JOURNAL
OF SURGICAL
RESEARCH
52,65-70
(1992)
Effect of Transforming Growth Factor ,6 on Postoperative Formation and Intact Peritoneum
Adhesion
R. STAN WILLIAMS, M.D., ANN MARIE ROSSI,M.S., NASSER CHEGINI, PH.D., AND GREGORY SCHULTZ, PH.D. Department
of Obstetrics and Gynecology, Submitted
University
for publication
Press,
December
Gainesville,
Florida
32610
17, 1990
jury to the peritoneum at the time of surgery causes a rapid release of platelets, fibrinogen, histamine, and vasoactive kinins. With increased capillary permeability, the site of injury continues to release these products and fibrinous adhesions form between adjacent surfaces [ 11. If the fibrinolytic mechanisms are not inhibited by tissue ischemia or foreign bodies, the fibrinous adhesions may be digested and no permanent adhesion is formed. In the absence of adequate fibrinolytic mechanisms, the fibrin bands soon become infiltrated with fibroblasts which produce collagen and subsequently, become vascularized. The fibrin adhesion is transformed into a thick, vascular adhesion and becomes permanent. Most investigators have focused efforts to prevent postoperative adhesion formation on the early events of fibrin formation. Although this approach has had some success in preventing postoperative adhesion formation, at least 50% of patients still develop significant adhesions. Another avenue which may be complimentary is the prevention of fibroblast invasion of fibrinous adhesions with subsequent collagen formation and allowing natural fibrinolysis additional time to resolve fibrin bands. Peptide growth factors and their receptors appear to play important regulatory roles in healing other types of wounds and may also be involved in the development of adhesions at the sites of injured peritoneum postoperatively. Platelets contain several peptide growth factors which are released at the site of an injury includingplatelet derived growth factor, epidermal growth factor-like peptides, and transforming growth factor /3 (TGFP). TGFP is an extremely potent chemoattractant for macrophages, mononuclear leukocytes [2], and fibroblasts [3]. It also stimulates collagen and fibronectin synthesis by fibroblasts [4] and inhibits most epithelial cell growth [5]. These properties of TGF@ as well as the release of TGFP from platelets at sites of peritoneal injury make it a candidate for playing a pivotal role in the formation of permanent postoperative adhesions. Our hypothesis is that TGFP would have minimal effect on uninjured peritoneum, but following peritoneal injury, TGFfi released from platelets and other wound cells
Transforming growth factor @ (TGFB) is an extremely potent chemoattractant for macrophages, mononuclear leukocytes, and fibroblasts. It also acts as a potent stimulant for collagen and fibronectin synthesis and inhibits epithelial cell growth. TGF/3 plays an important role in healing many types of wounds, but its role in peritoneal adhesion formation is not known. These studies were performed to determine if TGF/3 could affect postoperative wound healing in a rat model. In the first experiment, 20 rats were divided into two groups and received either 2 pg TGFj3 or control diluent IP daily for 5 days after surgical injury to the uterine horns. The severity of the adhesions were graded 2 weeks postoperatively using a score of O-3. The TGF@ group showed a higher adhesion score at 2 weeks compared to control, 2.9 * 0.34 and 1.6 f 0.61, respectively (P < 0.001). On H&E stained sections of the adhesions, there was an increase in the number of both inflammatory cells and fibroblasts in the TGF/3treated animals. A comparison trial of bone-derived TGFB (a gift from Collagen Corporation, Palo Alto, CA) versus recombinant TGFB (a gift from Oncogen, Seattle, WA) versus control using the same protocol as above showed that both sources of TGF/3 were more effective in promoting postoperative adhesions when compared to controls, and there was no difference between TGFB groups, 3.0 f 0 for both TGFj3 groups, and 2.2 f 0.91 for control (P < 0.001). TGFj3 did not induce adhesions when given in the same dosage for 5 days ip in unoperated rats, although it did induce fibrosis if injected into the abdominal wall. These data suggest that TGF@ may play a role in stimulating fibrosis in postoperative fibrinous adhesions, but has no effect on intact mesothelium. Prevention of TGF/3 production postoper0 1992 atively may help prevent adhesion formation. Academic
of Florida,
Inc.
INTRODUCTION
The peritoneum is composed of surface mesothelial cells and submesothelial stroma containing fibroblasts, collagen, mast cells, macrophages, and blood vessels. In65
Copyright 0 1992 by All rights of reproduction
0022404/92 $1.50 Academic Press, Inc. in any form reserved.
66
JOURNAL
OF SURGICAL
RESEARCH:
would promote formation of severe adhesions by stimulating fibroblasts, attracting inflammatory cells, and stimulating the synthesis of collagen and angiogenesis. These studies were performed to determine any effect exogenous TGFP has on uninjured peritoneum and to measure the effect of TGFP on postoperative peritoneal wound healing in a rat model.
VOL.
52, NO. 1, JANUARY
1992
of the rats received PBS with RSA ip for 5 days. Rats were sacrificed and evaluated as described above. In the last experiment, a neutralizing antibody to TGFP (R&D, Minneapolis, MN) was tested to determine if postoperative adhesion formation could be altered by interfering with endogenous TGFfl action. This rabbit antiporcine TGF@ antibody neutralizes both TGFP-1 and TGF@-2 according to the manufacturers. A group of 20 rats underwent surgery and postoperatively MATERIALS AND METHODS half received 10 pg of TGFP-blocking antibody ip daily Female Sprague-Dawley rats weighing approximately for 5 days, and the remainder received PBS injections. 250 g were housed in an environmentally controlled vi- The animals were sacrificed and evaluated as described varium and maintained with water and rat chow ad Eibi- above. turn. Experimental use of these animals was approved by After grading the adhesions, the uterine horns with the University of Florida Institutional Animal Care and the adherent tissues were removed and fixed with Use Committee. Animals were anesthetized with intraBouins solution overnight. The tissue was then dehyperitoneal injection of Rompun 5 mg/kg and Ketaset 40 drated, paraffin embedded, and 5- to 6-pm sections were mg/kg. A lower midline incision was made exposing the cut and counterstained with hematoxylin and eosin Y. uterine horns and each uterine horn was injured with a All materials for light microscopy were purchased from l-cm burn by bipolar cautery, a l-cm crush injury, and a Polysciences (Warrington, PA). l-cm scrape until bleeding occurred. The midline inciStatistical analysis of adhesion scores between groups sion was closed with a single layer closure using 3-O was performed with the use of nonparametric tests. Dexon. When two groups were compared, a nonparametric The first experiment was performed to determine the Mann-Whitney rank test was used and when more than effect of TGFP on postoperative adhesion formation. A two groups were compared a Kruskal-Wallis One-Way group of 20 rats were divided into two treatment groups Anova by Rank test was performed. Statistical analysis of 10 rats each. One group received 500 ,ul of phosphate- was performed with a computer statistical program, buffered saline (PBS) containing 2 gg TGFP (bone de- Statgraphics (Statistical Graphics Corporation, Rockrived, a gift from Collagen Corporation, Palo Alto, CA) ville, MD). (BTGFP) combined with 500 pg rat serum albumin (RSA) as a carrier, injected intraperitoneally (ip) each RESULTS day for 5 days postoperatively. The remaining 10 rats were injected ip with 500 ~1PBS containing 500 pg RSA Effect of Bone-Derived TGFfl on Postoperative Adhesion daily for 5 days postoperatively. Formation Two weeks postoperatively rats were sacrificed, and Exogenous TGF/3 significantly increased postoperaadhesions were graded using a severity score of O-3 by an tive adhesion formation. Adhesions were more dense, observer who was masked to the treatment. A score of 0 vascular, and more extensive after TGFP treatment in was given for no adhesions; grade 1 represented thin, rats with surgically induced trauma to the peritoneal filmy adhesions; grade 2, moderately thick adhesions; surface. TGFP-treated animals showed a significantly and grade 3 for dense, vascular adhesions. higher adhesion score postoperatively as compared to In the second experiment, 30 rats were used to comthe control group, mean scores 2.9 k 0.34 (standard depare the effects of bone-derived TGF/3 and recombinant viation) for TGF/?-treated rats and 1.6 f 0.61 for control TGFP on postoperative adhesion formation. The rats rats (P < 0.001) (Fig. 1). were divided into three groups with 10 rats per group. Although boned-derived TGF@ was highly purified, it Treatment groups included (1) BTGF/3 with RSA + surwas possible that minor contaminants extracted from gery, (2) recombinant TGF/3 (a gift from Oncogen, Seatbone could have influenced the biologic effect seen in tle, WA) (RTGFP) with RSA + surgery, and (3) RSA in this experiment. To further evaluate the effect of TGFfi PBS + surgery. Dosages of TGFP and RSA, the injection regimen, and evaluation were the same as described alone, recombinant TGFB was used as another source of TGF/3 and compared with the effect of bone-derived above. TGF@. In a third experiment, a group of 30 rats which did not undergo surgery were divided into three groups to test Comparison of Bone-Derived TGFfl and Recombinant whether TGFP would stimulate adhesion formation in TGF/3 on Postoperative Adhesion Formation the absence of peritoneal injury. Ten rats were injected intraperitoneally for 5 days with 500 ~1of PBS with 2 pg In the comparison of BTGFB and RTGFP, both RTGFP combined with RSA. Another 10 rats received groups showed a significantly higher rate of adhesion BTGFB combined with RSA for 5 days. The remainder formation postoperatively when compared to control (P
WILLIAMS
ET AL.: TGFP
IN POSTOPERATIVE
ADHESION
67
FORMATION
FIG. 1. Adhesion formation in control (A) and TGFP-treated (B) animals 2 weeks after surgical demonstrates a fibrotic nodule from an inadvertant injection of TGFp into the abdominal wall.
injury
(See text for details).
Panel C
68
JOURNAL
OF SURGICAL
RESEARCH:
< 0.001) and TGF/3 from either source was equally effective in promoting adhesions. Mean adhesion scores were 3.0 +- 0.0 in both TGF@ groups and 2.2 f 0.91 in the control group. This experiment confirmed the results of the first experiment and demonstrated that adhesion formation postoperatively was stimulated by TGF/3 and not any contaminating substance. In this system, recombinant TGFP was seen to be just as active as natural bone-derived TGFP. Effect of TGFP on Adhesion Rats
Formation
in Unoperated
Although TGFB increased adhesion formation if previous peritoneal injury had occurred, no stimulation of adhesions was seen in unoperated rats given ip TGF/3 regardless of the source of TGFP. No difference was seen between groups in the unoperated rats when TGF/3 was compared with control. No adhesions could be identified in either group. Effect of TGF@ Blocking Antibody Formation
on Adhesion
Daily injections of a neutralizing antibody to TGFP was unable to prevent postoperative adhesion formation. TGFP-blocking antibody treatment postoperatively showed no decrease in adhesion formation when compared to controls, 1.85 + 0.87 and 1.7 + 0.92, respectively (P > 0.05). Light Microscopy Figs. 2A and 2B show a portion of uterine horn with tissue adhesions attached from injured control (A) and injured, TGFP-treated (B) rats. The type and cellular adhesion was similar in both cases, but there was an obvious increase in both inflammatory cells and fibroblasts present in the TGFP-treated animals compared to the control animals. DISCUSSION
The role of peptide growth factors in peritoneal wound healing has not been previously evaluated in viuo. It is hypothesized that TGFB may play an important role in peritoneal wound healing and adhesion formation by attracting inflammatory cells [2] and fibroblasts [3] to the site of injury. The subsequent stimulation of collagen and fibronectin synthesis by TGFP [4] would promote the formation of permanent adhesions at the sites of early fibrinous adhesions. This study was designed to measure the effect of exogenous TGFP on postoperative peritoneal wound healing as well as intact peritoneum. The comparison between these groups is important as any effect seen on intact peritoneum could not represent a physiologic effect, as fibrosis normally does not occur in the peritoneal cavity except at sites of injury.
VOL.
52, NO. 1, JANUARY
1992
This study demonstrated that TGFP increases the severity of adhesion formation after surgical injury in a rat model, but has no effect on uninjured peritoneum. Histologically, there were increased inflammatory cells and fibroblasts present in the adhesions of TGFP-treated animals. These results support the hypothesis that TGFfi stimulated chemotaxis of inflammatory cells and fibroblasts and subsequently stimulated collagen synthesis inducing fibrosis in fibrinous exudates postoperatively. A possible explanation of the different affect of TGFP on injured vs intact peritoneum might lie in TGFps effect on fibroblasts vs mesothelial cells. With peritoneal injury, fibroblasts and macrophages of the underlying mesothelium are exposed to substances which are present in the peritoneal cavity. However, with uninjured peritoneum, only the mesothelial cells of the peritoneum are exposed to peritoneal fluid. Since TGFP is stimulatory to fibroblasts [4] and inhibitory to most epithelial cells [5], the TGFP injections into the peritoneal cavity were effective in promoting fibroblast proliferation and collagen synthesis only after surgical injury had exposed the underlying submesothelial tissues. It has been shown previously that the mitogenic activity of mesothelial cells from the site of peritoneal injury is inhibited by TGFP in vitro [63. There are several sources of TGF/3 in healing wounds. Platelets are the most concentrated source of TGFB, 20 mg/kg [7] identified in humans. Following surgical injury to the peritoneum, one of the first physiologic events to occur is platelet aggregation and degranulation. Although TGFP is secreted in a latent form [8], it is activated by changes in pH [9], such as would occur in ischemic tissue adjacent to sites of surgical injury. In addition, activated macrophages that occur in healing wounds secrete sialidase [lo] and proteases that are also capable of activating TGFfl [ll]. In skin incisions, after attracting macrophages [2] and fibroblasts [3], TGFP enhances extracellular matrix formation by increasing transcription of genes for collagen and fibronectin, decreasing secretion of proteases, increasing secretion of protease inhibitors, and increasing transcription of cellular receptors for matrix proteins [4]. In addition, TGFP has been shown to stimulate fibroblast contraction of the wound [ 121. Thus, TGFP could act to convert the filmy adhesions to thick fibrous adhesions by attracting macrophages and fibroblasts to the sites of injury and then stimulating fibroblasts to produce extracellular matrix proteins. The source of TGFP, either bone-derived or recombinant, did not alter the stimulation of adhesion formation postoperatively in these experiments. Both of these sources are primarily TGF/31. The presence of TGF/3 has been demonstrated in other wound healing models. TGFP has been assayed in subcutaneous rat wound chambers from Days 3-14 after wounding. Peak levels of TGFP were seen Days 7-11 after wounding, although levels were not measured im-
WILLIAMS
ET AL.:
TGF@ IN POSTOPERATIVE
ADHESION
FORMATION
69
FIG. 2. Light microscope photographs of adhesion formation (AD) to the uterine horns (UT) 2 weeks after induction of surgical injury in control (A) and TGFP-treated animals (B). The adhesions in the TGFP-treated animals contain more fibroblasts and inflammatory cells (darkly stained nuclei) than control (A). Mag. = X16; AD, adhesion; UT, uterine horns; F, fascia and muscle; P, peritoneum.
mediately after wounding when platelet degranulation would occur [ 131. These investigators demonstrated, however, that TGFP production continues after the initial injury has occurred. Activated lymphocytes and macrophages are a continuing source for TGFP in the healing wound after initial platelet degranulation has released TGFP. TGFP has previously been shown to induce the formation of granulation tissue containing inflammatory cells and proliferating fibroblasts when it is injected subcutaneously in mice [ 141. This phenomenon was also seen in some of the rats in this study. In two rats, one of the ip injections was erroneously injected intramuscularly resulting in the formation of localized fibrosis (Fig. 1C). In
contrast, this study demonstrates that TGFP is not active on intact peritoneal surfaces. Our inability to affect adhesion formation with a neutralizing TGFP antibody may be secondary to inadequate neutralizing activity of the antibody or more probably inadequate delivery of antibody to the site of injury. With only once a day drug administration, it is probable that TGFP produced during the remainder of the day was not neutralized. A continuous administration of TGFP antibody may circumvent this problem. These studies have demonstrated the ability of TGFP to increase postoperative adhesions in a rat model while not affecting intact peritoneum. Thus, TGFP may play a role in the formation of postoperative adhesion forma-
70
JOURNAL
OF SURGICAL
RESEARCH:
tion. Neutralization of TGFP activity postoperatively may present another avenue of preventing adhesion formation after surgery. By allowing the fibrinolytic pathways to digest the fibrin adhesions before invasion by fibroblasts, permanent adhesions may be prevented. REFERENCES 1. 2.
3.
4. 5.
6.
Hickey, M. J., and DiZerega, G. S. Recent advances in adhesion prevention. Contemp. OB/GYN. Update on Surgery. p. 14,199O. Wahl, S. M., Hunt, D.A., Wakefield, L. M., McCartney-Francis, N., Wahl, L. M., Roberts, A. B., and Sporn, M. B. Transforming growth factor @ (TGF-0) induces monocyte chemotaxis and growth factor production. Proc. Natl. Acad. Sci. USA 84: 5188, 1987. Postlethwaite, A. E., Keski-Oja, J., Moses, H. L., and Kang, A. H. Stimulation of the chemotactic migration of human fibroblasts by transforming growth factor j3. J. Exp. Med. 16’75: 251, 1987. Roberts, A. B., and Sporn, M. B. The transform- growth factor-b’s. Handbook Exp. Phurm. 95: 419, 1990. Coffey, R. J., Sipes, N. J., Bascom, C. C., Graves-Deal, R.. Pennington, C. Y., Weissman, B. E., and Moses, H. L. Growth madulation of mouse keratinocytes by transforming growth factors. Cancer Res 48: 1596,198s. Fukassawa, M., Yanogihara, D. L., Rodgers, K. E., and DiZerega, G. S. The mitogenic activity of peritoneal tissue repair cells: Control by growth factors. J. Surg. Res. 47: 42, 1989.
VOL.
I.
52, NO. 1, JANUARY
1992
Van Den Eijnden-van Raaij, A. J. M., Koornneef, I., and van Zoelen, E. J. J. A new method for high yield purification of type (3 transforming growth factor from human platelets. Biochem. Biophys. Res. Comm. 157: 16,1988. 8. Wakefield, L. M., Smith, D. M., Flanders, K. C., and Spom, M. B. Latent transforming growth factor-b from human platelets. J. Biol. Chem. 263: 7646, 1988. 9. Lawrence, D. A., Pircher, R., and Jullien, P. Conversion of a high weight latent P-TGF from chicken embryo fibroblasts into a low molecular weight active &TGF under acidic conditions. Biochem. Biophys. Res. Comm. 133: 1026,1985. 10. Pilatte, Y., Bignon, J., and Lambre, C. R. Lysosomal and cytosolic sialidases in rabbit alveolar macrophages: demonstration of increase lysosomal activity after in vivo activation with bacillus Biochem. Biophys. Acta 923: 150,1987. Calmette-Guerin. 11. Lyons, R. M., Keski-Oja, J., and Moses, H. L. Proteolytic activation of latent transforming growth factor-b from fibroblast-conditioned medium. J. Cell. Biol. 106: 1659, 1988. 12. Montesano, R., and Orci, L. Transforming growth factor b stimulates collagen-matrix contraction by fibroblasts: Implications for wound healing. Proc. Natl. Acad. Sci. USA 05: 4894, 1988. 13. Cromack, D. T., Sporn, M. B., Roberts, A. B., Merino, M. J., Dart, L. L., and Norton, J. A. Transforming growth factor @levels in a rat wound chambers. J. Surg. Res. 42: 622,1987. 14. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., Kehrl, J. H., and Fauci, A. S. Transforming growth factor type-@: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc. Natl. Acad. Sci. USA 83: 4167,1986.
OF SURGICAL
RESEARCH
52,65-70
(1992)
Effect of Transforming Growth Factor ,6 on Postoperative Formation and Intact Peritoneum
Adhesion
R. STAN WILLIAMS, M.D., ANN MARIE ROSSI,M.S., NASSER CHEGINI, PH.D., AND GREGORY SCHULTZ, PH.D. Department
of Obstetrics and Gynecology, Submitted
University
for publication
Press,
December
Gainesville,
Florida
32610
17, 1990
jury to the peritoneum at the time of surgery causes a rapid release of platelets, fibrinogen, histamine, and vasoactive kinins. With increased capillary permeability, the site of injury continues to release these products and fibrinous adhesions form between adjacent surfaces [ 11. If the fibrinolytic mechanisms are not inhibited by tissue ischemia or foreign bodies, the fibrinous adhesions may be digested and no permanent adhesion is formed. In the absence of adequate fibrinolytic mechanisms, the fibrin bands soon become infiltrated with fibroblasts which produce collagen and subsequently, become vascularized. The fibrin adhesion is transformed into a thick, vascular adhesion and becomes permanent. Most investigators have focused efforts to prevent postoperative adhesion formation on the early events of fibrin formation. Although this approach has had some success in preventing postoperative adhesion formation, at least 50% of patients still develop significant adhesions. Another avenue which may be complimentary is the prevention of fibroblast invasion of fibrinous adhesions with subsequent collagen formation and allowing natural fibrinolysis additional time to resolve fibrin bands. Peptide growth factors and their receptors appear to play important regulatory roles in healing other types of wounds and may also be involved in the development of adhesions at the sites of injured peritoneum postoperatively. Platelets contain several peptide growth factors which are released at the site of an injury includingplatelet derived growth factor, epidermal growth factor-like peptides, and transforming growth factor /3 (TGFP). TGFP is an extremely potent chemoattractant for macrophages, mononuclear leukocytes [2], and fibroblasts [3]. It also stimulates collagen and fibronectin synthesis by fibroblasts [4] and inhibits most epithelial cell growth [5]. These properties of TGF@ as well as the release of TGFP from platelets at sites of peritoneal injury make it a candidate for playing a pivotal role in the formation of permanent postoperative adhesions. Our hypothesis is that TGFP would have minimal effect on uninjured peritoneum, but following peritoneal injury, TGFfi released from platelets and other wound cells
Transforming growth factor @ (TGFB) is an extremely potent chemoattractant for macrophages, mononuclear leukocytes, and fibroblasts. It also acts as a potent stimulant for collagen and fibronectin synthesis and inhibits epithelial cell growth. TGF/3 plays an important role in healing many types of wounds, but its role in peritoneal adhesion formation is not known. These studies were performed to determine if TGF/3 could affect postoperative wound healing in a rat model. In the first experiment, 20 rats were divided into two groups and received either 2 pg TGFj3 or control diluent IP daily for 5 days after surgical injury to the uterine horns. The severity of the adhesions were graded 2 weeks postoperatively using a score of O-3. The TGF@ group showed a higher adhesion score at 2 weeks compared to control, 2.9 * 0.34 and 1.6 f 0.61, respectively (P < 0.001). On H&E stained sections of the adhesions, there was an increase in the number of both inflammatory cells and fibroblasts in the TGF/3treated animals. A comparison trial of bone-derived TGFB (a gift from Collagen Corporation, Palo Alto, CA) versus recombinant TGFB (a gift from Oncogen, Seattle, WA) versus control using the same protocol as above showed that both sources of TGF/3 were more effective in promoting postoperative adhesions when compared to controls, and there was no difference between TGFB groups, 3.0 f 0 for both TGFj3 groups, and 2.2 f 0.91 for control (P < 0.001). TGFj3 did not induce adhesions when given in the same dosage for 5 days ip in unoperated rats, although it did induce fibrosis if injected into the abdominal wall. These data suggest that TGF@ may play a role in stimulating fibrosis in postoperative fibrinous adhesions, but has no effect on intact mesothelium. Prevention of TGF/3 production postoper0 1992 atively may help prevent adhesion formation. Academic
of Florida,
Inc.
INTRODUCTION
The peritoneum is composed of surface mesothelial cells and submesothelial stroma containing fibroblasts, collagen, mast cells, macrophages, and blood vessels. In65
Copyright 0 1992 by All rights of reproduction
0022404/92 $1.50 Academic Press, Inc. in any form reserved.
66
JOURNAL
OF SURGICAL
RESEARCH:
would promote formation of severe adhesions by stimulating fibroblasts, attracting inflammatory cells, and stimulating the synthesis of collagen and angiogenesis. These studies were performed to determine any effect exogenous TGFP has on uninjured peritoneum and to measure the effect of TGFP on postoperative peritoneal wound healing in a rat model.
VOL.
52, NO. 1, JANUARY
1992
of the rats received PBS with RSA ip for 5 days. Rats were sacrificed and evaluated as described above. In the last experiment, a neutralizing antibody to TGFP (R&D, Minneapolis, MN) was tested to determine if postoperative adhesion formation could be altered by interfering with endogenous TGFfl action. This rabbit antiporcine TGF@ antibody neutralizes both TGFP-1 and TGF@-2 according to the manufacturers. A group of 20 rats underwent surgery and postoperatively MATERIALS AND METHODS half received 10 pg of TGFP-blocking antibody ip daily Female Sprague-Dawley rats weighing approximately for 5 days, and the remainder received PBS injections. 250 g were housed in an environmentally controlled vi- The animals were sacrificed and evaluated as described varium and maintained with water and rat chow ad Eibi- above. turn. Experimental use of these animals was approved by After grading the adhesions, the uterine horns with the University of Florida Institutional Animal Care and the adherent tissues were removed and fixed with Use Committee. Animals were anesthetized with intraBouins solution overnight. The tissue was then dehyperitoneal injection of Rompun 5 mg/kg and Ketaset 40 drated, paraffin embedded, and 5- to 6-pm sections were mg/kg. A lower midline incision was made exposing the cut and counterstained with hematoxylin and eosin Y. uterine horns and each uterine horn was injured with a All materials for light microscopy were purchased from l-cm burn by bipolar cautery, a l-cm crush injury, and a Polysciences (Warrington, PA). l-cm scrape until bleeding occurred. The midline inciStatistical analysis of adhesion scores between groups sion was closed with a single layer closure using 3-O was performed with the use of nonparametric tests. Dexon. When two groups were compared, a nonparametric The first experiment was performed to determine the Mann-Whitney rank test was used and when more than effect of TGFP on postoperative adhesion formation. A two groups were compared a Kruskal-Wallis One-Way group of 20 rats were divided into two treatment groups Anova by Rank test was performed. Statistical analysis of 10 rats each. One group received 500 ,ul of phosphate- was performed with a computer statistical program, buffered saline (PBS) containing 2 gg TGFP (bone de- Statgraphics (Statistical Graphics Corporation, Rockrived, a gift from Collagen Corporation, Palo Alto, CA) ville, MD). (BTGFP) combined with 500 pg rat serum albumin (RSA) as a carrier, injected intraperitoneally (ip) each RESULTS day for 5 days postoperatively. The remaining 10 rats were injected ip with 500 ~1PBS containing 500 pg RSA Effect of Bone-Derived TGFfl on Postoperative Adhesion daily for 5 days postoperatively. Formation Two weeks postoperatively rats were sacrificed, and Exogenous TGF/3 significantly increased postoperaadhesions were graded using a severity score of O-3 by an tive adhesion formation. Adhesions were more dense, observer who was masked to the treatment. A score of 0 vascular, and more extensive after TGFP treatment in was given for no adhesions; grade 1 represented thin, rats with surgically induced trauma to the peritoneal filmy adhesions; grade 2, moderately thick adhesions; surface. TGFP-treated animals showed a significantly and grade 3 for dense, vascular adhesions. higher adhesion score postoperatively as compared to In the second experiment, 30 rats were used to comthe control group, mean scores 2.9 k 0.34 (standard depare the effects of bone-derived TGF/3 and recombinant viation) for TGF/?-treated rats and 1.6 f 0.61 for control TGFP on postoperative adhesion formation. The rats rats (P < 0.001) (Fig. 1). were divided into three groups with 10 rats per group. Although boned-derived TGF@ was highly purified, it Treatment groups included (1) BTGF/3 with RSA + surwas possible that minor contaminants extracted from gery, (2) recombinant TGF/3 (a gift from Oncogen, Seatbone could have influenced the biologic effect seen in tle, WA) (RTGFP) with RSA + surgery, and (3) RSA in this experiment. To further evaluate the effect of TGFfi PBS + surgery. Dosages of TGFP and RSA, the injection regimen, and evaluation were the same as described alone, recombinant TGFB was used as another source of TGF/3 and compared with the effect of bone-derived above. TGF@. In a third experiment, a group of 30 rats which did not undergo surgery were divided into three groups to test Comparison of Bone-Derived TGFfl and Recombinant whether TGFP would stimulate adhesion formation in TGF/3 on Postoperative Adhesion Formation the absence of peritoneal injury. Ten rats were injected intraperitoneally for 5 days with 500 ~1of PBS with 2 pg In the comparison of BTGFB and RTGFP, both RTGFP combined with RSA. Another 10 rats received groups showed a significantly higher rate of adhesion BTGFB combined with RSA for 5 days. The remainder formation postoperatively when compared to control (P
WILLIAMS
ET AL.: TGFP
IN POSTOPERATIVE
ADHESION
67
FORMATION
FIG. 1. Adhesion formation in control (A) and TGFP-treated (B) animals 2 weeks after surgical demonstrates a fibrotic nodule from an inadvertant injection of TGFp into the abdominal wall.
injury
(See text for details).
Panel C
68
JOURNAL
OF SURGICAL
RESEARCH:
< 0.001) and TGF/3 from either source was equally effective in promoting adhesions. Mean adhesion scores were 3.0 +- 0.0 in both TGF@ groups and 2.2 f 0.91 in the control group. This experiment confirmed the results of the first experiment and demonstrated that adhesion formation postoperatively was stimulated by TGF/3 and not any contaminating substance. In this system, recombinant TGFP was seen to be just as active as natural bone-derived TGFP. Effect of TGFP on Adhesion Rats
Formation
in Unoperated
Although TGFB increased adhesion formation if previous peritoneal injury had occurred, no stimulation of adhesions was seen in unoperated rats given ip TGF/3 regardless of the source of TGFP. No difference was seen between groups in the unoperated rats when TGF/3 was compared with control. No adhesions could be identified in either group. Effect of TGF@ Blocking Antibody Formation
on Adhesion
Daily injections of a neutralizing antibody to TGFP was unable to prevent postoperative adhesion formation. TGFP-blocking antibody treatment postoperatively showed no decrease in adhesion formation when compared to controls, 1.85 + 0.87 and 1.7 + 0.92, respectively (P > 0.05). Light Microscopy Figs. 2A and 2B show a portion of uterine horn with tissue adhesions attached from injured control (A) and injured, TGFP-treated (B) rats. The type and cellular adhesion was similar in both cases, but there was an obvious increase in both inflammatory cells and fibroblasts present in the TGFP-treated animals compared to the control animals. DISCUSSION
The role of peptide growth factors in peritoneal wound healing has not been previously evaluated in viuo. It is hypothesized that TGFB may play an important role in peritoneal wound healing and adhesion formation by attracting inflammatory cells [2] and fibroblasts [3] to the site of injury. The subsequent stimulation of collagen and fibronectin synthesis by TGFP [4] would promote the formation of permanent adhesions at the sites of early fibrinous adhesions. This study was designed to measure the effect of exogenous TGFP on postoperative peritoneal wound healing as well as intact peritoneum. The comparison between these groups is important as any effect seen on intact peritoneum could not represent a physiologic effect, as fibrosis normally does not occur in the peritoneal cavity except at sites of injury.
VOL.
52, NO. 1, JANUARY
1992
This study demonstrated that TGFP increases the severity of adhesion formation after surgical injury in a rat model, but has no effect on uninjured peritoneum. Histologically, there were increased inflammatory cells and fibroblasts present in the adhesions of TGFP-treated animals. These results support the hypothesis that TGFfi stimulated chemotaxis of inflammatory cells and fibroblasts and subsequently stimulated collagen synthesis inducing fibrosis in fibrinous exudates postoperatively. A possible explanation of the different affect of TGFP on injured vs intact peritoneum might lie in TGFps effect on fibroblasts vs mesothelial cells. With peritoneal injury, fibroblasts and macrophages of the underlying mesothelium are exposed to substances which are present in the peritoneal cavity. However, with uninjured peritoneum, only the mesothelial cells of the peritoneum are exposed to peritoneal fluid. Since TGFP is stimulatory to fibroblasts [4] and inhibitory to most epithelial cells [5], the TGFP injections into the peritoneal cavity were effective in promoting fibroblast proliferation and collagen synthesis only after surgical injury had exposed the underlying submesothelial tissues. It has been shown previously that the mitogenic activity of mesothelial cells from the site of peritoneal injury is inhibited by TGFP in vitro [63. There are several sources of TGF/3 in healing wounds. Platelets are the most concentrated source of TGFB, 20 mg/kg [7] identified in humans. Following surgical injury to the peritoneum, one of the first physiologic events to occur is platelet aggregation and degranulation. Although TGFP is secreted in a latent form [8], it is activated by changes in pH [9], such as would occur in ischemic tissue adjacent to sites of surgical injury. In addition, activated macrophages that occur in healing wounds secrete sialidase [lo] and proteases that are also capable of activating TGFfl [ll]. In skin incisions, after attracting macrophages [2] and fibroblasts [3], TGFP enhances extracellular matrix formation by increasing transcription of genes for collagen and fibronectin, decreasing secretion of proteases, increasing secretion of protease inhibitors, and increasing transcription of cellular receptors for matrix proteins [4]. In addition, TGFP has been shown to stimulate fibroblast contraction of the wound [ 121. Thus, TGFP could act to convert the filmy adhesions to thick fibrous adhesions by attracting macrophages and fibroblasts to the sites of injury and then stimulating fibroblasts to produce extracellular matrix proteins. The source of TGFP, either bone-derived or recombinant, did not alter the stimulation of adhesion formation postoperatively in these experiments. Both of these sources are primarily TGF/31. The presence of TGF/3 has been demonstrated in other wound healing models. TGFP has been assayed in subcutaneous rat wound chambers from Days 3-14 after wounding. Peak levels of TGFP were seen Days 7-11 after wounding, although levels were not measured im-
WILLIAMS
ET AL.:
TGF@ IN POSTOPERATIVE
ADHESION
FORMATION
69
FIG. 2. Light microscope photographs of adhesion formation (AD) to the uterine horns (UT) 2 weeks after induction of surgical injury in control (A) and TGFP-treated animals (B). The adhesions in the TGFP-treated animals contain more fibroblasts and inflammatory cells (darkly stained nuclei) than control (A). Mag. = X16; AD, adhesion; UT, uterine horns; F, fascia and muscle; P, peritoneum.
mediately after wounding when platelet degranulation would occur [ 131. These investigators demonstrated, however, that TGFP production continues after the initial injury has occurred. Activated lymphocytes and macrophages are a continuing source for TGFP in the healing wound after initial platelet degranulation has released TGFP. TGFP has previously been shown to induce the formation of granulation tissue containing inflammatory cells and proliferating fibroblasts when it is injected subcutaneously in mice [ 141. This phenomenon was also seen in some of the rats in this study. In two rats, one of the ip injections was erroneously injected intramuscularly resulting in the formation of localized fibrosis (Fig. 1C). In
contrast, this study demonstrates that TGFP is not active on intact peritoneal surfaces. Our inability to affect adhesion formation with a neutralizing TGFP antibody may be secondary to inadequate neutralizing activity of the antibody or more probably inadequate delivery of antibody to the site of injury. With only once a day drug administration, it is probable that TGFP produced during the remainder of the day was not neutralized. A continuous administration of TGFP antibody may circumvent this problem. These studies have demonstrated the ability of TGFP to increase postoperative adhesions in a rat model while not affecting intact peritoneum. Thus, TGFP may play a role in the formation of postoperative adhesion forma-
70
JOURNAL
OF SURGICAL
RESEARCH:
tion. Neutralization of TGFP activity postoperatively may present another avenue of preventing adhesion formation after surgery. By allowing the fibrinolytic pathways to digest the fibrin adhesions before invasion by fibroblasts, permanent adhesions may be prevented. REFERENCES 1. 2.
3.
4. 5.
6.
Hickey, M. J., and DiZerega, G. S. Recent advances in adhesion prevention. Contemp. OB/GYN. Update on Surgery. p. 14,199O. Wahl, S. M., Hunt, D.A., Wakefield, L. M., McCartney-Francis, N., Wahl, L. M., Roberts, A. B., and Sporn, M. B. Transforming growth factor @ (TGF-0) induces monocyte chemotaxis and growth factor production. Proc. Natl. Acad. Sci. USA 84: 5188, 1987. Postlethwaite, A. E., Keski-Oja, J., Moses, H. L., and Kang, A. H. Stimulation of the chemotactic migration of human fibroblasts by transforming growth factor j3. J. Exp. Med. 16’75: 251, 1987. Roberts, A. B., and Sporn, M. B. The transform- growth factor-b’s. Handbook Exp. Phurm. 95: 419, 1990. Coffey, R. J., Sipes, N. J., Bascom, C. C., Graves-Deal, R.. Pennington, C. Y., Weissman, B. E., and Moses, H. L. Growth madulation of mouse keratinocytes by transforming growth factors. Cancer Res 48: 1596,198s. Fukassawa, M., Yanogihara, D. L., Rodgers, K. E., and DiZerega, G. S. The mitogenic activity of peritoneal tissue repair cells: Control by growth factors. J. Surg. Res. 47: 42, 1989.
VOL.
I.
52, NO. 1, JANUARY
1992
Van Den Eijnden-van Raaij, A. J. M., Koornneef, I., and van Zoelen, E. J. J. A new method for high yield purification of type (3 transforming growth factor from human platelets. Biochem. Biophys. Res. Comm. 157: 16,1988. 8. Wakefield, L. M., Smith, D. M., Flanders, K. C., and Spom, M. B. Latent transforming growth factor-b from human platelets. J. Biol. Chem. 263: 7646, 1988. 9. Lawrence, D. A., Pircher, R., and Jullien, P. Conversion of a high weight latent P-TGF from chicken embryo fibroblasts into a low molecular weight active &TGF under acidic conditions. Biochem. Biophys. Res. Comm. 133: 1026,1985. 10. Pilatte, Y., Bignon, J., and Lambre, C. R. Lysosomal and cytosolic sialidases in rabbit alveolar macrophages: demonstration of increase lysosomal activity after in vivo activation with bacillus Biochem. Biophys. Acta 923: 150,1987. Calmette-Guerin. 11. Lyons, R. M., Keski-Oja, J., and Moses, H. L. Proteolytic activation of latent transforming growth factor-b from fibroblast-conditioned medium. J. Cell. Biol. 106: 1659, 1988. 12. Montesano, R., and Orci, L. Transforming growth factor b stimulates collagen-matrix contraction by fibroblasts: Implications for wound healing. Proc. Natl. Acad. Sci. USA 05: 4894, 1988. 13. Cromack, D. T., Sporn, M. B., Roberts, A. B., Merino, M. J., Dart, L. L., and Norton, J. A. Transforming growth factor @levels in a rat wound chambers. J. Surg. Res. 42: 622,1987. 14. Roberts, A. B., Sporn, M. B., Assoian, R. K., Smith, J. M., Roche, N. S., Wakefield, L. M., Heine, U. I., Liotta, L. A., Falanga, V., Kehrl, J. H., and Fauci, A. S. Transforming growth factor type-@: Rapid induction of fibrosis and angiogenesis in vivo and stimulation of collagen formation in vitro. Proc. Natl. Acad. Sci. USA 83: 4167,1986.
