- Email: [email protected]
Healing of experimental defects on the stomach: Influence of autologous and heterologous materials
JOURNAL
OF SURGICAL
RESEARCH
36,
35-49 (1984)
Healing of Experimental Defects on the Stomach: Influence Autologous and Heterologous Materials GIOVANNI Department
B. RATTO AND GIOVANNI
of Semeiotica
Chirurgica
I, University
of
MOTTA’
of Genoa,
Italy
Submitted for publication December 15, 1982 This study was undertaken to evaluate the influence of various autologous and heterologous structures on gastric defect healing. In 2 10 rats, standard gastric defects were closed using the following structures as patches: (1) pedicled colonic patches, (a) mucosal layer to lumen, and (b) peritoneal surface to lumen; (2) leiomuscular colonic buttons; and (3) lyophilized bovine tendon collagen sponges. The specimens were removed from the defect arcas at standard time intervals ranging from 2 to 12 months after surgery; the samples were then examined by light microscopy (hematoxylin-eosin; Azan Mallory; Grimelius methods) and by trasmission and scanning electronmicroscopy. The colonic mucosa appeared to be the most resistant of the structures to gastric secretions; it showed no mucosal disruption during the entire period of observation and also underwent a progressive gastric-like morphological transformation. In all the other cases,the healing process involved the initial development of granulation tissue followed by scar tissue with gastric mucosa growing over the scar tissue to fill the defects.
The gastric mucosa is a specialized structure which is highly resistant to gastric acidity even under conditions of maximal stimulation acting in this way as a limiting harrier to ion diffusion [5, 61. Surgical wounds are still known to heal rapidly in spite of the alterations which they cause in the mucosa layer [8-lo]. The repair process occurs during the first 14 to 21 days through the formation of a nonspecific granulation and scar tissue. The rapid healing of these gastric wounds is due to the high rate of collagen synthesis [ 121. The purpose of this study was to evaluate the influences of various autologous and heterologous structures on gastric defect healing in rats. In investigating this phenomenon, pedicled coionic patches (mucosal or peritoneal surface to lumen), leiomuscular colonic grafts, and lyophilized bovine tendon collagen sponges were used to close a standard gastric defect. In particular we tried to examine (1) the rate of healing of these standard gastric defects which were surgically closed, (2) the evolution of the healing processes, (3) the resistance of
the different structures to gastric secretions, and (4) the morphological changes of these materials after chronic exposure to the gastric acid environment. MATERIALS
AND METHODS
The experiments were performed on 2 10 male Wistar rats each weighing about 250 g. The animals were fasted for 24 hr but allowed access to water. Under ether anaesthesia the abdomen was opened through a midline incision. A standard hole ( 10 X 6 mm) was made on the posterior surface of the gastric body; the posterior wall was chosen to avoid traction on the vascular pedicles of the colonic patches used to plug the defects. The animals were divided into six groups and treated as follows: Group A. In 50 rats, a lo-mm segment of the left half of the transverse colon and its vascular supply were isolated. An end-to-end anastomosis was performed on the resected colon to restore its continuity. The isolated segment was opened longitudinally on its antimesocolic side and rinsed with saline. It was then sutured, mucosal surface to lumen, in order to close the gastric defect.
’ Present address: Direttore Cattedra di Semeiotica Chirurgica I Universita degli Studi, Viale Benedetto XV, 10, 16132 Genova, Italy. 35
0022-4804184 $1 SO Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.
36
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. I, JANUARY
Group B. Thirty rats underwent the same procedure as in group A but were chronically stimulated by daily subcutaneous injections of histamine in beeswax ( 10V3M) for 2 months [3]. Group C. In 40 rats the colonic mucosa was removed from the patch by blunt dissection; care was taken not to leave residual colonic mucosa. The patch was then anastomosed, leiomuscular layers to lumen, in order to close the standard gastric defect. Group D. In 40 rats the colonic segment was longitudinally opened near the insertion of the mesocolon; the patch was then reversed and sutured, peritoneal surface to lumen, over the gastric opening. The colonic mucosa, however, was first removed by blunt dissection to avoid massive secretion of mucus into the peritoneal cavity.
1984
Group E. In 40 rats standard gastric defects were plugged using a piece of native bovine collagen sponge extracted from the Achilles tendon using the Schor and Court, and Elsdale and Bard methods [23-271. In a control group of 10 rats, standard gastric holes were directly sutured and a colonic segment was isolated and sutured back into its original location to assessany morphological changes related to the procedure.
All the anastomoses were performed using an interrupted single layer 7-O Prolene (polypropylene) extramucosal suture. Within each group, the animals were randomly divided into 4 subgroups which were reoperated at 2,4,6, and 12 months after the initial surgery. At this time the defect areas and their surrounding tissues were removed.
FIG. 1. Group A: mucosal surface of a colonic patch 2 months after implantation. It appears to be composed of polygonal units, each with a central crypt surrounded by epithelial cells. The units are demarcated by shallow furrows. This superficial epithelial layer seems to be grossly intact (SEM X 1500).
RATTO
AND MOTTA:
GASTRIC
DEFECT HEALING
37
In group B all the samples were collected 2 the critical point method [ 191, and coated with months after implantation. The specimens gold by sputtering. For TEM, the specimens were first cut in half and then divided into were rinsed with water and stained in alcoholic four strips along the minor axis [ 15, 161. Each uranyl acetate, dehydrated through ethanols strip was then examined by serial section to and propylene oxide, and then embedded in quantify the extension of the morphological Epon Araldite. changes to be investigated. The first half was A statistical analysis of the mortality rate studied under light microscopy (LM) and the was carried out by using the 2 X 2 chi-square other under trasmission (TEM) and scanning test. The samples were divided into four ar(SEM) electronmicroscopy. In preparing for bitrary subgroups based on their morphology LM, the tissues were fixed in 10% formalin in order to statistically evaluate the evolution for routine staining (hematoxylin-eosin and of the healing process. The patches in group Azan Mallory); the same tissues were then A were classified in the following manner: (1) treated by Grimelius method to assess the no changes, (2) changes involving less than presence of diffise endocrine system cells [ 111. 50% of the patch mucosa, (3) changes inFor TEM and SEM, the specimens were fixed volving more than 50% but not the entire muin a 2% glutaraldehyde buffered solution and cosal layer, and (4) changes involving the enpostfixed in a 1% sodium cacodylate osmium tire mucosal layer. The patches in groups C, tetroxide. For SEM, the samples were dehy- D, and E were instead classified as ( 1) no drated in a graded series of alcohols, dried by changes, (2) loss of the original architecture
FIG. 2. Group A: a tubular mucous gland of the colonic patch 2 months after implantation. The columnar cells exhibit an active mucus production and some cells are seen discharging mucus into the crypt. (TEM X3700).
38
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
wit.hout
any mucosal overgrowth, (3) loss of the : original architecture with incomplete mucosd overgrowth, (4) loss of the original ar-
FIG. 3. Group A: a colonic patch 2 months except for an abundant chronic inflammatory
1984
chitecture with complete mucosal overgrow These phenomena were analyzed statistic; by means of a 5 X 4 chi-square test.
after implantation. The graft maintains infiltrate (LM X 100).
its original
structure
RATTO
AND
MOTTA:
GASTRIC
RESULTS
At the time of removal, all the defects appeared to be closed by vital tissues. One death was observed in group C and three deaths in group E, all within the first week after surgery. These deaths were caused by dilluse peritonitis due to patch perforation. One death was observed in group D (24 hr after surgery) caused by volvulus of a jejunal loop which became necrotic. A statistically significant difference in survival rate was observed only between groups A and E (P < 0.05). Group A. The mucosal layer of all the patches removed 2 months after implantation displayed a typical colonic mucosal morphology as seen under SEM (Fig. l), TEM (Fig. 2), and LM (Fig 3). Scattered necrosis or erosions of the epithelial layer were not observed. Under LM, however, initial signs of gastric-like transformation were seen near the colon-gastric anastomoses as documented
FIG.4. rough
DEFECT
39
HEALING
by loss of muciparity and an increase in cytoplasmatic acidophilia. Four months after surgery this gastric-like transformation was clearly seen in one-third of the cases, but it never involved more than 50% of the patch mucosa and was generally confined to the superficial epithelial layers. Six months after implantation the gastric-like transformation involved more than 50% of the patch mucosa in 4 1% of the cases. Under TEM examination, the epithelial cells of the patches showed ultrastructural patterns similar to that of gastric mucosa, including a significant amount of zymogen granules (Fig. 4). Twelve months after implantation the patches were completely covered with gastric-like mucosa in 78% of the samples, as demonstrated by LM, TEM, and SEM examination (Fig. 5). Under LM a near complete loss of affinity for mucus staining and an increase in the number of parietallike cells, as well as an increase of the glandular crypts were observed (Fig. 6). In 22% of the
Group A: the epithelial patch cells 6 months afier implantation; surfaced endoplasmatic reticulum are evident. (TEM X8400).
secretory
granules
as well as
40
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG. 5. Group A: the mucosal surface of colonic patch 12 months after implantation. The superficial epithelial layer is comprised of mucussecreting cells with morphological patterns very similar to those of normal gastric mucosa (SEM X300).
cases the original colonic morphology was preserved but only in some deep residual mucosal islands. Over the entire period of observation the manifestation of this gastric-like transformation was statistically significant (P < 0.001). Changes in the distribution and number of cells belonging to the diffuse endocrine system normally found in the patch mucosa were not statistically relevant (Fig. 7). The so-called “tuft cells” [ 131 of the gastrointestinal mucosa, were no longer observed on the mucosal surface of the colonic patches after implantation on the stomach (Fig. 8). Group B. The specimens removed 2 months after implantation showed morphological characteristics similar to those of group A; the only difference was the presence of some scattered superficial epithelial erosions (P > 0.05). Group C. The inner surface of all the patches
removed 2 months after implantation appeared to be made up of an irregular fibrous mesh work (Fig. 9); the patch wall was completely replaced by granulation tissue (Fig. 10). An epithelialization process, which spread from the anastomoses, was constantly observed at 4 months after implantation. However, this process was found to be complete in only 20% of the cases. At 6 to 12 months the process was complete in all instances (P < 0.00 1). The defects were plugged under the mucosal layer by scar tissue. Group D. Two months after implantation the serosal surface appeared to be extensively altered as demonstrated by the complete destruction of the mesothelial cell layer. Here too, under LM the defects were seen to be closed by granulation tissue, which totally replaced the normal colonic wall components.
RAT-TO AND MOTTA:
GASTRIC
DEFECT HEALING
41
FIG.6. Group A: a colonic patch 12 months after implantation; the mucosal layer shows a loss of affinity for mucus staining and an increase in cytoplasmatic acidophilia. (LM X80).
Four months after implantation, an overgrowth of gastric mucosa starting from the site of the colon-gastric anastomosis was observed. This neomucosal layer succeeded in closing the defect in only one case. Six and twelve months after surgery the patches appeared completely covered by gastric mucosa (P < 0.001). This neomucosa displayed an ultrastructural morphology (TEM) and cells of the diffuse endocrine system (Grimelius method) which were identical to those found in the normal gastric mucosa. The patch wall under the mucosal layer consisted mainly of scar tissue (Fig. 11). Group E. Two months after surgery all the defects were plugged by a granulation tissue rich in collagen fibers and new vessels. The gastric epithelium was seen growing over this granulation tissue into the gap. The new gastric mucosa had restored complete continuity to the defect in 80% of the cases 6 months after
surgery and in all samples 12 months after implantation (P < 0.001). In the control group the only change was the formation of foreign body granulomas around the sutures. DISCUSSION
The present study demonstrated that the healing process restored in all cases, a continuous and morphologically normal mucosal layer over the surgical defects regardless of the graft material employed. It also revealed, however, that the steps involved in the healing process were not identical for each material used. This finding was due to the different biological properties inherent in each of the materials employed, providing them with varying degrees of resistance and adaptability to the gastric acid environment. The colonic mucosal layer was found to be the best struc-
42
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG. 7. Group A: the colonic mucosa 6 months after implantation. (A) The presence of numerous argentin cells is evident (LM X40); (B) a detailed portion of the above picture, showing the intracytoplasmatic argentaffin granules (LM X 1000).
ture used to close the defects. It showed the highest degree of resistanceagainst gastric secretions. It was the only material capable of conserving its structural integrity. In fact no damage was seen in the mucosa exposed to basalsecretory levels, while under chronic histamine stimulation, it presented only some scattered superficial epithelial erosions. The leiomuscular colonic buttons and the peritoneal surface to lumen patches were instead all found to be completely digestedeven under basalsecretory conditions; and for this reason they were not exposed to chronic histamine stimulation. Finally, the colonic mucosa ap-
peared to be more resistant to perforation; this finding, however, was statistically significant only when compared to the heterologous collagen sponge.How the colonic mucosa was able, in our experimental model, to withstand gastric secretions without development of ulcerations and interstitial hemorrhage remains an unanswered question. Studies done by Ritchic [20, 2 l] and Lawson [ 141on the ability of the gastric mucosa to resist ulcer formation showed that its high degree of resistance to acid-peptic digestion was related to the morphological or functional makeup of the mucosa. Because the colonic mucosa presented
RATTO
AND
MOTTA:
GASTRIC
DEFECT
HEALING
43
FIG. l-Continued
no signs of histological damage we must assume that its functional or morphological makeup confers a high degree of resistance. Since the blood supply to the patch was carefully preserved, it may be possible that the absence of lesions on the colonic patches is in part the result of the ability of the well vascularized tissue to effectively clear or neutralize any H+ which might enter the graft. We also observed both grossly and microscopically, a highly active production of mucus by the patch mucosa. However, this mucus gel alone is not likely to directly protect the epithelial surface from the acid [l-25]. In fact, the mechanism through which the colonic
mucosa deals with hydrogen ions is very complex and probably consists of a combined process of absorption and neutralization [4-221. In an experimental study, Chung [4] stated that the colonic mucosa is intolerant to acid at any pH below 2 if exposed for 90 min or more. Under these conditions, he observed structural changes in the form of edema, sloughing, and coagulative necrosis in the coionic mucosa. Nevertheless, Chung concluded that it is not possible to determine whether such adaptation occurs in the colonic mucosa as a result of chronic exposure to an acid environment. The discrepancy between our results and those of Chung may be attributed
44
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG.8. Control group: the mucosal surface of a colonic segment 2 months after isolation and reimplantation in its original site. A tuft cell is clearly visible, its characteristic “tuft” of microvilli stands out against the shorter projections of the surrounding cells. (SEM X7000).
to either the difference in the length of the observation periods, which would allow a progressiveadaptation of the colonic mucosa to the gastric environment, or to the difference in the intensity of the acid insult on the colon. In Chung’s study a segmentof colon was cannulated and perfused with acid solutions (pH ranging between 6 and 1.3) while our colonic patches were chronically exposed to physiological pHs. We measuredthe gastric acid output in a separategroup of 10 anesthetized rats fasted for 48 hr both under basal conditions and after histamine administration ( lop3 M). In this instance our results were similar to those reported by Bunce et al. [2]. Furthermore, the colonic mucosa was also seen to provide adequateprotection for the underlying muscolar layers for the entire period of study.
Our findings demonstrated that the colonic mucosa was not only highly resistant to the gastric secretionsbut also underwent a process of morphological adaptation to its new environment. This adaptation consisted of a gastric-like transformation which initially began near the anastomosesand involved only the most superficial epithelial layers. This observation would seemto indicate that the gastric-like transformation depends on a direct contact with gastric secretions. Even though the gastric mucosa is known to repair mucosal erosions rapidly, we were lead to interpret our findings to represent gastric-metaplasiafor the following reasons: (1) signs of mucosal sloughing or atrophy were not observed at any time in group A, (2) we also found intermediate morphological stages(lossof muciparity
RATTO
FIG. 9. Group after implantation.
C: The luminal The surface
AND
MOTTA:
GASTRIC
DEFECT
HEALING
45
surface of a colonic patch denuded of its original mucosal layer 2 months is made up of an irregular fibrous meshwork (SEM X600).
and ! an increase in cytoplasmic acidophilia), and (3) the time period necessary for gastric rep; tir through mucosal overgrowth, as seen ing roups C, D and E, was considerably shorter thal n that of Group A. In group A, complete rest oration of gastric-like mucosal continuity
was consistently observed but only in tl lose patches which were examined 1 year after implantation (P < 0.001). The meaning of this gastric-like transformation remains uncl lear. In this regard we have initiated further investigations to evaluate the possibility of s 3me
46
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
36, NO.
1, JANUARY
1984
FIG. 10. Group C: a colonic patch denuded of its original mucosal layer 2 months after The wall of the patch is composed of an inflammatory granulation tissue (LM X 100).
chal ages in functional activity associated with the morphological transformations we observed. Presently our observations under LM
implantation.
and TEM would seem to indicate that the cells of the patch mucosa do assume a gast.SiClike functional capacity as demonstrated by
RATTO
AND
MOTTA:
GASTRIC
DEFECT
HEALING
FIG. 11. Group D: B colonic patch 12 months after implantation. The graft appears to be covered by gastric-like mucosa, while the underlying layers are made up of a partially hyalinized scar tissue (LM X 100).
the presence of chief cells containing zymogen gra nules. The distribution of cells belonging to1 :he diffuse endocrine system remained con-
stant even after chronic gastric exposure. ? ‘his finding supports the presence of some en’docrine function in the patch mucosa, reveal ing
48
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
36, NO.
1, JANUARY
1984
another area which calls for further investi- the subsequent loss of this property due to its digestion by tissue collagenases [24]. gation. The serosal and leiomuscular colonic strucACKNOWLEDGMENTS tures presented a minor resistive capacity to the gastric secretions. They were completely We express our most sincere thanks to Professor A. digested and substituted by granulation and Riva, head of the Institute of Human Anatomy at the scar tissues. These structures were therefore University of Cagliari, Italy, and Professor A. Badini and less apt to favor the healing of the gastric de- Dr. M. Truini of the Institute of Pathological Anatomy the University of Genoa, Italy, for their helpful colfects with complete restitutio ad integrum to oflaboration with this study. all the layers of the gastric wall. These materials did, however, appear to be able to plug gastric REFERENCES holes in almost all cases. In both groups, the granulation and scar tissues acted as good I. Allen, A., and Garner, A. Mucus and bicarbonate secretion in the stomach and their possible role in supports for the complete overgrowth of the mucosal protection. Gut 21: 249, 1980. gastric mucosa which closed the gaps. The 2. Bunce, K. T., Grewal, M., and Parsons, M. E. Pepsin high rate of healing of the gastric wounds in secretion in the isolated rat stomach preparation. groups C and D probably depends on the Physiol. Sot.. 55P, July 1979. characteristics of the collagen metabolism of 3. Code, C. F., and Varco, R. L. Chronic Histamine action. Proc. Sot. Exp. Biol. Med. 44: 475, June 1940. the rat stomach. In an experimental study on 4. Chung, R. S. K., Johnson, G. M., and Denbeste, L. rats Gottrup [8] demonstrated that wound The handling of H+ by the colon: The biophysical healing is more rapid in the stomach than in aspect of esophageal replacement. J. Surg Rex 33: most other organs, for example skin or tendon. 538, 1977. He also stated that the glandular oxyntic part, 5. Davenpot, H. W., Warner, H. A., and Code, C. F. Functional significance of gastric mucosal barrier to the area in which we created the defects, is sodium. Gastroenterology 47: 142, 1964. the most biochemically active zone of the Davenport, H. W. Salicylate damage to the gastric stomach. He also found that the concentration mucosal barrier. N. Engl. J. Med. 276: 1307, 1967. of collagen in this zone increased significantly Elsdale, T., and Bard, J. Collagen substrata for studies within 40 days after surgery and superceded on cell behaviour. J. Cell. Biol. 54: 626, 1972. Gottrup, F. Healing of incisional wounds in stomach the presurgical level by 90%. Finally it was and duodenum. A biomechanical study. Amer. J. also shown that the rate of collagen synthesis Surg. 14th 296, 1980. correlates well with the gain in mechanical 9. Gottrup, F. Healing of incisional wounds in stomach strength of wounds [9]. These findings could and duodenum. Collagen distribution and relation to provide an explanation for the rapid develmechanical strength. Amer. J. Surg. 141: 222, 198 I. opment of adequate mechanical resistance to 10. Gottrup, F. Healing of incisional wounds in stomach and duodenum. The influence of aging. Acta Chir. our surgical lesions. Stand. 147: 363, 1981. The heterologous collagen sponge allowed Il. Grimelius, L. The argyrophil reaction in islet cells of a similar healing process even though it apadult human pancreas studied with a new silver nitrate peared less resistant to gastric secretions (P procedure. Acta Sot. Med. Upsal. 73(56): 27 I, 1968. 12. Hastings, J. C., Van Winkle, W., Barker, E., Hines, < 0.05). In a previous study [ 171 we demD., and Nichols, W. Effect of suture materials on onstrated that the collagen sponge was prohealing wounds of stomach and colon. Surg. Gynecol. gressively and completely digested during the Obstet. 140: 701, 1975. first month after implantation. We also ob13. Isomiki, A. M. A new cell type (tuft cell) in the gasserved on the first days after surgery that fitrointestinal mucosa of the rat. A trasmission and scanning electron microscopy study. Acta Pathol. Mibroblast migration and collagen fiber synthesis crobiol. Stand. (A), Suppl. 240: I, 1973. were already enhanced by the presence of the 14. Lawson, H. H. Gastritis and gastric ulceration. Brit. heterologous collagen. The bovine collagen J. Surg. 53: 493, 1966. was never antigenically rejected, probably be- 15. Lehy, T., Bonnefond, A., Dubrasquet, M., Nasca, S., cause of its initial low antigenicity [ 181 and Lewin, M., and Bonfils, S. Comparative effects of
RATTO
16.
17.
18. 19.
AND MOTTA:
GASTRIC
antrccolic transposition and antrectomy on fundic mucosa and acid secretion of the rat. Gastroenterology 64: 421, 1973. Lehy, T., Voillemot, N., Dubrasquet, M., and Dufougerai, F. Gastrin cell hyperplasia in rats with chronic antral stimulation. Gastroenterology 68: 7 1, 1975. Motta, G., Ratto, G. B., De Barbie& A., Carte, G., Zardi, L., Sacco, A., and Castagnola, M. Can heterologous collagen enhance the granulation tissue growth? An experimental study. Poster section VII Congr. FECTS., Copenhagen, Aug. 10-13, 1982. Peacock, E. E., Jr., and Petty, J. M. Biological reactions to collagen transplantations. Surg. Forum 17: 815, 1958. Polliack, A., Lampen, N., and De Harven, E. Comparison of air drying and critical point drying procedures for the study of human blood cells by scanning electron microscopy. S.E.M./IITRI 6: 529, 1973.
DEFECT HEALING
49
20. Ritchie, W. P., and Delaney, J. P. Susceptibility of experimental atrophic gastritis to ulceration. Gastroenterology 60(4): 554, 197 1. 21. Ritchie, W. P. Acute gastric mucosal damage induced by bile salts acid and ischemia. Gastroenterology 68(4): 699, 1975. Russeau, B., and Sladen, G. E. Effect of luminal pH 22. on the absorption of water, Nat and Cl- by rat intestine in vivo. Biochim. Biophys. Acta 233: 59 1, 197 1. 23. Schor, S. L., and Court, J. Different mechanism in the attachment of cells to native and denaturated collagen. J. Cell. Sci. 38: 267, 1979. 24. Watson, R., Rothbord, S., and Vonamee, P. The antigenicity of rat collagen. J. Exp. Med. 99: 535, 1959. 25. Williams, S. E., and Tumberg, L. A. Retardation of acid diffusion by pig gastric mucus: A potential role in mucosal protection. Gastroenterology 79: 299,198O.
OF SURGICAL
RESEARCH
36,
35-49 (1984)
Healing of Experimental Defects on the Stomach: Influence Autologous and Heterologous Materials GIOVANNI Department
B. RATTO AND GIOVANNI
of Semeiotica
Chirurgica
I, University
of
MOTTA’
of Genoa,
Italy
Submitted for publication December 15, 1982 This study was undertaken to evaluate the influence of various autologous and heterologous structures on gastric defect healing. In 2 10 rats, standard gastric defects were closed using the following structures as patches: (1) pedicled colonic patches, (a) mucosal layer to lumen, and (b) peritoneal surface to lumen; (2) leiomuscular colonic buttons; and (3) lyophilized bovine tendon collagen sponges. The specimens were removed from the defect arcas at standard time intervals ranging from 2 to 12 months after surgery; the samples were then examined by light microscopy (hematoxylin-eosin; Azan Mallory; Grimelius methods) and by trasmission and scanning electronmicroscopy. The colonic mucosa appeared to be the most resistant of the structures to gastric secretions; it showed no mucosal disruption during the entire period of observation and also underwent a progressive gastric-like morphological transformation. In all the other cases,the healing process involved the initial development of granulation tissue followed by scar tissue with gastric mucosa growing over the scar tissue to fill the defects.
The gastric mucosa is a specialized structure which is highly resistant to gastric acidity even under conditions of maximal stimulation acting in this way as a limiting harrier to ion diffusion [5, 61. Surgical wounds are still known to heal rapidly in spite of the alterations which they cause in the mucosa layer [8-lo]. The repair process occurs during the first 14 to 21 days through the formation of a nonspecific granulation and scar tissue. The rapid healing of these gastric wounds is due to the high rate of collagen synthesis [ 121. The purpose of this study was to evaluate the influences of various autologous and heterologous structures on gastric defect healing in rats. In investigating this phenomenon, pedicled coionic patches (mucosal or peritoneal surface to lumen), leiomuscular colonic grafts, and lyophilized bovine tendon collagen sponges were used to close a standard gastric defect. In particular we tried to examine (1) the rate of healing of these standard gastric defects which were surgically closed, (2) the evolution of the healing processes, (3) the resistance of
the different structures to gastric secretions, and (4) the morphological changes of these materials after chronic exposure to the gastric acid environment. MATERIALS
AND METHODS
The experiments were performed on 2 10 male Wistar rats each weighing about 250 g. The animals were fasted for 24 hr but allowed access to water. Under ether anaesthesia the abdomen was opened through a midline incision. A standard hole ( 10 X 6 mm) was made on the posterior surface of the gastric body; the posterior wall was chosen to avoid traction on the vascular pedicles of the colonic patches used to plug the defects. The animals were divided into six groups and treated as follows: Group A. In 50 rats, a lo-mm segment of the left half of the transverse colon and its vascular supply were isolated. An end-to-end anastomosis was performed on the resected colon to restore its continuity. The isolated segment was opened longitudinally on its antimesocolic side and rinsed with saline. It was then sutured, mucosal surface to lumen, in order to close the gastric defect.
’ Present address: Direttore Cattedra di Semeiotica Chirurgica I Universita degli Studi, Viale Benedetto XV, 10, 16132 Genova, Italy. 35
0022-4804184 $1 SO Copyright 0 1984 by Academic Press, Inc. All rights of reproduction in any form reserved.
36
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. I, JANUARY
Group B. Thirty rats underwent the same procedure as in group A but were chronically stimulated by daily subcutaneous injections of histamine in beeswax ( 10V3M) for 2 months [3]. Group C. In 40 rats the colonic mucosa was removed from the patch by blunt dissection; care was taken not to leave residual colonic mucosa. The patch was then anastomosed, leiomuscular layers to lumen, in order to close the standard gastric defect. Group D. In 40 rats the colonic segment was longitudinally opened near the insertion of the mesocolon; the patch was then reversed and sutured, peritoneal surface to lumen, over the gastric opening. The colonic mucosa, however, was first removed by blunt dissection to avoid massive secretion of mucus into the peritoneal cavity.
1984
Group E. In 40 rats standard gastric defects were plugged using a piece of native bovine collagen sponge extracted from the Achilles tendon using the Schor and Court, and Elsdale and Bard methods [23-271. In a control group of 10 rats, standard gastric holes were directly sutured and a colonic segment was isolated and sutured back into its original location to assessany morphological changes related to the procedure.
All the anastomoses were performed using an interrupted single layer 7-O Prolene (polypropylene) extramucosal suture. Within each group, the animals were randomly divided into 4 subgroups which were reoperated at 2,4,6, and 12 months after the initial surgery. At this time the defect areas and their surrounding tissues were removed.
FIG. 1. Group A: mucosal surface of a colonic patch 2 months after implantation. It appears to be composed of polygonal units, each with a central crypt surrounded by epithelial cells. The units are demarcated by shallow furrows. This superficial epithelial layer seems to be grossly intact (SEM X 1500).
RATTO
AND MOTTA:
GASTRIC
DEFECT HEALING
37
In group B all the samples were collected 2 the critical point method [ 191, and coated with months after implantation. The specimens gold by sputtering. For TEM, the specimens were first cut in half and then divided into were rinsed with water and stained in alcoholic four strips along the minor axis [ 15, 161. Each uranyl acetate, dehydrated through ethanols strip was then examined by serial section to and propylene oxide, and then embedded in quantify the extension of the morphological Epon Araldite. changes to be investigated. The first half was A statistical analysis of the mortality rate studied under light microscopy (LM) and the was carried out by using the 2 X 2 chi-square other under trasmission (TEM) and scanning test. The samples were divided into four ar(SEM) electronmicroscopy. In preparing for bitrary subgroups based on their morphology LM, the tissues were fixed in 10% formalin in order to statistically evaluate the evolution for routine staining (hematoxylin-eosin and of the healing process. The patches in group Azan Mallory); the same tissues were then A were classified in the following manner: (1) treated by Grimelius method to assess the no changes, (2) changes involving less than presence of diffise endocrine system cells [ 111. 50% of the patch mucosa, (3) changes inFor TEM and SEM, the specimens were fixed volving more than 50% but not the entire muin a 2% glutaraldehyde buffered solution and cosal layer, and (4) changes involving the enpostfixed in a 1% sodium cacodylate osmium tire mucosal layer. The patches in groups C, tetroxide. For SEM, the samples were dehy- D, and E were instead classified as ( 1) no drated in a graded series of alcohols, dried by changes, (2) loss of the original architecture
FIG. 2. Group A: a tubular mucous gland of the colonic patch 2 months after implantation. The columnar cells exhibit an active mucus production and some cells are seen discharging mucus into the crypt. (TEM X3700).
38
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
wit.hout
any mucosal overgrowth, (3) loss of the : original architecture with incomplete mucosd overgrowth, (4) loss of the original ar-
FIG. 3. Group A: a colonic patch 2 months except for an abundant chronic inflammatory
1984
chitecture with complete mucosal overgrow These phenomena were analyzed statistic; by means of a 5 X 4 chi-square test.
after implantation. The graft maintains infiltrate (LM X 100).
its original
structure
RATTO
AND
MOTTA:
GASTRIC
RESULTS
At the time of removal, all the defects appeared to be closed by vital tissues. One death was observed in group C and three deaths in group E, all within the first week after surgery. These deaths were caused by dilluse peritonitis due to patch perforation. One death was observed in group D (24 hr after surgery) caused by volvulus of a jejunal loop which became necrotic. A statistically significant difference in survival rate was observed only between groups A and E (P < 0.05). Group A. The mucosal layer of all the patches removed 2 months after implantation displayed a typical colonic mucosal morphology as seen under SEM (Fig. l), TEM (Fig. 2), and LM (Fig 3). Scattered necrosis or erosions of the epithelial layer were not observed. Under LM, however, initial signs of gastric-like transformation were seen near the colon-gastric anastomoses as documented
FIG.4. rough
DEFECT
39
HEALING
by loss of muciparity and an increase in cytoplasmatic acidophilia. Four months after surgery this gastric-like transformation was clearly seen in one-third of the cases, but it never involved more than 50% of the patch mucosa and was generally confined to the superficial epithelial layers. Six months after implantation the gastric-like transformation involved more than 50% of the patch mucosa in 4 1% of the cases. Under TEM examination, the epithelial cells of the patches showed ultrastructural patterns similar to that of gastric mucosa, including a significant amount of zymogen granules (Fig. 4). Twelve months after implantation the patches were completely covered with gastric-like mucosa in 78% of the samples, as demonstrated by LM, TEM, and SEM examination (Fig. 5). Under LM a near complete loss of affinity for mucus staining and an increase in the number of parietallike cells, as well as an increase of the glandular crypts were observed (Fig. 6). In 22% of the
Group A: the epithelial patch cells 6 months afier implantation; surfaced endoplasmatic reticulum are evident. (TEM X8400).
secretory
granules
as well as
40
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG. 5. Group A: the mucosal surface of colonic patch 12 months after implantation. The superficial epithelial layer is comprised of mucussecreting cells with morphological patterns very similar to those of normal gastric mucosa (SEM X300).
cases the original colonic morphology was preserved but only in some deep residual mucosal islands. Over the entire period of observation the manifestation of this gastric-like transformation was statistically significant (P < 0.001). Changes in the distribution and number of cells belonging to the diffuse endocrine system normally found in the patch mucosa were not statistically relevant (Fig. 7). The so-called “tuft cells” [ 131 of the gastrointestinal mucosa, were no longer observed on the mucosal surface of the colonic patches after implantation on the stomach (Fig. 8). Group B. The specimens removed 2 months after implantation showed morphological characteristics similar to those of group A; the only difference was the presence of some scattered superficial epithelial erosions (P > 0.05). Group C. The inner surface of all the patches
removed 2 months after implantation appeared to be made up of an irregular fibrous mesh work (Fig. 9); the patch wall was completely replaced by granulation tissue (Fig. 10). An epithelialization process, which spread from the anastomoses, was constantly observed at 4 months after implantation. However, this process was found to be complete in only 20% of the cases. At 6 to 12 months the process was complete in all instances (P < 0.00 1). The defects were plugged under the mucosal layer by scar tissue. Group D. Two months after implantation the serosal surface appeared to be extensively altered as demonstrated by the complete destruction of the mesothelial cell layer. Here too, under LM the defects were seen to be closed by granulation tissue, which totally replaced the normal colonic wall components.
RAT-TO AND MOTTA:
GASTRIC
DEFECT HEALING
41
FIG.6. Group A: a colonic patch 12 months after implantation; the mucosal layer shows a loss of affinity for mucus staining and an increase in cytoplasmatic acidophilia. (LM X80).
Four months after implantation, an overgrowth of gastric mucosa starting from the site of the colon-gastric anastomosis was observed. This neomucosal layer succeeded in closing the defect in only one case. Six and twelve months after surgery the patches appeared completely covered by gastric mucosa (P < 0.001). This neomucosa displayed an ultrastructural morphology (TEM) and cells of the diffuse endocrine system (Grimelius method) which were identical to those found in the normal gastric mucosa. The patch wall under the mucosal layer consisted mainly of scar tissue (Fig. 11). Group E. Two months after surgery all the defects were plugged by a granulation tissue rich in collagen fibers and new vessels. The gastric epithelium was seen growing over this granulation tissue into the gap. The new gastric mucosa had restored complete continuity to the defect in 80% of the cases 6 months after
surgery and in all samples 12 months after implantation (P < 0.001). In the control group the only change was the formation of foreign body granulomas around the sutures. DISCUSSION
The present study demonstrated that the healing process restored in all cases, a continuous and morphologically normal mucosal layer over the surgical defects regardless of the graft material employed. It also revealed, however, that the steps involved in the healing process were not identical for each material used. This finding was due to the different biological properties inherent in each of the materials employed, providing them with varying degrees of resistance and adaptability to the gastric acid environment. The colonic mucosal layer was found to be the best struc-
42
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG. 7. Group A: the colonic mucosa 6 months after implantation. (A) The presence of numerous argentin cells is evident (LM X40); (B) a detailed portion of the above picture, showing the intracytoplasmatic argentaffin granules (LM X 1000).
ture used to close the defects. It showed the highest degree of resistanceagainst gastric secretions. It was the only material capable of conserving its structural integrity. In fact no damage was seen in the mucosa exposed to basalsecretory levels, while under chronic histamine stimulation, it presented only some scattered superficial epithelial erosions. The leiomuscular colonic buttons and the peritoneal surface to lumen patches were instead all found to be completely digestedeven under basalsecretory conditions; and for this reason they were not exposed to chronic histamine stimulation. Finally, the colonic mucosa ap-
peared to be more resistant to perforation; this finding, however, was statistically significant only when compared to the heterologous collagen sponge.How the colonic mucosa was able, in our experimental model, to withstand gastric secretions without development of ulcerations and interstitial hemorrhage remains an unanswered question. Studies done by Ritchic [20, 2 l] and Lawson [ 141on the ability of the gastric mucosa to resist ulcer formation showed that its high degree of resistance to acid-peptic digestion was related to the morphological or functional makeup of the mucosa. Because the colonic mucosa presented
RATTO
AND
MOTTA:
GASTRIC
DEFECT
HEALING
43
FIG. l-Continued
no signs of histological damage we must assume that its functional or morphological makeup confers a high degree of resistance. Since the blood supply to the patch was carefully preserved, it may be possible that the absence of lesions on the colonic patches is in part the result of the ability of the well vascularized tissue to effectively clear or neutralize any H+ which might enter the graft. We also observed both grossly and microscopically, a highly active production of mucus by the patch mucosa. However, this mucus gel alone is not likely to directly protect the epithelial surface from the acid [l-25]. In fact, the mechanism through which the colonic
mucosa deals with hydrogen ions is very complex and probably consists of a combined process of absorption and neutralization [4-221. In an experimental study, Chung [4] stated that the colonic mucosa is intolerant to acid at any pH below 2 if exposed for 90 min or more. Under these conditions, he observed structural changes in the form of edema, sloughing, and coagulative necrosis in the coionic mucosa. Nevertheless, Chung concluded that it is not possible to determine whether such adaptation occurs in the colonic mucosa as a result of chronic exposure to an acid environment. The discrepancy between our results and those of Chung may be attributed
44
JOURNAL OF SURGICAL RESEARCH: VOL. 36, NO. 1, JANUARY
1984
FIG.8. Control group: the mucosal surface of a colonic segment 2 months after isolation and reimplantation in its original site. A tuft cell is clearly visible, its characteristic “tuft” of microvilli stands out against the shorter projections of the surrounding cells. (SEM X7000).
to either the difference in the length of the observation periods, which would allow a progressiveadaptation of the colonic mucosa to the gastric environment, or to the difference in the intensity of the acid insult on the colon. In Chung’s study a segmentof colon was cannulated and perfused with acid solutions (pH ranging between 6 and 1.3) while our colonic patches were chronically exposed to physiological pHs. We measuredthe gastric acid output in a separategroup of 10 anesthetized rats fasted for 48 hr both under basal conditions and after histamine administration ( lop3 M). In this instance our results were similar to those reported by Bunce et al. [2]. Furthermore, the colonic mucosa was also seen to provide adequateprotection for the underlying muscolar layers for the entire period of study.
Our findings demonstrated that the colonic mucosa was not only highly resistant to the gastric secretionsbut also underwent a process of morphological adaptation to its new environment. This adaptation consisted of a gastric-like transformation which initially began near the anastomosesand involved only the most superficial epithelial layers. This observation would seemto indicate that the gastric-like transformation depends on a direct contact with gastric secretions. Even though the gastric mucosa is known to repair mucosal erosions rapidly, we were lead to interpret our findings to represent gastric-metaplasiafor the following reasons: (1) signs of mucosal sloughing or atrophy were not observed at any time in group A, (2) we also found intermediate morphological stages(lossof muciparity
RATTO
FIG. 9. Group after implantation.
C: The luminal The surface
AND
MOTTA:
GASTRIC
DEFECT
HEALING
45
surface of a colonic patch denuded of its original mucosal layer 2 months is made up of an irregular fibrous meshwork (SEM X600).
and ! an increase in cytoplasmic acidophilia), and (3) the time period necessary for gastric rep; tir through mucosal overgrowth, as seen ing roups C, D and E, was considerably shorter thal n that of Group A. In group A, complete rest oration of gastric-like mucosal continuity
was consistently observed but only in tl lose patches which were examined 1 year after implantation (P < 0.001). The meaning of this gastric-like transformation remains uncl lear. In this regard we have initiated further investigations to evaluate the possibility of s 3me
46
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
36, NO.
1, JANUARY
1984
FIG. 10. Group C: a colonic patch denuded of its original mucosal layer 2 months after The wall of the patch is composed of an inflammatory granulation tissue (LM X 100).
chal ages in functional activity associated with the morphological transformations we observed. Presently our observations under LM
implantation.
and TEM would seem to indicate that the cells of the patch mucosa do assume a gast.SiClike functional capacity as demonstrated by
RATTO
AND
MOTTA:
GASTRIC
DEFECT
HEALING
FIG. 11. Group D: B colonic patch 12 months after implantation. The graft appears to be covered by gastric-like mucosa, while the underlying layers are made up of a partially hyalinized scar tissue (LM X 100).
the presence of chief cells containing zymogen gra nules. The distribution of cells belonging to1 :he diffuse endocrine system remained con-
stant even after chronic gastric exposure. ? ‘his finding supports the presence of some en’docrine function in the patch mucosa, reveal ing
48
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
36, NO.
1, JANUARY
1984
another area which calls for further investi- the subsequent loss of this property due to its digestion by tissue collagenases [24]. gation. The serosal and leiomuscular colonic strucACKNOWLEDGMENTS tures presented a minor resistive capacity to the gastric secretions. They were completely We express our most sincere thanks to Professor A. digested and substituted by granulation and Riva, head of the Institute of Human Anatomy at the scar tissues. These structures were therefore University of Cagliari, Italy, and Professor A. Badini and less apt to favor the healing of the gastric de- Dr. M. Truini of the Institute of Pathological Anatomy the University of Genoa, Italy, for their helpful colfects with complete restitutio ad integrum to oflaboration with this study. all the layers of the gastric wall. These materials did, however, appear to be able to plug gastric REFERENCES holes in almost all cases. In both groups, the granulation and scar tissues acted as good I. Allen, A., and Garner, A. Mucus and bicarbonate secretion in the stomach and their possible role in supports for the complete overgrowth of the mucosal protection. Gut 21: 249, 1980. gastric mucosa which closed the gaps. The 2. Bunce, K. T., Grewal, M., and Parsons, M. E. Pepsin high rate of healing of the gastric wounds in secretion in the isolated rat stomach preparation. groups C and D probably depends on the Physiol. Sot.. 55P, July 1979. characteristics of the collagen metabolism of 3. Code, C. F., and Varco, R. L. Chronic Histamine action. Proc. Sot. Exp. Biol. Med. 44: 475, June 1940. the rat stomach. In an experimental study on 4. Chung, R. S. K., Johnson, G. M., and Denbeste, L. rats Gottrup [8] demonstrated that wound The handling of H+ by the colon: The biophysical healing is more rapid in the stomach than in aspect of esophageal replacement. J. Surg Rex 33: most other organs, for example skin or tendon. 538, 1977. He also stated that the glandular oxyntic part, 5. Davenpot, H. W., Warner, H. A., and Code, C. F. Functional significance of gastric mucosal barrier to the area in which we created the defects, is sodium. Gastroenterology 47: 142, 1964. the most biochemically active zone of the Davenport, H. W. Salicylate damage to the gastric stomach. He also found that the concentration mucosal barrier. N. Engl. J. Med. 276: 1307, 1967. of collagen in this zone increased significantly Elsdale, T., and Bard, J. Collagen substrata for studies within 40 days after surgery and superceded on cell behaviour. J. Cell. Biol. 54: 626, 1972. Gottrup, F. Healing of incisional wounds in stomach the presurgical level by 90%. Finally it was and duodenum. A biomechanical study. Amer. J. also shown that the rate of collagen synthesis Surg. 14th 296, 1980. correlates well with the gain in mechanical 9. Gottrup, F. Healing of incisional wounds in stomach strength of wounds [9]. These findings could and duodenum. Collagen distribution and relation to provide an explanation for the rapid develmechanical strength. Amer. J. Surg. 141: 222, 198 I. opment of adequate mechanical resistance to 10. Gottrup, F. Healing of incisional wounds in stomach and duodenum. The influence of aging. Acta Chir. our surgical lesions. Stand. 147: 363, 1981. The heterologous collagen sponge allowed Il. Grimelius, L. The argyrophil reaction in islet cells of a similar healing process even though it apadult human pancreas studied with a new silver nitrate peared less resistant to gastric secretions (P procedure. Acta Sot. Med. Upsal. 73(56): 27 I, 1968. 12. Hastings, J. C., Van Winkle, W., Barker, E., Hines, < 0.05). In a previous study [ 171 we demD., and Nichols, W. Effect of suture materials on onstrated that the collagen sponge was prohealing wounds of stomach and colon. Surg. Gynecol. gressively and completely digested during the Obstet. 140: 701, 1975. first month after implantation. We also ob13. Isomiki, A. M. A new cell type (tuft cell) in the gasserved on the first days after surgery that fitrointestinal mucosa of the rat. A trasmission and scanning electron microscopy study. Acta Pathol. Mibroblast migration and collagen fiber synthesis crobiol. Stand. (A), Suppl. 240: I, 1973. were already enhanced by the presence of the 14. Lawson, H. H. Gastritis and gastric ulceration. Brit. heterologous collagen. The bovine collagen J. Surg. 53: 493, 1966. was never antigenically rejected, probably be- 15. Lehy, T., Bonnefond, A., Dubrasquet, M., Nasca, S., cause of its initial low antigenicity [ 181 and Lewin, M., and Bonfils, S. Comparative effects of
RATTO
16.
17.
18. 19.
AND MOTTA:
GASTRIC
antrccolic transposition and antrectomy on fundic mucosa and acid secretion of the rat. Gastroenterology 64: 421, 1973. Lehy, T., Voillemot, N., Dubrasquet, M., and Dufougerai, F. Gastrin cell hyperplasia in rats with chronic antral stimulation. Gastroenterology 68: 7 1, 1975. Motta, G., Ratto, G. B., De Barbie& A., Carte, G., Zardi, L., Sacco, A., and Castagnola, M. Can heterologous collagen enhance the granulation tissue growth? An experimental study. Poster section VII Congr. FECTS., Copenhagen, Aug. 10-13, 1982. Peacock, E. E., Jr., and Petty, J. M. Biological reactions to collagen transplantations. Surg. Forum 17: 815, 1958. Polliack, A., Lampen, N., and De Harven, E. Comparison of air drying and critical point drying procedures for the study of human blood cells by scanning electron microscopy. S.E.M./IITRI 6: 529, 1973.
DEFECT HEALING
49
20. Ritchie, W. P., and Delaney, J. P. Susceptibility of experimental atrophic gastritis to ulceration. Gastroenterology 60(4): 554, 197 1. 21. Ritchie, W. P. Acute gastric mucosal damage induced by bile salts acid and ischemia. Gastroenterology 68(4): 699, 1975. Russeau, B., and Sladen, G. E. Effect of luminal pH 22. on the absorption of water, Nat and Cl- by rat intestine in vivo. Biochim. Biophys. Acta 233: 59 1, 197 1. 23. Schor, S. L., and Court, J. Different mechanism in the attachment of cells to native and denaturated collagen. J. Cell. Sci. 38: 267, 1979. 24. Watson, R., Rothbord, S., and Vonamee, P. The antigenicity of rat collagen. J. Exp. Med. 99: 535, 1959. 25. Williams, S. E., and Tumberg, L. A. Retardation of acid diffusion by pig gastric mucus: A potential role in mucosal protection. Gastroenterology 79: 299,198O.