In vitro and in vivo analysis of the inability of fetal rabbitwounds to contract

THOMASM. KRUMMEL, MD*; H. PAUL EHRLICH, PhD°; JEFFREY M. NELSON, MD, BARBARA A. MICHNA, MD,BRIAN L. THOMAS, MD; JEFFREY H. HAYNES, MD; l. KELMAN COHEN, MD, ROBERT F. DIEGELMANN, PhD

Fetal rabbit wounds that are sutured show excellent repair without obvious scarring . In contrast, an unsutured wound in a rabbitfetus does not close, and it appears that the process of wound contraction does not occur. Experiments were carried out toillustrate the mechanisms responsible for the noncontraction of open fetal rabbit wounds . Results showed that the lack ofwound contraction was not an artifact caused by rapid fetal growth . With regard to the ability of cultured fetal fibroblasts toshow cytoplasmic muscle-induced cell contraction, we found that, in cultured fetal fibroblasts, cell contraction was inducedby adenosine triphosphate . Contractile abilities of fetal-derived fibroblasts were equivalent to those of adult-derivedfibroblasts. The fetal fibroblasts also demonstrated the generation of superior contractile activity when examined in afibroblast-populated collagen lattice model . Finally, the ability of amniotic fluid to alter wound contraction was addressed bymeans of the fibroblast-populated collagen lattice in vitro model. Increasing concentrations of amniotic fluid inhibited fetalfibroblast lattice contraction . Therefore, rabbit amniotic fluid contains an inhibitor that may be partially responsible for thenoncontraction of fetal rabbit wounds in utero. (WOUND REP REG 1993;1 :15-21)

In most adult animals, the contraction of open cuta-neous wounds restores a protective surface. However,an increasing body of literature" demonstrates thatthe fetus's response to injury is quite different fromthat ofthe adult. Contraction of open wounds does notappear to occur in the fetuses of many species,

From the Wound Healing Center of the Division of Plastic andReconstructive Surgery Department of Surgery,Medical College of VirginialVirginia Commonwealth University, Richmond, Va. ; and the ShrinersBurns Institute,° Massachusetts General Hospital,Harvard Medical School, Boston, Mass.

Presented in part at the Forty-fifth Annual Forum on Funda-mental Surgical Problems, American College ofSurgeons, Oct. 18, 1989, Atlanta, Ga .

Reprint requests : Robert F. Diegelmann, PhD, Medical Col-lege of Virginia, Box 117, MCVStation, Richmond, VA23298-0117.

*Present address: Department of Pediatric Surgery, Pennsyl-vania State College ofMedicine, Hershey, PA 11733.

Copyright C 1993 by The Wound Healing Society.1067-1927193 $1.00 + .10 36/1/42814

ATP

Adenosine triphosphateDMEM

Dulbecco's modification of Eagle's mediumFBS

Fetal bovine serumFPCL

Fibroblast-populated collagen lattice

including those of the human being,' monkey's andrabbit'- ', whereas contraction has been reported in thefetal lamb.'° Several hypotheses have been formulatedto explain the observed lack of contraction ofopen fetalrabbit wounds .

Somasunduram and Prathap' have suggested thata lack of adult-like granulation tissue in fetal woundsmay be responsible for the lack of fetal wound contraction . In addition, there may be other contributingfactors that further explain the inability of fetalwounds to contract . Our study was designed to explorethree additional hypotheses: (1) fetal wounds docontract, but contraction is notapparent because oftherapid growth of the fetus in utero during the studyperiod ; (2) there may be an intrinsic defect in the

15

16

KRUMMEL ET AL .

Figure 1 Fetal rabbit wounded at 24 days with two parallelparavertebral linear incisions, harvested after 5 days . Unsutured (A)and partially sutured (B) incisions have gaped open, with no grossevidence of wound contraction .

development of contractile forces by fetal fibroblasts;(3) amniotic fluid, which continuously bathes thewounds, may contain factors that inhibit contraction.These three hypotheses were tested sequentially .

MATERIALS AND METHODS

Fetal surgical techniquesTime-dated pregnant New Zealand White rabbits werepurchased (Hazelton Research Products, Denver, Pa.)and housed for 3 to 5 days before surgery. Fetalwounding was done on day 24 of gestation (term = 31days) with anesthetic techniques and operative princi-ples of Adzick and Harrison." After the fetuses werereturned to the amnionic sac, any lost amniotic fluidwas replaced with Plasmalyte solution . Fetuses wereharvested by repeat cesarean section 1 hour to 6 daysafter wounding . After assurance of viability, fetusesanddoes were killed with an overdose of pentobarbital .Amniotic fluid was harvested from gravid rabbit uteriat 28 and 31 days of gestation. Amniotic fluids werepooled and centrifuged at 300g for 10 minutes toremove cells and debris . The cleared amniotic fluid wasconcentrated fivefold by vacuum dialysis, filter steril-ized, and stored at 4° C.

Fetal wounding. Paired paravertebral incisionswere made on the backs of 10 fetuses to analyze thelack of contraction of open wounds. One incision ofeach pair was sutured closed ; the other wound was notsutured. The wounds were examined after 5 days .

Full-thickness 4 mm circular punch wounds weremade on the backs of 10 fetal rabbits from two does,

WOUNDREPAIR ANDREGENERATIONJANUARY-MARCH 1993

and 4 mm circular India ink tattoo patterns wereplaced adjacent to the punch wounds to measurecorresponding skin growth to test the hypothesis thatrapid fetal growth masks actual wound contraction.Fetal wound expansion was determined by comparisonof the planimetrically measured area of the tattoopatterns with the areas of the circular punch woundsfrom 1 hour to 6 days after wounding . To ensure thatthe exact size of the excised skin was retained after itsremoval for analysis, a.2 x 2 cm square was marked onthe fetal skin before excision. The square of skin wasthen placed on a flat surface and adjusted to the 2 x2 cm markings before planimetry.

Assurance of the humane and ethical treat-ment of animals. All animals that were involved inthis research study received humane care, and thestudy protocols were in compliance with the VirginiaCommonwealth University Institutional Animal Careand Use Committee (Protocol No. 9010-1312) . Theseguidelines meet the National Research Council's crite-ria for humane care as outlined in "Guide for the Careand Use of Laboratory Animals" prepared by theInstitute of Laboratory Animal Resources and pub-lished by the National Institutes of Health (NIHPublication No. 86-23, Revised 1985) .

Cell culture techniquesFetal and adult rabbit fibroblasts were established intissue culture by means of outgrowth from dermalexplants . These primary cell lines were maintained inDulbecco's modification of Eagle's medium (DMEM),which was supplemented with 10% fetal bovine serum(FBS), penicillin, and streptomycin (100 wg/ml each),and amphotericin B (Fungizone) (0.25 wg/ml) . Onlyfirst- or second-passage fibroblasts were used in thesestudies.

Adenosine triphosphate-induced cellcontraction studiesAdenosine triphosphate (ATP)-induced cell contrac-tion of glycerol-permeabilized monolayer cell prepara-tion was measured to ascertain whether differencesexisted between the cytoplasmic contractile units offetal and adult fibroblasts." One 22 x 22 mm glasscoverslip wasplaced in a35mm Petri dish with 5 x 103fibroblasts in 2 ml of DMEM with 10% FBS andgentamicin, 10 wg/ml. Four dishes had fetal fibroblasts,and four had adult fibroblasts. The dishes were incu-bated for 24 hours. The culture medium was removed,1 ml of 50% glycerol in 50 mmol/L of KCI, 5 mmol/LMgCl,, 10 mmol/L Tris-HCl (pH 7.5) buffer wasadded, and dishes were incubated for 30 minutes at

WOUND REPAIR AND REGENERATIONVOL. 1, N0. 1

Figure 2 Rapid growth of fetus during 6-day gestation of wound study . Tattoo patterns were placed adjacent to4 mm punch wounds to determine whether observed noncontraction was artifact of rapid fetal growth .

room temperature. Separate 30-minute washes with25%, 10%, and 5% glycerol in the same buffer followedthe initial incubation .

To measure cell contraction the 5% glycerol wasremoved, and 1 ml of 1 mmol/L ATP in 5 mmol/LMgCl2, 30 mmol/L KCI, 0.1 mmol/L CaC12 , and 10mmol/L Tris-HCI (pH 7 .0) was added to each dish . 13,14Immediately after the addition of ATP, the ATPcontraction buffer mixture was removed from twodishes with adult cells andtwowith fetal cells, and cellswere fixed in 2 ml of 4% paraformaldehyde . In theremaining dishes, theATP contraction buffer mixturewas replaced with fixative after 10 minutes of incuba-tion . The length of each fibroblast was measuredbefore incubation and at 10 minutes after the additionof ATP with a micrometer secured in the eyepiece ofa Zeiss IM-35 inverted microscope with phase con-trast and fluorescent optics (Carl Zeiss Inc., Thorn-wood, N.Y.) .

Coverslips were stained with rhodamine phalloidinand viewed with fluorescent optics to evaluate thechanges in F-actin in microfilaments of fixed cellpreparations." Rhodamine phalloidin, which was sup-plied by Molecular Probes, Eugene, Ore., wasdiluted toa ratio of 200 :1 with phosphate-buffered saline solu-tion . Coverslips were rinsed with phosphate-bufferedsaline solution after fixation for 5 minutes withbuffered 4% paraformaldehyde . Thediluted rhodaminephalloidin (0.2 ml) was added to each coverslip, andthey were incubated for 30 minutes at room tempera-ture in a moist atmosphere . The coverslips were rinsed

KRUMMEL ET AL .

7%

Figure 3 Fetal wound areas along a time course compared withtattooed controls . At harvest, all wounds were significantly larger(p < 0.01) than tattooed controls (n = 8 or greater) .

three times with phosphate-buffered saline solution,mounted, and viewed by epifluorescence .

Fibroblast-populated collagen lattice . Toevaluate whether fetal fibroblasts have the capacity toproduce contractile forces such as those necessary forwound contraction, we used an in vitro model ofwoundcontraction, the contraction of fibroblast-populatedcollagen lattice (FPCL), and compared fetal cells withadult cells.16,17 For the construction of the FPCL,8.5 x 104 fibroblasts were rapidly mixed in 1 .5 ml ofmedium with 0.5 ml ofcollagen from rat tails (5 mg/ml)in a 35 mm Petri dish . The dish was placed in a 37° Cincubator with a mixture of 95% air and 5% carbon

18

KRUMMEL ETAL.

dioxide gas in a water-saturated atmosphere . Thecollagen polymerized within 90 seconds, trapping cellsin the rapidly forming lattice .

Collagen. Rat tail tendons were excised from adultmale Sprague-Dawley rats, which weighed 300 gm.Collagen was extracted from the tendons by stirring at4° C for 2 days in acetic acid, 0.5 mol/L. The collagenwas precipitated from acetic acid by adding NaCl to thesolution to achieve 10% wt/vol . The collagen wascollected by centrifugation at 39,000g for 10 minutes,andthe pellet was resuspended in 1 mmol/L HCl. Thecollagen solution was dialyzed against 1 mmol/L HCl,with four changes for the removal of the salt . Thesalt-free collagen solution was frozen and lyophilized,and the collagen was resuspended in sterile 1 mmol/LHCl and stored at 4° C.ls

WOUND REPAIR AND REGENERATIONJANUARY-MARCH 1993

Figure 4 ATP-induced fibroblast contraction . Fetal (A) and adult (B) fibroblasts derived from dermis were plated onglass coverslips and permeabilized with glycerol . Fetal (C) and adult (D) fibroblasts were greatly contracted andcytoplasmic actin filaments were aggregated 10 minutes after addition of ATP. (Original magnification x400.)

Fetal and adult fibroblasts were suspended inDMEM with 10% FBS and gentamicin (10 wg/ml) at1 .7 x 105 cells/ml after the release of confluent monolayer cultured fibroblasts from 150 cm' tissue cultureflasks by trypsin-Versene solution with exposure for 4minutes. The cell suspension (0.5 ml) was mixed with1 .0 ml of DMEM with 10% FBS and 0.5 ml of collagenin a 35 mm Petri dish andtransferred to the incubator.We measured the diameters of the FPCLs at 1, 2, 4 and7 days after their manufacture. The area for eachFPCL was calculated, and FPCLs with adult fibro-blasts were compared with FPCL with fetal fibroblasts .The assessment of significance of difference betweenfetal and adult FPCLs was made with Student's t test .

Amniotic fluid was concentrated fivefold by vacuumdialysis versus phosphate-buffered saline solution to

WOUND REPAIR AND REGENERATIONVOL. 1, N0. 1

Table l. ATP-induced fibroblast contraction

Fetal and adult fibroblast areas at 0 and 10 minutes after administrreduction from original surface area .

test the effect of rabbit fetal amniotic fluid on FPCLcontraction. FPCLscontained 0ml (0%), 0.004 ml (1%),0.02 ml (5%), 0.08 ml (20%), 0.2 ml (50%), or 0.4 ml(100%) of concentrated amniotic fluid. Three FPCLswith each concentration of amniotic fluid were madewith either adult fibroblasts or fetal fibroblasts . FPCLareas were calculated at 2 days, andthe percentage ofdifference between those with the addition of amnioticfluid and controls wasreported . Statistical significancebetween controls and supplemented FPCLs was as-sessed with Student's t test.

RESULTS

Fibroblast Source

Adult

Fetal

In vivo studiesAfter 5 days in utero, linear incisions that are fullysutured heal completely, as seen by histologic analysisand gross observation (data not shown) . In contrast,the unsutured wounds (Fig . 1, A) and even partiallysutured wounds (Fig. 1, B) gaped open and showed noevidence of wound closure by wound contraction. Wetested the hypothesis that the absence of fetal wound

masks actual wound contraction, by comparing thepunchwounds with uninjured identical tattoo patterns(Fig. 2) . The area of a fetal wound at 1 hour afterwounding was larger than the control tattoo pattern(Fig . 3) . The fetal wound gaped open at each observa-tion point, demonstrating a significant (p < 0.01 byStudent's t test) increase in the area of the punchwound compared with that of the control tattoopattern. There was arelative decline in the percentageofincrease of wound size by the sixth day. Thus, fetalwounds not only fail to contract, but actually expandbeyond the area accounted for by rapid fetal growth .

ATP-Induced fibroblast contraction studiesFetal and adult rabbit fibroblasts were tested forATP-induced cell contraction. The structure of actinfilaments was evaluated by -rhodamine-phalloidinstaining (Fig. 4) . Both fetal andadult rabbit fibroblasts

KRUMMEL ET AL .

19

ofATP to culture media, Percentage of contraction is expressed as percentage of

exhibited identical staining patterns of cytoplasmicactin filaments. After the addition of ATP, fetalfibroblasts contracted by 70%, to 30% of their initialsize. Adult fibroblasts contracted by 81%, to 19% oftheir initial size (Table 1) . Although adult fibroblastswere more contractile on days 2 and 4, compared withfetal fibroblasts, this difference was not observed ondays 1 and 7 and probably does not account for theabsence of fetal wound contraction in utero.

FPCL studiesWe prepared FPCL, as described in the Materials andMethods section, to examine whether fetal fibroblastscan organize collagen fibers. The fetal fibroblasts werebetter at producing lattice contraction than were adultfibroblasts (Fig. 5) . The fetal fibroblasts reduced thesize of FPCL significantly (p < 0.01) better than didadult cells.

We explored a soluble inhibitor of wound contrac-tion, possibly found in rabbit amniotic fluid, with thisFPCL model. Amniotic fluid in concentrations thatranged from 1% to 100% were tested in the FPCLmodel. Increasing the concentration of amniotic fluidto more than 20% resulted in a significant inhibition ofFPCL contraction by fetal fibroblasts (Fig . 6) . Therewas also a reduction of FPCL contraction by adultfibroblasts in the presence of high concentrations ofamniotic fluid, butthe differences were not statisticallysignificant compared with controls . The inhibitor didnot cause cell death. Morphologically, fibroblasts inFPCL with added amniotic fluid were elongated andspread on the culture plate. Therewas no morphologicevidence of cell death or altered morphology by theinclusion of amniotic fluid in the FPCL.

DISCUSSIONThe process of wound contraction has been studied inmany animal species, and its importance from ateleologic pointofview is obvious. By themechanismofcontraction, wounds involving tissue loss or nonappo-sition of wound edges are healed with the restoration

Minutes Area (Run 2) Contraction

0 13.76 ± 3.7 wm2 (n = 26)10 2.58 ± 0.2 wm2 (n = 20) 81%0 24.17 ± 5.2 ~tm2 (n = 20)

10 7.4 ± 2.6 Wm2 (n = 28) 70%

ZO

KRUMMEL ET AL .

Figure 5 Collagen lattice contraction byfetal and adult fibroblasts .After 1 day, fetal fibroblasts had contracted lattice to a significantlygreater degree than had adult fibroblasts (p < 0.01). Each timepoint represents measurement of three FPCLs.

of prewound integrity, usually with minimal scarringand subsequent restoration of normal function . Al-though the contractile process is not completely un-derstood, it is generally believed that granulationtissue is formed," and then resident cells producedforces 19,2 o that cause wound contraction to take place.With experimental evidence derived from FPCL stud-ies, we concluded that fibroblasts with a cellularlocomotion mechanism that acts on the connectivetissue collagen fibers appear responsible for woundcontraction." It has been suggested that a cell-contraction mechanism that involves a specialized cell,the myofibroblast, which undergoes cell contraction, isresponsible for wound contraction in vivo . ls,2o How-ever, that concept has not been supported experimen-tally with the FPCL model." ," The supposition thewound contraction does not occur in the fetal rabbithas been reported by others' and has been demon-strated in our study. However, the closure ofopen fetalrabbit wounds by wound contraction has been reportedby Somasunduram and Prathap' after experimentalmanipulations.' They observed that amniotic fluidcould inhibit wound contraction and that the processoccurred after the open wounds were covered by asilicone elastomer patch to avoid contact with amnioticfluid.' We have confirmed these observations and haveruled out the possibility that the protective patchcauses wound splinting (unpublished observations).

Three approaches were taken to analyze these inutero observations . In the first study, skin punchwound areas were compared with tattoo patterns atvarious time points, and it was concluded that rapid

WOUND REPAIR AND REGENERATIONJANUARY-MARCH 1993

Figure 6 Dose-dependent inhibition of FPCL contraction by inclu-sion ofamniotic fluid . Atconcentrations ofamniotic fluid greater than20%, there was a significant inhibition of lattice contraction . Resultsare expressed as percentage of contraction relative to control (0%amniotic fluid) (n = 3) .

fetal growth did not explain the lack of fetal woundcontraction. The second study addressed the possibilitythat a defect existed in the ability of the fibroblast togenerate contractile forces . When low-passage fetal andadult dermal-derived fibroblasts were analyzed, it wasclear that fetal fibroblasts, like those ofthe adult, werefully capable of cell contraction.Adefect in the ability offetal fibroblasts to organize

and contract a collagen matrix was then investigated .The results of the FPCL studies demonstrate that fetalrabbit fibroblasts like those of the adult, are capable ofcontracting a collagen lattice .

The third area of investigation addressed thepossible role of amniotic fluid as an inhibitor of FPCLcontraction. Amniotic fluid, at concentrations of lessthan 20%, produced no change in lattice contraction,whereas concentrations greater than 20% inhibitedlattice contraction. Although other factors in the fetussuch as low oxygen tension, a lack of inflammatorymediators, andreduced levels ofpeptide growth factorsmay have some importance, it is clear that with theFPCL model, amniotic fluid exhibited a dose-dependent inhibition of both fetal and adult rabbitfibroblast-populated lattice contraction . There was noevidence to suggest that amniotic fluid was toxic tofibroblasts because the cells maintained viability in itspresence . The active component in the amniotic fluidwas retained during vacuum dialysis and would beexpected to have a molecular size greater than 10,000d. Therefore, in addition to the fact that fetal woundsfail to contract because of an apparent lack of granu-lation tissue, there also appears to be an inhibitor

WOUND REPAIRAND REGENERATIONVOL. 1, N0 . 1

present in amniotic fluid that may further contributeto the inhibition of wound contraction .

In contrast, amnionic fluid from sheep, an animalmodel in which fetal wounds do contract,l° has beenreported to enhance FPCL contraction ."

ACKNOWLEDGMENTSSupported by an American College of Surgeons Fel-lowship and National Institutes of Health grants GM41343, GM 47566, and GM 32705.

REFERENCES1. Mast BA, Diegelmann RF, Krummel TM, Cohen IK. Scarless

wound healing in the mammalian fetus. Surg Gynecol Obstet1992;174:441-51.

2. Longaker MT,Adzick NS . The biology of fetal woundhealing: areview . Plast Reconstr Surg 1991;87:788-98.

3. LongakerMT, Chiu ES, Adzick NS, SternM, Harrison MR,SternR. Studies in fetal wound healing. V. A prolonged presence ofhyaluronic acid characterizes fetal wound fluid. Ann Surg1991;213:292-6 .

4. Siebert JW, Burd AR, McCarthy JG, Weinzweig J, Ehrlich HP .Fetal wound healing: a biochemical study of scarless healing.Plast Reconstr Surg 1990;85:495-502 .

5. Rowlatt U. Intrauterine wound healing in a 20 week humanfetus . Virchows Arch A [Pathol Anat] 1979;381:353-61.

6. Sopher D. A study of wound healing in the fetal tissues of thecynomolgus monkey. Lab Anim 1975 ;6 :327-35.

7. Somasunduram K, Prathap K. Intrauterine healing of skinwounds in rabbit fetuses. J Pathol 1970;100:81-6.

8. Somasunduram K, Prathap K. The effects of exclusion ofamniotic fluid on intra-uterine healing ofskin wounds in rabbitfoetuses . J Pathol 1972;107:127-30.

9. Krummel TM, Nelson JM, Diegelmann RF, Lindblad WJ,Salzberg AM,Greenfield LJ, Cohen IK . Fetal response to injuryin the rabbit. J Pediatr Surg 1987;22:640-4 .

10. Burrington JD . Wound healing in the fetal lamb . J Pediatr Surg1971 ;6 :523-8 .

KRUMMEL ET AL.

2'I

11 . Adzick NS, Harrison MR. Surgical techniques in the fetal rabbit.In : Nathanielsz P, editor . Animal models in fetal medicine.Amsterdam: Elsevier North-Holland, 1980:67-103.

12 . Ehrlich HP, Rajaratnam JBM, Griswold TR . ATP-induced cellcontraction in dermal fibroblasts: effects of cAMP and myosinlight-chain kinase . J Cell Physiol 1986;128:223-30.

13 . Goldman RD, SchlossJA, Starger JM . Organizational changes ofactin like microfilaments during animal cell movement . CellMotil Cytoskeleton . 1976;A:217-45.

14 . Kreiss TE, Birchmeier W. Stress fiber sarcomeres of fibroblastsare contractile. Cell 1980;22:555-61.

15. Barak LS, Yocum RR, Nothnasel EA, Webb WW. Fluorescencestaining of actin cytoskeleton in living cells with 7-nitrobenz-2-oxa-1,3-diazole phallicidin. Proc Natl Acad Sci U S A 1980;77:980-4.

16. Bell E, Ivarsson B, Merrill C. Production of a tissue-likestructure by contraction of collagen lattices by human fibro-blasts of different proliferative potential in vitro. Cell BiolToxical 1979;76:1274-8.

17. Ehrlich HP, Wyler DJ. Fibroblast contraction ofcollagen latticesin vitro: inhibition by chronic inflammatory cell mediators.J Cell Physiol 1983;116:345-51.

18 . Watts GT, Grillo HC, Gross J. Studies in wound healing. 11. Therole of granulation tissue in contraction. Ann Surg 1958;148 :153-60 .

19 . Gabbiani G. The role of contractile proteins in wound healingand fibro-contractive diseases . Methods Achiev Exp Pathol1979;9:187-206 .

20. Gabbiani G, Hirschel BJ, Ryan GB, Statkou PR, Msjno G.Granulation tissue as a contractile organ: a study of structureand function . J Exp Med 1972;135:719-34.

21 . Ehrlich HP . The role of connective tissue matrix in woundhealing. Prog Clin Biol Res 1988;266:243-58.

22 . Ehrlich HP, Rajaratnam JBM. Cell locomotion forces versus cellcontraction forces for collagen lattice contraction : an in vitromodel of wound contraction. Tissue Cell 1990;22:407-17.

23 . Rittenberg T, Longaker MT, Adzick NS, Ehrlich HP . Sheepammonic fluid has a protein factor which stimulates humanfibroblast populated collagen lattice contraction. J Cell Physiol1991;149:444-50.