ARTHRITIS & RHEUMATISMVol. 58, No. 1, January 2008, pp 292–301DOI 10.1002/art.23168© 2008, American College of Rheumatology

Investigation of Sensory Neurogenic Components in aBleomycin-Induced Scleroderma Model Using

Transient Receptor Potential Vanilloid 1 Receptor– andCalcitonin Gene-Related Peptide–Knockout Mice

Arpad Szabo,1 Laszlo Czirjak,1 Zoltan Sandor,1 Zsuzsanna Helyes,1 Terezia Laszlo,1

Krisztian Elekes,1 Tamas Czompoly,1 Anna Starr,2 Susan Brain,2

Janos Szolcsanyi,1 and Erika Pinter1

Objective. Along with their classic afferent func-tion (nociception), capsaicin-sensitive transient recep-tor potential vanilloid 1 (TRPV1) receptor–expressingsensory nerve terminals exert local and systemic effer-ent activities. Activation of TRPV1 causes sensory neu-ropeptide release, which modulates the inflammationprocess. The aim of the present study was to examinethe role of this modulatory role of TRPV1 receptor andthat of calcitonin gene-related peptide (CGRP) inbleomycin-induced scleroderma, using transgenic mice.

Methods. Cutaneous sclerosis was induced withdaily subcutaneous injections of bleomycin for 30 days.Control groups were treated with phosphate bufferedsaline (PBS). TRPV1 receptor gene–deficient(TRPV1�/�) mice and CGRP-knockout (CGRP�/�)mice and their wild-type (WT) counterparts were inves-tigated. A composite sclerosis score was calculated onthe basis of thickening, leukocyte infiltration, and theamount/orientation of collagen bundles. Dermal thick-

ness and the number of �-smooth muscle actin (�-SMA)–positive cells were also determined. The quantityof the collagen-specific amino acid hydroxyproline wasmeasured by spectrophotometry.

Results. Bleomycin treatment induced markedcutaneous thickening and fibrosis compared with thatobserved in control mice treated with PBS. The compos-ite sclerosis score was 18% higher, dermal thickness was19% higher, the number of �-SMA–positive cells was47% higher, and the amount of hydroxyproline was 57%higher in TRPV1�/� mice than in their WT counter-parts. Similarly, the composite sclerosis score was 47%higher, dermal thickness was 29% higher, the number of�-SMA–positive cells was 76% higher, and the amountof hydroxyproline was 30% higher in CGRP�/� micethan in the respective WT groups.

Conclusion. These results suggest that activationof the TRPV1 receptor by mediators of inflammationinduces sensory neuropeptide release, which might exertprotective action against fibrosis. We confirmed theprotective role of CGRP in the development of cutane-ous sclerosis.

Scleroderma (systemic sclerosis [SSc]) is a com-plex autoimmune disease. It is characterized by excessivecollagen production by activated fibroblasts, pathologicremodeling of connective tissue resulting in collagendeposition in the dermis, as well as vascular injury andimmune abnormalities (1). In localized scleroderma,these pathologic changes are limited to the skin andsubcutaneous tissue, but various internal organs are alsoaffected in SSc (2). Although several studies have beenperformed, the pathogenesis of scleroderma is complexand remains largely unknown.

Supported by the Wellcome Trust International ResearchDevelopment Award, Hungarian Research grants (OTKA-T-046729and ETT 05-598/2003), and the Peter Pazmany Programme of theHungarian National Office of Research and Technology. Dr. Starr’swork was supported by the Biotechnology and Biological SciencesResearch Council, UK.

1Arpad Szabo, MD, Laszlo Czirjak, MD, PhD, DSc, ZoltanSandor, MD, PhD, Zsuzsanna Helyes, MD, PhD, Terezia Laszlo, MD,PhD, Krisztian Elekes, MSc, Tamas Czompoly, MD, PhD, JanosSzolcsanyi, MD, PhD, DSc, Erika Pinter, MD, PhD, DSc: University ofPecs, Pecs, Hungary; 2Anna Starr, PhD, Susan Brain, PhD: King’sCollege London, London, UK.

Address correspondence and reprint requests to Erika Pinter,MD, PhD, DSc, Department of Pharmacology and Pharmacotherapy,University of Pecs, Faculty of Medicine, Szigeti u. 12, H-7624 Pecs,Hungary. E-mail: erika.pinter@aok.pte.hu.

Submitted for publication February 15, 2007; accepted inrevised form September 21, 2007.

292

Potent profibrotic cytokines (soluble factors)such as transforming growth factor �, interleukin-4,platelet-derived growth factor, monocyte chemoattrac-tant protein 1, and connective tissue growth factor areup-regulated in SSc (3). Earlier studies have shown thatscleroderma fibroblasts express �-smooth muscle actin(�-SMA), and their number and distribution coincidewith the localization and progression of the scleroticprocess (4). In SSc, the vasculopathy includes fibrointi-mal proliferation and episodes of vasospasms that leadto ischemia and consequent obliterative fibrosis (1). Thevasospastic episodes, called Raynaud’s phenomenon,occur following exposure to cold or stress and veryfrequently are the first manifestation of the disease (5).Previous studies suggested that activation of the immunesystem plays an important role in pathogenesis of SSc;however, it is not clear how autoimmunity and tissuefibrosis interact with each other (2).

Although numerous animal models of SSc havebeen developed, murine models are used most exten-sively. No animal model has been described that repro-duces all manifestations of SSc precisely (6). In thisstudy, we used the bleomycin-induced model developedby Yamamoto et al in 1999 (7). Bleomycin is an antibi-otic obtained from Streptomyces venticillus. It possessesantitumor activity and is frequently used to treat variouscancers (8). Bleomycin binds to DNA through its amino-terminal peptide, and the activated complex generatesfree radicals (by interacting with O2 and Fe2�) that areresponsible for scission of the deoxyribose backbone ofthe DNA (9). In vitro studies indicate that bleomycincauses accumulation of cells in the G2 phase of the cellcycle (10). Bleomycin is degraded by a specific hydrolasethat is found in a variety of normal tissue, and hydrolaseactivity is low in the skin and lung (11). Lung fibrosis isa well-known side effect of bleomycin treatment (12);therefore, bleomycin is frequently used to induce exper-imental pulmonary fibrosis in rodents. In addition,scleroderma has been reported in patients with cancerafter they received bleomycin therapy (13), andYamamoto et al observed that local injection of bleo-mycin causes skin fibrosis (7).

Capsaicin, the active ingredient in hot peppers,selectively excites and then desensitizes a major sub-population of nociceptive sensory nerve fibers, whichcontain the above-mentioned sensory neuropeptidesand thus are classified as “capsaicin-sensitive afferents”(14). In recent studies, the receptor for capsaicin, firstcalled vanilloid receptor 1 and now called transientreceptor potential vanilloid 1 receptor (TRPV1), wasidentified and cloned (15). The TRPV1 receptor is

associated with a nonselective cation channel that can beactivated by noxious heat, protons, vanilloids such ascapsaicin, and mediators of inflammation, e.g., the5-lipoxygenase product 12-hydroperoxyeicosatetraenoicacid; however, its endogenous ligand has not yet beenidentified (16).

Besides causing the classic afferent function (no-ciception), activation of TRPV1 receptors causes therelease of sensory neuropeptides. Among these, calcito-nin gene-related peptide (CGRP) mediates vasodilata-tion, and tachykinins, for example substance P, evokeplasma protein extravasation. Furthermore, somatosta-tin with antiinflammatory and antinociceptive action isalso released. It is known that TRPV1 receptor–expressing sensory neurons play an important modula-tory role in the pathomechanism of several diseases,such as bronchial asthma (17), rheumatoid arthritis(18–20), eczema (21), dermatitis (22), and migraine (23).As an important integrator molecule of pain and inflam-mation, the TRPV1 receptor (24) may play a role inrepair mechanisms and the chronic fibrotic phase ofinflammatory processes.

Previous studies have established that the num-ber of CGRP-immunoreactive C fibers is significantlydecreased in skin samples from patients with sclero-derma (25–27). The importance of vasculopathy andsubsequent obliterative fibrosis in the pathogenesis ofSSc is well known (1). Based on these facts, we proposethat the vasodilator neuropeptide CGRP (28) may exerta protective action in scleroderma. The aim of thepresent study was to examine the potential modulatoryrole of the TRPV1 receptor and CGRP in an experi-mental animal model of bleomycin-induced sclero-derma, using genetically manipulated mice.

MATERIALS and METHODS

Animals. Experiments were performed on 4–6-week-old female TRPV1 receptor gene–knockout mice (TRPV1�/�)and their wild-type (WT) counterparts (TRPV1�/�); all miceweighed 20–25 gm. The mice were successfully bred at theLaboratory Animal Centre of the University of Pecs, understandard pathogen-free conditions at 24–25°C, and had hadaccess to standard chow and water ad libitum. Alpha CGRP–knockout and WT mice were bred at the animal house ofKing’s College London.

Generation of transgenic mice. The generation ofTRPV1 receptor–knockout mice was achieved by homologousrecombination in embryonic stem cells (129 ES) to generate amouse lacking transmembrane domains 2–4 of the murineTRPV1 gene. Germline chimeras were crossed onto femaleC57BL/6 mice to generate heterozygotes, which were inter-crossed, giving rise to healthy homozygous mutant offspring in

ROLE OF TRPV1 RECEPTOR IN SCLERODERMA 293

the expected Mendelian ratio, as described by Davis et al (29).TRPV1 receptor–knockout mice were fully backcrossed ontoC57BL/6 mice, and these mice were used to generate WT andTRPV1 receptor–knockout colonies.

Alpha CGRP–knockout mice were created by disrup-tion of exon 5 (specific to �CGRP) of the calcitonin/�CGRPgene, using a cassette containing lacZ/cytomegalovirus/neomycin resistance genes (30). We received a pair of mice (1WT mouse and 1 CGRP-knockout mouse) that had previouslybeen fully backcrossed with C57BL/6 mice. We then used thesemice to generate WT and CGRP-knockout colonies.

Induction of cutaneous sclerosis with bleomycin. Cu-taneous sclerosis was induced by daily 0.1-ml subcutaneousinjections of bleomycin (100 �g/ml; Pharmachemise, Haarlem,The Netherlands) for 30 days, with a 27-gauge needle on thedorsal skin of the animals. The control group was treated withthe solvent (phosphate buffered saline [PBS]). Mice werekilled by cervical dislocation, under anesthesia with ketamine(100 mg/kg intraperitoneally; Richter Gedeon, Budapest, Hun-gary) and xylazine (5 mg/kg intramuscularly; Lavet Ltd.,Budapest, Hungary). The excised skin samples were investi-gated by histologic, biochemical, and molecular biologic meth-ods (7).

Groupings of mice. For each experiment, 4 groups ofmice were used, as follows: for the TRPV1 study, PBS-treatedTRPV1�/� mice, PBS-treated TRPV1�/� mice, bleomycin-treated TRPV1�/� mice, and bleomycin-treated TRPV1�/�

mice (n � 10–12 animals/group); for the CGRP study, PBS-treated CGRP�/� mice, PBS-treated CGRP�/� mice,bleomycin-treated CGRP�/� mice, and bleomycin-treatedCGRP�/� mice (n � 8 animals/group).

Histologic analysis. The day after mice received thefinal injections, the shaved dorsal skin was removed, fixed in4% paraformaldehyde, and embedded in paraffin. The generalhistologic appearance of the tissue was examined by hematoxy-lin and eosin and collagen-specific picrosyrius staining. Skinspecimens were assessed and scored using a semiquantitativecomposite sclerosis score. The composite histologic sclerosisscore was calculated on the basis of dermal inflammation (0 �none, 1 � little, 2 � mild, 3 � moderate, and 4 � severe),thickened collagen bundles (0 � normal, 1 � little, 2 � mild,3 � moderate, and 4 � severe), and dermal thickness com-pared with normal skin (0 � �125%, 1 � 125–149%, 2 �150–174%, 3 � 175–200%, and 4 � �200%). All parameterswere scored on a scale of 0 to 4, and the values were added (31).

Measurement of dermal thickness. Dermal thicknessat the injection sites was analyzed with an Olympus BX-51microscope (Tokyo, Japan) at 40� magnification using theSoft Imaging system (Olympus). The distance between theepidermal–dermal junction and the dermal–subcutaneous fatjunction was measured in 3 consecutive skin sections from eachanimal. In each group of mice, dermal thickness was expressedin micrometers.

Detection of myofibroblasts. Paraffin-embedded tissuesections from injected skin were used to quantify the numberof myofibroblasts, by staining for �-SMA. After deparaffiniza-tion, skin sections were immunostained with monoclonal anti-body against �-SMA (clone 1A4; DakoCytomation, Carpinte-ria, CA), according to the manufacturer’s instructions, usingthe Dako Autostainer Universal Staining System. Sectionswere visualized with diaminobenzidine and counterstained

with hematoxylin. In each section, �-SMA–positive cells werecounted in 3 randomly chosen high-power fields (32).

Measurement of hydroxyproline content. The aminoacid hydroxyproline is a major component of the proteincollagen; therefore, it can be used as an indicator to determinethe amount of collagen. Full-thickness, 6-mm–diameter punchbiopsy specimens were obtained from the shaved dorsal skin ofeach animal after the 4-week treatment and stored at –80°C.Collagen deposition was estimated by determining the totalcontent of hydroxyproline in the skin. The stored skin pieceswere hydrolyzed with 6M hydrochloric acid at 130°C for 3hours, according to the method previously described (33).After neutralization with sodium hydroxide, the hydrolysateswere diluted with distilled water and oxidated with chloramineT (Sigma, Munich, Germany), and staining was performedwith p-dimethylaminobenzaldehyde (Ehrlich’s reagent;Sigma). The absorbance at 557 nm was determined spectro-photometrically, and the quantity of hydroxyproline was cal-culated from a standard curve. Results were expressed asmicrograms of hydroxyproline per 6-mm–diameter skin pieces.

Reverse transcriptase–polymerase chain reaction (RT-PCR). Shaved dorsal skin was incised and stored in 1 ml ofRNAlater solution (Ambion, Cambridge, UK) at �20°C untilprocessed further. Total RNA was isolated using a GenEluteMammalian Total RNA Kit (Sigma) with proteinase K (Fluka,Buchs, Switzerland), according to the manufacturer’s instruc-tions. RNA yield and purity were determined by spectropho-tometry (NanoDrop Technologies, Wilmington, DE) and werealso analyzed by electrophoresis on 1% agarose gels. Specificmessenger RNA (mRNA) levels were quantified by the Light-Cycler RNA Master SYBR Green I quantitative real-timeRT-PCR assay on a LightCycler system (Roche, Mannheim,Germany).

Type I collagen �1 chain mRNA (GenBank accessionno. NM007742.2) was amplified by sense primer 5�-TCTACTGCAACATGGAGACAG-3� at position 3932 andantisense primer 5�-GCTGTTCTTGCAGTGATAGGTG-3�at position 4185. The housekeeping gene GAPDH mRNA(GenBank accession no. BC083080) was used as a control,amplified with sense primer 5�-GCAGTGGCAAAG-TGGAGATT-3� at position 122 and antisense primer 5�-TCTCCATGGTGGTGAAGACA-3� at position 370. The20-�l reaction mixture contained 250 ng total RNA, 5 mMMgCl2, 0.5 �M primers, plus reaction buffers, according to themanufacturer’s recommendation. The LightCycler PCR pro-gram consisted of the initial reverse transcription step at 55°Cfor 20 minutes, followed by a denaturation step at 95°C for 30seconds and 45 cycles of amplification for 10 seconds at 95°C,5 seconds at 58°C, and 15 seconds at 72°C.

In order to verify the purity of the products, meltingcurve analysis was performed at the end of the experiment.Quantification of results was accepted only when a singledominant peak was present in the melting analyses. In order tofurther confirm the purity and size of the PCR products, thereactions were also analyzed by electrophoresis on 1% agarosegels. The results were evaluated using LightCycler3 DataAnalysis software version 3.5.28 (Roche).

To compare the different RNA transcription levels,threshold cycle (Ct) values were compared directly. The Ct isdefined as the number of cycles needed for the fluorescencesignal to reach a specific threshold level of detection and is

294 SZABO ET AL

inversely correlated with the amount of template nucleic acidpresent in the reaction (34). First, expression of type I collagen�1 chain mRNA was quantified relative to that of the house-keeping gene GAPDH mRNA of the same sample, by calcu-lating the corrected difference in Ct (�Ct) value according tothe following formula: �Ct � Ct collagen � Ct GAPDH (35). Next,the differences between the variously treated animals wereanalyzed by calculating the x-fold difference compared with thePBS-treated WT control animals, according to the formulax-fold � �2.7 � (�Ct treated � �Ct PBS-treated WT control),because a 1-cycle �Ct difference corresponded to a 2.7-foldchange in the mRNA level (see Results).

Ethics considerations. All experimental procedureswere carried out according to the 1998/XXVIII Act of theHungarian Parliament on Animal Protection and Consider-ation Decree of Scientific Procedures of Animal Experiments(243/1988) and the Animals (Scientific Procedures) Act 1986(Great Britain). The studies were approved by the EthicsCommittee on Animal Research of Pecs University according

to the Ethical Codex of Animal Experiments, and a license wasgiven (license no. BA 02/200-6-2001).

Statistical analysis. Results are expressed as themean SEM. Statistical analysis was carried out with thenonparametric Mann-Whitney U test to determine significantdifferences between histologic scores, hydroxyproline content,and type I collagen mRNA levels in different groups. P valuesless than 0.05 were considered significant.

RESULTS

Establishment of bleomycin-induced dermalsclerosis in TRPV1�/� mice. In the initial studies previ-ously performed, subcutaneous injections of 0.01–1.0 mg/ml bleomycin for 18–24 days induced markeddermal sclerosis around the injection site in CH3 andBALB/c mice, but not in PBS-treated mice (15,36).

Figure 1. Composite histologic sclerosis score, dermal thickness, number of �-smooth muscle actin (�-SMA)–positive cells, and collagen-specific hydroxyproline content in transient receptor potential vanilloid 1 receptor–knockout (KO) (TRPV1�/�) and TRPV1�/� (wild-type [WT]) mice. a–c, Bleomycin treatment significantlyincreased the composite histologic score, dermal thickness, and the number of �-SMA–positive cells in bothTRPV1�/� and TRPV1�/� mice compared with the respective phosphate buffered saline (PBS)–treated groups.Furthermore, each measured histologic value for bleomycin-treated TRPV1�/� mice was significantly higher thanthat for bleomycin-injected TRPV1�/� animals, while no differences were observed between the groups ofPBS-treated mice. d, In both TRPV1�/� and TRPV1�/� mice, bleomycin treatment resulted in a significantincrease in the hydroxyproline content in the 6-mm skin patches compared with the respective PBS-treatedcontrol groups. Although there was no difference between the PBS-injected TRPV1�/� and TRPV1�/� mice, thebleomycin-induced increase in hydroxyproline content was significantly greater in TRPV1�/� mice than inTRPV1�/� mice. Values are the mean and SEM (n � 10–12 mice per group). � � P � 0.05, TRPV1�/� versusWT, by Mann-Whitney U test.

ROLE OF TRPV1 RECEPTOR IN SCLERODERMA 295

Because TRPV1 receptor–knockout mice were gener-ated from a C57BL/6 mouse strain, we first validated thesclerotic effect of bleomycin treatment in TRPV1�/�

animals. Histologic analysis showed thickened and ho-mogeneous collagen bundles, thickening of the dermis,replacement of subcutaneous fat by collagen bundles,and moderate inflammatory infiltrates in bleomycin-treated TRPV1�/� mice compared with PBS-treatedmice (Figures 1a and 2).

The composite sclerosis score was 58% higher inbleomycin-treated TRPV1�/� mice compared with thatin PBS-treated TRPV1�/� mice (mean SEM 6.33 0.19 versus 4.00 0.31) (Figure 1a). Dermal thicknesswas 42% higher in bleomycin-treated TRPV1�/� micecompared with that in PBS-treated TRPV1�/� mice(393.05 15.41 �m versus 278.62 11.38 �m) (Figure1b). The number of �-SMA–positive cells was increasedby 75% in bleomycin-treated TRPV1�/� mice comparedwith that in PBS-treated TRPV1�/� control mice(16.33 3.31 cells/field versus 9.3 1.34 cells/field)(Figure 1c). Consistent with the histologic changes, thelevel of the collagen-specific amino acid hydroxyprolinewas 47.5% higher in bleomycin-treated WT mice com-pared with that in PBS-treated WT mice (118.5 6.70�g/skin site versus 80.3 10.20 �g/skin site) (Figure 1d).PBS treatment itself did not induce significant dermalsclerotic changes compared with the naive skin of un-treated TRPV1�/� mice. There was no detectable dif-ference in the microscopic structure of the skin betweennaive TRPV1�/� mice and TRPV1�/� mice (results notshown).

Involvement of TRPV1 receptors in bleomycin-induced dermal sclerosis. In TRPV1�/� mice, the his-tologic changes were more pronounced, and the com-posite sclerosis score was 18% higher than that inbleomycin-treated TRPV1�/� mice (mean SEM7.46 0.20 versus 6.33 0.19) (Figure 1a). Dermalthickness was increased 19% in bleomycin-treatedTRPV1�/� mice compared with TRPV1�/� mice(464.86 10.15 �m versus 393.05 15.41 �m) (Figure1b). The number of �-SMA–positive cells in bleomycin-treated TRPV1�/� mice was 47% higher than that inbleomycin-treated TRPV1�/� mice (24.03 3.07 cells/field versus 16.33 3.31 cells/field) (Figure 1c). Simi-larly, the hydroxyproline content after bleomycin treat-ment was 57% greater in the knockout animals than inthe respective WT group (186.60 8.40 �g/skin siteversus 118.50 6.70 �g/skin site) (Figure 1d). Therewere no significant differences between composite scle-rosis scores, dermal thickness, hydroxyproline content,

Figure 2. Histopathologic evaluation of dermal sclerosis inTRPV1�/� mice and their WT counterparts. a and b, Hematoxylinand eosin (H&E)–stained sections from a TRPV1�/� mouse (a) anda TRPV1�/� mouse (b) treated with PBS for 4 weeks. Neither fi-brosis nor sclerosis was noted. c, H&E-stained section from aTRPV1�/� mouse treated with bleomycin (100 �g/ml) for 4 weeks.Dermal sclerosis was induced in the subcutaneous tissue. d, H&E-stained section from a TRPV1�/� mouse treated with bleomycin(100 �g/ml) for 4 weeks. Homogeneous collagen bundles, thickeningof the dermis, and total replacement of subcutaneous fat by collagenbundles were observed. e and f, Collagen-specific picrosyrius–stainedsections from a TRPV1�/� mouse (e) and a TRPV1�/� mouse (f)treated with bleomycin. g and h, Sections from a TRPV1�/� mouse (g)and a TRPV1�/� mouse (h) treated with bleomycin labeled withanti–�-SMA antibody (visualized with diaminobenzidine and counter-stained with hematoxylin). An increased number of �-SMA–positivecells was observed. (Original magnification � 40 in a–f; � 100 in gand f). See Figure 1 for other definitions.

296 SZABO ET AL

and numbers of �-SMA–positive cells in the PBS-treatedTRPV1�/� mice and the PBS-treated TRPV1�/� mice(Figures 1a–d).

Analysis of type I collagen �1 chain mRNAexpression in sclerotic skin, using quantitative RT-PCR.The quantitative PCR was calibrated by amplifyingcollagen mRNA, using a 2-fold serial dilution of totalRNA (33 ng, 66.5 ng, 125 ng, and 250 ng per reaction)isolated from mouse skin as starting material, andpresented as the raw Ct collagen values (Figure 3a). Linearregression analysis showed a good fit (r � 0.956) with aslope of 2.7, indicating that a 1-unit decrease in the levelof Ct collagen corresponds to a 2.7-fold increase in theconcentration of collagen mRNA. Alternatively, 2-foldmore collagen mRNA is represented by a 0.74-unitlower Ct collagen value.

The results of quantitative PCR measurements oftype I collagen �1 mRNA levels in skin biopsy speci-mens are presented in Figure 3b. Two groups of skinsamples (obtained 2 weeks and 4 weeks after initialtreatment) were examined. The mRNA levels of PBS-treated TRPV1�/�, bleomycin-treated TRPV1�/�, andbleomycin-treated TRPV1�/� mouse skin samples werecalculated relative to the levels in corresponding PBS-treated WT control samples. Surprisingly, in most of thesamples, the level of collagen mRNA was slightly de-creased compared with control samples. However, thedifferences were not statistically significant in any of thegroups examined. These results show that bleomycin

treatment did not significantly alter the type I collagen�1 mRNA level compared with the control GAPDHmRNA level.

Involvement of CGRP in bleomycin-inducedscleroderma model. Initially, the effect of bleomycininjection in the C57BL/6-derived CGRP�/� mice wascharacterized, and histologic investigations showedmarked dermal sclerosis in the bleomycin-treated mice(Figures 4a–c and Figure 5). Among WT mice, thecomposite sclerosis score was 42% higher (mean SEM4.25 0.38 versus 2.98 0.51), dermal thickness wasincreased 52% (434.49 16.41 �m versus 285.85 17.36 �m), and the number of �-SMA–positive cells wasaugmented by 62% (15.83 3.60 cells/field versus9.75 0.88 cells/field) in those treated with bleomycincompared with those that received PBS (Figures 4a–c).In bleomycin-treated WT mice, hydroxyproline contentincreased by 47% compared with that in PBS-treatedWT mice (137.60 14.58 �g/skin site versus 93.40 10.65 �g/skin site) (Figure 4d). Lack of the CGRP genecaused increased sclerotic changes, whereby the com-posite sclerosis score in bleomycin-treated CGRP�/�

mice was 47% higher than that in the respective WTmice (6.21 0.58 versus 4.25 0.47), dermal thicknesswas increased by 29% (557.23 18.38 �m versus434.49 16.41 �m), and the number of �-SMA–positivecells was augmented by 76% (27.83 4.58 cells/fieldversus 15.83 3.60 cells/field) (Figures 4a–c).

In accord with the histologic findings, the hy-

Figure 3. Measurement of the type I collagen �1 chain mRNA level by quantitative reverse transcription–polymerase chainreaction (PCR). a, The quantitative PCR was calibrated on a 2-fold dilution series of total mRNA isolated from skin biopsyspecimens. The raw threshold cycle (Ct collagen) values are presented as the mean and SEM results from 4 independentmeasurements. b, Type I collagen mRNA levels were measured in dorsal skin specimens from bleomycin (bleo)–treatedanimals. The amounts of mRNA in different samples relative to those in skin samples from PBS-treated WT mice (controlgroup) are shown. No significant differences were found between the different groups. n � 4–6 mice/group. See Figure 1 forother definitions.

ROLE OF TRPV1 RECEPTOR IN SCLERODERMA 297

droxyproline content increased 30% (mean SEM179.30 18.90 �g/skin site versus 137.60 14.58�g/skin site) in bleomycin-treated CGRP�/� animals(Figure 4d). Surprisingly, we observed that the compos-ite sclerosis score (3.96 0.49 versus 2.98 0.51),dermal thickness (386.25 10.26 �m versus 285.85 17.36 �m), the number of �-SMA–positive cells(15.04 2.39 cells/field versus 9.75 0.88 cells/field),and the hydroxyproline content (126.56 13.10 �g/skinsite versus 93.40 10.65 �g/skin site) were elevated inthe PBS-treated CGRP�/� mice compared with PBS-treated CGRP�/� mice (Figures 4a–d).

The severity of sclerotic changes in PBS-treatedCGRP�/� animals was similar to changes observed inbleomycin-treated CGRP�/� animals. Therefore, wehistologically analyzed naive dorsal skin samples from

CGRP-knockout and WT mice, and also measured thehydroxyproline content. The histologic structures of thenaive skin samples obtained from CGRP�/� mice andCGRP�/� mice did not differ. In addition, the hy-droxyproline concentration in samples of naive skinfrom CGRP�/� and CGRP�/� mice was not different(80.05 12.66 �g/skin site versus 100.59 15.41 �g/skinsite) (results not shown).

DISCUSSION

In this study, we demonstrated that a geneticdeficit of TRPV1 receptors or CGRP peptide increasesthe severity of sclerotic changes in bleomycin-induceddermal sclerosis in mice. On the basis of these results, wepresume that activation of TRPV1 receptors and subse-

Figure 4. Composite histologic sclerosis score, dermal thickness, number of �-smooth muscle actin (�-SMA)–positive cells, and collagen-specific hydroxyproline content in calcitonin gene-related peptide–knockout (KO)(CGRP�/�) and CGRP�/� (wild-type [WT]) mice. a–c, The composite histologic score, dermal thickness, andnumber of �-SMA–positive cells were significantly increased in bleomycin-treated CGRP�/� mice compared withbleomycin-treated CGRP�/� mice. However, these histologic parameters were similarly increased in thephosphate buffered saline (PBS)–treated CGRP�/� animals compared with their WT counterparts. d, Thehydroxyproline content in bleomycin-treated CGRP�/� mice was significantly higher than that in bleomycin-treated CGRP�/� mice. However, in accordance with histologic findings, the hydroxyproline content inPBS-treated CGRP�/� animals was significantly greater than the content in PBS-treated CGRP�/� mice. Valuesare the mean and SEM (n � 4–6 mice per group in a–c and 8 mice per group in d). � � P � 0.05 versus WTmice, by Mann-Whitney U test.

298 SZABO ET AL

quent neuropeptide release exert a protective modula-tory role in the pathogenesis of bleomycin-inducedscleroderma. Our experiments using CGRP�/� micehave provided the first direct evidence that a geneticdeficit of this peptide results in augmentation of thepathologic alterations in dermal sclerosis.

Over the last 5 years, the modulatory function ofthe TRPV1 receptor has been investigated in severalinflammatory diseases. Recent studies by our group andother investigators demonstrated the pronociceptive andproinflammatory roles that TRPV1 receptors play in themediation of acute pain and inflammation; however,their function in chronic inflammatory conditions hasnot yet been elucidated (36,37). Results of experimentsusing TRPV1 receptor–knockout mice established thatthe lack of TRPV1 receptors diminishes the symptomsof experimental arthritis (38). An accumulating numberof reports support the protective role of TRPV1 recep-tors in immune-mediated inflammatory diseases, such asoxazolone-induced allergic contact dermatitis (39), dini-trobenzene sulfonic acid–induced colitis (40), andendotoxin-induced shock (41).

Our study focused on the role of TRPV1 recep-tors in chronic fibrotic–sclerotic conditions. In thecourse of model validation, we measured 150% skinthickening, marked dermal sclerosis, an increased num-ber of �-SMA–positive cells, and a 148.5% increase inhydroxyproline content in bleomycin-treated TRPV1�/�

C57BL/6 mice compared with the PBS-treated controls,which is consistent with previous data concerning theC57BL/6 strain. PBS treatment itself did not inducesignificant sclerotic changes compared with the naiveskin of untreated mice (42,43). However, in contrast toprevious studies (44,45), we were not able to demon-strate increased type I collagen �1 chain mRNA levelsafter bleomycin treatment in TRPV1�/� mice. The basisof this observation currently is not clear, but the elevatedlevel of collagen protein along with normal mRNAlevels could be the result of diverse processes, such asincreased mRNA translation, increased procollagen pro-tein processing, or decreased collagen turnover.

In the present study, we established that a geneticdeficiency of TRPV1 receptors increases the bleomycin-induced dermal sclerosis detected by histologic andbiochemical methods. The absence of TRPV1 receptorsexacerbated histologic parameters of sclerotic changes,such as dermal thickness, collagen bundles, inflamma-tion, and number of �-SMA–positive cells. In addition,after bleomycin treatment, the level of hydroxyprolinewas elevated in the knockout animals compared withthat in their WT counterparts. Because there was no

Figure 5. Histopathologic evaluation of dermal sclerosis in transientreceptor potential vanilloid 1 receptor–knockout (TRPV1�/�) andCGRP�/� mice and their WT counterparts. a and b, Hematoxylin andeosin (H&E)–stained sections from an untreated (naive) CGRP�/�

mouse (a) and a naive CGRP�/� mouse (b). Neither fibrosis norsclerosis was noted. c, H&E-stained section from a CGRP�/� mousetreated with PBS for 4 weeks. Dermal sclerosis was not observed. d,H&E-stained section from a CGRP�/� mouse treated with PBS for 4weeks. Surprisingly, detectable sclerosis had developed. e and f,H&E-stained sections from a CGRP�/� mouse (e) and a CGRP�/�

mouse (f) treated with bleomycin (100 �g/ml) for 4 weeks. Sclerosisdeveloped in the subcutaneous tissue of both mice and was marked inthe knockout mouse. g and h, Picrosyrius-stained sections from aCGRP�/� mouse (g) and a CGRP�/� mouse (h) treated with bleomy-cin (100 ıg/ml) for 4 weeks. Sclerosis developed in the subcutaneoustissue of both mice and was marked in the knockout mouse. (Originalmagnification � 40.) See Figure 4 for other definitions.

ROLE OF TRPV1 RECEPTOR IN SCLERODERMA 299

detectable difference in the microscopic structure of theskin between the naive TRPV1�/� and the TRPV1�/�

mice (i.e., those that did not receive bleomycin), weconcluded that TRPV1 deficiency itself does not inducefibrotic changes. The development of fibrotic reactionsrequires the presence of a well-defined profibrotic agent,such as bleomycin. Presumably, mediators of inflamma-tion (e.g., bradykinin, prostanoids, or lipoxygenase prod-ucts) released during the development of dermal sclero-sis may activate or sensitize TRPV1 receptors (16).Activation of TRPV1 causes sensory neuropeptide re-lease, which subsequently exerts a protective influenceon pathologic alterations. The tachykinin substance Pmediates plasma protein extravasation (46), CGRP pos-sesses well-known, long-lasting vasodilatory effects (28),and somatostatin reduces the release of other neuropep-tides and directly inhibits inflammatory and immunecells (47).

Among neuropeptides released by TRPV1 recep-tor activation, only CGRP has been described as apotential factor in the pathogenesis of scleroderma andRaynaud’s phenomenon. Normal skin of the digits isrichly innervated by CGRP-containing nerve fibers thatplay a role in nociception (48). A significant reduction inthe number of CGRP-immunoreactive neurons has beenobserved in the skin of patients with primary Raynaud’sphenomenon and those with SSc (25–27). In Raynaud’sdisease, a deficiency of CGRP-containing nerves maylimit the manifestation of cold vasodilatation (49). Thesemorphologic findings have actuated investigation of thefunctional role of CGRP in in vivo sclerotic conditions,using genetically manipulated CGRP-knockout mice ina model of bleomycin-induced scleroderma. To ourknowledge, our results provide the first evidence that alack of CGRP exacerbates both histopathologic signs ofdermal sclerosis and elevates the collagen-specific hy-droxyproline content of skin after long-term treatmentwith bleomycin.

In the background of these findings, the absentvasodilator effect of CGRP can be supposed. The con-sequent increase in vasospasm leads to ischemia. Periodsof ischemia alternating with reperfusion cause endothe-lial dysfunction, formation of free radicals, and activa-tion of immune cells, which eventually provoke vascu-lopathy and obliterative fibrosis. We presume thatTRPV1 receptor–dependent CGRP release is responsi-ble for the protective role of TRPV1 receptor. Based onour findings, the role of TRPV1 receptor–dependentrelease of other neuropeptides (substance P, somatosta-tin) cannot be excluded.

Unexpectedly, we observed significant sclerotic

changes in the dorsal skin of PBS-treated CGRP�/�

mice compared with that in the skin of PBS-treatedCGRP�/� mice. Because there was no difference in theskin structure of naive CGRP�/� mice and CGRP�/�

mice, we assume that CGRP�/� mice are more sensitiveto physical injury induced by needles. Evidence existsthat physical injury such as that induced by vibration,trauma, or radiation therapy can provoke factors in-volved in idiopathic SSc (50). However, the exact mech-anism of increased fibrotic response in PBS-treatedCGRP�/� mice has not been elucidated.

According to the present findings, we can sup-pose that either TRPV1 receptor–mediated or TRPV1receptor–independent release of CGRP has a protectiverole against sclerotic changes, by acting on vascularpathophysiologic factors. CGRP receptor would be apotential target for the development of novel drugs totreat scleroderma.

AUTHOR CONTRIBUTIONS

Dr. Pinter had full access to all of the data in the study andtakes responsibility for the integrity of the data and the accuracy of thedata analysis.Study design. Szabo, Czirjak, Brain, Pinter.Acquisition of data. Szabo, Sandor, Laszlo, Elekes, Czompoly, Starr,Pinter.Analysis and interpretation of data. Szabo, Czirjak, Sandor, Helyes,Brain, Szolcsanyi, Pinter.Manuscript preparation. Szabo, Czirjak, Sandor, Helyes, Starr, Brain,Pinter.Statistical analysis. Szabo, Sandor.Breeding experimental procedures relevant to CGRP-knockout mice.Brain.

REFERENCES

1. Abraham DJ, Varga J. Scleroderma: from cell and molecularmechanisms to disease models [review]. Trends Immunol 2005;26:587–95.

2. Denton CP, Black CM, Abraham DJ. Mechanisms and conse-quences of fibrosis in systemic sclerosis [review]. Nat Clin PractRheumatol 2006;2:134–44.

3. Trojanowska M. Molecular aspects of scleroderma. Front Biosci2002;7:608–18.

4. Sappino AP, Masouye I, Saurat JH, Gabbiani G. Smooth muscledifferentiation in scleroderma fibroblastic cells. Am J Pathol1990;137:585–91.

5. Sakkas LI. New developments in the pathogenesis of systemicsclerosis [review]. Autoimmunity 2005;38:113–6.

6. Jimenez SA, Christner PJ. Murine animal models of systemicsclerosis [review]. Curr Opin Rheumatol 2002;14:671–80.

7. Yamamoto T, Takagawa S, Katayama I. Animal model of scleroticskin. I. Local injections of bleomycin induce sclerotic skin mim-icking scleroderma. J Invest Dermatol 1999;112:456–62.

8. Umezawa H, Maeda K, Takeuchi T, Okami Y. New antibiotics,bleomycin A and B. J Antibiot (Tokyo) 1966;19:200–9.

9. Grollman AP, Takeshita M, Pillai KM, Johnson F. Origin andcytotoxic properties of base propenals derived from DNA. CancerRes 1985;45:1127–31.

300 SZABO ET AL

10. Twentyman PR. Bleomycin: mode of action with particular refer-ence to the cell cycle. Pharmacol Ther 1983;23:417–41.

11. Sebti SM, Jani JP, Mistry JS, Gorelic E, Lazo JS. Metabolicinactivation: a mechanism of human tumor resistance to bleomy-cin. Cancer Res 1991;51:227–32.

12. Adamson IV, Bowden DH. The pathogenesis of bleomycin-induced pulmonary fibrosis. Am J Pathol 1974;77:185–98.

13. Finch WR, Rodnan GP, Buckingham RB, Prince RK, WinkelsteinA. Bleomycin-induced scleroderma. J Rheumatol 1980;7:651–9.

14. Szolcsanyi J. Capsaicin-sensitive sensory nerve terminals with localand systemic efferent functions: facts and scopes of an unorthodoxneuroregulatory mechanism. Prog Brain Res 1996;113:343–9.

15. Caterina MJ, Schumacher MA, Tominaga M, Rosen TA, LevineJD, Julius D. The capsaicin receptor: a heat-activated ion channelin the pain pathway. Nature 1997;389:816–24.

16. Hwang SW, Cho H, Kwak J, Lee SY, Kang CJ, Jung J, et al. Directactivation of capsaicin receptors by products of lipoxygenases:endogenous capsaicin-like substances. Proc Natl Acad Sci U S A2000;97:6155–60.

17. Barnes PJ. Asthma as an axon reflex. Lancet 1986;1:242–5.18. Ferrel WR, Lam FY. Sensory neuropeptides in arthritis. In:

Geppetti P, Holzer P, editors. Neurogenic inflammation. BocaRaton (FL): CRC Press; 1996. p. 211–27.

19. Helyes Z, Szabo A, Nemeth J, Jakab B, Pinter E, Banvolgyi A, etal. Antiinflammatory and analgesic effects of somatostatin re-leased from capsaicin-sensitive sensory nerve terminals in aFreund’s adjuvant–induced chronic arthritis model in the rat.Arthritis Rheum 2004;50:1677–85.

20. Keeble J, Russell F, Curtis B, Starr A, Pinter E, Brain SD.Involvement of transient receptor potential vanilloid 1 in thevascular and hyperalgesic components of joint inflammation.Arthritis Rheum 2005;52:3248–56.

21. Naukkarinen A, Jarvikallio A, Lakkaporpi J, Harvima IT, HarvimaRJ, Horsmanheimo M. Quantitative histochemical analysis ofmast cells and sensory nerves in psoriatic skin. J Pathol 1996;180:200–5.

22. Gutwald J, Goebeler M, Sorg CJ. Neuropeptides enhance irritantand allergic contact dermatitis. J Invest Dermatol 1991;96:695–8.

23. Buzzi MG, Moskowitz MA. The antimigraine drug, sumatriptan(GR43175), selectively blocks neurogenic plasma extravasationfrom blood vessels in dura mater. Br J Pharmacol 1990;99:202–6.

24. Tominaga M, Tominaga T. Structure and function of TRPV1.Pflugers Arch 2005;451:143–50.

25. Bunker CB, Terenghi G, Springall DR, Polak JM, Dowd PM.Deficiency of calcitonin gene-related peptide in Raynaud’s phe-nomenon. Lancet 1990;336:1530–3.

26. Terenghi G, Bunker CB, Liu YF, Springall DR, Cowen T, DowdPM, et al. Image analysis quantification of peptide-immunoreac-tive nerves in the skin of patients with Raynaud’s phenomenon andsystemic sclerosis. J Pathol 1991;164:245–52.

27. Bunker CB, Goldsmith PC, Leslie TA, Hayes N, Foreman JC,Dowd PM. Calcitonin gene-related peptide, endothelin-1, thecutaneous microvasculature and Raynaud’s phenomenon. Br JDermatol 1996;134:399–406.

28. Brain SD, Williams TJ, Tippins JR, Morris HJ, MacIntyre I.Calcitonin gene-related peptide is a potent vasodilator. Nature1985;313:54–6.

29. Davis JB, Gray J, Gunthorpe MJ, Hatcher JP, Davey PT, OverendP, et al. Vanilloid receptor-1 is essential for inflammatory thermalhyperalgesia. Nature 2000;405:183–7.

30. Salmon AM, Damaj I, Sekine S, Picciotto MR, Marubio L,Changeux JP. Modulation of morphine analgesia in �CGRPmutant mice. Neuroreport 1999;10:849–54.

31. Yamamoto T, Takagawa S, Katayama I, Mizushima Y, NishiokaK. Effect of superoxide dismutase on bleomycin-induced dermal

sclerosis: implications for the treatment of systemic sclerosis.J Invest Dermatol 1999;113:843–7.

32. Distler JH, Jungel A, Huber LC, Schulze-Horsel U, Zwerina J,Gay RE, et al. Imatinib mesylate reduces production of extracel-lular matrix and prevents development of experimental dermalfibrosis. Arthritis Rheum 2007;56:311–22.

33. Woessener JF Jr. The determination of hydroxyproline in tissueand protein samples containing small proportions of this aminoacid. Arch Biochem Biophys 1961;93:440–7.

34. Walker NJ. A technique whose time has come. Science 2002;296:557–9.

35. Radonic A, Thulke S, Mackay IM, Landt O, Siegert W, Nitsche A.Guideline to reference gene selection for quantitative real-timePCR. Biochem Biophys Res Commun 2004;313:856–62.

36. Bolcskei K, Helyes Z, Szabo A, Sandor K, Elekes K, Nemeth J, etal. Investigation of the role of TRPV1 receptors in acute andchronic nociceptive processes using gene-deficient mice. Pain2005;117:368–76.

37. Szallasi A, Cruz F, Gepetti P. TRPV1: a therapeutic target fornovel analgesic drugs? Trends Mol Med 2006;12:545–54.

38. Szabo A, Helyes Z, Sandor K, Bite A, Pinter E, Nemeth J, et al.Role of transient receptor potential vanilloid 1 receptors inadjuvant-induced chronic arthritis: in vivo study using gene-defi-cient mice. J Pharmacol Exp Ther 2005;314:111–9.

39. Banvolgyi A, Palinkas L, Berki T, Clark N, Grant AD, Helyes Z,et al. Evidence for a novel protective role of the vanilloid TRPV1receptor in a cutaneous contact allergic dermatitis model. J Neu-roimmunol 2005;169:86–96.

40. Massa F, Sibaev A, Marsicano G, Blaudzun H, Storr M, Lutz B.Vanilloid receptor (TRPV1)-deficient mice show increased sus-ceptibility to dinitrobenzene sulfonic acid induced colitis. J MolMed 2006;84:142–6.

41. Okajima K, Harada M. Regulation of inflammatory responses bysensory neurons: molecular mechanism(s) and possible therapeu-tic applications. Curr Med Chem 2006;13:2241–51.

42. Yamamoto T, Kuroda M, Nishoika K. Animal model of scleroticskin. III. Histopathological comparison of bleomycin-inducedscleroderma in various mice strains. Arch Dermatol Res 2000;292:535–41.

43. Chan ES, Fernandez P, Merchant AA, Montesinos MC, Trzaska S,Desai A, et al. Adenosine A2A receptors in diffuse dermal fibrosis:pathogenic role in human dermal fibroblasts and in a murinemodel of scleroderma. Arthritis Rheum 2006;54:2632–42.

44. Yamamoto T, Nisshioka K. Increased expression of p53 and p21(Waf1/Cip1) in the lesional skin of bleomycin-induced sclero-derma. Arch Dermatol Res 2005;296:509–13.

45. Matsushita M, Yamamoto T, Nishioka K. Plasminogen activatorinhibitor-1 is elevated, but not essential, in the development ofbleomycin-induced murine scleroderma. Clin Exp Immunol 2004;139:429–38.

46. Lembeck F, Holzer P. Substance P as neurogenic mediator ofantidromic vasodilation and neurogenic plasma extravasation.Naunyn Schmiedebergs Arch Pharmacol 1979;310:175–83.

47. Pinter E, Helyes Z, Szolcsanyi J. Inhibitory effect of somatostatin oninflammation and nociception. Pharmacol Ther 2006;112:440–56.

48. Dalsgaard CJ, Jernbeck J, Stains W, Kjartansson J, HaegerstrandA, Hokfelt T, et al. Calcitonin gene-related peptide-like immuno-reactivity in nerve fibers in the human skin: relation to fiberscontaining substance P-, somatostatin- and vasoactive intestinalpolypeptide-like immunoreactivity. Histochemistry 1989;91:35–8.

49. Cooke JP, Marshall JM. Mechanisms of Raynaud’s disease [re-view]. Vasc Med 2005;10:293–307.

50. Komocsi A, Tovari E, Kovacs J, Czirjak L. Physical injury asprovoking factor in three patients with scleroderma. Clin ExpRheumatol 2000;18:622–4.

ROLE OF TRPV1 RECEPTOR IN SCLERODERMA 301