SCARLESS WOUND HEALING IN THE MAMMALIAN FETUS

Bruce A. Mast, M.D., Robert F. Diegelmann, PH.D., Richmond, Virginia,

Thomas M. Krummel, M.D., Hershey, Pennsylvania, and l. Kelman Cohen, M.D.,

Richmond, Virginia

From the Division of Plastic and Reconstructive Surgery and theWound Healing Center, Department of Surgery, Medical Collegeof Virginia, Richmond and the Division of Pediatric Surgery, De-partment of Surgery, the Milton S. Hershey Medical Center, HersheyMedical School, Hershey, Pennsylvania.

Some studies contained in this review have been supported byNIH Cranto; CM 41343 and CM 20298.

Reprint Requesto;: Dr. l. Kelman Cohen, Medical College of Vir-ginia, Division of Pla.o;tic and Reconstructive Surgery, Box 154, MCVStation, Richmond, Virginia 23298-0154.

SURGEONS are often required to manage patientswith atholo ic processes in ic bnormalhealine: occ~ uc conditions exist in virtuallyall surgical specialties and include inadequate tis-sue repair that causes dehiscence or chronic, non-healing oren wounds, cutaneous keloids and hy-pertrophic scars, adhesions limiting function aftertendon repair, failed nerve repair because of scar,biliary and intestinal tract strictures, intra-abdom-inal adhesions and visceral scarring, such as cir-rhosis and atherosclerotic vascular disease. AlIthese conditions are associated with significantmorbidity and mortality rates, or both, and areoften difficult and challenging to treat. A searchfor' improved management of these conditionsis a continuous endeavor of investigators ofwound healing and connective tissue metabolismo

Tissue repair in fue mammalian fetus is fun-damentally different than normal adult healing.In adult humans, injured tissue is repaired bycollagen deposition, collagen remodeling andeventual scar formation, whereas fetal woundhealing seems to be more of a regenerative pro-cess with minimal or no scar formation. There-fore, fetal healing provides an accessible mam-malian animal model of an optimal healingresponse. Understanding the mechanisms of suchscarless healing mar provide new treatment mo-dalities by which therapy of many of the afore-mentioned conditions can be greatly improved.Additionally, understanding fetal healing mar beinvaluable to the clinician as fue new age ofhuman fetal surgery is being entered. Currentknowledge concerning the remarkable process ofmammalian fetal wound healing is presented inthis study.

4.::

ASPECfS OF DEVELOPMENT THAT MAY AFFECf HEALING

The fetus is a developing organism in whichcellular migration, proliferation and differentia-tion occur at an unparalleled magnitude. Becausehealing of damaged tissue is a process of cellularinteraction, fue mechanisms underlying fetal tis-sue repair cannot be fully understood withoutan awareness of the unique physiologic and bio-chemical processes within fue fetus, as well asfue unique intrauterine environment in whichhealing occurs. Several aspects of fetal develop-ment have fue potential to affect fue cellularand extracellular processes necessary for healing.

The hormonal milieu of an organism has directeffects on fue cellular activity within various organsystems. In adults, the endocrine system is inte-grally involved in the response to traumatic injury(1). Compared with adults, the fetus contains afunctionally different endocrine ~"tP~ for ex-ample, the fetal adrenal gland is probably secondonly to fue liver in size. The majority of fuegland is composed of a "fetal zone" that quicklyinvolutes after birth. Fetal adrenal steroid secre-tion is enormous, approximately ten times thatof fue nonstressed adult (2). It is quite possiblethat fue different state of hormonal flux in fuefetus mar affect the processes of tissue repair.

The i!!!:mune sys~m and its associated inflam-matory mealators are intimately involved in adulttissue repair. The fetus is immunologically im-mature and less competent than fue adulto Neu-tropenia and a lack of antigenic stimulation marcontribute to fue "\IpIJ1~t\!,.rilY:' (3, 4). Humoralimmunity is comprised mostIy of immunoglobulinG obtained passively from the maternal cir-culation, but fue fetus is capable of producingimmunoglobulin M when provoked (2). The com-plement system is algO underdeveloped. Compo-nents are present in reduced quantities comparedwith adults (5, 6). It is possible that fue immu-nologic differences in the fetus mar relate tothe differences in wound healing observed inthe fetal system.

~~ is essential for normal adult woundhealing. The ~e~u~ ls J1w2xicz however, and has

~l

442 SURGERY, cynecology & Obstetrics. May 1992

an oxygen partial pressure of approximately 20torr, reflecting an oxygen partial pressure of 30to 35 torr of maternal blood in the placentalinteIVillous spaces (2). Obviously, this factor haspotencial effects on tissue repair, although thefetus seems to compensate fairly weli-; the lacticacid level (an indicator of anaerobic glycolysis)is only slightly higher in umbilical cord blood,compared with maternal blood (2).

The external fetal environment consists of theplacenta and amniotic fluido Fetal suIVival is de-pendent on a symbiosis between mother and fetusthrough the placenta. As a transfer organ theplacenta acts as surrogate lungs, kidneys and in-testinal tract for the fetus (7). Functional im-pairment of the placenta, such as decreased utera-placental blood flow as encountered in maternalhypertension, mar directly impair fetal functions.This includes healing.

~niotic fluid provide~ a sterile- wei~bt1ess~n;vironment that is rotective and thermall stable.Factors Wlt in amniotic fluid can locally ectthe healing process. Systemic effects of amnioticfluid are algO possible because the fetus swallowsthe fluid (8), thus allowing the absorption of avariety of potential mediators. In fact, absorptionof various nutrients has be en demonstrated inisolated intestines of fetal rabbits (9). An effecton fetal cellular function has been demonstratedby the observation that amniotic fluid has a tra-phic effect in supporting fetal gastric cells inculture (10, 11). Obviously, such interactions maraffect tissue repair.

OVERVIEW OF ADULT (POSTNATAL) WOUND HEAIlNG

Normal healing is essentially the same at allpostnatal ages. Because the adult repair responseis the norm, it is essential to be familiar withthe biologic features of normal postnatal healingbefore discussion of fetal repair. Adult healingoccurs by a sequence of events in which cellularand matrix compOI)ents act in concert to repaira tissue defecto Based on Schilling's categoriza-tion, the healing response can be described infour broad phases-hemostasis, inflammation,proliferation and remodeling (12). At the timeof injury, hemostatic mechanisms, which includevasoconstriction, platelet aggregation and fibrindeposition, are initiated to control local hemor-rhage. The initial wound matrix of fibrin actsas a scaffold on which inflammatory cells enterthe wound (12, 13). Platelet alpha degranulationresults in the release of platelet derived growthfactor (PDGF) (14, 15) and transforming growth

r

¡,¡

~.

Volume 174

factor beta (TGF-~) (16), which are chemoat-tractants, mitogens and stimulators of collagendeposition for cells that will enter fue wound.

Acute inflammation is established 12 to 24hours postoperatively as vasoconstriction is fol-lowed by vasodilation and increased capillary per-meability. This results in fue influx of polymor-phonuclear leukocytes (neutrophils) into fuewound site. The neutrophils control bacterial con-tamination, but are probably not necessary forhealing, because neutropenic patients and ani-malsusually heal normally (17). Neutrophils aregradually replaced by macrophages and then lym-phocytes, as a chronic inflarnmatory infiltrate isestablished. Macrophages are thought to be es-sential because macrophage depletion inhibitsnormal fibroplasia (18). Similar to platelets, mac-rophages secrete various growth factors, such asPDGF and TGF-~, which enhance fibroplasia (18,19). The redundancy of fuese factors further stim-ulates fue influx of fibroblasts into fue woundwhere they are stimulated to proliferate and pro-duce collagen (16). In addition to fue enhancedfibroblast response, epithelialization occurs andneovascularization is prominent as endothelialcells proliferate. As fibroblast activity continues,the inicial fibrin matrix is replaced by collagen.

Remodeling of fue matrix occurs years afterinjury because there is continued synthesis anddegradation of collagen. The resultant collage-nous scar represents a fine balance between thetwo processes. Crosslinking and structural mod-ifications of collagen fibrils are largely responsiblefor fue increase in tensile strength of fue healingtissue during this time (20-22). The final resultof adult healing is a replacement ornormal tissueby a collagenous scar, which lacks fue orderedstructure of normal skin and is an imperfectprocess as exemplified by fue observation thathealed wounds never regain the full tensilestrength of uninjured tissue (20).

A distinct feature of postnatal open woundsis fue ability to concentrically decrease fue areaof lost tissue by fue process of contraction. Theforces generated during contraction are thoughtto arise from myofibroblasts within fue granula-tion tissue of fue wound. The specialized fibro-blasts contain fue contractile protein actin (23).Evidence algO exists that direct fibroblast-collagenmatrix interaction mar cause contraction of openwounds (24). By reducing fue area of tissue de-fect, contraction facilitates repair of open woundsin which tissue loss is a prominent feature.

Mast el al. SCARLESS WOUND HEALING IN MAMMALIAN FETUS 443

MODELS OF FETAL HEALING

Monkeys (25), rats (26-28), sheep (29) andrabbits (4, 30-32) have be en used in investiga-tions of fetal healing. The latter two have beenused most frequentIy. The South American opos-sum has also been studied in an attempt to gaineasier access to the fetus that continues to developon fue exterior of fue mother (33, 34). Theuse of an animal model is imperative becauseof unique physiologic features and environmentof fue fetus that would be nearly impossible toreplicate in a culture dish or other in vitro systems.Generally, fue procedures for studies of fetal heal-ing involve a maternal laparotomy during whicha hysterotomy is created, thus providing accessto fue fetus. Wounding takes place while fuefetus remains within fue uterus. Most commonly,wounds that are incisional and excisional are cre-ated, although burning and freezing have alsobeen performed. The hysterotomy and maternalincision are then closed. U sually, cesarean sectionis performed at specified times postoperativelyso that fetal wound tissue can be sampled beforeparturition and consequent violation of intrauter-ine conditions (35).

Healing wounds are evaluated by histologic andbiochemical analysis, or both, of wound tissue.There are many studies in which excisional biopsyspecimens of primarily closed wounds have beenanalyzed. Although useful data are obtained inthis manner, limitations arise in fue biopsy ofprimarily closed skin wounds because of fue in-ability to accurately identify the boundary be-tween healing and uninjured tissue. Additionally,"contamination" ofspecimenswith ~rmal tissuerenders biochemical assays less speciñc. Becauseof fue limitations, as well as fue relative paucityor difficulty in obtaining fetal wóund tissue,wound implant devices have be en used, includingthe Hunt-Schilling chamber, fue viscose cellulosesponge, Goodson's and Hunt's expanded poly-tetrafluoroethylene (Gore-tex TM) tube and fue poly-vinyl alcohol (PV A) implant (36). The ideal im-plant should be as inert as possible within fuewounds, so that observed cellu1ar and matrixevents closely correlate with those of wound bi-opsies.

seven days (32). Scar formation has been notedto be grossly and microscopically absent in bothexperimentally created wounds (29, 32, 37) andsurgical wounds after in utero repair of cleft lipsin mice (38), monkeys (39) and rabbits (40).Evaluation of the processes responsible for thesephenomena can be descriptively divided into cel-lular and extracellular matrix events.

Cellular events in fetal closed wounds. A strikingdifference between postnatal and fetal repair isthe absence of acute inflammation in fetalwounds of several species. Only a few neutrophilswere noted within incised fetal sheep wounds(29). Similar observations were made in the mon-key (25), while an absence of acute inflarnmationwas reported in fetal rat wounds (41). Tissuesarnpling with a Variety of implants has confirmedthe absence of acute inflarnmation within woundsof fetal rabbits (4, 32, 42).

Several hypotheses could explain the lack ofacute inflarnmation in fetal wounds. The sterile,intrauterine environment mar contribute by fail-ing to provide antigenic stimulation for polymor-phonuclear infiltration (25). However, "sterile"wounds in germ-free adult guinea pigs demon-strated acute inflarnmation, although in lesserquantities than in control animals (43). This sug-gests that bacterial contamination mar not bethe sale initiator of acute inflarnmation in adultwounds. Immunologic immaturity and neutro-penia have also been implicated as being causative(4, 29), indicating that the fetus is unable tomount an inflarnmatory response despite provo-carian. In an experimental model for biliary atre-sia, acute inflarnmation within the liver resultedafter injection of periportal sodium morrhuatein fetal dogs (44). Thus, it seems that the un-derdeveloped fetal immune system mar be ca-rabIe of an adult-like acute inflammatory re-sponse after stimulation.

Similar to differences in acute inflammationin fetal and postnatal healing, the chronic in-flammatory response also differs. Although mac-rophages have been identified in fetal rabbitwounds by nonspecific esterase staining (4) anda slight in crease in macrophages has be en de-scribed in fetal sheep wounds (29), lymphocyteshave been rarely encountered in fetal wounds.Mesenchymal cell infiltration in fetal sheepwounds has been temporally associated with theappearance of macrophages (29) and likewise,mesenchymal cells have been noted in woundsof fetal rabbits (32). Mononuclear cells have alsobe en repeatedly observed in fetal rabbit wounds

CHARACTERISllCS OF THE FETAL HEALING RESPONSE

Closed linear wounds. The most striking featuresof primarily closed, linear wounds are the rapiditywith which they heal and the absence of obviousscarring. Mter injury, linear wounds in fetal rab-bits arrear macroscopically healed by five to

.

.H

444 SURGERY, Gynecology & Obstetncs. May 199;

using PVA implants, but fue specific identity isstill unknown (32).

The hallmark of fue proliferative phase in adulthealing is fibroplasia. Results of studies of dif-ferent fetal species suggest that there is inter-species variation in fue fetal wound fibroblasticresponse, although fue response seems to be sig-nificantly reduced when compare~ with adultwounds. Biopsies of healed fetal sheep woundswere observed to contain a mild fibroblastic in-filtration (29). Similarly, fetal rabbit wounds thatwere evaluated by simple biopsy (45) or PVAimplant histologic features (32) contained onlya scant presence of fibroblasts. However, whenGore-tex implants were used as fue means oftissue sampling in fetal rabbit wounds, fibroblasticencapsulation of fue implant was observed (4).The biomaterial differences between Gore-texand polyvinyl alcohol could account for thesevaried observations.

Epithelialization is also a prominent prolifer-ative response in postnatal wound healing. Thisprocess occurs quite rapidly in primarily closedfetal wounds. Within 72 hours of injury, fetalrat wounds were completely epithelialized withan epidermis that was similar in structure andthickness compared with that of adjacent unin-jured skin (41). The epidermis of fetal rat cheekwounds was considered completely normal withthe presence of rete pegs after 24 hours of heal-ing (27). Healed wounds in fetal rabbits werecovered with epithelium that appeared to havenormal cell density, depth and cell structure (37).Wounds in fetal sheep seemed grossly healed bythree days, including an epithelial covering, al-though the epithelium was slightly thickenedcompared with normal skin. As in adults, fuecrypts of hair follicles adjacent to fue woundsseemed to be fue origin of fue rapid epithelializa-tion (29). Fetal maturation seems to slow epitheli-alization in fetal sheep, although fue fetal rate(48 to 72 hours) remained significantly morerapid than adult sheep, which required sevendays for complete wound epithelialization (46).

Neovascularization and its associated endothe-lial proliferation is a conspicuous aspect of heal-ing tissue. Most studies have found that adult-likeangiogenesis is absent in primarily closed, linearfetal wounds. Angiogenesis is absent in woundbiopsies from fetal sheep (29), monkeys (25),rabbits (45) and rats (26, 27), as well as PV Aimplants removed from wounds in fetal rabbits(32). However, as encountered with observationsconcerning fibroplasia, apparently conflicting

2.

Volume 174

data exist from studies in which Gore-tex (4)or viscose cellulose (31, 42) implants were usedin fetal rabbits. In those experiments, granulatioTltissue containing capillary loop formation wasseen infiltrating the. wound devices.

Extracellular matrix of the fetal wound. Healingadult wounds have matrices of fibrin, fibronectin,proteoglycans and collagen. The healed adultwound matrix is composed almost entirely of col-lagen. Excessive, disorganized collagen depositionreplaces the normal dermis in wounds as a maturescar is formed. The absence of appreciable scar-ring in fetal wounds suggests that marked dif-ferences exist in the fetal wound matrix, com-pared with the adult matrix. Histologically, fetalsheep wounds arrear to be lacking an abundanceof collagen (29) and wounds through muscle infetal rats also are lacking collagen (26). PVAimplants removed from wounds of mid-third tri-mester fetal rabbits after five to seven days ofhealing contained no infiltrating collagen by his-tologic examination using Masson's trichromestaining, whereas adult rabbit wound implantscontained abundant collagen. High pressure liq-uid chromatography with a sensitivity to the levelof 40 picamoles, also failed to detect collagenhydroxyproline within the fetal implants (32).The extracellular matrix of the fetal wound tissuewithin the implants, however, stained intenselywith alcian blue, which was eliminated after pre-treatment with hyaluronidase. This suggests thatthe glycosaminoglycan (GAG) hyaluronic acid(HA) was the predominant noncollagenous com-ponent of the matrix of wounds in fetal rabbittissue (32).-;)

Further ~haracterization of the GAG presentwithin the fetal wound matrix was achieved afterbiochemical extraction of the GAG within PV Aimplants removed from wound sites in fetal rab-bits (47). It was shown that there was a progressiveaccumulation of GAG within fetal wounds andconfirmation was provided that HA was the pre-dominant GAG. Fetal wounds contained signifi-cantly more GAG conipared with adult wounds,and HA predominated as the sale GAG in fetalwounds for a longer period (47). Subsequently,HA was detected in wound fluid from fetal sheepas long as three weeks after wounding, whereasHA could not be detected in adult sheep woundfluid at seven days after injury (48). The resultsof these studies demonstrate that a striking fea-ture of the fetal wound matrix is a persistentabundance of HA.

Clearly, histologic examination shows that col-

Mast el al... SCARLESS WOUND HEAilNG IN MAMMALIAN FETUS 445

lagen is present in fetal wounds, although infil-tration into PV A implants does not seem to occur.Therefore, fue absence of obvious fetal scarringmar algO be attributed to a highly organized de-position of collagen, as well as a lack of excessivecollagen deposition. Light microscopic examina-tion of healed wounds in fetal rabbits showedthat fue dermal matrix contained a reticular col-lagenous pattem (32, 37), rather than bundlesof collagen and fibroblasts as seen in adultwounds. Type-specific collagen immunohisto-chemical staining of wounds from fetal sheepin fue second and early part of fue third trimesterhas shown that 14 days after wounding, fue con-formation of dermal collagen within fue woundswas similar to that of unwounded skin, in otherwords, there was no formation of scar. Interest-ingly, as fetal maturation occurred, a transitionto postnatal healing became apparent. Collagendeposition became more adult-like in wounds infetal sheep during fue miqdle part of fue thirdtrimester as demonstrated by fue deposition ofpacked collagen fibers resulting in minimal scarformation 14 days after injury (46).

Although collagen deposition in fetal woundsis highly organized' and nonexcessive, paradoxi-cally, collagen appears in fetal wounds more rap-idly than in adult wounds. Hydroxyprolineaccumulation, which is indicative of collagen pro-duction, is detectable earlier in wounds in fetalrabbits than in adult wounds (4). Similarly, inpolyvinyl alcohol sponges, hydroxyproline accu-mulation in wounds in fetal sheep is greater andmore rapid than in adult wounds (49). Immu-nohistochemical staining has algO demonstratedan earlier deposition of collagen in wounds infetal sheep compared with adult wounds (46).These observations corre late well with fue in vierafinding that fetal dermal fibroblasts have a greatercapacity for collagen production than do adultfibroblasts (50). The aforementioned findingssuggest that matrix formation in fue fetal woundis highly efficient and organized.

The adult wound heals by fue replacement ofnormal dermis with a scar that consists of ex-cessive and abnormally organized collagen. Inmarked contrast, fue fetal wound contains a per-sistent abundance of HA while collagen deposi-tion is rapid and nonexcessive. Additionally, itseems that collagen deposition is highly or-ganized, so that fue normal dermal structure isrestored and scarring does not occur. Fetal mat-uration seems to involve a transition to postna-

tal-like healing, but scar formation is only min-imal.

open wounds. The repair response initiated byopen fetal wounds is not as uniform as fue repairprocess in closed wounds, so that more species-related differences seem to existo In one of fueearliest studies (1958) of fetal wound healing inmarnmals, extensive wounds were created by fuetransection of fue vertebral column and all sur-rounding tissues of fetal guinea pigs. The ex-tensive wounds healed in utero, but in a postnatalmanner with rapid formation of granulation tis-sue, fibroplasia, collagen deposition and scar for-mation (51). Simple open wounds in fetal sheepalgO healed rapidly with only a scant acute in-flammatory infiltrate, followed by abundant fi-broblasts and undifferentiated mesenchymal cells.In contrast to fue guinea pig, exuberant gran-ulation tissue formation did not occur. Addition-ally, fue sheep wounds exhibited extensive con-traction and epithelialization (29). Contrary toguinea pigs and sheep, excisional wounds in fetalmonkeys (25) and rabbits (30, 32, 52-54) donot contract.

The healing of open wounds was studied infetal monkeys five to seven weeks before term(25). There was no contraction of fue woundsafter five days, although mild epidermal prolif-eration at fue wound edge was presento On post-wound dar eight, epidermal proliferation wasmarked, but epithelial migration into fue woundhad not occurred. The wound base containedill defined spindle shaped cells, but no granu-lation tissue. Mter 27 days, only parcial epitheli-alization had occurred. The tissue within fuewound was cellular and "rich in ground sub-stance," but did not have any resemblance togranulation tissue in postnatal open wounds.

Linear incisions in fetal rabbits that were notclosed primarily were observed to gape open with-out evidence of conu-action (32). Using skin tat-toas that were fue same size as fue open wounds,it was shown that fue wounds did not fail toclase as a consequence of an overpowering effectof rapid fetal growth (52). Therefore, not onlydo open wounds in fetal rabbits fail to contract,but they actually expand.

The absence of contraction of open woundsin fetal rabbits was further characterized usinglight and electron microscopy (53). Histologically,fue wounds showed no migration of epitheliumfrom fue wound edges, no acute inflammation,no fibroblastic response and minimal collagendeposition. In fue noncontracting wounds, anti-

.

446 SURGERY, Gynecology & Obstetrics. May 1992. Volume 174

actin antibody staining was present in the pan-niculus carnosus muscle, but not within thewound margino This suggests the absence of my-ofibroblasts. In addition, on transmission electronmicroscopy only a few fibroblasts were seen andno ultrastructural characteristics of myofibroblastswere identified. Therefore, noncontracting orenwounds in fetal rabbits are asso~ated with anabsence of myofibroblasts.

Noting the absence of myofibroblasts in fetaloren wounds and considering the theory of adultwound contraction centering around direct fi-broblast~matrix interaction, fetal fibroblasts wereevaluated for the ability to generate and exerta contractile force. The cell surface areas of cul-tured fetal and adult rabbit fibroblasts diminishedby 70 to 80 per cent after the addition of mediumcontaining adenosine triphosphate, indicatingthat the cells had contracted to an equal extent(52). Fibroblast populated collagen lattices(FPCL) were then established as an in vitro meansof studying fibroblast-matrix interaction. Latticescontaining fetal or adult rabbit fibroblasts con-tracted equally at 24 hours, but at two and fourdays, the fetal FPCL demonstrated a significantlygreater reduction in area. Upon addition of rabbitamniotic fluid in concentrations as low as 20per cent, contraction of fetal and adult FPCLwas significantly inhibited (52). The results ofthese studies provided the first in vitro evidencethat fetal rabbit fibroblasts are capable of gen-erating a contractile force and exerting that forceonto a collagenous matrix. Furthermore, it seemsthat the fetal environment of amniotic fluid hasan inhibitory effect on contraction.

The effect of the fetal environment on orenwound healing has also been studied in utero.Exclusion of oren excisional wounds from ex-posure to amniotic fluid in fetal rabbits was doneby coverage with a sutured silicone sheet. Whilecellular events within the wound remained un-changed compared with control wounds, the cov-ered wounds seemed to "contract" (30). In asimilar experiment in the same species, comput-erized morphometry documented a 26 per centreduction in the size of the covered wounds com-pared with control wounds (54). In both studies,the healing of oren wounds in neonatal rabbitswas algo evaluated. Excisional wounds in the ne-onates healed in a manner similar to adult openwounds with rapid contraction and formation ofgranulation tissue (30, 54). Therefore, these ex-periments provide further evidence suggestingthat amniotic fluid mar be an inhibitor of open

wound healing in the fetal rabbit and also showthat there is an abrupt transition from fetal non-closure of oren wounds to adult-like healingupon the natural removal of amniotic fluid afterparturition.

The interspecies variation that exists in orenwound healing in the fetus makes it difficult toextrapolate the findings to the human condition.So me observational data in the human fetus, how-ever, suggest that human fetal oren wounds marbehave more like liase of monkeys and rabbitsrather than sheep. A postmortem examinationof a spontaneously aborted human fetus of ap-proximately 20 weeks gestational age showed mul-tiple digit amputations, as well as an amputateddistal one-third of the right leg, presumedly fromamniotic band trauma. There was also a largefull thickness ulcer with exposed, underlyingbaile that appeared to be secondary to frictionfrom the baile of the contralateral leg stump.The leg stump and ulcer were oren, containingnecrotic tissue. There was no inflammatory re-sponse and no granulation tissue formation. Theamputated digits also contained necrotic debris.No inflammation or granulation tissue were pres-ent, although an abnormal, thin epithelial cov-ering was presento A mild mesenchymal prolif-eration was observed in so me wounds (55). Thesefindings closely resemble the observed repair pro-cesses in oren wound healing in fetal monkeysand rabbits. Therefore, compared with other spe-cies, these animals mar be a more accurate modelfor fetal oren wound healing in humans.

Bum wounds. The response initiated by elec-trical burn wounds was studied in fetal rats at16.5 through 21.5 days of gestation (term is 22days) (56). Microscopic examination of thewounds was performed up to 24 hours after in-jury. The wounds of younger gestational age fe-tuses lacked an acute inflammatory response.However, beginning at 19.5 days gestational age,acute inflammation was presento The degree ofneutrophilic infiltration was correlated with ges-tational age from that point onward. A mesen-chymal cell "filling in" was also observed in thewounds. Similar to observations in linear incisionsin fetal sheep, the older fetuses had more of apostnatal repair response within their wounds.

REGUlAnON OF FETAL REPAIR

Adult wound mediators. Polypeptide growth fac-tors are thought to be important mediators inembryogenesis and postnatal healing. Two suchfactors, TGF-¡3 and PDGF, accelerate healing in

:.

Mast el al.: SCARLESS WOUND HEAUNG IN MAMMALIAN FETUS 447

normallaboratory animals and are associated withthe initiation of fibrosis (16, 57). PDGF is achemoattractant for smooth muscle cells and fi-broblasts (14, 15) and TGF-j3 has a mitogeniceffect on fibroblasts and stimulates them to pro-duce collagen (16). Considering the absence offibroplasia and the lack of excessive collagen de-position in fetal wounds, the factors could bediminished or in active in the fetal wound. Al-ternatively, fetal cells mar become less responsiveto the factors. Indeed, the number of TGF-j3receptors on the surface of fibroblasts diminishesas fetal development proceeds, so that embryonicfibroblasts have the greatest number of receptors,fetal fibroblasts have fewer and adult fibroblastshave the least (58).

To test whether or not the fetal healing re-sponse can be altered by TGF-j3, the substancewas exogenously introduced into fetal rabbitwound sites using PVA implants (59). The TGF-j3treated wounds demonstrated an altered, adult-like healing response in which fibroplasia andcollagen deposition were markedly increasedcompared with untreated fetal wounds. A directinteraction of TGF-j3 with fetal fibroblasts wasimplied by the results of another study thatshowed cell surface binding of TGF-j3 to culturedfetal rabbit fibroblasts (58). The addition of PDGFto wounds in fetal rabbits similarly resulted ingreater fibroblast infiltration and collagen depo-sitian, as well as neovascularization (60).

The fetal repair response can by converted toadult-like healing after treatment with growth fac-tors that are considered to be vital in normalpostnatal healing. Thus, fetal cells are fully ca-rabIe of responding to the growth factors in amanner similar to that observed in adults. Theabsence of features typical of adult healing infetal wounds suggests that these growth factorsmar not be present within the wound sites orthat their activation is inhibited.

In another study, the significance of the ab-sence of neutrophils in fetal repair was evaluatedby the introduction of adult acute inflarnmatorycomponents into fetal wounds (61). PVA implantswere first placed subdermally into maternal rab-bits and then transferred into the respective fe-tuses of the does, six to 36 hours later. Mterfive days of fetal healing, the implants from thefetal wounds contained significantly more neu-trophils than present when they were removedfrom the maternal wounds. However, no fibro-blast infiltration or collagen deposition was pres-ent in the fetal implants. Therefore, despite the

apparent stimulation of neutrophilic infiltrationinto fue site of fue fetal wound by componentsremoved from an adult wound during fue phaseof acute inflammation, fibroplasia did not resulto

Therefore, fue absence of acute inflammationin fetal wounds mar not be directly related tothe lack of fibrosis.

Hyaluronic acid stimulating activity. One of fuesubstantial differences between fetal and adultwounds is fue progressive and prolonged pres-ence of HA within the fetal wound matrix. Severalinvestigations have dealt with fue hypothesis thatfetal wound cells are regulated to produce HApreferentially. The studies have used a rat fibro-sarcoma cell line that produces high quantitiesof HA as a model for fetal cells. Using fue model,it was determined that fetal sheep wound fluidsampled from subcutaneous wire mesh cylindersstimulated fue production of HA, whereas adultsheep wound fluid did not (62). Serum fromfetal sheep (63) and amniotic fluid (64), as wellas serum from fetal calves (65), were algO foundto stimulate fue production of HA by this cellline. Biochemical analyses of a factor withinserum from fetal calves that stimulates HA pro-duction has characterized it as a heat stable gly-coprotein that is inactivated at extremes of acid-ification or alkalinization (65).

The results of fuese studies suggest that a factorwithin fue fetus mar regulate fibroblasts to pro-duce HA rather than collagen. Furthermore, fuepresence of HA stimulating activity in serum andwound fluid indicates a possible systemic, as wellas local, availability. A factor containing fue stim-ulatory activity has been isolated in vitro and ap-pears to be a glycoprotein whose activity is stableat in vivo temperature and acid/base conditions.However, the actual extension of fuese findingsto the in vivo situation is unclear. In fuese studies,only HA production in a cell line that is knownto produce high quantities of this substance wasstudied. The possible production of other GAGshas not be en investigated. It is possible that agenerally increased production of several GAGsmar be observed in normal fetal fibroblasts orthat normal cells are not stimulated to increaseHA or GAG production. Nevertheless, fuese in-teresting findings warrant further study and po-tentially provide great insight into how fetal andadult wounds mar mechanistically differ.

Hyaluronic acid. Similar to fetal skin, fue fetalwound matrix is enriched in HA (66). As pre-viously stated, it is associated with a highly or-ganized healing response so that scarring is min-

448 SURGERY, Gynecology & Obstetncs' May 1992. Volume 174

by an experiment in which a full thickness skinallograft from a maternal ewe was transplantedto her fetus at 60 days of gestation (79). Repeatlaparotomy was performed after 40 days, dem-onstrating complete take of fue graft, but fuegraft retained qualities of adult skin and waswool bearing, with borders clearly distinguishablefrom the fetal skin. Primarily closed, full thicknessincisional wounds we~e then created in fetal skinand fue adult graft. The fetal wounds healedwith a collagen staining pattern that was identicalto uninjured skin, demonstrating fue absence ofscar formation. However, fue adult graft healedby gross and microscopic scar formation. Thesefindings show that adult wound healing is notchanged into a fetal repair response despite per-fusion of fue injured skin by fetal blood andexposure to amniotic fluid and suggests that fuedifferentiated state of fue cellular populationwithin fue skin is vitally important in determininghow fue tissue will respond to injury. For instance,it is possible that mature differentiated fibroblastscan only respond to tissue injury by proliferationand exuberant collagen deposition.

1I1

ti;

.1

I

~ I

APPUCATIONS OF FETAL REPAIR

There are two primary goals in elucidating fuebiologic processes responsible for scarless fetalhealing. First, fue processes of in utero repairmust be understood so that as fue fetus be comesa surgical patient (80), fue healing will be man-ageable. The second goal is to apply fue mech-anisms of scarless healing to pathologic, postnatalprocesses in an attempt to improve clinical treat-mento Such application has enormous potentialimpact because of reduction in fue significantmorbidity and mortality rates associated with con-ditions such as hypertrophic scars, keloids andintestinal strictures. Some examples of this ex-tension of fetal repair to adult healing existo Forinstance, because of fue association of HA withcellular migration and proliferation, fue effectof HA on adult wounds has be en evaluated. Top-ical application of HA has been shown to improvethe retardation in epithelialization of partialthickness linear wounds in rats that are diabetic

(81).Hyaluronic acid was also evaluated for fue effect

on scar fbrmation in full thickness, linear inci-sions in monkeys, rabbits and guinea pigs (82).In all animal groups, there seemed to be lescar formation in wounds treated with HA compared with those treated with saline solutionHowever, results in regards to improved healin

imal or nonexistent. Observations from other bi-ologic systems, including true regeneration (67-70), development (71) and neoplasia (72) haveimplicated HA as an affecter of the cellular andmatrix events within these tissues. Furthermore,HA has been demonstrated to directly bind tocells (73-76) and to complex with collagen andnoncollagen protein (77). Based upon these data,it was hypothesized that HA mar influence thecellular and matrix events, or both, in fetal heal-ing so that HA mar have a regulatory influenceon scarless repair.

This hypothesis was tested by studying the fetalhealing response in rabbits after a reduction infetal wound HA content by specific in vivo deg-radation with Streptomyces hyaluronidase intro-duced using PV A implants. A biochemically con-firmed reduction in the HA content of woundstreated with the enzyme was associated with anincrease in fibroplasia, collagen deposition andneovascularization (78). Therefore, the degrada-tion of HA within fetal wounds dramatically al-tered the healing response so that features morecharacteristic of adult healing appeared.

The results of the current study are similarto liase obtained after the treatment of fetalwounds with TGF-¡3 and PDGF. A potential elec-trostatic interaction between the polyanionic HAand the cationic growth factors is suggested, sothat a reduction in the content of HA withinthe wound could augment the effect of thegrowth factors, leading to the observed histologicresults (78). A direct physicochemical effect oncollagen organization by HA is algO possible, sothat an elevated content of HA would favor adispersed, hydrated conformation of collagen,rather than the dense, compact organization en-countered in scars (78).

The state o/ cellular differentiation. In regenerationand development, the accumulation and subse-quent degradation of HA is associated withchanges in the state of cellular differentiation.It has been proposed that HA mar similarly affectthe state of differentiation of immature cells in-volved in fetal repair. Elevated levels of HA infetal wounds mar prevent differentiation of im-mature mesenchymal cells present within the der-mis and thereby prevent fibroplasia and scar for-mation. A controlled differentiation of the cellswould allow the normal structure of the skin tobe reformed (66, 78).

Evidence supporting the importance to thehealing response of the differentiated state ofcells present within wounded skin was provided

~~~C".

,~i.."~

Mast et al.: SCARLESS WOUND HEALING IN MAMMALIAN FETUS 449

fue fetal system, and HA mar influence cellularand matrix events within the wound. As greaterknowledge of fue biologic factors of scarless heal-ing in fue fetus is gained, applications to al>-normal adult healing mar be developed. As fueage of fetal surgery is approaching, it is essentialto understand fue injury response in these newpatients.

with decreased scar formation are variable in stud-ies in which HA was used for ophthalmologicapplications (83, 84) or tendon repair (85). Fur-ther well-controlled studies are needed to evaluateany potential effect that HA mar have upon con-traction, epithelialization and collagen depositionin adult wounds.

Another approach has been to provide adultwounds with elements of fue fetal environment.The totally fluid environment in which fetal repairoccurs has be en hypothesized to be a prominentfactor in the lack of scarring and contractionof fetal wounds. To evaluate fue effect of a fluidenvironment on adult wound healing, primarilyclosed incisional wounds in pigs were continuallyexposed to saline solution by fue use of liquid-tight wound chambers (86). The wounds treatedwith fluid had less gross scarring, more normalepidermis and better cosmesis compared withcontrol subjects that were exposed to air. Therewas no difference in hydroxyproline content ofwounds from either group. The tensile strengthof the control wounds was significantly greaterthan fue wounds covered with fluid at postwounddar 14, although there was no significance inthe differences in ~trengths by dar 30. This studyindicates fue importance of fue fluid environmentin fetal repair and suggests that simple manip-ulation of adult wound conditions ¡;nay have adesired effect on healing. Further studies are re-quired to evaluate various fluid environments,such as amniotic fluid and fluids enriched withvarious growth factors.

SUMMARY

Fetal wound healing is a remarkable processthat is fundamentally different than postnatalhealing. Healing of primarily closed, linearwounds occurs rapidly and without scarring. Inlate gestational age fetal sheep, a transition toadult-like healing occurs as evidenced by minimalscar formation. Acute inflammation is not in-volved, fibroblast recruitment and proliferationis minimal, fue matrix of fue wounds is enrichedwith HA and collagen deposition is highly or-ganized so that scarring is minimal or nonexis-tent. Oren wounds in several species do not con-tract, an observation that mar be because of aninhibitor of contraction in amniotic fluido Theunderlying mechanisms that regulate fetal woundrepair are currently not well understood. An al-tered supply or activity of growth factors marbe instrumental. Fibroblasts mar preferentiallyproduce HA secondary to a factor present within

REFERENCES

l. Gann, D. S., and Amaral,]. F. Endocrine and metabolicresponses to injury. In: PrincipIes of Surgery. 5th ed.Pp. 1-68. Edited by S. l. Schwartz, G. T. Shires andF. C. Spencer. New York: McGraw-Hill Book Co., 1989.

2. Pritchard, ]. A., MacDonald, P. C., and Gant, N. F.The morphologic and functional development of fuefetus. In: Williams Obstetrics. Edited by]. A. Pritchard,P. C. MacDonald and N. F. Gant. Pp. 145-180. Norwalk:Aooleton-Century-Crofts, Inc., 1985.

3. Ylayfair,]. H. L., Wolfendalen, M. R., and Kay, H. E.M. The leukocytes of peripheral blood in fue humanfetus. Br. ]. Haematol., 1966, 9: 336-344.

4. Adzick, N. S., Harrison, M. R., Glick, P. Lo, and others.Comparison of fetal, newbom and adult wound healingby histologic enzyme-histochemical and hydroxyprolinedetermination. ]. Pediatr. Surg., 1985, 20: 315-319.

5. Kohler, P. F. Maturation of fue human complementsystem. l. Onset, time and sites of fetal Clq, C4, C3and C5 synthesis. ]. Clin. lnvest., 1973, 52: 671-677.

6. Adinolfi, M. Human complement: Onset and site ofsynthesis during fetal liCeo Am. ]. Dis. Child., 1977,131: 1015-1023.

7. Moore, K. L. The Developing Human. Clinically Ori-ented Embryology. 4th Ed. Pp. 65-129. Philadelphia:W. B. Saunders Co., 1988.

8. Sadler, T. W. Langman's Medical Embryology. 5th Ed.Pp. 19-108. Baltimore: Williams & Wilkins Co., 1985.

9. Phillips, f. D., Diamond,]. M., and Fonkalsrud, E. W.Fetal rabhit intestinal absorption: Implications for trans-amniotic fetal feeding. ]. Pediatr. Surg., 1990, 25: 909-913.

10. Mulvihill, S. ]., Stone, M. M., and Fonkalsrud, E. W.Trophic effects of amniotic fluid on fetal gastrointestinaldevelopment. ]. Surg. Res., 1986, 40: 291-296.

11. Mulvihill, S. ]., Halden, G., and Debas, H. T. Trophiceffect of amniotic fluid on cultured fetal gastric mucosalcells. ]. Surg. Res., 1989, 46: 327-329.

12. Schilling,]. A. Wound healing. Surg. Clin. North Am.,1976, 56: 869-874.

13. Weigel, P. H., Fuller, G. M., and LeBoeuf, R. D. Amodel for the Tole of hyaluronic acid and fibrin inthe early events during fue inflammatory response andwound healing. ]. Theor. Biol., 1986, 119: 219-234.

14. Seppa, H. E. f., Grotendorst, G. R., Seppa, S. l., andothers. The p1atelet-<lerived growth factor is a chemo-attractant for fibroblasts.]. Cell. Biol., 1982,92: 584-588.

15. Grotendorst, G. R., Chang, T., Seppa, H. E. ]., andothers. Platelet-<lerived growth factor is a chemoattract-ant for vascular smooth muscle cells. ]. Cell Physiol.,1983, 113: 261-266.

16. Roberts, A. B., Spom, M. R., Assoian, R. K., and others.Transforming growth factor type 13: Rapid inductionof fibrosis and angiogenesis in vivo and stimulation ofcollagen formation in vitro. Proc. Nat!. Acad. Sci. USA,1986, 83: 4167-4171.

17. Simpson, D. M., and Ross, R. The neutrophilic leukocytein wound repair. A study with antineutrophilic serum.]. Clin. lnvest., 1972, 51: 2009-2023.

18. Leibovich, D. S., and Ross, R. The Tole of fue mac-rophage in wound repair. A study with hydrocortisone

450 SURGERY, Gynecology & Obstetncs. May 1992. Volume 174

and antimacrophage serum. Am. J. Pathol., 1975, 78:71-100.

19. Diegelmann, R. F., Caben, l. K, Kaplan, A. M. Therole of macrophages in w.ound repair: A review. Plast.Reconstr. Surg., 1981, 68. 107-113.

20. Levenson, S. M., Geever, E. F., Crowley, L. V., andothers. The healing of rat skin wounds. Ann. Surg.,1965, 161: 293-308.

21. Madden, J. W., and Peacock, E. E. Studies on thebiology of colla~en during wound healing. l. Rate ofcollagen synthesls and deposition in cutaneous woundsof the rato Surgery, 1968, 64: 288-294,-

22. Doillon, C. J., Dunn, M. G., Bender, E., and others.Collagen fiber formation in repair tissue: Developmentof strength and toughness. Coll. Relat. Res., 1985, 5:481-492.

23. Gabbiani, G., Hirschel, B. J., Ryan, B., and others. Gran-ulation tissue as a contractile organ. J. Exp. Med., 1972,135: 719-734.

24. Ehrlich, H. P. Wound closure: Evidence of cooperationbetween fibroblasts and collagen matrix. Eye, 1988, 2:149-157.

25. Sopher, D. A study of wound healing in the fetal tissuesof the cynomolgus monkey. Laboratory Animal Hand-books, 1975, 6: 327-335.

26. Rowsell, A. R. The intra-uterine healing of foetal musclewounds: Experimental study in the rato Br. J. Pediatr.Surg., 1984, 37: 635-642.

27. Robinson, B. W., and Goss, A. N. Intra-uterine healingof fetal rat cheek wounds. Cleft Palate J., 1981, 18:251-255.

28. Sopher, D. The response of rat fetal membranes toinjury. Ann. R. Coll. Surg., Engl., 1972, 51: 240-249.

29. Burrington, J. D. Wound healing in the fetal lamb. J.Pediatr. Surg., 1971, 6: 523-528.

30. Somasun daram , K., and Prathap, K. The effect of ex-clusion of amniotic fluid on mtrauterine healing ofskin wounds in rabbit fetuses. J. Pathol., 1972, 107:127-130.

31. Thomasson, B., Vilijanto, J., Jaakelainen, A, and others.Enzyme histochemical observations on the formationof granulation tissue in rabbit fetuses and does. ActaChir. Scand., 1973, 139: 327-333.

32. Krummel, T. M., Nelson, J. M., Diegelmann, R. F.,and others. Fetal response to injury in the rabbit. J.Pediatr. Surg., 1987, 22: 640-644.

33. Block, M. Wound healing in the new-bom opossum(Didelphis virginianam). Nature, 1960, 187: 341.

34. Ditesheim, J. A., Ledbetter, M. S., Morykwas, M. J.,and others. An ex utero model of fetal wound healing.Surg. Forum, 1990, 41: 643-645.

35. Adzick, N. S., and Harrison, M. R. Surgical techniquesin the fetal rabbit. In: Animal Models in Fetal Medicme.Edited by P. Nathanielsz. Amsterdam: Elsevier, 1986.

36. Caben, l. K., and Mast, B. A. Models of wound healing.J. Trauma Suppl., 30 (12), 1990, 149-155.

37. Siebert,J. W., Burd, A R., McCarthy,J. G., and others.Fetal wound healing: A biochemical study of scarlesshealing. Plast. Reconstr. Surg., 85: 495-502.

38. Hallock, G. G. In utero cleft lip repair in A/J mice.Plast. Reconstr. Surg., 1985, 75: 785-788.

39. Hallock, G. G., Rice, D. C., and McClure, H. M. Inutero lip repair in the rhesus monkey: An update.Plast. Reconstr. Surg., 1987, 80: 855-858.

40. Longaker, M. T., Dodson, T. B., and Kaban, L. B. Arabbit model for fetal cleft lip repair.J. Oral Maxillofac.Surg., 1990, 48: 714-719.

41. Goss, A. N. Intra-uterine healing of fetal rat oral mu-casal, skin and cartilage wounds. J. Oral Pathol., 1977,6: 35-43.

42. Vilijanto, J., Thomasson, B., Pikkarainen, J., and others.Enzyme f'oetal connective tissue regeneration: A bio-

chemical study in rabbits. Acta. Chir. Scand., 1975,141: 85-89.

43. Tipton, J. B., and Dingman, R. O. Some aspects ofwound healing in the germ-free animal. Plast. Reconstr.Surg., 1966, 38: 499-511.

44. Holder, T. M., and Ashcroft, K. W. The effects of hileduct ligation and inflammation in fue fetus. J. Pediatr.Surg., 1967, 2: 35-40.

45. Chignier, E., Baguet, J., Dessapt, B., and others. Skinhealing and fibrin stabilizing factor blood levels inthe rabbit fetus. J. Surg. Res., 1981, 31: 415-432.

46. Longaker, M. T., Whitby, D. J., Adzick, N. S., andothers. Studies in fetal wound healing, VI. Second andearly third trimester fetal wounds demonstrate rapidcollagen deposition without scar formation. J. Pediatr.Surg., 1990, 25: 63-69.

47. DePaIma, R. L., Krummel, T. M., Durham, L. A., 111,and others. Characterization and quantitation of woundmatrix in fue fetal rabbit. Matrix, 1989, 9: 224-231.

48. Chiu, E. S., Longaker, M. T., Adzick, N. S., and others.Hyaluronic acid patterns in fetal and adult wound fluidoSurg. Forum, 1990, 41: 636-639.

49. Burd, D. A. R., Longaker, M. T., Adzick, N. S., andothers. Foetal wound healing in a large animal model:fue deposition .of collagen is confirmed. Br. J. Plast.Surg., 1990, 43. 571-577.

50. Thomas, B. L., Krummel, T. M., Melany, M., and others.Collagen synthesis and type expression of fetal fibro-blasts in vitro. Surg. Forum, 1988, 39: 642-644.

51. Hess, A. Reactions of mammalian fetal tissues to injury,11. Skin. Skin Anatomic Res., 1958, 119: 435-438.

52. Krummel, T. M., Ehrlich, H. P., Nelson, J. M., andothers. Fetal wounds do not contract in utero. Surg.Forum, 1989, 40: 613-615.

53. Haynes" J. H., Krummel, T. M., Schatzki, P. F., andothers. Histology of the open fetal rabbit wound. Surg.Forum, 1989, 40: 558-560.

54. Ditesheim,J. A., Delozier,J. B., Rees, R. S., and others.Cov,ered fetal excisional wounds heal by tissue regen-eratlon. Surg. Forum, 1989, 40: 615-617.

55. Rowlatt, U. Intrauterine wound healing in a 20 weekhuman fetus. Virchows Arch., 1979, 381: 353-361.

56. Dixon, J. B. Inflammation in the foetal and neonatalrat: The local reaction to skin burns. J. Pathol. Biol.,1960, 80: 73-82.

57. Pessa, M. E., Kirby, K. l., and Copeland, E. M. Growthfactors and determinants ofwound repair. J. Surg. Res.,1987, 42: 207-217.

58. Durham, L. A., 111, Krummel, T. M., Cawthorn, J. W.,and others. Analysis of transforming growth factor-betareceptor binding in embyronic, fetal and adult rabbitfibroblasts. J. Pediatr. Surg., 1989, 24(8): 784-788.

59. Krummel, T. M., Michna, B. A., Thomas, B. L., andothers. Transforming growth factor-¡:l induces fibrosisin a fetal wound modelo J. Pediatr. Surg., 1988, 23:647-652.

60. Haynes, J. H., Johnson, D. E., Flood, L. C., and others.Platelet-derived growth factor induces fibrosis at a fetalwound site. Surg. Forum, 1990, 41: 641-643.

61. Borchelt, B. D., Krummel, T. M., Cawthorn, J. W., andothers. Transposition of an early adult wound to afetal rabbit wound does not result in recruitment offibroblasts. Surg. Forum, 1989, 40: 555-557.

62. Longaker, M. T., Chiu, E. S., Harrison, M. R., andothers. Studies in fetal wound healing IV. Hyaluronic.acid-stimulating activity distinguishes fetal wound fluidfrom adult wound fluido Ann. Surg., 1989,210: 667-672.

63. Longaker, M. T., Harrison, M. R., Crombleholme, T.M., and others. Studies in fetal wound healing 1: Afactor in fetal serum that stimulates deposition of hy-aluronic acid. J. Pediatr. Surg., 1989, 24: 789-792.

64. Longaker, M. T., Adzick, N. S., Jackson, L. H., and