Mol Imaging Biol (2015)DOI: 10.1007/s11307-015-0875-z* World Molecular Imaging Society, 2015

RESEARCH ARTICLE

Non-Invasive Intravital Imaging of siRNA-Mediated Mutant Keratin Gene Repressionin SkinRobyn P. Hickerson,1,6 Tycho J. Speaker,1 Maria Fernanda Lara,1,7

Emilio González-González,2,3,8 Manuel A. Flores,1 Christopher H. Contag,2,3,4,5

Roger L. Kaspar1

1TransDerm Inc., 2161 Delaware Ave., Santa Cruz, CA, 95060, USA2Molecular Imaging Program at Stanford (MIPS), Stanford University School of Medicine, Stanford, CA, USA3Department of Pediatrics, Stanford University School of Medicine, Stanford, CA, USA4Department of Radiology, Stanford University School of Medicine, Stanford, CA, USA5Departments of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA, USA6Centre for Dermatology and Genetic Medicine, University of Dundee, Dundee, UK7Urology Research Unit Virgen de la Victoria and Regional Hospital, Malaga, Spain8Canvax Biotech S.L., Technological Park, Cordoba, Spain

AbstractPurpose: Small interfering RNAs (siRNAs) specifically and potently inhibit target geneexpression. Pachyonychia congenita (PC) is a skin disorder caused by mutations in genesencoding keratin (K) 6a/b, K16, and K17, resulting in faulty intermediate filaments. A siRNAtargeting a single nucleotide, PC-relevant mutation inhibits K6a expression and has beenevaluated in the clinic with encouraging results.Procedures: To better understand the pathophysiology of PC, and develop a model system tostudy siRNA delivery and visualize efficacy in skin, wild type (WT) and mutant K6acomplementary DNAs (cDNAs) were fused to either enhanced green fluorescent protein ortandem tomato fluorescent protein cDNA to allow covisualization of mutant and WT K6aexpression in mouse footpad skin using a dual fluorescence in vivo confocal imaging systemequipped with 488 and 532 nm lasers.Results: Expression of mutant K6a/reporter resulted in visualization of keratin aggregates, whileexpression of WT K6a/reporter led to incorporation into filaments. Addition of mutant K6a-specific siRNA resulted in inhibition of mutant, but not WT, K6a/reporter expression.Conclusions: Intravital imaging offers subcellular resolution for tracking functional activity ofsiRNA in real time and enables detailed analyses of therapeutic effects in individual mice tofacilitate development of nucleic acid-based therapeutics for skin disorders.

Key words: Genodermatosis, Gene therapy, Gene regulation, In vivo confocal fluorescencemicroscopy

IntroductionThe combination of potency and specificity of smallinterfering RNAs (siRNAs) to degrade targeted messengerRNAs (mRNAs) has made this class of inhibitors very

Electronic supplementary material The online version of this article(doi:10.1007/s11307-015-0875-z) contains supplementary material, whichis available to authorized users.

Correspondence to: Roger Kaspar; e-mail: roger.kaspar@transderminc.com

attractive as therapeutic agents. A number of siRNAs haveshown promise in clinical trials [1–3], including a trial usinga siRNA designed to target a causative keratin mutation inthe monogenic skin disease, pachyonychia congenita (PC)[4, 5]. These studies have shown quantitative and specificknockdown of mutant gene expression without affectingwild type expression [14]. The skin is an ideal organ toevaluate siRNA therapeutics due to its accessibility foradministration and visualization. We have recently reportedin vivo imaging of therapeutic effects in mouse skin using aconfocal microscope with subcellular resolution. This deviceis equipped with multiple lasers to capture both a fluores-cence channel (enhanced green fluorescent protein (EGFP)expression using a 488 nm laser with appropriate filters) anda reflectance channel (658 nm) through the same optical pathto provide complementary information regarding skinstructure and tissue depth of reporter gene expression [6].

PC is a dominant negative genetic skin disorder causedby mutations in the inducible keratins (K) including K6a,K6b, K16, and K17 [7–9]. Expression of the mutant proteinresults in characteristic intracellular aggregates and faultyintermediate filament formation in cultured keratinocytes,which is thought to compromise the structural integrity ofthe cells and result in the characteristic symptoms of PC.The pathologic sequela in PC patient skin, initiated with theexpression of mutant keratin protein and ultimately leadingto painful plantar keratoderma (the major patient complaint),is beginning to be elucidated [10]. We have previouslydemonstrated, in cultured cells expressing both wild type(WT) and mutant fluorescently labeled keratins, that siRNAknockdown of the mutant keratin is quantitative and reversesthe mutant aggregate phenotype, resulting in visualization ofnormal intermediate filaments [11]. Using a reverse tran-scription quantitative PCR (RT-qPCR) method that distin-guishes the single nt change between K6a WT and N171Kmutant mRNAs, we demonstrated 90 % inhibition of mutantmRNA with the specific siRNA (K6a.513.12) withoutaffecting WT expression in transiently-transfected PC-10cells [12]. Similarly, using human primary keratinocytesderived from a PC patient harboring the K6a N171Kmutation, RT-qPCR revealed a 98 % reduction in mutantmRNA relative to WT following treatment with K6a.513.12siRNA [5]. Furthermore, a small clinical trial using injectedTD101, a siRNA targeting the specific mutant keratin (K6aN171K), resulted in localized keratoderma clearing andreduction of pain, suggesting that the mutant keratin wasselectively inhibited and the pathology reduced [4]. Unfortunately,the intense pain associated with siRNA injection required oralpain medications and regional nerve blocks [4], underscoring theneed for improved nucleic acid delivery methods.

In the current study, we improve upon our earlierintravital skin imaging studies in animal models [6, 13] byadding a 532 nm fluorescence laser (to detect red fluorescentproteins) to the confocal imaging system [6, 14], allowingsequential detection of EGFP and tandem tomato fluorescentprotein (tdTFP) at subcellular resolution. This allows non-

invasive imaging of the functional activity of siRNA whendelivered to cells expressing an engineered reporter (e.g.,tdTFP) containing the K6a.513.12 siRNA target site whileusing a second reporter (e.g., EGFP) that lacks the targetsequence as an internal control. The confocal microscope isa modified version of a device used for clinical reflectanceimaging [15, 16]. The use of optical reporter genes withdifferential expression in response to therapeutic nucleicacids along with clinically relevant multiparametric, highresolution in vivo microscopy offers tremendous opportuni-ties for spatiotemporal evaluation of therapeutic effects thatcan be used to improve delivery strategies and optimizeefficacy. These approaches are ideal for genetic diseases ofthe skin and can be used to greatly enhance animal modelsfor PC and related diseases.

Materials and Methods

Animals

Swiss Webster mice (Charles River, Wilmington, MA) were usedaccording to the Guide for the Care and Use of Laboratory Animals(National Research Council) and with strict adherence to a protocolapproved by the TransDerm Institutional Animal Care and UseCommittee. All injections and imaging were performed under 2–3 % isoflurane anesthesia as previously described [11].

siRNA

siRNAs, including NSC4 (Dharmacon Products Catalog #D-001210, 5′-UAGCGACUAAACACAUCAAUU, inverted bacterialβ-galactosidase sequence), K6a_513c.12 (5′-CCCUCAAaAACAAGUUUGCUU, the single nucleotide change resulting in mutantkeratin protein is shown as a lowercase Ba^ [11, 17]), CD44 (5′-GGCGCAGAUCGAUUUGAAU [18]), and K6a-3′-UTR-1 (5′-GCACAAGUGACUAGUCCUAUU [19]) were provided by GEHealthcare Dharmacon Products (Lafayette, CO).

Cloning of Expression Plasmids

CMV Promoter Expression Plasmids

The plasmids pCMV-K6a(WT)-EGFP (pTD239) and pCMV-K6a(N171K)-EGFP (pTD240) were generated by replacing the740-bp Bam HI/Not I fragment containing EYFP from pTD100 andpTD101 [11], respectively, with the Bam HI/Not I flanked 727-bpEGFP complementary DNA (cDNA) sequence PCR-generatedfrom pTD166 (pUbc-luc2/eGFP, [6]) using the following primers:Forward-376: 5′-GGA TCC ACC GGT CGC CAC CAT GGTGAG CAA GGG CGA GG-3′ and Reverse-377: 5′-GCG GCCGCT TTA CTT GTA C-3′. Generation of pCMV-K6a(WT)-tdTFP(pTD128) was described in [11]. pCMV-K6a(N171K)-tdTFP(pTD129) was recloned by ligation of the 1528-bp BamHI/NotIfragment of pTD116 into the 6769-bp BamHI/NotI backbonefragment of pTD104.

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

UBC Promoter Expression Plasmids

The plasmids pUBC-K6a(WT)-EYFP (pTD200), pUBC-K6a(N171K)-EYFP (pTD201), and pUBC-K6a(WT)-tdTFP(pTD202) were generated by replacing the Luc2-EGFP sequenceof pTD166 with the K6a/reporter gene sequences from pTD103,pTD104, and pTD128 [11], respectively. The fragments generatedfrom Hind III digestion followed by Klenow fill-in to generateBblunt ends^ and further digestion with Not I of pTD103 (2506 bp),pTD104 (2506 bp), pTD128 (3235 bp), and pTD129 (3235 bp)were ligated into the 4190-bp backbone fragment generated bydigestion of pTD166 with Xba I followed by Klenow fill-in togenerate Bblunt ends^ and further digestion with Not I. pUBC-K6a(N171K)-tdTFP (pTD203) was generated by ligation of themutant K6a sequence from pTD201 into the backbone of pTD202.Specifically, the 1759-bp Eco RI/Bam HI fragment of pTD201 wasligated into the 5670-bp Eco RI/Bam HI backbone fragment ofpTD202. The plasmids pUBC-K6a(WT)-EGFP (pTD241) andpUBC-K6a(N171K)-EGFP (pTD242) were generated by replacingthe 740-bp Bam HI/Not I fragment containing EYFP from pTD200and pTD201, respectively, with the Bam HI/Not I flanked 716-bpEGFP cDNA sequence PCR-generated from pTD166 describedabove.

Cell Culture and Transfections

Human 293FT embryonic kidney cells were maintained in high-glucose DMEM supplemented with 10 % fetal bovine serum,2 mM L-glutamine, and 1 mM sodium pyruvate. Cells wereseeded in a 48-well plate at a density of 8×104 cells in 500 μlmedium 18–24 h prior to co-transfection with WT and mutantkeratin fluorescent protein fusion constructs with and withoutunmodified siRNA. K6a N171K-specific siRNA (K6a.513c.12) ornon-specific control siRNA (NSC4 or CD44) was co-transfectedat a final concentration of 1 nM with a total of 400 ng plasmidDNA consisting of 200 ng each of two of the following plasmids:K6a (N171K) fusion, K6a (WT) fusion, or pUC19. Cells wereimaged 48 h following transfection by inverted fluorescencemicroscopy as described [11].

Intradermal Injections and IntravitalFluorescence Imaging

Plasmids (20 μg in 75 μl PBS) were injected intradermally into thefootpad skin of anesthetized Swiss Webster mice as previouslydescribed [20]. Anesthetized mice were subsequently analyzed witha modified VivaScope 2500 System (previously described [6, 21])upgraded for dual fluorescence capabilities with a blue (488 nm)and newly added green (532 nm) laser (Caliber ID, formerly LucidInc., Rochester, NY) [22]. A standard microscope slide seated on acustom stage was coupled to the objective using UltrasoundTransmission gel, Aquasonic® 100 (Parker laboratories, INC.Fairfield, NJ). The mouse paw of interest was placed over theslide prepared with a drop of CrodamolTM STS (Croda Inc., Edison,NJ) as index matching fluid. Z-depth position was determined usingthe 630 nm laser, and data in all three channels were collected asdescribed [6].

Image Processing

VivaScope native grayscale image stacks were exported as TIFimages for processing using NIH ImageJ and FIJI public domainimage-processing software. Image stacks were despeckled toremove outlier noise pixels, processed with an unsharp mask filteralgorithm (sigma=1.0, weight=0.7), and false-colored in the case offluorescence data. Image intensity distributions (window and level)were set to common standard values across respective image sets tonormalize intensities for comparison. Still images from matcheddepths were extracted and assembled into overlays and composites.Finally, image stacks were exported as AVI format movies at7.0 fps using JPEG compression.

Skin Sectioning and Immunofluorescence

Following intravital imaging, mice were euthanized and the treatedskin tissues were excised and embedded in OCT compound(Tissue-Tek, Torrance, CA). Sections (10 μm) were processedand imaged by fluorescence microscopy as described [14].

Results

In Vitro Analysis of Mutant-Specific siRNA

Expression plasmids were designed and constructed toexpress WT and mutant (N171K) K6a cDNA fusedupstream of EGFP or tdTFP under the control of either theUBC or CMV promoter (Fig. 1; all future references toexpression plasmids are to those utilizing the UBC promoterunless otherwise noted). The simultaneous expression ofmutant and WT K6a fused to distinct reporter genes, EGFPand tdTFP, allowed rapid sequential detection of both formsof K6a in mouse skin using an intravital confocal imagingsystem equipped with dual fluorescence capabilities. Wehypothesized that co-delivery of two plasmids encodingreporter proteins, one fused to WT K6a and the other tomutant K6a cDNA, along with siRNA targeting the mutantkeratin mRNA, would result in selective inhibition of thefluorescent protein containing the mutant K6a, in recipientcells, while the fluorescent reporter fused to the WT keratinwould be unaffected. This would enable ratiometric imagingsuch that the inhibition would be internally controlled.

In vitro experiments performed in 293FT cells (Fig. 2)demonstrated that co-transfection of the pUBC-K6a(N171K)-tdTFP plasmid with K6a_513a.12 siRNA(targets the K6a N171K single nucleotide mutation [11])resulted in inhibition of the targeted tdTFP reporter (Fig. 2a),while co-transfection with non-specific siRNA (NSC4;Fig. 2) or CD44-specific siRNA (designed to inhibitCD44) did not show inhibition and resulted in similarexpression levels compared to transfection of plasmids (nosiRNA) alone (data not shown). Co-transfection of plasmidsexpressing K6a (WT)-EGFP (pTD241) and K6a (N171K)-tdTFP (pTD203) with K6a.513c.12 siRNA resulted in

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

specific knockdown of expression of K6a (N171K)-tdTFPwith little or no effect on levels of K6a (WT)-EGFP(Fig. 2b). Swapping the EGFP and tdTFP cDNAs in thereporter plasmids (i.e., pUBC-K6a(WT)-tdTFP and pUBC-K6a(N171K)-EGFP) yielded inhibition of target gene ex-pression (EGFP signal in this configuration) when co-delivered with K6a.513c.12 siRNA (Fig. 2c). Unexpectedly,the K6a mutant (N171K)-EGFP-expressed fluorescencesignal was not reduced to the expected level (Fig. 2d).

Subsequent analysis of the filter sets revealed bleed-throughof light emitted from tdTFP (red) into the EGFP channel(green), resulting in spurious green signal when red signallevels were high. Therefore, the pUBC-K6a(WT)-EGFP andpUBC-K6a(N171K)-tdTFP combination of expression plas-mids was used in all future in vivo experiments to preventbleed-through signal from potentially masking the effects ofsiRNA on gene expression. These transfection experimentswere repeated with the matched plasmid set containing theCMV promoter in place of the UBC promoter (Fig. 1) andsimilar results were obtained (data not shown).

Separate Intradermal Injections of EGFP andtdTFP Reporter Plasmids

In order to determine if successive intradermal injectionsresult in plasmid delivery and subsequent expression in thesame cells or in distinct subsets of cells, two separateinjections (12 h apart) of plasmids expressing differentially-labeled reporter proteins were performed. The pUBC-K6a(N171K)-tdTFP expression construct was injected intothe footpad at time 0. At 12 h, pUBC-K6a(WT)-EFGP wasinjected into the same footpads, as close to the originalinjection site as possible. The footpad skin was imagedin vivo using the VivaScope at 36 h. As seen in Fig. 3 andSupplementary Video 1, a different, yet overlapping, set ofkeratinocytes express the two different reporter proteins.This is in contrast to co-injection of the different plasmids inthe same bolus, which results in an apparent completeoverlap of cells expressing each reporter (see SupplementaryVideo 2). For this reason, siRNAs and reporter plasmidswere co-injected in the following experiments to maximizeco-delivery to the same cells.

Fig. 1 Schematic representations of the expression plas-mids used.

Fig. 2 In vitro analysis of siRNA activity. K6a N171K-specific siRNA (K6a_513a.12) or non-specific control siRNA (NSC4) wasco-transfected with a the pUBC-K6a(N171K)-tdTFP expression construct (pTD203) in 293FT tissue culture cells or b pTD203and a similar plasmid, pTD241, expressing K6a (WT)-EGFP. K6a_513c.12 siRNA or non-specific control (NSC4) siRNA was co-transfected with c pUBC-K6a(N171K)-EGFP expression construct (pTD242) or d pTD242 and pTD202 expressing K6a (WT)-tdTFP. Scale bar=200 μm.

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

siRNA-mediated Inhibition of Mutant KeratinExpression In Vivo

Unmodified K6a_513c.12 siRNA or non-specific controlsiRNA (K6a-3′-UTR-1) was co-injected with a combinationof two expression plasmids (pUBC-K6a(WT)-EGFP andpUBC-K6a(N171K)-tdTFP) into mouse footpads (Fig. 4a).After 24 h, the footpads were analyzed by whole bodyimaging, which revealed localized fluorescence in the centerof the injected footpads (Fig. 4b). Next, fluorescence at thecellular level was evaluated by confocal in vivo imaging.The footpad was imaged at 40-μm depth using the 488 nmlaser. As with whole animal imaging, a single region ofEGFP fluorescence was localized to the center of eachinjected paw. A Z-stack of images and multiple duplicate

images at a fixed intermediate depth were captured using the488 nm (fluorescence), 532 nm (fluorescence), and 630 nm(reflectance) lasers. In Fig. 4c, an averaged overlaycomposite of the multiple images shows equivalent EGFPexpression in both treatment sites, while tdTFP expressionwas reduced to near baseline in the footpad co-injected withthe inhibitory K6a_513c.12 siRNA. The entire Z-stack canbe seen as a movie in the Supplementary Video 2.

Following intravital imaging, the mice were euthanizedand the footpads were embedded in OCT and frozen forhistological analysis. The levels of EGFP and tdTFPobserved in skin sections derived from each footpad usingconventional fluorescence microscopy supported the intravi-tal imaging findings showing decreased tdTFP expression inthe footpad co-injected with K6a_513c.12 siRNA (Fig. 4d).

Fig. 3 Separate intradermal injections of distinct reporter expression plasmids to mouse footpads results in distinct subsets ofkeratinocytes that express each reporter protein. To determine if successive intradermal injections result in delivery andexpression of reporter genes in the same or different subsets of cells, two separate injections (12 h apart) of plasmidsexpressing differentially-labeled reporter proteins were performed. The expression construct pUBC-K6a(N171K)-tdTFP wasinjected into a mouse footpad at time 0. At 12 h, pUBC-K6a(WT)-EFGP was injected into the same footpad, as close to theoriginal injection site as possible. The footpad skin was imaged at 36 h using the intravital scope. a Reflectance was used tovisualize skin structure and create a reference image. Fluorescence using b 488 and c 532 nm lasers are overlaid d to show thesubsets of cells expressing the reporter protein from each injection. Scale bar=100 μm.

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

Visualization of Keratin Aggregates in MouseFootpads after Injection with the K6a (N171K)Expression Plasmid

Plasmids expressing K6a WT-EGFP (pTD239) or K6a(N171K)-EGFP (pTD240) under the control of the CMVpromoter were injected intradermally into the footpads ofmice. Fluorescence imaging was performed at 24 h follow-ing injection (Fig. 5). Mouse footpads were imaged for bothfluorescence (488 nm, pseudo colored green) and reflectanceat 630 nm (data not shown). The presence of keratinaggregation, that may mimic the pathology observed in PCskin, is observed (denoted with arrows) in the footpadexpressing mutant N171K K6a, while little or no keratinaggregation is observed in footpad cells expressing WT K6a.

DiscussionsiRNAs can be designed to specifically and potently targetand silence genes that are overexpressed in disease states, ormutant alleles (including point mutations [11, 23–25])whose expression causes pathology. This offers an opportu-nity for therapeutic intervention for a variety of skindisorders with etiologies based on specific mutations inskin-expressed proteins [4, 5, 26]. Several siRNAs haveentered clinical trials [1–3], including a mutant-specificsiRNA (K6a.513.12) designed for local administration toskin for the treatment of PC [4]. Although a multitude ofskin nucleic acid delivery technologies have been reportedwith varying effectiveness [4, 22, 27–31], the realization ofsiRNA skin therapeutics as a new drug class will requiredevelopment of more efficient and patient-friendly deliverytechnologies. Since trials with human volunteers are limitedby ethical and practical considerations (e.g., the limitednumber of PC patients), suitable pre-clinical animal modelsare required for testing and optimizing of therapeuticapproaches. We demonstrated the ability to evaluate expres-sion of both WT and mutant forms of keratin K6a in mouseskin by tagging each with distinguishable fluorescencemarkers (EGFP or tdTFP) followed by visualization with amultiwavelength fluorescence intravital imaging system,which is able to detect both EGFP and tdTFP fluorescenceand a reflectance reference image. Furthermore, the ability ofsiRNA (K6a_513a.12), which has been shown to selectivelyinhibit mutant K6a expression (with little or no effect onuntargeted K6a WT expression [4, 5, 11, 12]), can bemonitored in vivo in real time, and relative levels of targetand wild type gene expression assessed at microscopicresolution (see Fig. 4) in living tissues.

The demonstration that K6a.513.12 siRNA potentlyand specifically targets mutant K6a N171K gene expres-sion with little or no effect on wild type (see Fig. 2) isconsistent with the previous experiments and publicationsleading up to a phase 1b trial using a clinical-gradeversion of the K6a_513a.12 siRNA (TD101, [4]). Forexample, in experiments using the wild type or mutant

K6a N171K cDNA linked to yellow fluorescent protein(YFP, precursor constructs to those used in this study) orto firefly luciferase, greater than 80 % reduction wasobserved in mutant (but not wild type) gene expressionwith cells treated with K6a.513.12 siRNA as assayed byFACS analysis or luciferase light output and similarresults were observed in mouse footpad skin [17].Furthermore, a RT-qPCR method was developed todistinguish the single nt change between K6a wild typeand N171K mutant mRNAs. In transfection experimentsin which equimolar amounts of mRNAs encoding wildtype and mutant K6a were expressed in transiently-transfected PC-10 cells, 90 % inhibition of mutantmRNA was observed under conditions in which wildtype expression was unaffected [12]. Even more dramaticRT-qPCR results were observed following transfection ofhuman primary keratinocytes derived from a PC patientharboring the K6a N171K mutation, in which a 98 %reduction in mutant mRNA was observed with respect towild type following treatment with K6a.513.12 siRNA[5]. Taken together, these results (and additional unpub-lished data) demonstrate the remarkable quantifiablepotency and discriminatory activity of this siRNA.

One model of PC pathophysiology (and other kerati-nization disorders such as epidermolysis bullosa simplexand epidermolytic hyperkeratosis) is that mutant keratinsare not able to assemble properly into intermediatefilaments, and this results in faulty structures [32, 33].With fluorescence intravital imaging as described here,aggregates can be observed following expression of the

Fig. 4 Intravital visualization of siRNA-mediated targeting ofmutant gene expression in mouse footpad skin. a Schematicof the relevant plasmids expressing WT K6a-EGFP (pTD241)and mutant K6a (N171K)-tdTFP (pTD203) under the controlof the UBC promoter. b Equivalent amounts of unmodifiedsiRNA (K6a_513a.12 or a non-specific control siRNA de-signed to target the K6a 3′-UTR) and WT (pTD241) andmutant (pTD203) K6a plasmids (20 μg each) were intrader-mally injected into the footpad of an anesthetized mouse.Whole animal imaging (IVIS Lumina II, Ex 535 nm, Em DsRed)after 24 h revealed decreased red fluorescence in the centerof the left paw (co-injected with K6a_513a.12) compared tothe right paw (co-injected with non-specific control K6a 3′-UTR-1 siRNA). c Following whole animal imaging, mousepaws were further imaged with the confocal imagingVivaScope. Images were collected every 2 μm from the skinsurface to a depth of 130 μm. En face images usingreflectance and green and red fluorescence from both pawsare shown at 70 μm. Note that the left panel in c isreflectance microscopy to show general skin structure. Scalebars=100 μm. The full movie dataset is provided in Supple-mentary Video 2. d Fluorescence microscopy of frozen skinsections prepared from footpad skin of mice euthanized afterthe intravital imaging confirms that specific siRNA(K6a_513a.12), but not the non-specific control siRNA,inhibits mutant K6a-tdTFP expression.

b

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

fluorescently-tagged mutant K6a N171K protein. To ourknowledge, this is the first time aggregates, whichpresumably include mutant keratins that have notassembled properly (a presumed PC pathogenesis molec-ular marker), have been observed in vivo.

Intravital imaging will be useful for evaluating, andoptimizing, siRNA delivery technologies including topicalformulations or mechanical delivery with microneedles. Forexample, topical application of siRNA could be adminis-tered prior to or after introduction of reporter plasmid(s).Alternatively, transgenic reporter mice could be preparedthat stably and simultaneously express both K6a (WT)-EGFP and K6a (N171K)-tdTFP for a more stable, andpossibly more versatile, model. Application of siRNAtargeting the N171K mutation would result in inhibition ofthe tdTFP expression but not control EGFP expression, andas transgenes, the two reporters would operate similarly. Avariety of different siRNA delivery technologies could thenbe screened to determine which most effectively deliversfunctional siRNA to different regions of the skin and not belimited to the footpad. As a first step towards this goal, wehave shown specific inhibition of a targeted reporter proteinrelative to a control in this model by co-injection of siRNAwith the relevant expression plasmids.

The development of an effective, topical, and patient-friendly siRNA delivery approach would be a boon forpatients suffering from a variety of skin disorders. Theability to design, synthesize, and identify effective siRNAsis now routine in many laboratories, suggesting delivery isthe last major hurdle to reaching full clinical potential forskin diseases. The availability of imaging strategies thatallow detection and discrimination of distinct siRNA-targeted and non-targeted control reporters noninvasively inliving skin will facilitate development of more effective

delivery systems by providing a rapid readout of efficacy.By combining macroscopic and microscopic in vivo imagingto reveal siRNA delivery and then validating the in vivoresults with tissue analysis post-necropsy, a maximumamount of high-value information can be obtained from asmaller number of animals than would be required if thestudies were performed via serial euthanization studies. Themodels and tools described here should accelerate and refinesuch delivery tools.

ConclusionsFluorescence confocal imaging can be used to visualizereporter gene expression with subcellular resolution. Dualfluorescence capabilities can be utilized to discriminategreen and red fluorescent protein expression and monitorin real time repression of one reporter following inhibitionwith a specific siRNA tailored to that gene and using thesecond non-targeted reporter as an internal control. Thismodel can be used to screen for effective and potent siRNAsas well as evaluate delivery technologies to facilitatedevelopment of siRNAs as therapeutics.

Acknowledgments. The authors would like to thank Conor Cox for hisefforts in acquiring the confocal microscopy and fluorescenthistological data. We also thank Robert Kaspar, Heini Ilves, and JedHumphries for technical support and Andrea Burgon for administrativesupport. This work was supported by NIH grants R44AR056559 (RLK,CHC) and the Chambers Family Foundation (CHC).

Conflict of Interest. Roger Kaspar and Robyn Hickerson have patentsissued and pending on using siRNA to treat PC and siRNA deliverytechnologies.

Fig. 5 Intravital visualization of mutant and wild type K6a expression in skin. The expression plasmids pUBC-K6a(WT)-EGFP(pTD239, a) or pUBC-K6a(N171K)-EGFP (pTD240, b) were intradermally injected (40 μg in 70 μL PBS) into mouse footpads andimaged (24 h). a WT K6a expression is generally uniformly expressed throughout the cytoplasm of each cell leaving a darknuclear region in the center of each cell (arrows depict examples). b. Mutant N171K K6a/EGFP expression results inpronounced (as compared to WT expression) aggregation (see arrows). Scale bar=100 μm. The full data set is shown as amovie in Supplementary Video 3.

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model

Electronic Supplementary MaterialBelow is the link to the electronic supplementarymaterial.Supplementary Video 1(Movie showing full dataset fromFig. 3.) Separate intradermal injections of distinct reporter expressionplasmids to mouse footpads results in distinct subsets of keratinocytes thatexpress each reporter protein. To determine if successive intradermalinjections would result in delivery and expression of reporter genes in thesame, or different, subsets of cells, two separate injections (12 h apart) ofplasmids expressing differentially-labeled reporter proteins were performed.The expression construct pUBC-K6a(N171K)-tdTFP was injected into amouse footpad at time 0. At 12 h, pUBC-K6a(WT)-EFGP was injected intothe same footpad, as close to the original injection site as possible. Thefootpad skin was imaged at 36 h using the intravital scope. a Reflectancewas used to visualize skin structure and create a reference image.Fluorescence using b 488 and c 532 nm lasers are d overlaid to show thesubsets of cells expressing the reporter protein from each injection. (AVI10221 kb)Supplementary Video 2As described in Fig. 4, mouse paws wereimaged with the confocal imaging VivaScope. Movie shows all imagescollected every 2 μm from the skin surface to a depth of 130 μm. En faceimages using reflectance and green and red fluorescence from both paws areshown. Note that the left panel in (c) is reflectance microscopy to showgeneral skin structure. (AVI 32240 kb)Supplementary Video 3(Movieshowing full dataset from Fig. 5.) Intravital visualization of mutant andwildtype K6a expression in skin. The expression plasmids pUBC-K6a(WT)-EGFP (pTD239, a) or pUBC-K6a(N171K)-EGFP (pTD240, b) wereintradermally injected (40 μg in 70 μL PBS) into mouse footpads andimaged (24 h). a WT K6a expression is generally uniformly expressedthroughout the cytoplasm of each cell leaving a dark nuclear region in thecenter of each cell (arrows depict examples). b Mutant N171K K6a/EGFPexpression results in pronounced (as compared to WT expression)aggregation. (AVI 3714 kb)

References1. Zhou Y, Zhang C, Liang W (2014) Development of RNAi technology

for targeted therapy—a track of siRNA based agents to RNAitherapeutics. J Control Release 193:270–281

2. Haussecker D (2014) Current issues of RNAi therapeutics delivery anddevelopment. J control Release : Off J Control Release Soc 195:49–54

3. Kubowicz P, Zelaszczyk D, Pekala E (2013) RNAi in clinical studies.Curr Med Chem 20:1801–1816

4. Leachman SA, Hickerson RP, Schwartz ME et al (2010) First-in-humanmutation-targeted siRNA phase Ib trial of an inherited skin disorder.Mol Ther 18:442–446

5. Leachman SA, Hickerson RP, Hull PR et al (2008) Therapeutic siRNAsfor dominant genetic skin disorders including pachyonychia congenita.J Dermatol Sci 51:151–157

6. Gonzalez-Gonzalez E, Kim YC, Speaker TJ et al (2011) Visualizationof plasmid delivery to keratinocytes in mouse and human epidermis. SciRep 1:158

7. Leachman SA, Kaspar RL, Fleckman P et al (2005) Clinical andpathological features of pachyonychia congenita. J Investig DermatolSymp Proc 10:3–17

8. McLean WH, Hansen CD, Eliason MJ, Smith FJ (2011) The phenotypicand molecular genetic features of pachyonychia congenita. J InvestigDermatol 131:1015–1017

9. Smith FJD, Hansen CD, Hull PR et al (2014) Pachyonychia congenita.GeneReviews:http://www.ncbi.nlm.nih.gov/books/NBK1280/

10. Cao YA, Hickerson RP, Seegmiller BL et al (2015) Gene expressionprofiling in pachyonychia congenita skin. J Dermatol Sci 77:156–165

11. Hickerson RP, Smith FJ, Reeves RE et al (2008) Single-nucleotide-specific siRNA targeting in a dominant-negative skin model. J InvestigDermatol 128:594–605

12. Hickerson RP, Leachman SA, Pho LN et al (2011) Development ofquantitative molecular clinical end points for siRNA clinical trials. JInvestig Dermatol 131:1029–1036

13. Ra H, Piyawattanametha W, Gonzalez-Gonzalez E et al (2011) In vivoimaging of human and mouse skin with a handheld dual-axis confocalfluorescence microscope. J Investig Dermatol 131:1061–1066

14. Hickerson RP, Gonzalez-Gonzalez E, Vlassov AV et al (2012) Intravitalfluorescence imaging of small interfering RNA-mediated gene repres-sion in a dual reporter melanoma xenograft model. Nucl Acid Ther22:438–443

15. Diaconeasa A, Boda D, Neagu M et al (2011) The role of confocalmicroscopy in the dermato-oncology practice. J Med Life 4:63–74

16. Calzavara-Pinton P, Longo C, Venturini M, Sala R, Pellacani G (2008)Reflectance confocal microscopy for in vivo skin imaging. PhotochemPhotobiol 84:1421–1430

17. Hickerson RP, Vlassov AV, Wang Q et al (2008) Stability study ofunmodified siRNA and relevance to clinical use. Oligonucleotides18:345–354

18. Lara MF, Gonzalez-Gonzalez E, Speaker TJ et al (2012) Inhibition ofCD44 gene expression in human skin models, using self-delivery shortinterfering RNA administered by dissolvable microneedle arrays. HumGene Ther 23:816–823

19. Smith FJ, Hickerson RP, Sayers JM et al (2008) Development oftherapeutic siRNAs for pachyonychia congenita. J Investig Dermatol128:50–58

20. Wang Q, Ilves H, Chu P et al (2007) Delivery and inhibition of reportergenes by small interfering RNAs in a mouse skin model. J InvestigDermatol 127:2577–2584

21. Gonzalez-Gonzalez E, Ra H, Hickerson RP et al (2009) siRNAsilencing of keratinocyte-specific GFP expression in a transgenic mouseskin model. Gene Ther 16:963–972

22. Hickerson RP, Wey WC, Rimm DL et al (2013) Gene silencing in skinafter deposition of self-delivery siRNA with a motorized microneedlearray device. Mol Ther Nucl Acids 2, e129

23. Schwarz DS, Ding H, Kennington L et al (2006) Designing siRNA thatdistinguish between genes that differ by a single nucleotide. PLoSGenet 2, e140

24. Allen EH, Atkinson SD, Liao H et al (2013) Allele-specific siRNAsilencing for the common keratin 12 founder mutation in Meesmannepithelial corneal dystrophy. Invest Ophthalmol Vis Sci 54:494–502

25. Ohnishi Y, Tamura Y, Yoshida M, Tokunaga K, Hohjoh H (2008)Enhancement of allele discrimination by introduction of nucleotidemismatches into siRNA in allele-specific gene silencing by RNAi.PLoS One 3, e2248

26. Leslie Pedrioli DM, Fu DJ, Gonzalez-Gonzalez E et al (2012) Genericand personalized RNAi-based therapeutics for a dominant-negativeepidermal fragility disorder. J Investig Dermatol 132:1627–1635

27. Hegde V, Hickerson RP, Nainamalai S et al (2014) In vivo genesilencing following non-invasive siRNA delivery into the skin using anovel topical formulation. J Control Release 196:355–362

28. Gonzalez-Gonzalez E, Speaker TJ, Hickerson RP et al (2010) Silencingof reporter gene expression in skin using siRNAs and expression ofplasmid DNA delivered by a soluble protrusion array device (PAD).Mol Ther 18:1667–1674

29. Steinstraesser L, Lam MC, Jacobsen F et al (2014) Skin electroporationof a plasmid encoding hCAP-18/LL-37 host defense peptide promoteswound healing. Mol Ther 22:734–742

30. Zheng D, Giljohann DA, Chen DL et al (2012) Topical delivery ofsiRNA-based spherical nucleic acid nanoparticle conjugates for generegulation. Proc Natl Acad Sci 109:11975–11980

31. Chen M, Zakrewsky M, Gupta V et al (2014) Topical delivery ofsiRNA into skin using SPACE-peptide carriers. J Control Release179:33–41

32. Cheng J, Syder AJ, Yu QC et al (1992) The genetic basis ofepidermolytic hyperkeratosis: a disorder of differentiation-specificepidermal keratin genes. Cell 70:811–819

33. Irvine AD, McLean WH (1999) Human keratin diseases: the increasingspectrum of disease and subtlety of the phenotype-genotype correlation.Br J Dermatol 140:815–828

R.P. Hickerson et al.: Visualization of siRNA Effectiveness in a Skin Disease Model