the tumor microenvironment and 3-d tumor modelsq-bio.org/w/images/6/62/freyer.pdfnew in vitro model...

The Tumor Microenvironment and 3-D Tumor

Models

James Freyer

Bioscience Division

Los Alamos National Laboratory

Outline

• The Tumor MicroenvironmentChronic versus acute changesConsequences of tumor microenvironmentAdvances in measuring the tumor microenvironmentDifficulties with in vivo models and clinical tumors

• 3-D Experimental Tumor Model SystemsTypes of model systemsThe multicellular spheroid tumor modelExample of application of spheroidsRecent developments and future work

• Mathematical Modeling in Tumor BiologyTumor microenvironmentGenetic/proteomic/metabolic networksTumor growth and development

• Questions?

Malignant Progression of Cancer

NormalCell

mutations

loss ofgrowthcontrol

mutationsCancerCell

MalignantCell

survival

invasion

angiogenesis

metastasis

therapyresistance

Important to realize: all of this happens in a 3-D context within a tissue!

Differences: Tumor and Normal Tissue Vasculature

Brown & Giaccia, Cancer Res. 58: 1408, 1998

Chronic Changes in Tumor Microenvironment

Brown & Giaccia., Cancer Res. 58: 1408, 1998

• Tumor cells grow faster thanvasculature: cells located far fromvessels

• Gradients in biochemistry ofextracellular space

Nutrients (oxygen, glucose)Metabolic wastes (pH, lactate)Signaling molecules (promotors,inhibitors)

• Gradients in cell physiologyProliferationMetabolismViabilityMotility, invasiveness

• Gradients in gene/protein expression

• Gradients in therapy response

• Generally occur over ~200 m

Transient Changes in Tumor Microenvironment

Kimura et al., Cancer Res. 56: 5522, 1996

• No organization to architecture ofvasculature: driven by semi-randomprocesses

Long, tortuous vesselsA-V shuntsBlockages

• Disorganized functionNo smooth muscle or nerve cellsVarying pressure gradientsTrapping of white/red cells

• Transient microregional variations inflow

Slowed, stopped, reversed flow~10-20 minute period most frequent

• Time-varying nutrient supply andwaste removal

• Superimposed on chronic gradients

• Altered by therapy

Both Chronic and Transient Hypoxia

Gilles et al., J. Magnet. Reson. Imag. 16: 430, 2002

Microenvironment Involved in Tumor Progression

Bindra & Glazer., Mutat. Res. 569: 75, 2005

Microenvironment Involved in Metastasis

Sabarsky & Hill., Clin. Exper. Metast. 20: 237, 2003

Therapeutic Impact of Tumor Microenvironment

• Hypoxia causes radiation resistanceMajor explanation for radiotherapy failureMajor focus of drug development and imaging

• Cell cycle arrested cells more resistantResistant to most common chemotherapies, radiationAble to repopulate tumor after treatment

• Limited drug deliveryPoor penetration (chronic) & limited delivery (transient)Problem for new therapies (antibodies, nonparticles)

• Induction of drug resistance and genetic instabilityGene expression and protein modificationsMutations: drug resistance, survival phenotypes

• Stimulation of angiogenesis and metastatic spreadInduction of pro-angiogenic factorsIncreased local invasion and distant metastases

Effect of Hypoxia on Therapy

Fyles et al., J. Clin. Oncol. 20: 680, 2000

H&N Cancer

pO2 > 10 mm Hg

pO2 < 10 mm Hg

Brizel et al., Radiother. Oncol. 53:113, 1999

Cervical Cancer

Imaging in Window Chamber Tumors

Sorg et al., J. Biomed. Optics 10: 044004, 2005

Day 3 Day 4

Day 5 Day 8

Oxygenated

Hypoxic

Imaging in Human Tumor Sections

Janssen et al., Int. J. Radiat. Biol. Phys. 62: 1169, 2005

Blood vesselsPerfusion markerProliferation marker

Metabolic Analysis of Tumor Microenvironment

Wallenta et al., Biomol. Engineer. 18: 249, 2002

Advanced MRI of Tumor Microenvironment

Gilles et al., J. Magnet. Reson. Imag. 16: 430, 2002

Histology

Vascular volume

Vascular permeability

V & P

V & P & pH

Advanced MRI of Human H&N Tumor

Padhani et al., Eur. Radiol. 17: 861, 2007

Limitations to in Vivo Tumor Biology

• Enormous complexity and heterogeneity both withinand between tumors

• Non-reproducibility of even the best rodent tumormodel systems

• Poor understanding of extent and control of transientvariations: basically chaos

• Inability to control experimental parameters

• Inability to perform mechanistic experiments onhumans

• Therefore, advances in basic understanding of tumorbiology (and progress in therapy?) require in vitro

experimental models of tumor

In Vitro Experimental Tumor Models

• Most basic: monolayer or suspension cell culturesUseful for very basic studiesA very poor model of a 3-D tissueDo not mimic any aspect of the tumor microenvironment

• Several different 3-D in vitro models have beendeveloped

Cells embedded in external matrix materialBioreactors: cells within artificial capillary structure‘Sandwich’ culture: cells trapped between two platesMulticell layers: 3-D layers of cells on a membraneEx vivo explants of tumor piecesMulticellular aggregates: spherical 3-D cultures(‘spheroids’)

Multicellular Tumor Spheroids

wastesnutrients

HK03 Wild Type HK03 Null

106

107

108

109

1010

1011

HK03-Tr Wild Type

Sp

he

roid

Vo

lum

e

(m

3)

HK03-Tr Null


0

10

20

30

40

50

60

HK03Tr Wild Type

S-P

ha

se

Fra

cti

on

(pe

rce

nt)

HK03Tr Null


0

50

100

150

200

250

300

HK03TR Wild Type

0 10 20 30 40 50 60 70

Via

ble

Rim

Th

ick

ne

ss

(m

)

Time of Growth(days)

HK03TR Null

0 10 20 30 40 50 60 70

Time of Growth(days)

Proliferating cells

Quiescent cells

Similarities: Spheroids and Tumors

• 3-D, tissue-like structureCell-cell contactsExtracellular matrixMicroenvironment develops spontaneously

• Heterogeneous microenvironmentGradients in extracellular biochemistryGradients in cellular physiologyGradients in cellular metabolismGradients in gene/protein expression

• Therapy resistanceRadiation (ionizing, UV, microwave)Many forms of chemotherapyHyperthermiaPhotodynamic therapyBiologicals (antibodies, liposomes, nanoparticles)

Advantages: Spheroids vs Tumors

• Highly reproducibleVery small inter-spheroid variabilityExcellent long-term ‘stability’ (decades)

• SymmetricalGradients are radially distributedVarious gradients are tightly correlatedEnables some unique experimental manipulationsIdeal for mathematical modeling

• Experimental controlExternal environment controlledReproducible manipulation of experimental conditionsEasy to manipulate individual spheroidsHigh ‘data density’

Research applications of spheroids

• Therapy testing and mechanistic studies

• Basic tumor biologyCell cycle regulationMetabolic regulationCellular physiologyCell-cell interactionsRegulation of gene/protein expressionMalignant progression

• Co-culturesTumor-normal cell mixturesAngiogenesis models

• Non-cancer applicationsArtificial organ researchDrug productionNormal tissue models

Example: Cell Cycle Regulation

• Despite common (mis)conception that malignant cellshave escaped growth control, majority of tumor cells ina solid tumor are not proliferating

• Common (mis)dogma is that cell cycle arrest in tumorsis due to lack of nutrients, specifically oxygen

• Although recent imaging and molecular techniqueshave documented spatial distribution of proliferation inrodent and human tumors, controlled manipulation andmechanistic experiments are not possible

• Actual molecular mechanism of cell cycle arrest intumors is currently unknown

• Spheroids are a good in vitro model to performmechanistic studies on this question

Multicellular Tumor Spheroids

0

0.2

0.4

0.6

0.8

1

0 6 12 18 24 30 36

Frac

tion

of C

ells

Rem

aini

ng in

Sph

eroi

d

Time of Dissociation(minutes)

0600Distance from Surface

(μm)

Fraction 3Fraction 4Necrosis

Fraction 2Fraction 1

nutrients wastes

250,000 cells/spheroid

Cell Cycle Proteins in Spheroids

Fraction Number1 2 3 4

p27

p21

p18

CKIs

CDK6

CDK4

CDK2

CDKs

cycD1

cycE

cycA

cyclins

Cyclin ACyclin D1Cyclin E

0 50 100 150 2000

0.5

1

1.5

Rel

ativ

e C

yclin

Pro

tein

(frac

tion

1 =

1)

Distance from Surface(μm)

CDK2CDK4CDK6

0

0.5

1

1.5

Rel

ativ

e C

DK

Pro

tein

(frac

tion

1 =

1)

p18p21p27

0

1

2

3

4

5

Rel

ativ

e C

KI P

rote

in(fr

actio

n 1

= 1)

G1- Versus S-phase Arrest

Fraction Number

1 2 3 4

EMT6

Mel28

outer inner

0

10

20

30

40

50

60

DNA contentBrdU Uptake

S-ph

ase

Frac

tion

(per

cent

)0

10

20

30

40

50

60

0 50 100 150 200

DNA ContentBrdU Uptake


Cell Cycle Arrest After Acute Oxygen Deprivation

N2

O2

OxygenNitrogen

0.5

1.5

2.5

3.5

0 5 10 15 20 25

Time of Culture(hours)

Rel

ativ

e C

ell N

umbe

r

O2 N2

O2N2O2N2

0

25

50

75

100

0 5 10 15 20 25


Fra

ctio

n of

Cel

ls(p

erce

nt)

G1

G2S

O2 N2O2 N2O2 N2

1

2

3

4

0 5 10 15 20 25


Rel

ativ

e P

rote

in L

evel

(0 h

r =

1.0)

p18

p27p21

Regulation of Proliferation in Spheroids

• Initial arrest is an active process regulated by acyclin/CDK mechanism

Little change in CDKs, loss of cyclin D1Upregulation of p18 and p27, loss of p21CKI binding to and inhibition of CDK activityBypassing initial G1-arrest allows S-phase arrest

• Interior arrested cells continue to undergo alterations incell cycle regulatory machinery

Loss of all regulatory molecules: CDKs, cyclins, CKIsMay explain prolonged recovery lag time: unable toresume without rebuilding?

• Inducers of initial arrest currently unknownSeveral CKIs, up- and down-regulated: multiple signals?Initiated relatively close to surface (~50 m)Unlikely to be related to oxygen deprivationGrowth factor or inhibitor? Pressure sensing?

Limitations to Current Spheroid Model Systems

• Only mimics chronic nutrient deprivation

• Difficult for in situ assay of microenvironmentalgradients (microelectrodes, histology)

• Separation of cells from different locations involvesrelatively long enzymatic treatment (complicates geneand protein expression data)

• Only applicable to adherent cells and those thatproliferate in aggregate culture

• Difficult to use for controlled, reproducible experimentswith co-cultures

Transient Deprivation System for Spheroids

20%oxygen

0%oxygen

return

0

25

50

75

100

125

150

0 1 2 3 4

Oxy

gen

Part

ial P

ress

ure

(mm

Hg)

Time After O2 to N2(minutes)

0

25

50

75

100

125

150

0 1 2 3 4 5 6O

xyge

n Pa

rtia

l Pre

ssur

e(m

m H

g)Time of Culture

(hours)

Effects of Transient Oxygen Deprivation

30 minutesoxygen

30 minutesnitrogen

30 minutesoxygen

0

0.5

1

1.5

Total Volume Cell Number Viable Rim

Start6 hr cycle12 hr cycle

Rel

ativ

e V

alue

(tim

e 0

= 1)

0

20

40

60

80

100

0 50 100 150 200

Fra

ctio

n of

cel

ls(p

erce

nt)


G1

G2

S

0

1

2

3

4

5

6

0 50 100 150 200R

elat

ive

CK

I P

rote

in(f

ract

ion

1 =

1)Distance from Surface

(μm)

p18

p27

p21

Transient Nutrient Deprivation in Spheroids

• New culture system developed and validated fortransient deprivation experiments

Compact, portable culture chamberAbility to rapidly alter nutrient conditionsImposes external transient supply on pre-existing chronicgradients: more like tumor in vivo

• Preliminary experiments show essentially no effect ofcyclic oxygen supply for up to 12 hours

No change in spheroid growth rate or cell numberNo increase in central necrosisNo alteration in cell cycle or CKI induction

• Preliminary experiments show remarkable resistance tonutrient deprivation

Complete nutrient deprivation causes total loss of ATP andextremely acidic intracellular pHComplete recovery of normal cellular energetics after nutrientrestoration

New In Vitro Model of Tumor Microenvironment

mediuminput

mediumoutput

top cap

mediumreservoirmembrane

cells inmatrixglass

cylinder

bottomcap

0

0.2

0.4

0.6

0.8

1

1.2

7.05

7.1

7.15

7.2

7.25

7.3

7.35

7.4

7.45

0 2 4 6 8 10

Rel

ativ

e C

once

ntra

tion

pH

Distance from Membrane(mm)

pH

Oxygen

Lactate

0

0.2

0.4

0.6

0.8

1

1.2

0

10

20

30

40

50

0 2 4 6 8 10R

elat

ive

Con

cent

ratio

n


S-phase

Protein

Gene

S-phase Percent

Preliminary Data with 1st Generation System

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6

4°25°37°

Cel

l C

on

cen

trat

ion

(x 1

0-7

cel

ls p

er c

m3)


0

10

20

30

40

50

0 1 2 3 4 5 6

4°25°37°

S-p

has

e F

ract

ion

(per

cen

t)


0

20

40

60

80

100

0 1 2 3 4 5 6

4°25°37°C

lon

og

enic

Eff

icie

ncy

(per

cen

t)


0

1

2

3

4

5

6

0 1 2 3 4 5 6

4°25°37°

Rel

ativ

e p

27

Pro

tein

(4o @

0.4

mm

= 1

.0)


Current State of New Model System

• Demonstration of feasibility of designSpatial correlation of microenvironment and biologyPotential for real-time, in situ measurement by NMRAllows rapid isolation of cells from different regionsExperimental control over many parameters

• Produces physiological gradients similar to those seenin spheroids and tumors

Cell proliferation and cell cycle distributionCell deathInduction of CKIs

• 1st generation system has problemsDifficult and non-reproducible separation of cells fromdifferent regions, still requires matrix digestionNo control over internal supply conditionsRelatively low cell number to get extended gradients

Theoretical Modeling of Tumors

• Overwhelming majority of literature based onmathematical models of tumor growth and development(~1200 papers since 1970)

• Interestingly, spheroid growth data very often used to‘test’ models

• Limited development in other areasInteractions with immune systemRegulation of cellular metabolismExtracellular biochemical environmentCellular invasionTherapy response (radiation, chemo)Protein regulatory networks

• Recent focus on developing biologically-based modelsof tumor growth and malignant progression

Modeling Hypoxia in Tumors

Kirkpatrick et al., Radiat. Res. 159: 336, 2003

Modeling Hypoxia in Tumors

Secomb et al., Annal. Biomed. Engineer. 32: 1519, 2004

Modeling Angiogenesis in Tumors

Stephanou et al., Math. Comput. Model. 41: 1137, 2005

Penetration of Chemotherapy Agent

Modak et al., Eur. J. Cancer. 42: 4204, 2006

Protein Network Model of Tumor Cell Invasion

Athale et al., J. Theor. Biol. 233: 469, 2004

Nested Deterministic Models of Tumor Growth

Marusic et al., Cell Prolif. 27: 73, 1994

Generic Models

Two-parameter Models

Functional Models

Fits of 15 Models to 15 Independent Data Sets


Fits of 15 Models to 15 Independent Data Sets


Doubling Time

Thickness of Viable Cell Rim

Deterministic Tumor Models

• Wide variety available and more being developed

• Most can do a good job of fitting basic tumor (spheroid)growth data

• Useful for graphing, comparing and extrapolating data

• Most do a poor job of predicting any biologicalparameters

• Not really useful for advancing our understanding oftumor biology

Generally not predictiveMany not directly connected to biologyThose that are have a very large number of parametersDifficult to distinguish one from the other

• The future of this field is in biologically-based models

Conceptual Model of Spheroid Growth Regulation

Freyer & Sutherland, Cancer Res. 46: 3504, 1986

Multi-Scale Mathematical Tumor Model

• Starts with single cell on 3-D lattice‘Programmed’ with metabolic, generegulation, cell cycle, volumegrowth rate, adhesion and celldeath parametersAssumes limited inward growthfactor penetration and internalgrowth inhibitor productionSimulation runs until lattice is filledor spheroid saturates: nothing ‘fit’or constrained

• Three scales consideredCellular (lattice Monte Carlo)Gene regulation (Boolean network)Extracellular (reaction-diffusionequations)

Final Conclusions

• Solid tumors are perhaps the most unique, complex,dynamic and chaotic biological system

• The tumor microenvironment is extremelyheterogeneous, both spatially and temporally

• This microenvironmental complexity explains mosttherapy failures, as well as promotes the progression ofmalignancy itself

• Actual tumors in vivo are poorly suited to mechanisticexperimentation

• Many 3-D in vitro experimental tumor models areavailable and important for advancing tumor biology

• Spheroids are an excellent tumor model system, buthave limitations

• Theoretical modeling of tumors is in its infancy, but cancontribute significantly in cancer research

Acknowledgements

• Spheroid projectsDr. Karen LaRueAntoinette TrujilloAnabel GuerraRebecca AlbertiniJeffery DietrichSusan CarpenterDr. Yi JiangJelena Pjesivac-GrbovicJames Coulter

• New tumor modelDr. Joseph HickeyAntoinette Trujillo

• Flow cytometrySusan CarpenterAntoinette TrujilloTravis Woods

• ExternalDr. Bert van der KogelMr. Hans PetersDr. Keith Laderoute

• FundingNIH: CA-71898, CA-80316, CA-89255, RR-01315NSF: PUSH ProgramLDRD: Los Alamos internal funding

the tumor microenvironment and 3-d tumor modelsq-bio.org/w/images/6/62/freyer.pdfnew in vitro model...

Documents