platelet-derived growth factor c plays a role in the branchial arch malformations induced by...

Platelet-Derived Growth Factor C Plays a Role in theBranchial Arch Malformations Induced by Retinoic Acid

Jing Han,1 Li Li,1 Zhaofeng Zhang,1 Ying Xiao,1 Jiuxiang Lin,2 Liping Zheng,1 and Yong Li1*1Department of Food Science and Nutrition, School of Public Health, Peking University, Beijing, China

2School of Stomatology, Peking University, Beijing, China

Received 25 May 2006; Revised 12 September 2006; Accepted 25 September 2006

BACKGROUND: All-trans-retinoic acid (RA) can produce branchial arch abnormalities in postimplantation rodentembryos cultured in vitro. Platelet-derived growth factor C (PDGF-C) was recently identified as a member of thePDGF ligand family. Many members of the PDGF family are essential for branchial arch morphogenesis and canbe regulated by RA. The roles of PDGF-C in branchial arch malformations induced by RA and possible mecha-nisms were investigated. METHODS: In whole embryo culture (WEC), mouse embryos were exposed to RA at 0,0.1, 0.4, 1.0, or 10.0 lM, PDGF-C at 25, 50, or 75 ng/mL, or PDGF-C at 25, 50, or 75 ng/mL containing 0.4 lMRA. After 48 h of culture, mouse embryos were examined for dysmorphogenesis, and whole-mount immuno-histochemistry was applied to PDGF-C. In explant cultures, explants were exposed to the same doses of RA andPDGF-C as WEC. Semiquantitative RT-PCR, zymography, and reverse zymography were used to evaluate theexpressions and activities of matrix metalloproteinase (MMP)-2, MMP-14, and tissue inhibitor of metalloprotei-nase (TIMP)-2. RESULTS: PDGF-C was reduced by RA, and exogenous PDGF-C rescued the branchial arch mal-formations induced by RA. Moreover, PDGF-C prevented RA-induced inhibition of the migratory ability ofmesenchymal cells in the first branchial arch, by regulating the expressions of MMP-2, MMP-14, and TIPM-2.CONCLUSIONS: Our results suggest that RA exposure reduces the expression of PDGF-C. The branchial archmalformations resulting from fetal RA exposure are caused at least partially by loss of PDGF-C and subsequentmisregulations of the expressions of MMP-2, MMP-14, and TIMP-2. Birth Defects Research (Part A) 79:221–230,2007. � 2006 Wiley-Liss, Inc.

Key words: PDGF-C; all-trans-retinoic acid; branchial arch morphogenesis; MMP-2; MMP-14; TIMP-2; craniofa-cial development; cell migration

INTRODUCTION

Platelet-derived growth factor C (PDGF-C), a recentlyidentified member of the PDGF ligand family, is a multido-main protein with a C-terminal domain capable of bindingto and activating PDGFR-a/a homodimers and PDGFR-a/b heterodimers (Gilbertson et al., 2001; Kazlauskas,2000; Li et al., 2000). The classical PDGF members, PDGF-A and PDGF-B, which are able to form three dimeric struc-tures, AA, AB, and BB, were discovered more than twodecades ago. All three forms of PDGF have been identifiedas key regulators for cell proliferation, survival, and migra-tion, as well as deposition and turnover of the extracellularmatrix (ECM) (Hoch and Soriano, 2003). Only recentlywere PDGF-C and PDGF-D discovered (Bergsten et al.,2001; LaRochelle et al., 2001; Li et al., 2000). Since theexpansion of the PDGF family, several investigators havereported that PDGF-C contributes to normal developmentof the heart, ear, central nervous system, and kidney(Campbell et al., 2005; Ponten et al., 2003; Reigstad et al.,

2005; Zhuo et al., 2004). Knockout studies of PDGF-C inmice clearly demonstrate a specific role for PDGF-C in pal-atogenesis (Ding et al., 2004). However, the functions andmechanisms of PDGF-C in embryonic development havenot been fully investigated.During early vertebrate embryogenesis, the cranial neural

crest cells (CNCs) originate from the posterior midbrain-hindbrain region and migrate to the first three branchialarches. This neural crest–derived mesenchyme, termed ecto-

The data in this manuscript have not been presented in part at any meeting.Grant sponsor: National Natural Science Foundation of the People’s Republic ofChina; Grant number: 30371224; Grant sponsor: Major Basic Research Develop-ment Program of the People’s Republic of China; Grant number: 2001CB510305.Grant sponsor: Beijing Natural Science Foundation; Grant number: 7052040.*Correspondence to: Yong Li, School of Public Health, Peking University,Beijing, 100083, China. E-mail: [email protected] online 20 December 2006 in Wiley InterScience (www.interscience.wiley.com).DOI: 10.1002/bdra.20329

Birth Defects Research (Part A): Clinical andMolecular Teratology 79:221–230 (2007)

� 2006 Wiley-Liss, Inc. Birth Defects Research (Part A) 79:221–230 (2007)

mesenchyme, interacts with epithelial and mesodermal cellpopulations within the arch and differentiates into cartilage,bone, and other connective tissues of the head and neck(Graham, 2003; Noden and Trainor, 2005). Migration ofCNCs, and/or the ability of their mesenchymal derivativesto remodel developing structures, is affected by the ECMand the matrix metalloproteinases (MMPs) (Chin and Werb,1997; Robbins et al., 1999; Xu et al., 2005). PDGF signaling isknown to regulate the deposition of ECM and tissue remod-eling factors and to participate in branchial arch morphogen-esis, which has been shown by previous studies (Ding et al.,2004; Jinnin et al., 2005; Robbins et al., 1999). In the PDGFR-anull and patch (Ph) mouse, which carries a deletion encom-passing the PDGFR-a gene, severe first branchial arch mal-formations were observed, which may result from faultymigration and/or differentiation of specific populations ofCNCs (Ding et al., 2004; Soriano, 1997; Tallquist and Soriano,2003). A previous study demonstrated that a PDGF-A–de-pendent role of MMP-2 contributed to the branchial archmalformations in Ph/Ph embryos (Robbins et al., 1999).Recently, it was reported that PDGF-C mediated increasedmRNA and protein levels of MMP-1 and its tissue inhibitorof metalloproteinase (TIMP)-1 (Jinnin et al., 2005). However,little is known about the role of PDGF-C and its effects onMMP expressions in the branchial arch morphogenesis.

Retinoic acid (RA) plays an important role in normalembryogenesis, by regulating morphogenesis, cell prolifer-ation, differentiation, and extracellular matrix production(Axel et al., 2001; Clagett-Dame and DeLuca, 2002). It isalso teratogenic at high concentrations and induces abnor-mal morphology in vertebrate embryos, including malfor-mations of the central nervous system, heart, ear, limb, andcraniofacial primordia (Campbell et al., 2004; Cuervo et al.,2002; Emmanouil-Nikoloussi et al., 2000). The branchialarches are particularly sensitive target tissues, based onample experimental and clinical evidence (Menegola et al.,2004; Mulder et al., 2000; Niederreither et al., 2000; Van-muylder et al., 2004). The mechanisms of RA-inducedbranchial arch malformations have been investigated bynumerous studies; however, they are not fully understood(Mark et al., 2004; Matt et al., 2003; Tahayato et al., 2003).Studies have shown that some members of the PDGF fam-ily can be regulated by RA (Liebeskind et al., 2000; Pinget al., 1999; Snyder et al., 2005). For example, RA up-regu-lated the expression of mRNAs of PDGF-A, PDGFR-a, andPDGFR-b in rat lung fibroblasts (Liebeskind et al., 2000).Due to the critical roles of PDGFs and RA in branchial archmorphogenesis and the relationship between them, wetested the hypothesis that RA may have effects on PDGF-Cin branchial arch malformation.

In this study, we demonstrated the effects of RA onPDGF-C expression during the critical period of CNCmigra-tion. Our results showed that exogenous PDGF-C rescuedRA-induced branchial arch malformations. Moreover,PDGF-C prevented RA-induced inhibition of the migratoryability of mesenchymal cells in the first branchial arch byregulating the expressions of MMP-2, MMP-14, and TIMP-2.

MATERIALS ANDMETHODS

Whole Embryo Culture (WEC)

Virgin female ICR mice were housed under controlledconditions of temperature (22 6 0.58C), humidity (50 610%), and light (12:12 h light/dark cycle) and were pro-

vided with food and water ad libitum. Both animal careguidelines and protocols for this study were in accordancewith Institutional Animal Care and Use Committee(IACUC) and OECD guidelines. The mice were matedovernight with males of proven fertility. The morning aftervaginal plug formation was designated as embryonic day(E) 0.5. Pregnant females were killed at E8.5, and embryoswere explanted by the technique of New (1978). It has beenproven that the growth of rodent embryos in vitro is com-parable to growth in vivo, based on the somite number andgeneral morphology (Lee et al., 1995; New, 1978). Embryos(three to five somites) with intact yolk sacs and ectoplacen-tal cones were randomly placed in 50 mL sealed culturebottles (three embryos/bottle) containing 3 mL of heat-inactivated sterile rat serum supplemented with 100 IU/mL of penicillin G (Sigma, St. Louis, MO) and 100 lg/mLof streptomycin (Sigma).All-trans-RA (Sigma) was dissolved in dimethyl sulfox-

ide (DMSO) and added to the culture media to final con-centrations of 0.1, 0.4, 1.0, or 10 lM. PDGF-C (R & D Sys-tems, Minneapolis, MN) was added to the culture media toconcentrations of 25, 50, or 75 ng/mL. Meanwhile, PDGF-C was added to the culture media containing 0.4 lM RA, toconcentrations of 25, 50, or 75 ng/mL. Control culturescontained 0.1% DMSO (v/v). All embryonic mouse cul-tures were terminated at 48 h and the embryos were eval-uated for viability by the presence of yolk sac circulationand heartbeat. The first branchial arches used, includingboth the maxillary and the mandibular processes, were dis-sected from embryos, frozen on dry ice, and stored at�708C until further analysis.

Whole-Mount Immunohistochemistry (WMI)

Embryos were immunostained according to the WMImethod previously described (Wei et al., 1999). Briefly, af-ter 48 h of culture, embryos (at least six per experimentalgroup from at least three different bottles) were fixed inDent’s fixative (1:4 in volume DMSO:methanol) overnightat 48C. After washing in methanol and incubation with 5%H2O2 in methanol, embryos were hydrated and incubatedfor 3 days at 48C with specific antibodies to PDGF-C (SantaCruz, CA). As negative controls, embryos were incubatedwith isotype-matched nonimmune IgG. 3,30-Diaminobenzi-dine tetrahydrochloride (DAB) was used to visualize thereaction product. Stained cells appeared dark brownthrough the light microscope.

Explant Cultures

First branchial arch tissues from E10.5 embryos wereminced into small (�1 mm2) pieces, and explant cultureswere established by culturing the pieces on gelatin-coatedcoverslips for motility analysis as previously described (Rob-bins et al., 1999). The motility analysis was done by main-taining the tissue in DMEM/F12 (1:1, v/v) medium (GibcoBRL, Gaithersburg, MD) supplemented with 1% FBS (GibcoBRL). RA was added to the cultures to final concentrationsof 0.1, 0.4, 1.0, or 10.0 lM. PDGF-C was added to the culturesto concentrations of 25, 50, or 75 ng/mL. Meanwhile, PDGF-C was added to the cultures containing 0.4 lM RA, to con-centrations of 25, 50, or 75 ng/mL. O-phenanthroline (o-PE;Sigma), a nonselective inhibitor of all metalloproteinases,was added to the cultures to concentrations of 1, 1.5, or2 mM to determine the roles of MMPs. Following a 48-h cul-ture, images of the explants and associated cells were

222 HAN ET AL.

Birth Defects Research (Part A) 79:221–230 (2007)

obtained and quantitated for the extent of cellular outgrowthusing NIH ImageJ. The extent of cell migration was deter-mined by measuring the radial distance from the edge of theexplant to the farthest cells at six consistent points aroundthe circumference of the explant. Data were expressed as av-erage number of micrometers from the edge of the explants.The number of branchial arch explants reflected the numberof multiple minced pieces.

RT-PCR

Total RNA was prepared from the first branchial arches.The frozen tissues were thawed in Trizol reagent (Gibco BRL),and total RNA was prepared utilizing the manufacturer’sinstructions. cDNA was prepared from 1–2 lg of total RNAusing an oligo (dT) primer and M-MLV Reverse Transcriptase(Gibco BRL). b-Actin was used as a loading control for eachPCR reaction, and primers were (sense primer first): b-actin,50-CCAAGGCCAACCGCGAGAAGATGAC-30 and 50-AGG-GTACATGGTGGTGCCGCCAGAC-30 (587 bp). The primerpairs for MMP-2, MMP-14 and TIMP-2, were previously pub-lished (Guyot et al., 2003). The relative amount of cDNA ineach matched set was normalized for the presence of the ubiq-uitous gene b-actin, using NIH ImageJ software. The analysiswas repeated three times in separate RNA isolations.

Zymography and Reverse Zymography

Branchial arches were homogenized in 0.1 M phosphatebuffer, pH 7.4, containing 0.1% Triton X-100 to a concentra-tion of 50 mg/mL. Homogenates were centrifuged at 5,000rpm for 10 min at 48C. Protein levels were measured by theBradford method, and the supernatant was used for zymog-raphy (Brown et al., 2002; Robbins et al., 1999) with reversezymography analyses as previously described (Oliver et al.,1997). Aliquots of each sample were subjected to electropho-resis in 10% SDS-polyacrylamide gels into which gelatin (1mg/mL) had been crosslinked. Following electrophoresis,the gels were soaked for 15 min in 2.5% Triton X-100 andrinsed with water. The gels were incubated for 20 h in LSCBbuffer (50 mM Tris, pH 7.6, 0.2 M NaCl, 5 mM CaCl2, 0.02%Brij 35, and 0.02% NaN3) at 378C. Following incubation, thegels were stained for 1 h with 0.125% Coomassie blue anddestained with 35% ethanol/10% acetic acid. Clear zones ofgelatin lysis against a blue background stain indicatedenzyme activity. A synthetic inhibitor of MMPs, o-PE at 2

mM treated the parallel gels to verify the zones of lysis wereproduced byMMPs. MMP-2 (gelatinase-A) andMMP-9 (gel-atinase-B) (Sigma) served as the standard. The stained gelswere digitized, and the zones of proteolysis, correspondingto the presence of proteases in the gel, were quantitatedusing NIH ImageJ. Reverse zymography analysis was per-formed as zymography described above except the 15%polyacrylamide gels contained 1 mg/mL gelatin and 160ng/mL MMP-2 (Sigma). The TIMP activities appeared asdark blue bands against a pale blue background.

Statistical Analysis

Data from WEC and explant cultures were evaluated bya one-way analysis of variance (ANOVA) followed by leastsignificant difference (LSD) test as a posthoc test or Dun-nett’s T3 test. The frequencies were analyzed with the Fish-er’s exact test. Statistical significance was at P < .05.

RESULTS

Expression Level of PDGF-C Reduced by RA

To determine whether RA affected expression of PDGF-C, WMI was used. The expression level of PDGF-C in bran-chial arches was suppressed by 0.4 lM RA. We found thatimmunoreactivity of PDGF-C was detected in the epitheliallining of the branchial arches, the branchial pouches andmembranes, the primitive oral cavity, and the nasal (olfac-tory) placode. Immunoreactivity of PDGF-C was weaker inthe first branchial arch of 0.4 lM RA-treated embryos thanin that of controls (Fig. 1B,C). Negative controls showed nospecific immunostaining (Fig. 1A).

Effects of Exogenous PDGF-C on RA-InducedMalformations of the Branchial Arches

Failure to maintain an appropriate expression level ofPDGF-C may be the cause of the branchial arch abnormal-ities of RA-treated embryos. To test this hypothesis, weconducted a rescue experiment. To determine the effects ofPDGF-C on branchial arch development and whether itcould rescue the RA-induced impairment of the branchialarch morphogenesis, PDGF-C was added to the culturemedium alone and the culture medium containing 0.4 lMRA, respectively.

Figure 1. RA-induced change in the expres-sion of PDGF-C in branchial arches usingwhole-mount immunohistochemistry. Theimmunoreactivity of PDGF-C was detected inthe epithelial lining of the branchial arches,the branchial pouches and membranes, theprimitive oral cavity, and the nasal (olfactory)placode. Immunoreactivity of PDGF-C wasweaker in the first branchial arch of 0.4 lMRA-treated embryos (C) than that of controls(B). Negative controls showed no specific im-munostaining (A). Embryos were photo-graphed at the same magnification. ol, olfac-tory placode; or, oral cavity; ba, branchialarches; h, heart. Scale bar¼ 1 mm.

223PDGF-C IN BRANCHIAL ARCH MORPHOGENESIS


Table 1Frequency and Distribution of the Observed Branchial Arch (BA) Abnormalities

Control

RA (lM) 0.4 lM RA þ PDGF-C (ng/mL)

0.1 0.4 1 10 25 50 75

Examinedembryos 33 20 38 36 20 25 26 22

Abnormal BAs 0 8 (40.0%)* 28 (73.7%)* 36 (100%)* 20 (100%)* 16 (64.0%)* 8 (30.8%)y{ 10 (45.5%)*{

I-II BA fused 7 20 22 15 16 6 7III BA absent 1 10 20 10 2 2 3No. of BAs 3.006 0.00 2.606 0.30 2.166 0.37* 1.936 0.28* 0.406 0.16* 2.126 0.60* 2.826 0.4§ 2.696 0.6Length of

BA (mm) 0.406 0.04 0.326 0.06* 0.276 0.05* 0.236 0.04* 0.066 0.02* 0.316 0.05*§ 0.336 0.06*§ 0.296 0.03*First BA-body

ratio (%) 9.066 1.26 7.976 1.16y 7.366 1.47* 6.926 1.19* 1.986 0.93* 8.026 1.67* 8.846 1.72§ 7.576 0.72*Crown-rump

length (mm) 4.416 0.46 3.996 0.22y 3.716 0.64* 3.326 0.70* 2.386 0.14* 3.916 0.35* 3.956 0.45* 3.796 0.24*Somite no. 5.006 0.00 4.956 0.22 4.866 0.41 4.706 0.73 2.806 0.90* 4.966 0.20 5.006 0.00 5.006 0.00

Values are expressed as mean6 SD for number of BA, length of BA, first BA-body ratio, crown-rump length, and somite number.*P < .01 compared to control.yP < .05 compared to control.{P < .05 compared to 0.4 lM RA.§P < .01 compared to 0.4 lM RA.

Figure 2. Morphological appear-ance of embryos after 48 h of cul-ture. (A) Control. Note the threedistinct branchial arches (arrows).(B) PDGF-C (50 ng/mL) showednormally shaped branchial archesbut did not result in a significantincrease in branchial arch growthcompared to that of controls. In (C),0.4 lM RA resulted in fusion of thefirst and second branchial arches(arrow). (D) PDGF-C (50 ng/mL)rescued the branchial arch malfor-mations caused by 0.4 lM RA. Notethe three distinct branchial arches(arrows). Black lines in (A) indicate(1) crown-rump length; (2) the firstbranchial arch length. Embryoswere photographed at the samemagnification. Scale bar ¼ 1 mm.

224 HAN ET AL.


After 48 h of culture, embryonic growth was inhibited.The crown-rump lengths and somite numbers werereduced by RA (Table 1). The branchial arch abnormal-ities included hypoplasia, agenesis, and the fused bran-chial arch (FBA) (Fig. 2C). The incidence of branchialarch abnormalities, the number of branchial arches, andthe length of the first branchial arch were all decreased indose-dependent manners (Table 1). We expected that asthe overall body size changes, a concomitant change inthe first branchial arch size should also occur. To accountfor this trend in changed body size and to determinewhether branchial arch size changed more than expectedwhen we measured the branchial arch growth, we nor-malized the branchial arch length to body size by divid-ing each measurement of the first branchial arch by thecrown-rump length of the embryo, which was called thefirst branchial arch–body ratio (Fig. 2A). The first bran-chial arch to body ratio was dose-dependently decreasedby RA (Table 1).

PDGF-C did not result in a significant increase in the firstbranchial arch morphogenesis at any dose (Fig. 2B) com-pared to controls (Fig. 2A). However, it could counteractthe teratogenic effect of RA. PDGF-C prevented the bran-chial arch defects induced by 0.4 lM RA (Fig. 2D). Supple-mentation of PDGF-C at doses of 25, 50, and 75 ng/mLcontaining medium with 0.4 lM RA reduced the incidenceof branchial arch malformations to 64.0, 30.8, and 45.5%,respectively. The number of branchial arches, the length ofthe first branchial arch, and the first branchial arch bodyratio were all increased by PDGF-C compared to 0.4 lMRA (Table 1).

Effects of Exogenous PDGF-C on RA-InducedInhibition of the Migratory Ability of

Mesenchymal Cells

To study the effects of RA and PDGF-C on the migratoryability of mesenchymal cells, the first branchial arch tissues

Figure 3. Effects of RA, PDGF-C,and o-phenanthroline on mesenchy-mal cell migration from branchialarch explants. (A) Vehicle control;(B) 0.1 lM RA; (C) 0.4 lM RA; (D) 1lM RA; (E) 10 lM RA; (F) 25 ng/mLPDGF-C; (G) 50 ng/mL PDGF-C;(H) 75 ng/mL PDGF-C; (I) 0.4 lMRA þ 25 ng/mL PDGF-C; (J) 0.4 lMRA þ 50 ng/mL PDGF-C; (K) 0.4lM RA þ 75 ng/mL PDGF-C; (L)1 mM o-phenanthroline; (M) 1.5 mMo-phenanthroline; (N) 2 mM o-phe-nanthroline. Explants were photo-graphed at the same magnification.Scale bar¼ 200 lm.



were minced into small pieces and cultured on gelatin-coated coverslips for 48 h. After culture, a substantial num-ber of cells had migrated out from the tissue from controlsand formed large ‘‘halos’’ around the explants (Fig. 3A).While the tissue from the RA-treated embryos, many fewercells migrated out of the explants (Fig. 3B–E). The extent ofmigration was quantified by measuring the average dis-tance the cells moved from the edge of the explant in a48-h period. The migration capabilities of branchial archcells from RA-treated embryos were reduced (Fig. 3B–E,Table 2). PDGF-C alone did not significantly stimulate cellmigration compared to controls (Fig. 3F–H). However, inthe presence of 0.4 lM RA, PDGF-C at 50 ng/mL could sig-nificantly increase cell migration compared to 0.4 lM RA(Fig. 3I–K, Table 2). In order to assess whether theimpaired migratory ability of RA-treated embryos was dueto inadequate MMP activities, explants of branchial archtissues from normal embryos were cultured in the presenceof metalloproteinase inhibitor, o-PE. Addition of o-PE tothe cultures decreased cell migratory ability in a dose-response fashion (Fig. 3L–N, Table 2).

Effects of RA and PDGF-C on the mRNAExpression of MMPs in the First Branchial Arch

Because MMPs are believed to participate in cell migra-tion and tissue remodeling, we evaluated the effects of RAand PDGF-C on the expression of MMPs in the first bran-chial arch. Total RNA prepared from the first branchialarch tissues was transcribed into cDNA and subsequentlyused in PCR utilizing primers for MMP-2, MMP-14 (themembrane-bound MMP that proteolytically activatesMMP-2), and TIMP-2. The expression levels of MMP-2 andMMP-14 were reduced, and the expression level of TIMP-2was increased by RA (Fig. 4). Following treatment at anyconcentration of PDGF-C in the presence of 0.4 lM RA, sig-nificant differences in MMP-2, MMP-14, and TIMP-2mRNA were not observed (data not shown).

Effects of RA and PDGF-C on the ProteinExpression of MMPs in the First Branchial Arch

To evaluate MMP-2 expression and activity, extracts ofthe first branchial arch were analyzed by gelatin zymogra-phy. MMP-2 (gelatinase A) was the only MMP detected,being present in the inactive 72-kDa proform and the pro-teolytically activated 62-kDa form. Other MMPs that canalso be detected with this analysis, including MMP-9 (gela-tinase B), were not observed in tissue at this stage of devel-opment. Compared to controls, significantly diminishedlevels of the pro- and active forms of MMP-2 in RA-treatedembryos were demonstrated. PDGF-C alone did not alterMMP-2 expression at any dose. However, PDGF-C couldresist the inhibition of 0.4 lM RA. In the presence of 0.4 lMRA, a significant increase of MMP-2 expression wasobserved in 50 ng/mL of PDGF-C compared with 0.4 lMRA (Fig. 5A,C).To determine tissue inhibitors of metalloproteinase,

reverse zymography was performed. TIMP-2, tissue inhibi-tor of metalloproteinase-2, was the only inhibitor detected.Figure 5 shows a reverse zymogram and demonstrates sig-nificantly increased levels of TIMP-2 expression in RA-treated embryos compared to controls. PDGF-C alone didnot alter TIMP-2 expression at any dose. However, in thepresence of 0.4 lM RA, both 25 and 50 ng/mL of PDGF-Csignificantly decreased TIMP-2 expression compared with0.4 lM RA (Fig. 5B,D).

DISCUSSION

This study suggests a link between the loss of PDGF-Csignaling and the branchial arch malformations induced byRA. Down-regulation of PDGF-C was caused by RA expo-sure. The specificity of the effect was first demonstrated bythe reduction of the incidence of branchial arch abnormal-ities and the improvement of branchial arch growth by

Table 2Effects of RA, Anti-PDGF-C Antibody, PDGF-C, and O-Phenanthroline on

Mesenchymal Cell Migration from Branchial Arch Explants

Embryo treatment/conditionNo. of explants

examined

Average distance (lm)of cell migration fromedge of explant (6SD)

Control 35 438.4 (26.7)0.1 lM RA 25 311.9 (30.9)*0.4 lM RA 30 254.5 (37.9)*1 lM RA 27 172.6 (19.5)*10 lM RA 28 46.8 (12.8)*25 ng/mL PDGF-C 25 421.1 (20.4)50 ng/mL PDGF-C 28 425.2 (17.6)75 ng/mL PDGF-C 22 419.9 (17.9)0.4 lM RA þ 25 ng/mL PDGF-C 23 276.9 (28.3)*0.4 lM RA þ 50 ng/mL PDGF-C 25 316.7 (20.2)*§

0.4 lM RA þ 75 ng/mL PDGF-C 24 286.7 (29.4)*1 mM phenanthroline 28 178.1 (24.4)*1.5 mM phenanthroline 26 96.6 (20.8)*2 mM phenanthroline 30 41.7 (8.17)*

*P < .01 compared to control.§P < .01 compared to 0.4 lM RA.

226 HAN ET AL.


application of PDGF-C. The specificity of the effect was fur-ther demonstrated by the stimulating effect of PDGF-C onthe migratory ability of mesenchymal cells impaired by RA.Regulating MMPs in the mesenchyme during branchial archdevelopment may be one of the mechanisms of PDGF-C sig-naling in RA-induced branchial arch malformations.

We found that PDGF-C was down-regulated by RA inWMI. The result is consistent with those of previous studies,which described that other members of PDGFs can be regu-lated by RA (Liebeskind et al., 2000; Ping et al., 1999; Snyderet al., 2005). In a recent study, Pdgfc�/� mouse embryosshowed cleft palate, demonstrating that PDGF-C has a spe-cific role in palatogenesis (Ding et al., 2004). Several studieshave shown that cleft palate can also be induced by RA(Cuervo et al., 2002; Emmanouil-Nikoloussi et al., 2000; Yuet al., 2005). Because palates form from the maxillary pro-cesses included to the first branchial arch (Kerrigan et al.,2000), the knockout study on PDGF-C indirectly supportsour findings that PDGF-C was down-regulated by RA andplayed a role in branchial arch morphogenesis.

We further demonstrated that exogenous PDGF-C res-cued the branchial arch malformations induced by RA.Similar rescue experiments have been reported (Ahlgrenet al., 2002; Schneider et al., 2001). Perturbing RA appearedto affect both Shh and Fgf8 genes and resulted in forebrainand facial defects, which were rescued by Shh and Fgf8proteins (Schneider et al., 2001). The ethanol-induced cra-nial neural crest cell death and the associated craniofacialgrowth defect could be rescued by the application of Shh(Ahlgren et al., 2002). In this study, 50 ng/mL PDGF-Cwas the most potent dosage that rescued branchial arch de-velopment that had been significantly impaired by 0.4 lMRA. Of interest, effects were reduced by a higher dosage ofPDGF-C, at 75 ng/mL. The results are consistent withthose of previous studies. Low concentrations (1 ng/mL)of PDGF-AB up-regulated the expression of pro-type I col-lagen mRNA, whereas this effect changed to the oppositeunder high concentrations (30 ng/mL) of PDGF-AB (Tanet al., 1995). Other research demonstrated that the amountof PDGF-A that stimulated neural crest outgrowth most

Figure 4. RA-induced changes in the expression of MMP-2,MMP-14, and TIMP-2 in the first branchial arch using RT-PCR. (A) The expression of MMP-2, MMP-14, TIMP-2, andb-actin transcripts. Lane 1, vehicle control; lane 2, 0.1 lM;lane 3, 0.4 lM; lane 4, 1 lM; lane 5, 10 lM; lane M, marker.(B) The ratio between the tested target gene and internalstandard b-actin band was calculated. Within each experi-ment, ratios of the samples were compared to those of con-trols, which were set at 1. Data were from three independentexperiments and are shown as mean 6 SD. *Values aresignificantly different from the vehicle control (P < .05).**Values are significantly different from the vehicle control(P < .01).



significantly was 10 ng/mL, but not 5 or 20 ng/mL (Liet al., 2001). A more recent study suggested that PDGF-Cup-regulated the protein levels of MMP-1 in humandermal fibroblasts more significantly by treatment with10 ng/mL than with 20 ng/mL (Jinnin et al., 2005).It is believed that matrix-degrading proteinases play an

important role in tissue remodeling (Basbaum and Werb,1996; Werb, 1997). Among those are the MMPs, a com-plex family of proteinases secreted as proenzymes(Curran and Murray, 2000; Vu and Werb, 2000). MMPsfunction primarily at the cell surface or in the extracellu-lar space, and their proteolytic activities are controlledthrough zymogen activation and inhibition by endoge-nous proteinase inhibitors known as the TIMPs (Bakeret al., 2002; Woessner, 2002). It is generally believed thatthe balance between MMPs and TIMPs is among the criti-cal determinants that control the integrity of the ECMand subsequently affect cell fate (Guyot et al., 2003). Onceectomesenchyme precursors have arrived in the branchialarches, ECM molecules are synthesized and secreted bydifferentiating cells. As the branchial arches enlarge andchange shape, the ECM must undergo changes in itsstructure and composition (Robbins et al., 1999). Duringsuch tissue remodeling, MMPs and TIMPs have beenshown to be critical to these alterations as morphogenesisproceeds (Chin and Werb, 1997; Parsons et al., 1997).They may function to degrade existing ECM componentsand also to regulate cell-ECM interactions as ECM struc-ture changes. Inhibition of MMP activity in the embry-onic mouse mandible disrupted its development (Chinand Werb, 1997). Similarly, perturbing the formation ofthe ECM in the CNC-derived odontogenic and palatalmesenchyme may contribute to the dental cusp growthdefect and cleft palate in Pdgfra�/� mice (Xu et al., 2005).Just as reduction of MMP-2 in the Ph/Ph mouse mayaffect remodeling of the first branchial arch causing thecharacteristic cleft face and small mandible (Robbinset al., 1999), down-regulation of MMP-2 and MMP-14 aswell as up-regulation of TIMP-2 may contribute to thebranchial arch malformations observed in RA-exposedembryos. In the same way, alterations of the expressionsof MMP-2, MMP-14, and TIMP-2 to normal levels may bethe reason for rescue of RA-induced branchial arch mal-formations by exogenous PDGF-C. Although to ourknowledge, there is no previous report that RA regulatesMMP and TIMP expressions in branchial arch cells, stud-ies have demonstrated that RA regulated MMP andTIMP expressions in other cell types (Ho et al., 2005;Lateef et al., 2004; Mao et al., 2003; Murphy et al., 2004).The origins of the cell types in each study were different,but the results of these studies indirectly support ourfindings. MMP-14 is a component of the pro-MMP-2 acti-vation cascade (Hernandez-Barrantes et al., 2000), but nota gelatinase or a collagenase. Therefore, it cannot bedetected by zymography, used in this study. A drawbackin the present study is the absence of the effects of RAand PDGF-C on protein expression and activity of MMP-14, which will be studied in the future.To further identify the roles of MMPs in branchial

arch development, a nonselective inhibitor of all MMPs,o-PE, was used. In a previous study, BB3103, anothergeneral MMP inhibitor, was used in palate organ cultureand caused cleft palate, implying a role for MMPs innormal palatogenesis (Brown et al., 2002). In the explantculture study, addition of o-PE to the cultures decreased

Figure 5. Effects of RA and PDGF-C on MMP-2 and TIMP-2expression and activity. (A) Representative zymogram and (B)reverse zymogram of first branchial arch extracts from RA- andPDGF-C–treated embryos. The 72-kDa pro- and 62-kDa activeforms of MMP-2 and the 21-kDa active form of TIMP-2 weredetected. (A,B) Lane 1, vehicle control; lane 2, 0.1 lM RA; lane 3,0.4 lM RA; lane 4, 1 lM RA; lane 5, 10 lM RA; lane 6, 25 ng/mLPDGF-C; lane 7, 50 ng/mL PDGF-C; lane 8, 75 ng/mL PDGF-C;lane 9, 0.4 lM RA þ 25 ng/mL PDGF-C; lane 10, 0.4 lM RA þ 50ng/mL PDGF-C; lane 11, 0.4 lM RA þ 75 ng/mL PDGF-C; (A)lane 12, 2 mM o-phenanthroline; (A) lane 13, standard MMP-2. (C)MMP-2 and (D) TIMP-2 activities were graphed as the fold changewhen the control MMP activity was set at 1. The pictures are repre-sentatives of three independent experiments. *Values are signifi-cantly different from the vehicle control (P < .05). **Values are sig-nificantly different from the vehicle control (P < .01). §Values aresignificantly different from the 0.4 lM RA (P < .05). §§Values aresignificantly different from the 0.4 lM RA (P < .01).

228 HAN ET AL.


cell migratory ability in a dose-dependent manner,reflecting the importance of MMPs in branchial arch tis-sue remodeling.

In conclusion, our results demonstrated that PDGF-Csignaling was decreased by RA exposure. ExogenousPDGF-C rescued branchial arch defects induced by RA, viaregulating MMP-2, MMP-14, and TIMP-2 syntheses andactivities. We suggest that PDGF-C signaling may be a newmechanism of branchial arch malformations induced byRA. The results of the current study may lay the founda-tion for research efforts aimed at understanding in greaterdetail new regulatory mechanisms involving the interplaybetween RA and PDGFs in craniofacial development.

REFERENCES

Ahlgren SC, Thakur V, Bronner-Fraser M. 2002. Sonic hedgehog rescues cra-nial neural crest from cell death induced by ethanol exposure. ProcNatl Acad Sci USA 99:10476–10481.

Axel DI, Frigge A, Dittmann J, et al. 2001. All-trans retinoic acid regulatesproliferation, migration, differentiation, and extracellular matrix turn-over of human arterial smooth muscle cells. Cardiovasc Res 49:851–862.

Baker AH, Edwards DR, Murphy G. 2002. Metalloproteinase inhibitors: bio-logical actions and therapeutic opportunities. J Cell Sci 115:3719–3727.

Basbaum CB, Werb Z. 1996. Focalized proteolysis: spatial and temporal reg-ulation of extracellular matrix degradation at the cell surface. CurrOpin Cell Biol 8:731–738.

Bergsten E, Uutela M, Li X, et al. 2001. PDGF-D is a specific, protease-acti-vated ligand for the PDGF beta-receptor. Nat Cell Biol 3:512–516.

Brown NL, Yarram SJ, Mansell JP, et al. 2002. Matrix metalloproteinaseshave a role in palatogenesis. J Dent Res 81:826–830.

Campbell JL Jr, Smith MA, Fisher JW, et al. 2004. Dose-response for retinoicacid-induced forelimb malformations and cleft palate: a comparison ofcomputerized image analysis and visual inspection. Birth Defects Res BDev Reprod Toxicol 71:289–295.

Campbell JS, Hughes SD, Gilbertson DG, et al. 2005. Platelet-derivedgrowth factor C induces liver fibrosis, steatosis, and hepatocellular car-cinoma. Proc Natl Acad Sci USA 102:3389–3394.

Chin JR, Werb Z. 1997. Matrix metalloproteinases regulate morphogenesis,migration and remodeling of epithelium, tongue skeletal muscle andcartilage in the mandibular arch. Development 124:1519–1530.

Clagett-Dame M, DeLuca HF. 2002. The role of vitamin A in mammalianreproduction and embryonic development. Annu Rev Nutr 22:347–381.

Cuervo R, Valencia C, Chandraratna RA, et al. 2002. Programmed cell deathis required for palate shelf fusion and is regulated by retinoic acid. DevBiol 245:145–156.

Curran S, Murray GI. 2000. Matrix metalloproteinases: molecular aspects oftheir roles in tumour invasion and metastasis. Eur J Cancer 36:1621–1630.

Ding H, Wu X, Bostrom H, et al. 2004. A specific requirement for PDGF-C inpalate formation and PDGFR-alpha signaling. Nat Genet 36:1111–1116.

Emmanouil-Nikoloussi EN, Goret-Nicaise M, Foroglou CH, et al. 2000. Cra-niofacial abnormalities induced by retinoic acid: a preliminary histo-logical and scanning electron microscopic (SEM) study. Exp ToxicolPathol 52:445–453.

Gilbertson DG, Duff ME, West JW, et al. 2001. Platelet-derived growth fac-tor C (PDGF-C), a novel growth factor that binds to PDGF alpha andbeta receptor. J Biol Chem 276:27406–27414.

Graham A. 2003. Development of the pharyngeal arches. Am J Med GenetA 119:251–256.

Guyot R, Magre S, Leduque P, et al. 2003. Differential expression of tissueinhibitor of metalloproteinases type 1 (TIMP-1) during mouse gonaddevelopment. Dev Dyn 227:357–366.

Hernandez-Barrantes S, Toh M, Bernardo MM, et al. 2000. Binding of active(57 kDa) membrane type-1 matrix metalloproteinase (MT1-MMP) totissue inhibitor of metalloproteinase (TIMP-2) regulates MT1-MMPprocessing and proMMP-2 activation. J Biol Chem 275:12080–12089.

Ho LJ, Lin LC, Hung LF, et al. 2005. Retinoic acid blocks pro-inflammatory cyto-kine-induced matrix metalloproteinase production by down-regulating JNK-AP-1 signaling in human chondrocytes. Biochem Pharmacol 70:200–208.

Hoch RV, Soriano P. 2003. Roles of PDGF in animal development. Develop-ment 130:4769–4784.

Jinnin M, Ihn H, Mimura Y, et al. 2005. Regulation of fibrogenic/fibrolyticgenes by platelet-derived growth factor C, a novel growth factor, inhuman dermal fibroblasts. J Cell Physiol 202:510–517.

Kazlauskas A. 2000. A newmember of an old family. Nat Cell Biol 2:E78–79.

Kerrigan JJ, Mansell JP, Sengupta A, et al. 2000. Palatogenesis and potentialmechanisms for clefting. J R Coll Surg Edinb 45:351–358.

LaRochelle WJ, Jeffers M, McDonald WF, et al. 2001. PDGF-D, a new prote-ase-activated growth factor. Nat Cell Biol 3:517–521.

Lateef H, Stevens MJ, Varani J. 2004. All-trans-retinoic acid suppresses ma-trix metalloproteinase activity and increases collagen synthesis in dia-betic human skin in organ culture. Am J Pathol 165:167–174.

Lee YM, Osumi-Yamashita N, Ninomiya Y, et al. 1995. Retinoic acid stage-dependently alters the migration pattern and identity of hindbrain neu-ral crest cells. Development 121:825–837.

Li J, Molkentin JD, Colbert MC. 2001. Retinoic acid inhibits cardiac neuralcrest migration by blocking c-Jun N-terminal kinase activation. DevBiol 232:351–361.

Li X, Ponten A, Aase K, et al. 2000. PDGF-C is a new protease-activatedligand for the PDGF alpha-receptor. Nat Cell Biol 2:302–309.

Liebeskind A, Srinivasan S, Kaetzel D, et al. 2000. Retinoic acid stimulatesimmature lung fibroblast growth via a PDGF-mediated autocrinemechanism. Am J Physiol Lung Cell Mol Physiol 279:L81–90.

Mao JT, Tashkin DP, Belloni PN, et al. 2003. All-trans retinoic acid modu-lates the balance of matrix metalloproteinase-9 and tissue inhibitor ofmetalloproteinase-1 in patients with emphysema. Chest 124:1724–1732.

Mark M, Ghyselinck NB, Chambon P. 2004. Retinoic acid signalling in thedevelopment of branchial arches. Curr Opin Genet Dev 14:591–598.

Matt N, Ghyselinck NB, Wendling O, et al. 2003. Retinoic acid-induced de-velopmental defects are mediated by RARbeta/RXR heterodimers inthe pharyngeal endoderm. Development 130:2083–2093.

Menegola E, Broccia ML, Di Renzo F, et al. 2004. Relationship between hind-brain segmentation, neural crest cell migration and branchial archabnormalities in rat embryos exposed to fluconazole and retinoic acidin vitro. Reprod Toxicol 18:121–130.

Mulder GB, Manley N, Grant J, et al. 2000. Effects of excess vitamin A on de-velopment of cranial neural crest-derived structures: a neonatal andembryologic study. Teratology 62:214–226.

Murphy KA, Villano CM, Dorn R, et al. 2004. Interaction between the arylhydrocarbon receptor and retinoic acid pathways increases matrix met-alloproteinase-1 expression in keratinocytes. J Biol Chem 279:25284–25293.

New D. 1978. Whole embryo culture and the study of mammalian embryoduring organogenesis. Biol Rev 53:81–122.

Niederreither K, Vermot J, Schuhbaur B, et al. 2000. Retinoic acid synthesis andhindbrain patterning in the mouse embryo. Development 127:75–85.

Noden DM, Trainor PA. 2005. Relations and interactions between cranialmesoderm and neural crest populations. J Anat 207:575–601.

Oliver GW, Leferson JD, Stetler-Stevenson WG, et al. 1997. Quantitativereverse zymography: analysis of picogram amounts of metalloprotei-nase inhibitors using gelatinase A and B reverse zymograms. Anal Bio-chem 244:161–166.

Parsons SL, Watson SA, Brown PD, et al. 1997. Matrix metalloproteinases.Br J Surg 84:160–166.

Ping D, Boekhoudt G, Boss JM. 1999. Trans-Retinoic acid blocks platelet-derived growth factor-BB-induced expression of the murine monocytechemoattractant-1 gene by blocking the assembly of a promoter proxi-mal Sp1 binding site. J Biol Chem 274:31909–31916.

Ponten A, Li X, Thoren P, et al. 2003. Transgenic overexpression of platelet-derived growth factor-C in the mouse heart induces cardiac fibrosis,hypertrophy, and dilated cardiomyopathy. Am J Pathol 163:673–682.

Reigstad LJ, Varhaug JE, Lillehaug JR. 2005. Structural and functional spe-cificities of PDGF-C and PDGF-D, the novel members of the platelet-derived growth factors family. FEBS J 272:5723–5741.

Robbins JR, McGuire PG, Wehrle-Haller B, et al. 1999. Diminished matrixmetalloproteinase 2 (MMP-2) in ectomesenchyme-derived tissues ofthe Patch mutant mouse: regulation of MMP-2 by PDGF and effects onmesenchymal cell migration. Dev Biol 212:255–263.

Schneider RA, Hu D, Rubenstein JL, et al. 2001. Local retinoid signalingcoordinates forebrain and facial morphogenesis by maintaining FGF8and SHH. Development 128:2755–2767.

Snyder JM, Jenkins-Moore M, Jackson SK, et al. 2005. Alveolarization in reti-noic acid receptor-beta-deficient mice. Pediatr Res 57:384–391.

Soriano P. 1997. The PDGF alpha receptor is required for neural crest celldevelopment and for normal patterning of the somites. Development124:2691–2700.

Tahayato A, Dolle P, Petkovich M. 2003. Cyp26C1 encodes a novel retinoicacid-metabolizing enzyme expressed in the hindbrain, inner ear, firstbranchial arch and tooth buds during murine development. Gene ExprPatterns 3:449–454.

Tallquist MD, Soriano P. 2003. Cell autonomous requirement for PDGFRal-pha in populations of cranial and cardiac neural crest cells. Develop-ment 130:507–518.

Tan EM, Qin H, Kennedy SH, et al. 1995. Platelet-derived growth factors-AA and -BB regulate collagen and collagenase gene expression differ-entially in human fibroblasts. Biochem J 310:585–588.



Vanmuylder N, Evrard L, Glineur R, et al. 2004. Histologic and ultrastruc-tural features of cell death both physiological and induced by two dif-ferent teratogens in branchial arches of mouse embryo. Rev Med Brux25:14–20.

Vu TH, Werb Z. 2000. Matrix metalloproteinases: effectors of developmentand normal physiology. Genes Dev 14:2123–2133.

Wei X, Makori N, Peterson PE, et al. 1999. Pathogenesis of retinoic acid-induced ear malformations in primate model. Teratology 60:83–92.

Werb Z. 1997. ECM and cell surface proteolysis: regulating cellular ecology.Cell 91:439–442.

Woessner JF, Jr. 2002. MMPs and TIMPs—an historical perspective. MolBiotechnol 22:33–49.

Xu X, Bringas P, Jr., Soriano P, et al. 2005. PDGFR-alpha signaling is criticalfor tooth cusp and palate morphogenesis. Dev Dyn 232:75–84.

Yu Z, Lin J, Xiao Y, et al. 2005. Induction of cell-cycle arrest by all-trans reti-noic acid in mouse embryonic palatal mesenchymal (MEPM) cells. Tox-icol Sci 83:349–354.

Zhuo Y, Zhang J, Laboy M, et al. 2004. Modulation of PDGF-C and PDGF-Dexpression during bleomycin-induced lung fibrosis. Am J Physiol LungCell Mol Physiol 286:L182–188.

230 HAN ET AL.


platelet-derived growth factor c plays a role in the branchial arch malformations induced by...

Documents