Transforming Growth Factor Beta (TGF-/~) Induces Fibrosis in a Fetal Wound Model

By Thomas M. Krummel, Barbara A. Michna, Brian L. Thomas, Michael B. Sporn, Jeffrey M. Nelson,

Arnold M. Salzberg, I. Kelman Cohen, and Robert F. Diegelmann

Richmond, Virginia and Bethesda, Mary land

�9 The adult cellular response to t issue in jury is charac- terized by acute inflammation followed eventually by fibro- blast proliferation and collagen synthesis. Fetal tissue responses to injury differ markedly from those of the adult; an early acute inflammatory response is absent, few f ibro- blasts participate, and no collagen is deposited. The object of the present study was to analyze the effects of trans- forming growth factor beta [TGF-/~), an important regula- tory molecule in adult healing events, on the fetal tissue response following wounding. Fetal cellular and extracel lu- lar matrix responses to in jury were evaluated by placing subcutaneous wound implants containing TGF-/~ (0.01 to 10 ng) in fetal rabbits at 24 days gestation ( term = 31 days). Histologic responses one to seven days later were compared with fetal and adult control implants wi thou t TGF-/~. The histology of the adult implant was character- ized by an early acute inflammatory response: by day 7 f ibroblasts and collagen were predominant. In contrast, control implants removed from fetal rabbits had no histo- logic evidence of acute inflammation or fibroblast penetra- tion and no collagen was deposited. When implants con- taining 1.0 ng TGF-/~ were remnved from fetal rabbits at seven days, a grossly fibrotic reaction was observed: histology confirmed marked fibroblast penetration wi th collagen deposition. Fetal implants containing 0.01 ng or 10 ng TGF-/~ showed few fibroblasts but had increased num- bers of inflammatory cells compared with controls. These observations demonstrate that the fetal response becomes adult l ike wi th f ibroblast proliferation and collagen accumu- lation when TGF-/~ is added, thus documenting the respon- siveness of the fetal system to adult repair signals. Such responsiveness thus suggests a critical difference in the fetal wound environment. Fetal repair may proceed in the absence of trophic factors l ike TGF-/~, thus accounting for optimal "healing" in the absence of excessive fibrosis. �9 1988 by Grune & Strat ton, Inc.

INDEX WORDS: Wound healing; transforming growth fac- tor beta.

DULT M A M M A L I A N cellular and extracellular matrix responses to injury are characterized by

an orderly cellular cascade beginning with chemoat- tracted neutrophils followed by lymphocytes and macrophages, and leading to eventual fibroblast ingrowth with the synthesis of a collagenous scar. ~ The fetal tissue responses are dramatically different; only mononuclear inflammatory cells respond initially, ~9 fibroblasts do not appear to be critical to the reparative process, 27"9 and the extracellular matrix deposited is composed largely of proteoglycans rich in hyaluronic acid rather than collagen. ~~ In the adult, peptide growth factors are known to play an important role in the regulation of this cascade of localized recruitment,

proliferation, and subsequent replacement of the dam- " aged connective tissue matrix following injury. 11

Transforming growth factor beta (TGF-/3) is perhaps the single most important growth factor, and is released by a number of cells involved in inflammation and healing. 12 TGF-/3 has been shown to increase total protein, collagen and DNA content in wound cham- bers in vivo, 13 and to stimulate biosynthetic rates of collagen and fibronectin in cultured fibroblasts. ~4 In newborn mice, TGF-~ injected subcutaneously has been shown to stimulate the local formation of granu- lation tissue. 15

Since the fetal response to injury lacks the charac- teristic adult processes of abundant fibroblast migra- tion, proliferation, and collagen synthesis, this suggests that either TGF-fl is not a participant in the fetal response or that fetal target cells are not responsive to TGF-/3. Therefore, this study was designed to answer the latter questions by providing exogenous TGF-/3 at a fetal wound site to assess in utero responsiveness of fetal cells to this "panregulin."

MATERIALS AND METHODS

o r G F - ~

TGF-/3 was purified from human platelets and homogeneity was confirmed by electrophoresis on sodium dodecyl sulfate/polyacryl- amide gels. ~6 Standard solutions were made in 4 mmol /L HC1 containing 1 m g / m L bovine serum albumin; volumetric dilutions to final concentrations of 0.01, 0.1, 1.0, and 10.0 ng TGF-/3/5 #L were then made.

Implants Wound implants containing polyvinyl alcohol (PVA) sponge were

miniaturized, 9 but otherwise made as described by Diegelmann et al, 17 by placing preweighed PVA sponge within a hollow perforated

From the Divisions of Pediatric and Plastic Surgery, Department of Surgery, Medical College of Virginia, Virginia Commonwealth University, Richmond; and the National Cancer Institute, Bethesda, MD.

Supported by An American College of Surgeons Fellowship, a Virginia Commonwealth University Faculty Grant-in-Aid and National Institute of Health Grant No. 20298.

Presented at the 36th Annual Meeting of the Surgical Section of the American Academy of Pediatrics, New Orleans, October 31 to November 2, 1987.

Address reprint requests to Thomas M. Krummel, MD, Medical College of Virginia, 401 N 11t h St, Richmond, VA 23298-0015.

�9 1988 by Grune & Stratton, Inc. 0022-3468/88/2307-0013503.00/0

Journal of Pediatric Surgery, Vol 23, No 7 (July), 1988: pp 647-652 647

648 KRUMMEL ET AL

12 • 2 mm silicone tube. Five microliters study concentration containing either 0.01, 0.1, 1.0 or 10.0 ng of TGF-/3 was then injected and absorbed into the PVA sponge of a given implant. Control implants received 5 ~L of the TGF-/3 vehicle only. Implants were then lyophilized and sterilized by gamma irradiation.

Wounding Techniques

Fetal rabbits. Time-dated pregnant New Zealand white rabbits were commercially obtained (Myrtle's Rabbitry, Nashville, TN). Subcutaneous PVA implants were placed in fetal rabbits on the 24th day of gestation (term = 31 days) and removed following cesarean section delivery of viable fetuses one to seven days postoperatively, replicating operative techniques of Adzick and Harrison. ~s Fetal survival was approximately 70%. Littermates received implants with equivalent concentrations of TGF-fl. Following harvest, PVA sponge inserts were removed from the silicone carrier, fixed in 10% buffered formalin, processed, and examined with standard histologic tech- niques. Three implants for each concentration of TGF-/~ (control, 0.01, 0.1, 1.0, 10.0 ng TGF-~) were obtained for each study day; thus a total of 105 implants from 105 viable fetuses were analyzed.

Adult rabbits. Control implants were placed subcutaneously in adult rabbits using the same anesthetic and operative techniques. They were then harvested from one to seven days postoperatively and used to provide a mature response for comparison. Since the adult response is normally characterized by inflammation, fibroblast proliferation, and collagen deposition, the effect of exogenous TGF-/3 was not analyzed in this study.

Fig 1. Adult control implant shows extensive fibroblast infil- tration of the PVA sponge seven days following implantation (original magnification, x40).

RESULTS

Neutrophi ls were present within 24 hours of control implant p lacement in the adult and were then largely replaced by mononuclear cells within two to four days. Thereafter , fibroblasts were visible and predominant by seven days with deposition of an extracellular matr ix composed of collagen identified by Masson's t r ichrome stain (Fig 1).

In contrast , only mononuclear cells (monocytes or lymphocytes) were present in the control implants harvested on days 1 to 7 from the fetus (Fig 2). Nei ther neutrophils nor fibroblasts were seen up to day 7. Extracel lular matr ix was devoid of collagen by tri- chrome staining. Positive staining with Alcian blue demonstra ted proteoglycans. Implan t pre t rea tment with hyaluronidase abolished Alcian blue staining, thus suggesting a matr ix rich in hyaluronic acid, consistent with earlier observations. 1~

There was an apparent increase in mononuclear cells at days 1 and 2, with occasional fibroblasts present by days 5 to 7 in the implants containing 0.01 ng of TGF-/~ harvested from the fetuses (data not shown). When the TGF-f l concentrat ion was increased to 0.1 ng, there were increased mononuclear cells and

Fig 2. Fetal control implant at seven days demonstrates a complete absence of fibroblasts (original magnification, •

TGF-fl MODULATION OF FETAL WOUND HEALING EVENTS 649

Fig 3. Fetal implant with 0.1 ng TGF-~ is remarkable for (A) increased mononuclear cell with (B) occasional fibroblasts (original magnification, x401.

some neutrophils at days 1 to 3 with fibroblasts defi- nitely present from days 5 to 7 in the implants taken from the fetuses (Fig 3).

In marked contrast, neutrophils and mononuclear cells were seen in implants containing 1.0 ng TGF-fl at days 1 and 2, with progressive increases in fibroblasts up to day 7, at which time the implants were impossible to remove from the subcutaneous pockets in the fetuses (Fig 4). Definite trichrome-positive material was seen. The implants containing 1 0 ng of TGF-fl and placed in the fetuses remained relatively acellular except for RBCs until days 5 to 7 when mononuclear cells were seen. Neither fibroblasts nor collagen was observed (data not shown).

DISCUSSION

TGF-fl was first shown to cause phenotypic trans- formation of rat fibroblastsfl Since then it has been shown to control migration, 2~ proliferation, ]9 and differentiation, ~2 as well as other cellular functions. ]2 It is found in high concentrations in platelets, ~6 macro- phages, ~2 and lymphocytes. 22 TGF-B is ubiquitous; many cells synthesize TGF-fl and all cells tested pos- sess high affinity receptors. 23 Thus, TGF-fl is a critical

regulatory peptide that acts not through endocrine mechanisms, but through paracrine 24 or autocrine mechanisms 25 in the extracellular space. It is truly multifunctional, capable of stimulating or inhibiting proliferation and differentiation, as well as other cellu- lar processes, 12 depending on concentration 26 and the presence or absence of other growth factors.

TGF-fl plays an important regulatory role in adult tissue repair as presently understood. 11"16'19"21'23 It is released at the injured site from platelets, lympho- cytes, and macrophages, and acts as a chemoattractant for monocytes and fibroblasts, and may be stimulatory or inhibitory depending on local conditions. Recently, local application of TGF-fl has been shown to acceler- ate healing and increase tensile strength of incisional wound in rats. 27

The lack of the adult traits of fibroplasia and collagen deposition observed in the fetal model 9 raised interesting questions about regulatory mechanisms in fetal healing. First, is TGF-fl present in the fetal milieu?; second, are fetal ceils responsive to TGF-fl? The present study was therefore designed to test the responsiveness of the fetal system to exogenous TGF-fl provided at an in utero wound site.

650 KRUMMEL ET AL

Fig 4. Fetal implant with 1.0 ng TGF-/~ with density fibrotic fibroblast infiltration at seven days with fibroblasts streaming into the implant (arrow), Note hair follicles of overlying adherent skin (original magnification, •

The cellular and matrix events following injury, as sampled by the PVA implant technique, are dramati- cally different in fetal controls as compared with adult controls. Classic mammalian responses to injury are seen in adult control implants with an acute inflamma- tory phase, characterized by neutrophil infiltration of the fibrin clot 24 hours after injury. Chronic inflam- matory cells, lymphocytes and macrophages, arrive by the second or third day; by six to seven days, all inflammatory cells are replaced by rapidly proliferat- ing fibroblasts synthesizing the extracellular matrix of repair, collagen.

Control implants in the fetus do not demonstrate the acute inflammatory infiltration with neutrophils as seen in the adult. Instead, only mononuclear cells, either monocytes or lymphocytes, but as yet uncharac- terized, appear from one to seven days. No fibroblasts are seen, and histologic examination of the extracellu- lar matrix showed that it is composed not of collagen,

but of proteoglycans, ~~ another class of matrix pro- teins.

The addition of TGF-fl to the fetal wound substan- tially alters the usual cellular response. At the lowest dose, 0.01 ng TGF-tJ, an apparent increase in mononu- clear cells was seen in the first two days, with a few fibroblasts seen by days 6 and 7. There appeared to be a qualitative increase in this response with higher TGF-fl doses, peaking at an initial delivered dose of 1.0 ng (Fig 4). At this dose, the local fibroblastic infiltra- tion was so dense as to preclude simple removal of the silicone tube (which is usually the case), necessitating en bloc excision of the implant and surrounding skin. Thus, the fetal response becomes fibroblastic like the adult response when TGF-fl is added. This observation suggests that fetal cells, like their adult counterparts, are responsive to the chemoattractant, proliferative, and synthetic signals of TGF-/3. This intrinsic response suggests a critical difference in the fetal wound envi- ronment; an environment in which TGF-/3 may not participate.

The cellular events at 10 ng are of some interest. None of the TGF-fl effects at lower doses is seen until days 6 to 7 when mononuclear cells are again seen; fibroblasts remain notably absent. Though initially confusing, such a response is consistent with the pres- ent understanding of the bifunctional dose-dependent effects of TGF-/3. 26 Thus, the resulting local concentra- tion achieved with 10 ng could be inhibitory until diffusion and/or metabolism reduce the concentration to weakly stimulatory levels by six to seven days.

Tissue repair following injury is obviously essential to survival; control of such repair would be advan- tageous since quantitative defects of excess or defi- ciency can have disasterous consequences. Excessive deposition of collagen at an injury site results in pathologic fibrosis, clinically recognized as hypertro- phic scars, keloids, anastomotic strictures, pulmonary fibrosis, and cirrhosis. Inadequate healing can result in dehiscence or anastomotic leak. Thus, a better under- standing of the control mechanisms resulting in the "optimal healing" observed in the fetus, 2'9 may yield new strategies in the management of wound healing in the adult.

ACKNOWLEDGMENT

The authors wish to thank John Cawthorn, R. Lawrence DePalma and Dr Lucian A. Durham III for technical assistance; Frances Freund and Donald Pearce for histologic preparations; and Elaine Blair for manuscript preparation.

REFERENCES

1. Schilling JA: Wound healing. Surg Clin North Am 56:859- 2. Burrington JD: Wound healing in the fetal lamb. J Pediatr 874, 1976 Surg 6:523-528, 1971

TGF-/3 MODULATION OF FETAL WOUND HEALING EVENTS 651

3. Sopher D: A study of would healing in the foetal tissues of the cynomolgus monkey. Lab Anim Handbook 6:327-335, 1975

4. Goss AN: Intrauterine healing of fetal rat oral mucosal, skin and cartilage wounds. J Oral Pathol 6:35-43, 1977

5. Rowlatt U: Intrauterine wound healing in a 20-week human fetus. Virchows Arch 381:353-361, 1979

6. Chignier E, Baguet J, Dessapt B, et al: Skin healing and fibrin stabilizing factor blood levels in the rabbit fetus. J Surg Res 31:415-432, 1981

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

8. Adzick NS, Harrison MR, Glick PL, et al: Comparison of fetal newborn and adult wound healing by histologic, enzyme-histochemi- cal and hydroxyproline determinations. J Pediatr Surg 20:315-319, 1985

9. Krummel TM, Nelson JM, Diegelmann RF, et al: Fetal response to injury in the rabbit. J Pediatr Surg 22:640-644, 1987

10. DePalma RL, Krummel TM, Nelson JM, et al: Fetal wound matrix is composed of proteoglycan rather than collagen. Surg Forum 38:626-628, 1987

11. Pessa ME, Bland KI, Copeland EM III: Growth factors and determinants of wound repair. J Surg Res 42:207-217, 1987

12. Sporn MB, Roberts AB, Wakefield LM, et al: Transforming growth factor-~3: Biological function and chemical structure. Science 233:532-534, 1986

13. Sporn MB, Roberts AB, Shull JH, et al: Polypeptide trans- forming growth factors isolated from bovine sources and used for wound healing in vivo. Science 219:1329-1331, 1983

14. Ignotz RA, Massague J: Transforming growth factor-/3 stim- ulates the expression of fibronectin and collagen and their incorpora- tion into the extracellular matrix. J Biol Chem 261:4337-4345, 1986

15. Roberts, AB, Sporn MB, Assoian RK, et ah Transforming growth factor type-B: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc Nat] Acad Sci USA 83:4167-4171, 1986

16. Assoian RK, Komoriya A, Meyers CA, et al: Transforming growth factor-/3 in human platelets. J Biol Chem 258:7155-7160, 1983

17. Diegelmann RF, Lindblad W J, Cohen IK: A subcutaneous implant for wound healing studies in humans. J Surg Res 40:229- 237, 1986

18. Adzick NS, Harrison MS: Surgical techniques in the fetal rabbit, in Nathanielsz P (ed): Animal Models in Fetal Medicine. Amsterdam, North Holland, 1980

19. Roberts AB, Anzano MA, Lamb LC, et al: New class of transforming growth factors potentiated by epidermal growth factor: Isolation from non-neoplastic tissues. Proc Natl Acad Sci USA 78:5339-5343, 1981

20. Postlethwaite AE, Keski-Oja J, Moses HL, et al: Stimulation of the chemotactic migration of human fibroblasts by transforming growth factor-/3 J Exp Med 165:251,256, 1987

21. Wahl SM, Hunt DA, Wakefield LM, et al: Transforming growth factor type-/3 induces monocyte chemotaxis and growth factor production. Proc Natl Acad Sci USA 84:5788-5792, 1987

22. Kehrl JH, Wakefield LM, Roberts AB, et al: Production of transforming growth factor-/3 by human T-lymphocytes and its potential role in the regulation of T-cell growth. J Exp Med 163:1037-1050, 1986

23. Massague J, Like B: Cellular receptor for typeW3 transform- ing growth factor: Ligand binding and affinity labeling in human and rodent cell lines. J Biol Chem 260:2636-2645, 1985

24. James R, Bradshaw RA: Polypeptide growth factors. Ann Rev Biochem 53:259-292, 1984

25. Sporn MP, Todano G J: Autocrine secretion and malignant transformation of cells. N Engl J Med 303:878-880, 1980

26. Hill D J, Strain A J, Elstow SF, et al: Bifunctional action of transforming growth factor-/3 on DNA synthesis in early passage human fetal fibroblasts. J Cell Physiol 128:322-328, 1986

27. Mustoe TA, Pierce GF, Thomason A, et al: Accelerated healing of incisional wounds in rats induced by transforming growth factor-ft. Science 237:1333-1336, 1987

D i s c u s s i o n

Kevin Pringle (Iowa City): I have a question about your methods. Is it possible to dispense with the silicone tube carrier? Can you just put your matrix in without the carrier? If you put a 5 cm by 10 cm silastic patch in a fetal lamb when it is about 1 kg, and then look at that lamb at term when it is about 5 kilos, then it is about 1 by 2 cm. Interestingly enough, it you put in a Dexon mesh patch instead of a silastic patch, you do not get much in the way of healing at all. Clearly, the environment in which you demand your healing takes place is going to make a big difference in the results that you get. I think you have hit the nail on the head. I f we can work out what is going on with fetal healing, we may well be able to modify abnormal adult healing. I think that is a great approach. Clearly, we have a lot to learn about all types of healing and this is one of those good examples of cooperation between the basic cell biologist and the clinical investigator. I think this is the sort of study that we are going to need to do in the future if we are to understand some of the things we take for granted on a daily basis.

Steve Shochat (Stanford): Can you do this in vitro so that you can study a blocker or an ant i -TGF-~ factor to neutralize the TGF-fl factor? Obviously in the hypertrophied scar you need to find a blocker to this factor so that you can prevent the overgrowth of fibroblasts.

Krummel (closing): Dr Pringle, that is a fascinating demonstration of the difference in response in the fetus to various materials. The implant we used incorporates a silicone carrier because it is important for quantita- tion to be able to remove it and precisely define where the wound begins and ends. Hence the PVA implant was designed by a group of plastic surgeons and biochemists in Richmond so that precise quantitation of matrix (that is deposited within it) is feasible. I would agree totally with your observation that these findings do not really have anything to do with the possibility of fetal surgery. That capability has already been demonstrated. Our interest is in understanding what turns on cell growth and what turns it off. Dr Shochat asked a great question in terms of whether

652 KRUMMEL ET AL

this can be studied in vitro. We have embryonic fibroblasts from rabbits harvested at ten to 25 days' gestation growing in cultures as well as adult fibro- blasts. We are currently looking at cell replication and matrix synthesis in normal and TGF-C~-stimulated

conditions. Finally, I would emphasize that at this point we are just beginning to scratch the surface in terms of our understanding of the events of fetal wound healing. We would hope that more precise in-depth investigations will be forthcoming.