.,Carroll,J.,2013.Adaptationofthecentralretinaforhighacuityvision:cones,thefoveaandtheavascularzone.ProgRetinEyeRes35,63-81.

• RamónYCajal,S.,1988.Morphologiayconexionesdeloselementosdelaretinadelasaves.RevTrimHistolNormalPatol1,11-16.

• Rangasamy,S.,McGuire,P.G.,FrancoNitta,C.,Monickaraj,F.,Oruganti,S.R.,Das,A., 2014. Chemokine mediated monocyte trafficking into the retina: role ofinflammationinalterationoftheblood-retinalbarrierindiabeticretinopathy.PLoSOne9,e108508.

• Raposo, G., Stoorvogel, W., 2013. Extracellular vesicles: exosomes,microvesicles,andfriends.JCellBiol200,373-383.

• Reichenbach,A.,Bringmann,A., 2013.New functionsofMuller cells.Glia61,651-678.

• Reichenbach,A.,Stolzenburg,J.U.,Wolburg,H.,Hartig,W.,el-Hifnawi,E.,Martin,H.,1995.Effectsofenhancedextracellularammoniaconcentrationonculturedmammalianretinalglial(Muller)cells.Glia13,195-208.

• Richardson, P.R., Boulton, M.E., Duvall-Young, J., McLeod, D., 1996.Immunocytochemicalstudyofretinaldiodelaserphotocoagulationintherat.BrJOphthalmol80,1092-1098.

• Rivera,J.C.,Sitaras,N.,Noueihed,B.,Hamel,D.,Madaan,A.,Zhou,T.,Honore,J.C.,Quiniou, C., Joyal, J.S., Hardy, P., Sennlaub, F., Lubell,W., Chemtob, S., 2013.Microglia and interleukin-1beta in ischemic retinopathy elicitmicrovasculardegenerationthroughneuronalsemaphorin-3A.ArteriosclerThrombVascBiol33,1881-1891.

• Rohrer, B., Guo, Y., Kunchithapautham, K., Gilkeson, G.S., 2007. Eliminatingcomplement factor D reduces photoreceptor susceptibility to light-induceddamage.InvestigativeOphthalmology&VisualScience48,5282-5289.

• Rojas,J.C.,Gonzalez-Lima,F.,2011.Low-levellighttherapyoftheeyeandbrain.EyeBrain3,49-67.

• Romo,P.,Madigan,M.C.,Provis,J.M.,Cullen,K.M.,2011.DifferentialeffectsofTGF-betaandFGF-2oninvitroproliferationandmigrationofprimateretinalendothelialandMullercells.ActaOphthalmol89,e263-268.

• Roy,A.,Fung,Y.K.,Liu,X.,Pahan,K.,2006.Up-regulationofmicroglialCD11bexpressionbynitricoxide.JBiolChem281,14971-14980.

• Rozmyslowicz,T.,Majka,M.,Kijowski,J.,Murphy,S.L.,Conover,D.O.,Poncz,M.,Ratajczak,J.,Gaulton,G.N.,Ratajczak,M.Z.,2003.Platelet-andmegakaryocyte-derivedmicroparticlestransferCXCR4receptortoCXCR4-nullcellsandmakethemsusceptibletoinfectionbyX4-HIV.AIDS17,33-42.

• Ruby,N.F.,Brennan,T.J.,Xie,X.,Cao,V.,Franken,P.,Heller,H.C.,O'Hara,B.F.,2002.Roleofmelanopsin incircadianresponsesto light.Science298,2211-2213.

• Rutar,M.,Natoli, R., Albarracin, R., Valter, K., Provis, J., 2012a. 670-nm lighttreatmentreducescomplementpropagation followingretinaldegeneration. JNeuroinflammation9,257.

• Rutar, M., Natoli, R., Chia, R.X., Valter, K., Provis, J.M., 2015. Chemokine-mediated inflammation in the degenerating retina is coordinated byMullercells, activated microglia, and retinal pigment epithelium. Journal ofNeuroinflammation12.

177

• Rutar, M., Natoli, R., Kozulin, P., Valter, K., Gatenby, P., Provis, J.M., 2011a.Analysis of complement expression in light-induced retinal degeneration:synthesisanddepositionofC3bymicroglia/macrophages isassociatedwithfocalphotoreceptordegeneration.InvestOphthalmolVisSci52,5347-5358.

• Rutar, M., Natoli, R., Provis, J.M., 2012b. Small interfering RNA-mediatedsuppression of Ccl2 in Muller cells attenuates microglial recruitment andphotoreceptor death following retinal degeneration. J Neuroinflammation 9,221.

• Rutar,M.,Natoli,R.,Valter,K.,Provis,J.M.,2011b.Earlyfocalexpressionofthechemokine Ccl2 byMuller cells during exposure to damage-inducing brightcontinuouslight.InvestOphthalmolVisSci52,2379-2388.

• Rutar,M.,Provis,J.M.,Valter,K.,2010.Briefexposuretodamaginglightcausesfocal recruitment of macrophages, and long-term destabilization ofphotoreceptorsinthealbinoratretina.CurrEyeRes35,631-643.

• Saari, J.C., Bredberg, L., Garwin, G.G., 1982. Identification of the endogenousretinoidsassociatedwiththreecellularretinoid-bindingproteinsfrombovineretinaandretinalpigmentepithelium.JBiolChem257,13329-13333.

• Saika,R.,Sakuma,H.,Noto,D.,Yamaguchi,S.,Yamamura,T.,Miyake,S.,2017.MicroRNA-101a regulates microglial morphology and inflammation. JNeuroinflammation14,109.

• Saxena,K.,Rutar,M.V.,Provis,J.M.,Natoli,R.C.,2015.IdentificationofmiRNAsinaModelofRetinalDegenerations.InvestOphthalmolVisSci56,1820-1829.

• Schroder,K.,Tschopp,J.,2010.Theinflammasomes.Cell140,821-832.• Schroeder,P.,Pohl,C.,Calles,C.,Marks,C.,Wild,S.,Krutmann,J.,2007.Cellularresponse to infrared radiation involves retrograde mitochondrial signaling.FreeRadicalBiologyandMedicine43,128-135.

• Sethi, C.S., Lewis, G.P., Fisher, S.K., Leitner, W.P., Mann, D.L., Luthert, P.J.,Charteris,D.G.,2005.Glialremodelingandneuralplasticity inhumanretinaldetachmentwithproliferativevitreoretinopathy.InvestOphthalmolVisSci46,329-342.

• Sheedlo, H.J., Jaynes, D., Bolan, A.L., Turner, J.E., 1995. Mullerian glia indystrophic rodent retinas: an immunocytochemical analysis. Brain Res DevBrainRes85,171-180.

• Shen, W., Fruttiger, M., Zhu, L., Chung, S.H., Barnett, N.L., Kirk, J.K., Lee, S.,Coorey, N.J., Killingsworth,M., Sherman, L.S., Gillies,M.C., 2012. ConditionalMullercellablationcausesindependentneuronalandvascularpathologiesinanoveltransgenicmodel.JNeurosci32,15715-15727.

• Shichida,Y.,Matsuyama,T.,2009.Evolutionofopsinsandphototransduction.PhilosTransRSocLondBBiolSci364,2881-2895.

• Singh, P.K., Shiha, M.J., Kumar, A., 2014. Antibacterial responses of retinalMuller glia: production of antimicrobial peptides, oxidative burst andphagocytosis.JournalofNeuroinflammation11.

• Skog,J.,Wurdinger,T.,vanRijn,S.,Meijer,D.H.,Gainche,L.,Sena-Esteves,M.,Curry, W.T., Jr., Carter, B.S., Krichevsky, A.M., Breakefield, X.O., 2008.GlioblastomamicrovesiclestransportRNAandproteinsthatpromotetumourgrowthandprovidediagnosticbiomarkers.NatCellBiol10,1470-1476.

• Skokos, D., Botros, H.G., Demeure, C., Morin, J., Peronet, R., Birkenmeier, G.,Boudaly,S.,Mecheri,S.,2003.Mastcell-derivedexosomesinducephenotypicand functional maturation of dendritic cells and elicit specific immuneresponsesinvivo.JImmunol170,3037-3045.

178

• Spandau, U.H., Toksoy, A., Verhaart, S., Gillitzer, R., Kruse, F.E., 2003. HighexpressionofchemokinesGro-alpha(CXCL-1),IL-8(CXCL-8),andMCP-1(CCL-2)ininflamedhumancorneasinvivo.ArchOphthalmol121,825-831.

• Sprent, J.,2005.DirectstimulationofnaiveTcellsbyantigen-presentingcellvesicles.BloodCellsMolDis35,17-20.

• Srinivasan,V.J.,Ko,T.H.,Wojtkowski,M.,Carvalho,M.,Clermont,A.,Bursell,S.E.,Song,Q.H.,Lem,J.,Duker,J.S.,Schuman,J.S.,Fujimoto,J.G.,2006.Noninvasivevolumetric imaging andmorphometryof the rodent retinawithhigh-speed,ultrahigh-resolutionopticalcoherencetomography.InvestOphthalmolVisSci47,5522-5528.

• Sullivan, S.M., Lee,A.,Bjorkman, S.T.,Miller, S.M., Sullivan,R.K.,Poronnik,P.,Colditz, P.B., Pow, D.V., 2007. Cytoskeletal anchoring of GLAST determinessusceptibility tobraindamage: an identified role forGFAP. JBiolChem282,29414-29423.

• Swaroop, A., Kim, D., Forrest, D., 2010. Transcriptional regulation ofphotoreceptordevelopmentandhomeostasisinthemammalianretina.NatRevNeurosci11,563-576.

• Tang,J.,Du,Y.,Lee,C.A.,Talahalli,R.,Eells,J.T.,Kern,T.S.,2013.Low-intensityfar-red light inhibits early lesions that contribute todiabetic retinopathy: invivoandinvitro.InvestOphthalmolVisSci54,3681-3690.

• Tang,J.,Herda,A.A.,Kern,T.S.,2014.Photobiomodulationinthetreatmentofpatientswithnon-center-involvingdiabeticmacularoedema.BrJOphthalmol98,1013-1015.

• Tarallo, V., Hirano, Y., Gelfand, BradleyD., Dridi, S., Kerur, N., Kim, Y., Cho,WonG., Kaneko, H., Fowler, BenjaminJ., Bogdanovich, S., Albuquerque,RomuloJ.C., Hauswirth, WilliamW., Chiodo, VinceA., Kugel, JenniferF.,Goodrich, JamesA., Ponicsan, StevenL., Chaudhuri, G., Murphy, MichaelP.,Dunaief, JoshuaL., Ambati, BalamuraliK., Ogura, Y., Yoo, JaeW., Lee, D.-k.,Provost, P., Hinton, DavidR., Núñez, G., Baffi, JuditZ., Kleinman, MarkE.,Ambati, J., 2012. DICER1 loss and Alu RNA induce age-related maculardegenerationviatheNLRP3inflammasomeandMyD88.Cell149,847-859.

• Taylor,L.,Arnér,K.,Ghosh,F.,2015.Firstresponders:Dynamicsofpre-glioticmüllercellresponsesintheisolatedadultratretina.CurrentEyeResearch40,1245-1260.

• Tetta,C.,Ghigo,E.,Silengo,L.,Deregibus,M.C.,Camussi,G.,2013.Extracellularvesiclesasanemergingmechanismofcell-to-cellcommunication.Endocrine44,11-19.

• Tout,S.,Chan-Ling,T.,Hollander,H.,Stone,J.,1993.TheroleofMullercellsintheformationoftheblood-retinalbarrier.Neuroscience55,291-301.

• Trams,E.G.,Lauter,C.J.,Salem,N.,Jr.,Heine,U.,1981.Exfoliationofmembraneecto-enzymesintheformofmicro-vesicles.BiochimBiophysActa645,63-70.

• Trueblood, K.E., Mohr, S., Dubyak, G.R., 2011. Purinergic regulation of high-glucose-inducedcaspase-1activationintheratretinalMullercelllinerMC-1.AmJPhysiolCellPhysiol301,C1213-1223.

• Tschopp, J., Schroder, K., 2010. NLRP3 inflammasome activation: Theconvergence of multiple signalling pathways on ROS production? Nat RevImmunol10,210-215.

• Tseng,W.A.,Thein,T.,Kinnunen,K.,Lashkari,K.,Gregory,M.S.,D'Amore,P.A.,Ksander, B.R., 2013. NLRP3 inflammasome activation in retinal pigmentepithelial cells by lysosomal destabilization: implications for age-relatedmaculardegeneration.InvestOphthalmolVisSci54,110-120.

179

• Uckermann, O., Wolf, A., Kutzera, F., Kalisch, F., Beck-Sickinger, A.G.,Wiedemann, P., Reichenbach, A., Bringmann, A., 2006. Glutamate release byneuronsevokesapurinergicinhibitorymechanismofosmoticglialcellswellingintheratretina:activationbyneuropeptideY.JNeurosciRes83,538-550.

• Valter,K.,Bisti,S.,Gargini,C.,DiLoreto,S.,Maccarone,R.,Cervetto,L.,Stone,J.,2005.Timecourseofneurotrophicfactorupregulationandretinalprotectionagainstlight-induceddamageafteropticnervesection.InvestOphthalmolVisSci46,1748-1754.

• Vecino, E., Rodriguez, F.D., Ruzafa, N., Pereiro, X., Sharma, S.C., 2015. Glia-neuroninteractionsinthemammalianretina.ProgRetinEyeRes.

• Vejnar, C.E., Zdobnov, E.M., 2012. MiRmap: comprehensive prediction ofmicroRNAtargetrepressionstrength.NucleicAcidsRes40,11673-11683.

• Veleri, S., Lazar, C.H., Chang, B., Sieving, P.A., Banin, E., Swaroop, A., 2015.Biology and therapy of inherited retinal degenerative disease: insights frommousemodels.DisModelMech8,109-129.

• Verardo, M.R., Lewis, G.P., Takeda, M., Linberg, K.A., Byun, J., Luna, G.,Wilhelmsson,U.,Pekny,M.,Chen,D.F.,Fisher,S.K.,2008.AbnormalreactivityofmullercellsafterretinaldetachmentinmicedeficientinGFAPandvimentin.InvestOphthalmolVisSci49,3659-3665.

• Vickers, K.C., Remaley, A.T., 2012. Lipid-based carriers of microRNAs andintercellularcommunication.CurrOpinLipidol23,91-97.

• Voigt,J.,Grosche,A.,Vogler,S.,Pannicke,T.,Hollborn,M.,Kohen,L.,Wiedemann,P.,Reichenbach,A.,Bringmann,A.,2015.NonvesicularreleaseofATPfromratretinal glial (Muller) cells is differentially mediated in response to osmoticstressandglutamate.NeurochemRes40,651-660.

• Volland,S.,Esteve-Rudd,J.,Hoo,J.,Yee,C.,Williams,D.S.,2015.Acomparisonofsomeorganizationalcharacteristicsofthemousecentralretinaandthehumanmacula.PLoSOne10,e0125631.

• Walker, R.J., Steinle, J.J., 2007. Role of beta-adrenergic receptors ininflammatorymarkerexpressioninMullercells.InvestOphthalmolVisSci48,5276-5281.

• Wan,J.,Zheng,H.,Chen,Z.L.,Xiao,H.L.,Shen,Z.J.,Zhou,G.M.,2008.Preferentialregenerationofphotoreceptor fromMullerglia after retinaldegeneration inadultrat.VisionRes48,223-234.

• Wang,A.L.,Lukas,T.J.,Yuan,M.,Du,N.,Tso,M.O.,Neufeld,A.H.,2009.Autophagyand exosomes in the aged retinal pigment epithelium:possible relevance todrusenformationandage-relatedmaculardegeneration.PLoSOne4,e4160.

• Wang,M.,Ma,W.,Zhao,L.,Fariss,R.N.,Wong,W.T.,2011.AdaptiveMullercellresponses to microglial activation mediate neuroprotection and coordinateinflammationintheretina.JNeuroinflammation8,173.

• Wang,S.,Koster,K.M.,He,Y.,Zhou,Q.,2012.miRNAsaspotentialtherapeutictargetsforage-relatedmaculardegeneration.FutureMedChem4,277-287.

• Wang,Z.,Li,W.,Mitchell,C.K.,Carter-Dawson,L.,2003.ActivationofproteinkinaseCreducesGLASTintheplasmamembraneofratMullercellsinprimaryculture.VisNeurosci20,611-619.

• Whelan,H.T.,Smits,R.L.,Jr.,Buchman,E.V.,Whelan,N.T.,Turner,S.G.,Margolis,D.A.,Cevenini,V.,Stinson,H.,Ignatius,R.,Martin,T.,Cwiklinski,J.,Philippi,A.F.,Graf,W.R.,Hodgson,B.,Gould,L.,Kane,M.,Chen,G.,Caviness,J.,2001.EffectofNASAlight-emittingdiodeirradiationonwoundhealing.JClinLaserMedSurg19,305-314.

180

• Whitcup, S.M., Nussenblatt, R.B., Lightman, S.L., Hollander, D.A., 2013.Inflammationinretinaldisease.IntJInflam2013,724648.

• Wohl,S.G.,Reh,T.A.,2016.ThemicroRNAexpressionprofileofmouseMullergliainvivoandinvitro.SciRep6,35423.

• Wolfrum,U., Schmitt,A., 2000.Rhodopsin transport in themembraneof theconnectingciliumofmammalianphotoreceptorcells.CellMotilCytoskeleton46,95-107.

• Wong-Riley,M.T.,Liang,H.L.,Eells,J.T.,Chance,B.,Henry,M.M.,Buchmann,E.,Kane,M.,Whelan, H.T., 2005. Photobiomodulation directly benefits primaryneuronsfunctionallyinactivatedbytoxins:roleofcytochromecoxidase.JBiolChem280,4761-4771.

• Wu,K.H.,Madigan,M.C.,Billson,F.A.,Penfold,P.L.,2003.DifferentialexpressionofGFAPinearlyvlateAMD:aquantitativeanalysis.BrJOphthalmol87,1159-1166.

• Wysoczynski,M., Ratajczak,M.Z., 2009. Lung cancer secretedmicrovesicles:underappreciatedmodulatorsofmicroenvironmentinexpandingtumors.IntJCancer125,1595-1603.

• Xu,Y.,Cheng,Q.,Yang,B.,Yu,S.,Xu,F.,Lu,L.,Liang,X.,2015.IncreasedsCD200LevelsinVitreousofPatientsWithProliferativeDiabeticRetinopathyandItsCorrelationWithVEGFandProinflammatoryCytokines.InvestOphthalmolVisSci56,6565-6572.

• Yang,J.,Lu,Y.W.,Lu,M.M.,Leng,R.X.,Pan,H.F.,Ye,D.Q.,2013.MicroRNA-101,mitogen-activated protein kinases and mitogen-activated protein kinasesphosphatase-1insystemiclupuserythematosus.Lupus22,115-120.

• Yoneda, S.,Tanihara,H.,Kido,N.,Honda,Y.,Goto,W.,Hara,H.,Miyawaki,N.,2001.Interleukin-1betamediatesischemicinjuryintheratretina.ExpEyeRes73,661-667.

• Yoshimura,T.,Robinson,E.A.,Tanaka,S.,Appella,E.,Kuratsu,J.,Leonard,E.J.,1989. Purification and amino acid analysis of two human glioma-derivedmonocytechemoattractants.JExpMed169,1449-1459.

• Yue, L., Weiland, J.D., Roska, B., Humayun, M.S., 2016. Retinal stimulationstrategiestorestorevision:Fundamentalsandsystems.ProgRetinEyeRes53,21-47.

• Zhang,J.,Li,S.,Li,L.,Li,M.,Guo,C.,Yao,J.,Mi,S.,2015.ExosomeandexosomalmicroRNA: trafficking, sorting, and function. Genomics ProteomicsBioinformatics13,17-24.

• Zhao,L.,Zabel,M.K.,Wang,X.,Ma,W.,Shah,P.,Fariss,R.N.,Qian,H.,Parkhurst,C.N., Gan, W.B., Wong, W.T., 2015a. Microglial phagocytosis of livingphotoreceptorscontributestoinheritedretinaldegeneration.EMBOMolMed7,1179-1197.

• Zhao,M.,Bai,Y.,Xie,W.,Shi,X.,Li,F.,Yang,F.,Sun,Y.,Huang,L.,Li,X.,2015b.Interleukin-1betaLevelIsIncreasedinVitreousofPatientswithNeovascularAge-Related Macular Degeneration (nAMD) and Polypoidal ChoroidalVasculopathy(PCV).PLoSOne10,e0125150.