JOURNAL
OF SURGICAL
RESEARCH
52, 510-517 (19%)
Basement Membrane Components Enhance Isolated Enterocyte Growth’ BAO LIEN T. NGUYEN, M.D., JON S. THOMPSON, M.D., AND JOHN G. SHARP, PH.D. Surgical Services Omaha VAMC and the Departments of Surgery and Anatomy, University of Nebraska and Creighton University School of Medicine, Omaha, Nebraska 68198 Presented
at the Annual
Meeting
of the Association
for Academic
Inc.
INTRODUCTION Despite recent advances in immunosuppression, the results of intestinal transplantation remain disappoint’ This research was supported 0022.4804/92 $4.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
in part by NIH Grant AI25820. 510 Inc. reserved.
Springs, Colorado,
November
Center, 20-23, 1991
ing [ 11. While many technical problems have been overcome, rejection of the transplanted intestine and graft versus host disease remain significant obstacles to whole organ transplantation. These problems are related to the considerable amount of lymphoid tissue and passenger leukocytes in the small intestine. Efforts to reduce the amount of lymphoid tissue transplanted with the small intestine have met with variable success [a]. Selective cell transplantation has been investigated as an alternative to whole organ transplantation to replace lost function. The theoretical advantages of this approach include reduced antigenicity of the allograft, the potential for autografts from undiseased portions of the organ involved, multiple recipients for a single donor, and longterm preservation [3,4]. Hepatocyte and pancreatic islet cell transplantation have been studied extensively [ 3-81. More recently, intestinal epithelial cell transplantation has also been reported [3, 41. If isolated enterocyte transplantation proves useful in expanding functional intestinal surface area then this technique might be an attractive solution to the short bowel syndrome. Maintaining growth of intestinal epithelial cells in uitro and in. viuo is notoriously difficult [9, lo]. Our previous attempts at growing isolated enterocytes in viva were unsuccessful [ 111. The basement membrane components laminin, Type IV collagen, and heparan sulfate have been shown to enhance the growth of enterocytes by promoting adherence, migration, and differentiation [lo, 12-141. Matrigel, a solubilized extract of the basement membrane from the transplantable EngelbrothHolm-Swarm (EHS) mouse tumor which consists of 60% laminin, 30% Type IV collagen, and 3% heparan sulphate proteoglycan, stimulates growth and differentiation of epithelial cells in vitro and in. uiuo [12, 151. Laminin alone has also been shown to cause greater association and migration of plated intestinal cells [14]. These basement membrane components and other growth factors may be important in facilitating the growth of intestinal epithelial cells and thus improving the success of isolated enterocyte transplantation. The aim of this study was to determine whether basement membrane components would enhance enterocyte growth, in vitro and in uiuo, in serosa-lined pouches.
Isolated enterocyte transplantation may have a potential role in increasing intestinal surface area. However, enterocytes are notoriously difficult to grow. Basement membrane components (BMC) promote adherence, migration, and differentiation of enterocytes. Our aim was to determine if BMC enhance enterocyte growth. Twenty-nine rabbits had 5-cm ileal segments resected and serosal pouches (n =.22) created from the serosal surface of the colon. Harvesting of enterocytes by warm trypsinization resulted in 92 + 6% cell viability and yielded 5.0 +- 2.4 lo6 enterocytesjcm intestine. Enterocytes (10’) were cultured in vitro (n = 7) in 10 ml growth media in plain flasks and flasks coated with laminin alone or Matrigel (an extract of mouse basement membrane containing laminin plus Type IV collagen and heparan sulfate). The cells were subcultured at 2 weeks and examined after Geimsa staining at 4 weeks. Epithelial growth was confirmed by light microscopy and staining for cytokeratin and quantitated by image analysis (JAVA). Epithelial coverage of the flasks was greater with Matrigel (80 f 15%) than laminin (66 2 15%) which was greater than control (44 rt 20%) (P < 0.01). For in uiuo studies lo6 harvested cells were infused into the serosal pouches either in growth media (n = 9) or media + Matrigel (n = 13). Epithelial growth in the pouch was evaluated by qualitative scoring of cytokeratin staining. Cytokeratin staining was similar on control colon serosa (n = 5) and serosa after infusion of cells in media alone. However, Matrigel significantly enhanced the area, depth, and intensity of staining. These effects were apparent at 1 and 2 weeks, but diminished at 4 weeks. BMC enhance enterocyte growth in vitro. A combination of components had a greater effect than laminin alone. BMC increased cytokeratin staining in serosal pouches suggesting that they 0 1992 Academic also enhance enterocyte growth in vivo. Press,
Surgery, Colorado
Medical
NGUYEN,
MATERIAL
AND
THOMPSON,
AND
SHARP:
METHODS
Twenty-nine New Zealand white male rabbits (2.5-3.5 kg), which were housed and cared for in accordance with the American Association for the Accreditation of Laboratory Animal Care Guidelines, were included in the study. Enterocytes were harvested from resected ileal segments in each animal using warm trypsinization. In seven of the animals the cells were plated on plain, laminin-coated, and Matrigel (Collaborative Research Inc, Lexington, Massachusetts) coated culture flasks for in vitro studies. The presence of intestinal epithelial cells in the cultures was confirmed by light microscopy and immunohistologic staining for cytokeratin. Quantitative comparisons of enterocyte growth in the different culture flasks were made by determining the percentage of flask surface covered by cell growth using JAVA image analysis. In the remaining 22 rabbits intestinal continuity was restored and a serosal pouch was created from the serosal surface of adjacent cecal and colonic segments. Based on the results of the in vitro studies, the harvested enterocytes were infused into the serosal pouches of these animals, and for in vivo studies, either in suspension with growth media alone (n = 9) or with growth media plus Matrigel (n = 13). The serosal surface of the colon was studied in animals without cell infusion (n = 5) as a control. The rabbits were sacrificed at intervals up to 4 weeks and a qualitative comparison of cytokeratin staining in the serosal pouches was performed in a blinded fashion to evaluate enterocyte growth. Harvesting
Technique
Enterocytes were harvested using a warm trypsinization method [16]. The resected intestinal segment was cut lengthwise and rinsed with 0.1 M phosphate-buffered saline (PBS). The specimen was washed in 0.04% sodium hypochlorite (NaHCl) for 30 min at room temperature and then rinsed again with PBS (under a dissecting microscope). The mucosa was stripped from the underlying tissue, cut into small pieces, and incubated in a solution of 0.1% trypsin at 37°C for 30 min. Following the incubation period the supernatant was collected and mixed with 10 ml of fresh growth media (RPM1 1640, 10% fetal bovine serum, 1 mcg/ml hydrocortisone, 0.6 pg/ml insulin, lo3 M thioglycerol, 50 U/ml penicillin, and 50 U/ml streptomycin). This mixture was centrifuged at 700g for 10 min. The supernatant was discarded. The cell pellet which was composed of liberated enterocytes was resuspended in 10 ml of growth media and stored on ice. Fresh growth media was added to the original mucosa pieces, and the flask was vigorously hand shaken and returned to the incubator for 30 min. The process of recovering liberated enterocytes from the trypsin-incubated mucosa was repeated 5-6 times to ensure maximal disaggregation of the tissue. The collected
ISOLATED
ENTEROCYTE
GROWTH
511
enterocytes were washed twice with PBS and resuspended in 10 ml of growth media to form the final cell suspension. The total number of cells harvested was counted using a hemocytometer and the cell viability for each suspension was determined using the trypan blue dye exclusion method [12]. The mean number of cells harvested for each animal was 7.6 f 3.7 X lo7 with 92 + 6% viability. In Vitro Study Approximately lo5 cells were seeded in 25-cm2 plain, laminin-coated, and Matrigel-coated tissue culture flasks. Additionally, 5-cm2 plain, laminin-coated, and Matrigel-coated culture slide chambers were seeded with enterocytes from each animal. The vessels were incubated in 5% CO, at 37°C and refed with fresh media every 7 days. The cell cultures were subcultured at 14 days. Growth media was withdrawn from the flasks and discarded. Dispase (2 ml) was added to each culture flask which was then incubated at 37°C for 30 to 120 min until the cells round up. The cell clumps were dispersed by repeated aspiration with a pipette, transferred to a clean test tube, and then centrifuged at 700g for 5 min. The cell pellet was resuspended in 5 ml of fresh growth medium and placed in a new flask for further culture. Fourteen days after subculture the flasks were fixed and stained with Wright-Geimsa stain to evaluate cellular growth under light microscopy. The culture slide chambers were fixed in 95% cold ethanol and stained immunohistologically using a MAK-6 (Triton) cocktail containing monoclonal antibodies AEl and AE3 to confirm the presence of epithelial cells. Tritiated [3H]thymidine incorporation was determined in plain and matrigel-coated flasks (n = 3). Growth media was removed and replaced with media containing 1 &i [3H]thymidine. One hour later the flasks were washed with cold PBS and the cells liberated with Dispase. Absolute amounts of incorporated radioactivity were determined and expressed as counts per minute. In Vivo Study Operations were performed after an overnight fast using sterile technique. Anesthesia was achieved with intramuscular ketamine (35 mg/kg) and xylazine (7 mg/ kg) and halothane by inhalation. Following the resection of the distal 5 cm of the ileum and restoration of bowel continuity, construction of a serosa-lined pouch 7 cm in length and 1 cm in diameter was performed by anastomosing the serosa of adjacent colonic loops as previously described (Fig. 1) [ 111. The pouch was intubated with a short piece of Silastic tubing which was brought out through the abdominal wall and tunneled subcutaneously to exit the skin dorsally. The enterocytes were har-
512
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
52, NO. 5, MAY
1992
FIG. 1. (a) In uiuo model. An ileal segment is resected with restoration of intestinal continuity. Enterocytes are harvested for later infusion. A serosal pouch is constructed from adjacent colonic segments. A catheter is left in the pouch for infusion of cells.
vested from resected bowel segments and the isolated cells were infused into the pouch of the rabbits a few hours after the operation. The harvested enterocytes were either suspended in 10 ml growth media alone (n = 9) or in 10 ml growth media to which 1 ml Matrigel was added (n = 13). The animals were sacrificed at different intervals up to 28 days after operation. The pouches were rinsed and examined and then excised for immunohistologic analysis. At least three random histologic samples from each pouch were obtained and stained with the MAK-6 cocktail detecting cytokeratin. These sections were evaluated microscopically for epithelial cell growth based on immunocytochemical detection of cytokeratin. Two blinded observers graded the cytokeratin staining with respect to the percentage of basement membrane linear surface stained and the depth and intensity of the cytokeratin staining (Fig. 2). Control colonic serosa and small intestinal epithelium were also evaluated. The extent of staining was estimated as the percentage of the serosal surface which was cytokeratin positive. For depth of staining the following scoring was used: 1 = single cell layer, 2 = two to three cells deep, 3 = four cells deep, and 4 = > four cells deep. Intensity of staining was scored as 0 if none was present, 1 for golden, 2 for light brown, 3 for brown, and 4 for dark brown color. A composite score was calculated by multiplying the percent of cytokeratin staining times the depth and intensity scores. A score of 2 was arbitrarily chosen to represent substantial increase in cytokeratin staining as this represented approximately 50% coverage with depth and intensity scores of 2 which were clearly greater than the
mean and standard control serosa. Statistical
deviations
of these parameters
in
Methods
Data are expressed as the mean f SD. The data were analyzed using a x2 test and ANOVA and Bonferonni corrected t tests with P < 0.05 for significance. RESULTS
In Vitro Study The in vitro growth of epithelial cells was confirmed by light microscopy and staining with anticytokeratin antibody at both 14 and 28 days. Figure 3a demonstrates the presence of epithelial-like cells with large, oval nuclei growing tightly as polygonal, closely associated cells at 28 days. Figure 3b is a nearly pure epithelial subculture of the epithelial cells seen in Fig. 3a. As shown in Fig. 3d diffuse cytokeratin staining was found in virtually all of the cells of a subculture such as that illustrated in Fig. 3b. Growth was present in all flasks except one plain flask. The percentage of the surface of plain flasks covered with intestinal cell growth was significantly less than the laminin-coated flasks and the Matrigel-coated flasks (44% -t 18% vs 67% + 15% and 80% + 14%, P < 0.05). There was significantly more cell coverage in Matrigel-coated than laminin-coated flasks (P < 0.05) (Fig. 4). [3H]Thymidine incorporation was significantly greater in the cells cultured in Matrigel-coated flasks
NGUYEN,
THOMPSON,
AND
SHARP:
ISOLATED
ENTEROCYTE
GROWTH
FIG. 2. (a) Section of colon stained for cytokeratin. The epithelium is positive. A small extent (about 10%) of the serosal surface shows cytokeratin positivity. The depth of the stained layer was one cell and the intensity of staining was judged to be brown. The staining of serosal surface of pouches is recipient of in uitro cultured epithelial cells was similar to this example (X25). (b) Cytokeratin staining of the serosal surface of a pouch in a recipient 2 weeks after fresh cell only infusion. The surface was estimated to be 80% covered with cytokeratin-positive cells. The depth of the cell layer was judged to be one to two cells and the intensity was brown (X25). (c) Cytokeratin staining of the serosal surface of a pouch in a recipient of fresh cells in Matrige14 weeks after infusion. The surface was estimated to be 100% covered. The depth of the cell layer was judged to be one to three cells and the intensity was brown (X25). (d) Cytokeratin staining of the serosal surface of a pouch in a recipient of fresh cells in Matrige12 weeks after infusion. The surface was estimated to be 95% covered. The cell layer was judged to be up to four to five cells deep in places (two to three on average) and the staining intensity ranged from brown to dark brown (X25).
compared to plain cpm, P < 0.05).
flasks (4004 f 1148 vs 1810 f 496
In Vivo Study In all cases the pouch surface grossly had a granular appearance without evidence of luxurious epithelial growth. However, varying extents of epithelial cells were apparent by light microscopy and these stained positively with anticytokeratin antibody (Fig. 2). The epithelium had more of a squamous rather than a cuboidal or a columnar appearance and there was no good evidence of crypt or villus formation. There was very minimal cytokeratin staining on the serosal surface of control segments of colon (Fig. 2a and Table 1). Small intestinal epithelium was strongly cytokeratin positive with a gradation of positivity from lightly positive at the base of the crypts to strongly positive toward the tips of the villi (data not shown). Six rabbits were sacrificed 2 weeks following infusion of the isolated enterocytes in growth
media alone. There was no significant difference in cytokeratin staining compared to control serosa. Three other animals had cells cultured in vitro for 14 days. The cells were then liberated with dispase and infused into the serosal pouch in growth media alone. The percentage of basement membrane with cytokeratin staining was significantly less than those animals with cells infused directly (13% + 10% vs 35% + 21%, P < 0.05). The depth and density of staining were not significantly different from those of animals with cells infused directly or control colon serosa. Infusion of harvested enterocytes in growth media plus Matrigel resulted in significantly greater cytokeratin staining of pouch serosa than cells infused without Matrigel and control colon serosa (Figs. 2c and 2d and Table 2). There was greater surface stained and increased depth and intensity of staining 1 week after infusion compared to control serosa. The extent and depth of the staining diminished by 4 weeks, but the intensity was maintained (Fig. 3~). The composite score was sig-
514
JOURNAL
OF SURGICAL
RESEARCH:
VOL. 52, NO. 5, MAY 1992
FIG. 3. (a) Primary culture of rabbit small intestinal epithelial cells. In addition to a focus of epithelial cells and isolated epithelial cells, macrophages and fibroblasts are also evidence (stained with Wright-Giemsa, x120). (b) An almost completely pure culture of rabbit small intestinal epithelial cells obtained by Dispase treatment of a primary culture and replating in the presence of Matrigel (stained with WrightGiemsa, X250). (c) Background control immunocytochemical staining of cytokeratin in cells of a flask as in (b) (primary MAK-6 antibody cocktail omitted, x120). (d) Positive immunocvtochemical staining of cytokeratin in cells of a flask as in (b). The background staining is as in (c) (X120).
nificantly greater at 1 and 2 weeks than any of the other groups of animals. Sixty-seven and 71% of these animals, respectively, had scores greater than 2. DISCUSSION In the present study rabbit intestinal epithelial cells harvested by warm trypsinization and subcultured at 14 days remained viable in culture at 28 days. The cells were epithelial in appearance when observed by light microscopy, stained positive for cytokeratin, and incorporated [3H]thymidine. Similar in vitro growth of enterocytes has now been reported in several studies using a variety of harvesting techniques [3, 4, 9-11, 18, 191. It remains unclear, however, whether long-term culture is possible and if these cultured enterocytes can establish normal structure, function, and proliferative capacity [ 191. The basement membrane of the intestine is a thin continuous extracellular matrix that separates the epithelium from the underlying subepithelial tissues. Its major components, including Type IV collagen, the glycoprotein laminin, heparan sulphate proteoglycan, and other proteins, have been identified by immunofluorescence
and immunoelectron microscopy [20]. The basement membrane has a role in organogenesis and cellular attachment, polarization, and differentiation [ 13, 21-221. Hahn et al. [lo] found that there was a fourfold increase in the number of attached intestinal epithelial cells in cultures when these cells were plated on pepsin-soluble biomatrix compared to plain plastic surfaces. Presumably, cell attachment is the result of the interaction of specific receptors of the epithelial cell plasma membrane with components of the basement membrane. Although the presence of subepithelial tissue may be important in these processes [21, 221, laminin and basement membrane extract alone have been observed to promote cellular differentiation of fetal rat enterocytes, independent of mesenchymal cell influence [13, 141. Both a combination of basement membrane components and laminin alone enhanced epithelial cell growth in vitro in the present study. This observation has been made by others [13, 141. The combination of components was more effective in promoting epithelial growth than laminin alone. This is consistent with our previous observations in an in oivo model of intestinal regeneration in the rabbit and other in vitro studies [12, 13, 151.
NGUYEN,
THOMPSON,
AND
SHARP:
ISOLATED
FIG. 4. Photograph of plain (left), laminin-coated (middle), and Matrigel-coated lial growth in the flasks coated with basement membrane components.
Because Matrigel appeared to promote more growth than laminin in uitro, we chose to infuse this biomolecular preparation with harvested cells into the in uivo pouches. Maintaining viability of transplanted enterocytes in vivo has been extremely difficult. We previously harvested enterocytes with a chelation technique and infused the cells directly into a serosal pouch [ll]. There was no evidence of enterocyte growth. Similarly, in the
TABLE
1
Comparison of Cytokeratin Staining in Serosal Pouches 14 Days after Infusion of Enterocytes
Control serosa n=5 Percent surface with cytokeratin staining Thickness of cell layer stained Intensity of cytokeratin staining Composite score Composite score > 2
26
f12
1.0 +
Cells infused directly in growth media n=6
Cell cultured in vitro and infused n=3
35
13
f 21
f 10**
0.0
1.6 +
0.8
1.1 +
1.4 + 0.5 0.4 + 0.2 O/5
1.8 + 1.0 f
0.9 0.8*
1.5 + 1.1 0.5 + 0.7** O/3
* P < 0.05 vs control. ** P < 0.05 vs direct infusion.
O/6
0.7
ENTEROCYTE
515
GROWTH
(right) flasks demonstrating
greater confluence
of epithe-
present study harvested cells infused into serosal pouches either directly or after in vitro culture failed to increase the appearance of cytokeratin positive cells in the pouch. Vacanti et al. [3] reported that only 13% of animals having intestinal cells seeded in biodegradable polymers and implanted in the omentum showed evidence of sustained growth. Arnaout et al. [4] found that 50% of enterocyte transplants were successful for up to 3 weeks when placed intraperitoneally on collagen-coated dextron microcarriers. Basement membrane components added to the growth media significantly increased the quantity and density of cytokeratin-stained cells in the serosal pouches compared to control serosa and cells in growth media alone. Two thirds of the animals had substantial cytokeratin staining at both 1 and 2 weeks. At 1 week more than 50% of the basement membrane surface examined had cytokeratin positivity and the epithelial layer was routinely several cells deep with moderate staining intensity. These effects were less marked at 2 weeks and had diminished markedly at 4 weeks after infusion. This suggests that primarily differentiated cells were infused and because of limited proliferative capacity did not maintain long-term viability in the pouches. However, the presence of Matrigel did permit initial attachment and proliferation of these cells. The failure of the cells grown in vitro and subsequently seeded in the pouches to increase cytokeratin expression is potentially an important observation. The ability to amplify the harvested cells and perhaps repeat infusion on multiple occasions may be important for the
516
JOURNAL
OF SURGICAL
RESEARCH:
of Cytokeratin
Staining
in Serosal
Pouches
52, NO. 5, MAY
1992
2
TABLE Comparison
VOL.
at Different
Times
after
Infusion
of Cells with
Matrigel
Cells with Matrigel Control serosa n=5 Percent surface with cytokeratin Thickness of cell layer stained Intensity of cytokeratin staining Composite score Composite score > 2
staining
1 Week n=3 52 2.0 2.6 2.7 2/3
26 k 12 1.0 f 0.0 1.4 f 0.5 0.4 k 0.2 O/5
+ 28* f 0.5* f 0.9* f 2.0* (67%)*
2 Weeks n=7 36 2.2 2.3 2.3 5/7
t 20 f 0.8* k LO* zk 1.8* (71%)*
4 Weeks n=3 19 f lS# 1.5 i 0.7* 2.2 f 1.5 0.7 It 0.7# 0/3#
* P < 0.05 vs control serosa. ** P < 0.05 vs 1 and 2 weeks.
clinical application of this technique. Furthermore, Woehrle et al. [8] found that pancreatic islet cells grown in vitro prior to in vivo infusion underwent less vigorous rejection than those infused directly. Unfortunately, it appears that current in vitro techniques for small intestinal epithelial cells support amplification and differentiation rather than self-renewal characteristics [13]. There was insufficient epithelial growth in the present study to assess function. Arnaout et al. [4] found that transplanted enterocytes would replace bilirubin uridine diphosphoglucuronyl transferase activity in deficient Gunn rats. However, other metabolic activities have not been demonstrated and this will be an important aspect of future studies. While we have clearly demonstrated increased cellularity and cytokeratin expression in the pouches we have not shown conclusively that these changes represent growth of the isolated and transplanted enterocytes. Functional studies or other identifying markers will be necessary. We cannot exclude the possibility that infusion of the cells, media, and Matrigel altered the cytokeratin expression of existing mesothelial cells on the serosal surface. Although attachment and initial growth of enterocytes are presumed to have occurred in the serosal pouches of the Matrigel-supplemented animals in this study, enterocyte transplantation will be feasible only if long-term growth, differentiation, and function of intestinal epithelia occur. Several factors may be important in this process. The maturity of the transplanted cell, i.e., stem cells versus villus cells may influence proliferative capacity. Since the interaction between the epithelial cells and viable mesenchymal cells is important for cellular polarization and differentiation, transplantation of mesenchymal cells in conjunction with epithelial cells may be prerequisite for growing functionally absorptive cells [23]. Alternatively or additionally, preparation of the mesenchymal surface, e.g., stimulation of angiogenesis may be necessary. Various trophic factors are important in epithelial proliferation and differentiation. These have been employed in in vitro culture systems and may also be important in attempts to stimulate and
sustain growth of epithelial cells in vivo. The presence of luminal contents may also be important to promote growth and differentiation of cells in the serosal pouches. Thus, placing the serosal pouches in continuity with the gastrointestinal tract may stimulate more growth and differentiation of the transplanted enterocytes. In summary, basement membrane components enhance enterocyte growth in vitro. A combination of components promoted better growth than laminin alone. While growth of enterocytes was not conclusively demonstrated in vivo, basement membrane components did substantially increase the cellularity and cytokeratin expressivity of epithelial cells lining the serosal pouch. This beneficial effect was relatively short lived, however, suggesting that other factors will be needed to permit long-term growth of enterocytes in vivo. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical assistance of Susan K. Meyer, Don Daley, George Pallas, and Sally Mann and the secretarial support of Anita Solon.
REFERENCES 1. 2.
3.
4.
5.
Schraut, W. H. Current status of small bowel transplantation. Gastroenterology 94: 525, 1988. Williams, J. W., McClellan, T., Peters, T. G., Nog, S., Dean, P., Banner, B., Vera, S. R., and Stenz, F. Effect of pretransplant graft irradiation on canine intestinal transplantation. Surg. Gynecol. O&et. 167: 197, 1988. Vacanti, J. P., Morse, M. A., Saltzman, W. M., Domb, A. J., Perez-Atayde, A., and Langer, R. Selective cell transplantation using bioabsorbable artificial polymers as matrices. J. Pediatr. Surg. 23: 3, 1988. Arnaout, W. S., Moscioni, A. D., Felcher, A. H., Barbour, R. L., Brown, L. L., and Demetriou, A. A. Intraperitoneal transplantation of microcarrier attached enterocytes in rats. Am. J. Surg. 157: 85, 1989. Maganto, P., Cienfuegos, J. A., Santamaria, L., Rodriguez, V., Eroles, G., Andres, S., Castillo-Olivares, J. S., and Municio,
NGUYEN,
THOMPSON,
A. M. Auxiliary liver by transplanted cytes. J. Surg. Res. 48: 24, 1990.
AND
frozen-thawed
SHARP:
ISOLATED
hepato-
6.
Arnaout, W. S., Moscioni, A. D., Barbour, R. L., and Demetriou, A. A. Development of bioartificial liver: bilirubin conjugation in Gunn rats. J. Surg. Res. 48: 379, 1990. Gray, D. W. R. Islet isolation and transplantation techniques in the primate. Surg. Gynecol. Obstet. 170: 225, 1990. 8. Woehrle, M., Markmann, J. F., Silvers, W. K., Barker, C. F., and Naji, A. Transplantation of cultured pancreatic islets to BB rats. Surgery 100: 334, 1986. 9. Quaroni, A., Wands, J., Trelstad, R. L., and Isselbacher, K. J. Epithelioid cell cultures from rat small intestine. Characterization by morphologic and immunologic criteria. J. Cell. Biol. 80: 248, 1979. 10. Hahn, U., Cho, A., Schuppan, D., Hahn,E. G., Merker, H. J.,and Riecken, E. 0. Intestinal epithelial cells preferentially attach to a biomatrix derived from human intestinal mucosa. Gut 28(Sl): 153, 1987. 11. Nguyen, B. L. T., and Thompson, J. S. Growing small intestinal neomucosa in serosa lined pouches. Curr. Surg. 47: 269, 1990. 12. Thompson, J. S. Basement membrane components stimulate epithelialization of intestinal defects in vivo. Cell Tissue Kinetics 23: 443, 1990. 13. Hahn, U., Stallmach, A., Hahn, E. G., and Riecken, E. 0. Basement membrane components are potent promoters of rat intestinal epithelial cell differentiation in vitro. Gustroenterology 98: 322, 1990. 14. Olson, A. D., Davidson, M., Psycher, T., and Bienkoski, R. La-
15.
16. 17. 18.
19.
20.
21.
22. 23.
ENTEROCYTE
GROWTH
517
minin is a major factor in spreading and pattern formation of intestinal epithelial cells in culture. Gastroenterology 94: A333, 1988. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F. B., Laurie, G. W., and Martin, G. R. Basement membrane complexes with biological activity. Biochemistry 25: 312, 1986. Freshney, R. I. Culture of Animal Cells, 2nd ed., pp. 115-118. New York: A. R. Liss, 1987. Tennant, J. R. Evaluation of the trypan blue technique for determination of cell viability. Transplantation 2: 685, 1964. Whitehead, R. H., Brown, A., and Bhathal, P. S. A method for the isolation and culture of human colonic crypts in collagen gels. In Vitro 23(6): 436, 1987. Madara, J. L., Stafford, J., Dharmsathaphorn, K., and Carlson, S. Structural analysis of a human intestinal epithelial cell line. Gastroenterology 92: 1133, 1987. Timpl, R., and Dziadek, M. Structure, development and molecular pathology of basement membranes. Znt. Reu. Cell Biol. 2: 27, 1986. Hahn, U., Schuppan, D., Hahn, E., Merker, H. J., and Riecken, E. 0. Intestinal cells produce basement membrane proteins in vitro. Gut 28(Sl): 143, 1987. Ekblom, P. Developmentally regulated conversion of mesenthyme to epithelium. FASEB J. 3: 2131,1989. Kedinger, M., Simon-Assman, P., and Haffen, K. Growth and differentiation of intestinal endodermal cells in a coculture system. Gut 28(Sl): 237,1987.
OF SURGICAL
RESEARCH
52, 510-517 (19%)
Basement Membrane Components Enhance Isolated Enterocyte Growth’ BAO LIEN T. NGUYEN, M.D., JON S. THOMPSON, M.D., AND JOHN G. SHARP, PH.D. Surgical Services Omaha VAMC and the Departments of Surgery and Anatomy, University of Nebraska and Creighton University School of Medicine, Omaha, Nebraska 68198 Presented
at the Annual
Meeting
of the Association
for Academic
Inc.
INTRODUCTION Despite recent advances in immunosuppression, the results of intestinal transplantation remain disappoint’ This research was supported 0022.4804/92 $4.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
in part by NIH Grant AI25820. 510 Inc. reserved.
Springs, Colorado,
November
Center, 20-23, 1991
ing [ 11. While many technical problems have been overcome, rejection of the transplanted intestine and graft versus host disease remain significant obstacles to whole organ transplantation. These problems are related to the considerable amount of lymphoid tissue and passenger leukocytes in the small intestine. Efforts to reduce the amount of lymphoid tissue transplanted with the small intestine have met with variable success [a]. Selective cell transplantation has been investigated as an alternative to whole organ transplantation to replace lost function. The theoretical advantages of this approach include reduced antigenicity of the allograft, the potential for autografts from undiseased portions of the organ involved, multiple recipients for a single donor, and longterm preservation [3,4]. Hepatocyte and pancreatic islet cell transplantation have been studied extensively [ 3-81. More recently, intestinal epithelial cell transplantation has also been reported [3, 41. If isolated enterocyte transplantation proves useful in expanding functional intestinal surface area then this technique might be an attractive solution to the short bowel syndrome. Maintaining growth of intestinal epithelial cells in uitro and in. viuo is notoriously difficult [9, lo]. Our previous attempts at growing isolated enterocytes in viva were unsuccessful [ 111. The basement membrane components laminin, Type IV collagen, and heparan sulfate have been shown to enhance the growth of enterocytes by promoting adherence, migration, and differentiation [lo, 12-141. Matrigel, a solubilized extract of the basement membrane from the transplantable EngelbrothHolm-Swarm (EHS) mouse tumor which consists of 60% laminin, 30% Type IV collagen, and 3% heparan sulphate proteoglycan, stimulates growth and differentiation of epithelial cells in vitro and in. uiuo [12, 151. Laminin alone has also been shown to cause greater association and migration of plated intestinal cells [14]. These basement membrane components and other growth factors may be important in facilitating the growth of intestinal epithelial cells and thus improving the success of isolated enterocyte transplantation. The aim of this study was to determine whether basement membrane components would enhance enterocyte growth, in vitro and in uiuo, in serosa-lined pouches.
Isolated enterocyte transplantation may have a potential role in increasing intestinal surface area. However, enterocytes are notoriously difficult to grow. Basement membrane components (BMC) promote adherence, migration, and differentiation of enterocytes. Our aim was to determine if BMC enhance enterocyte growth. Twenty-nine rabbits had 5-cm ileal segments resected and serosal pouches (n =.22) created from the serosal surface of the colon. Harvesting of enterocytes by warm trypsinization resulted in 92 + 6% cell viability and yielded 5.0 +- 2.4 lo6 enterocytesjcm intestine. Enterocytes (10’) were cultured in vitro (n = 7) in 10 ml growth media in plain flasks and flasks coated with laminin alone or Matrigel (an extract of mouse basement membrane containing laminin plus Type IV collagen and heparan sulfate). The cells were subcultured at 2 weeks and examined after Geimsa staining at 4 weeks. Epithelial growth was confirmed by light microscopy and staining for cytokeratin and quantitated by image analysis (JAVA). Epithelial coverage of the flasks was greater with Matrigel (80 f 15%) than laminin (66 2 15%) which was greater than control (44 rt 20%) (P < 0.01). For in uiuo studies lo6 harvested cells were infused into the serosal pouches either in growth media (n = 9) or media + Matrigel (n = 13). Epithelial growth in the pouch was evaluated by qualitative scoring of cytokeratin staining. Cytokeratin staining was similar on control colon serosa (n = 5) and serosa after infusion of cells in media alone. However, Matrigel significantly enhanced the area, depth, and intensity of staining. These effects were apparent at 1 and 2 weeks, but diminished at 4 weeks. BMC enhance enterocyte growth in vitro. A combination of components had a greater effect than laminin alone. BMC increased cytokeratin staining in serosal pouches suggesting that they 0 1992 Academic also enhance enterocyte growth in vivo. Press,
Surgery, Colorado
Medical
NGUYEN,
MATERIAL
AND
THOMPSON,
AND
SHARP:
METHODS
Twenty-nine New Zealand white male rabbits (2.5-3.5 kg), which were housed and cared for in accordance with the American Association for the Accreditation of Laboratory Animal Care Guidelines, were included in the study. Enterocytes were harvested from resected ileal segments in each animal using warm trypsinization. In seven of the animals the cells were plated on plain, laminin-coated, and Matrigel (Collaborative Research Inc, Lexington, Massachusetts) coated culture flasks for in vitro studies. The presence of intestinal epithelial cells in the cultures was confirmed by light microscopy and immunohistologic staining for cytokeratin. Quantitative comparisons of enterocyte growth in the different culture flasks were made by determining the percentage of flask surface covered by cell growth using JAVA image analysis. In the remaining 22 rabbits intestinal continuity was restored and a serosal pouch was created from the serosal surface of adjacent cecal and colonic segments. Based on the results of the in vitro studies, the harvested enterocytes were infused into the serosal pouches of these animals, and for in vivo studies, either in suspension with growth media alone (n = 9) or with growth media plus Matrigel (n = 13). The serosal surface of the colon was studied in animals without cell infusion (n = 5) as a control. The rabbits were sacrificed at intervals up to 4 weeks and a qualitative comparison of cytokeratin staining in the serosal pouches was performed in a blinded fashion to evaluate enterocyte growth. Harvesting
Technique
Enterocytes were harvested using a warm trypsinization method [16]. The resected intestinal segment was cut lengthwise and rinsed with 0.1 M phosphate-buffered saline (PBS). The specimen was washed in 0.04% sodium hypochlorite (NaHCl) for 30 min at room temperature and then rinsed again with PBS (under a dissecting microscope). The mucosa was stripped from the underlying tissue, cut into small pieces, and incubated in a solution of 0.1% trypsin at 37°C for 30 min. Following the incubation period the supernatant was collected and mixed with 10 ml of fresh growth media (RPM1 1640, 10% fetal bovine serum, 1 mcg/ml hydrocortisone, 0.6 pg/ml insulin, lo3 M thioglycerol, 50 U/ml penicillin, and 50 U/ml streptomycin). This mixture was centrifuged at 700g for 10 min. The supernatant was discarded. The cell pellet which was composed of liberated enterocytes was resuspended in 10 ml of growth media and stored on ice. Fresh growth media was added to the original mucosa pieces, and the flask was vigorously hand shaken and returned to the incubator for 30 min. The process of recovering liberated enterocytes from the trypsin-incubated mucosa was repeated 5-6 times to ensure maximal disaggregation of the tissue. The collected
ISOLATED
ENTEROCYTE
GROWTH
511
enterocytes were washed twice with PBS and resuspended in 10 ml of growth media to form the final cell suspension. The total number of cells harvested was counted using a hemocytometer and the cell viability for each suspension was determined using the trypan blue dye exclusion method [12]. The mean number of cells harvested for each animal was 7.6 f 3.7 X lo7 with 92 + 6% viability. In Vitro Study Approximately lo5 cells were seeded in 25-cm2 plain, laminin-coated, and Matrigel-coated tissue culture flasks. Additionally, 5-cm2 plain, laminin-coated, and Matrigel-coated culture slide chambers were seeded with enterocytes from each animal. The vessels were incubated in 5% CO, at 37°C and refed with fresh media every 7 days. The cell cultures were subcultured at 14 days. Growth media was withdrawn from the flasks and discarded. Dispase (2 ml) was added to each culture flask which was then incubated at 37°C for 30 to 120 min until the cells round up. The cell clumps were dispersed by repeated aspiration with a pipette, transferred to a clean test tube, and then centrifuged at 700g for 5 min. The cell pellet was resuspended in 5 ml of fresh growth medium and placed in a new flask for further culture. Fourteen days after subculture the flasks were fixed and stained with Wright-Geimsa stain to evaluate cellular growth under light microscopy. The culture slide chambers were fixed in 95% cold ethanol and stained immunohistologically using a MAK-6 (Triton) cocktail containing monoclonal antibodies AEl and AE3 to confirm the presence of epithelial cells. Tritiated [3H]thymidine incorporation was determined in plain and matrigel-coated flasks (n = 3). Growth media was removed and replaced with media containing 1 &i [3H]thymidine. One hour later the flasks were washed with cold PBS and the cells liberated with Dispase. Absolute amounts of incorporated radioactivity were determined and expressed as counts per minute. In Vivo Study Operations were performed after an overnight fast using sterile technique. Anesthesia was achieved with intramuscular ketamine (35 mg/kg) and xylazine (7 mg/ kg) and halothane by inhalation. Following the resection of the distal 5 cm of the ileum and restoration of bowel continuity, construction of a serosa-lined pouch 7 cm in length and 1 cm in diameter was performed by anastomosing the serosa of adjacent colonic loops as previously described (Fig. 1) [ 111. The pouch was intubated with a short piece of Silastic tubing which was brought out through the abdominal wall and tunneled subcutaneously to exit the skin dorsally. The enterocytes were har-
512
JOURNAL
OF SURGICAL
RESEARCH:
VOL.
52, NO. 5, MAY
1992
FIG. 1. (a) In uiuo model. An ileal segment is resected with restoration of intestinal continuity. Enterocytes are harvested for later infusion. A serosal pouch is constructed from adjacent colonic segments. A catheter is left in the pouch for infusion of cells.
vested from resected bowel segments and the isolated cells were infused into the pouch of the rabbits a few hours after the operation. The harvested enterocytes were either suspended in 10 ml growth media alone (n = 9) or in 10 ml growth media to which 1 ml Matrigel was added (n = 13). The animals were sacrificed at different intervals up to 28 days after operation. The pouches were rinsed and examined and then excised for immunohistologic analysis. At least three random histologic samples from each pouch were obtained and stained with the MAK-6 cocktail detecting cytokeratin. These sections were evaluated microscopically for epithelial cell growth based on immunocytochemical detection of cytokeratin. Two blinded observers graded the cytokeratin staining with respect to the percentage of basement membrane linear surface stained and the depth and intensity of the cytokeratin staining (Fig. 2). Control colonic serosa and small intestinal epithelium were also evaluated. The extent of staining was estimated as the percentage of the serosal surface which was cytokeratin positive. For depth of staining the following scoring was used: 1 = single cell layer, 2 = two to three cells deep, 3 = four cells deep, and 4 = > four cells deep. Intensity of staining was scored as 0 if none was present, 1 for golden, 2 for light brown, 3 for brown, and 4 for dark brown color. A composite score was calculated by multiplying the percent of cytokeratin staining times the depth and intensity scores. A score of 2 was arbitrarily chosen to represent substantial increase in cytokeratin staining as this represented approximately 50% coverage with depth and intensity scores of 2 which were clearly greater than the
mean and standard control serosa. Statistical
deviations
of these parameters
in
Methods
Data are expressed as the mean f SD. The data were analyzed using a x2 test and ANOVA and Bonferonni corrected t tests with P < 0.05 for significance. RESULTS
In Vitro Study The in vitro growth of epithelial cells was confirmed by light microscopy and staining with anticytokeratin antibody at both 14 and 28 days. Figure 3a demonstrates the presence of epithelial-like cells with large, oval nuclei growing tightly as polygonal, closely associated cells at 28 days. Figure 3b is a nearly pure epithelial subculture of the epithelial cells seen in Fig. 3a. As shown in Fig. 3d diffuse cytokeratin staining was found in virtually all of the cells of a subculture such as that illustrated in Fig. 3b. Growth was present in all flasks except one plain flask. The percentage of the surface of plain flasks covered with intestinal cell growth was significantly less than the laminin-coated flasks and the Matrigel-coated flasks (44% -t 18% vs 67% + 15% and 80% + 14%, P < 0.05). There was significantly more cell coverage in Matrigel-coated than laminin-coated flasks (P < 0.05) (Fig. 4). [3H]Thymidine incorporation was significantly greater in the cells cultured in Matrigel-coated flasks
NGUYEN,
THOMPSON,
AND
SHARP:
ISOLATED
ENTEROCYTE
GROWTH
FIG. 2. (a) Section of colon stained for cytokeratin. The epithelium is positive. A small extent (about 10%) of the serosal surface shows cytokeratin positivity. The depth of the stained layer was one cell and the intensity of staining was judged to be brown. The staining of serosal surface of pouches is recipient of in uitro cultured epithelial cells was similar to this example (X25). (b) Cytokeratin staining of the serosal surface of a pouch in a recipient 2 weeks after fresh cell only infusion. The surface was estimated to be 80% covered with cytokeratin-positive cells. The depth of the cell layer was judged to be one to two cells and the intensity was brown (X25). (c) Cytokeratin staining of the serosal surface of a pouch in a recipient of fresh cells in Matrige14 weeks after infusion. The surface was estimated to be 100% covered. The depth of the cell layer was judged to be one to three cells and the intensity was brown (X25). (d) Cytokeratin staining of the serosal surface of a pouch in a recipient of fresh cells in Matrige12 weeks after infusion. The surface was estimated to be 95% covered. The cell layer was judged to be up to four to five cells deep in places (two to three on average) and the staining intensity ranged from brown to dark brown (X25).
compared to plain cpm, P < 0.05).
flasks (4004 f 1148 vs 1810 f 496
In Vivo Study In all cases the pouch surface grossly had a granular appearance without evidence of luxurious epithelial growth. However, varying extents of epithelial cells were apparent by light microscopy and these stained positively with anticytokeratin antibody (Fig. 2). The epithelium had more of a squamous rather than a cuboidal or a columnar appearance and there was no good evidence of crypt or villus formation. There was very minimal cytokeratin staining on the serosal surface of control segments of colon (Fig. 2a and Table 1). Small intestinal epithelium was strongly cytokeratin positive with a gradation of positivity from lightly positive at the base of the crypts to strongly positive toward the tips of the villi (data not shown). Six rabbits were sacrificed 2 weeks following infusion of the isolated enterocytes in growth
media alone. There was no significant difference in cytokeratin staining compared to control serosa. Three other animals had cells cultured in vitro for 14 days. The cells were then liberated with dispase and infused into the serosal pouch in growth media alone. The percentage of basement membrane with cytokeratin staining was significantly less than those animals with cells infused directly (13% + 10% vs 35% + 21%, P < 0.05). The depth and density of staining were not significantly different from those of animals with cells infused directly or control colon serosa. Infusion of harvested enterocytes in growth media plus Matrigel resulted in significantly greater cytokeratin staining of pouch serosa than cells infused without Matrigel and control colon serosa (Figs. 2c and 2d and Table 2). There was greater surface stained and increased depth and intensity of staining 1 week after infusion compared to control serosa. The extent and depth of the staining diminished by 4 weeks, but the intensity was maintained (Fig. 3~). The composite score was sig-
514
JOURNAL
OF SURGICAL
RESEARCH:
VOL. 52, NO. 5, MAY 1992
FIG. 3. (a) Primary culture of rabbit small intestinal epithelial cells. In addition to a focus of epithelial cells and isolated epithelial cells, macrophages and fibroblasts are also evidence (stained with Wright-Giemsa, x120). (b) An almost completely pure culture of rabbit small intestinal epithelial cells obtained by Dispase treatment of a primary culture and replating in the presence of Matrigel (stained with WrightGiemsa, X250). (c) Background control immunocytochemical staining of cytokeratin in cells of a flask as in (b) (primary MAK-6 antibody cocktail omitted, x120). (d) Positive immunocvtochemical staining of cytokeratin in cells of a flask as in (b). The background staining is as in (c) (X120).
nificantly greater at 1 and 2 weeks than any of the other groups of animals. Sixty-seven and 71% of these animals, respectively, had scores greater than 2. DISCUSSION In the present study rabbit intestinal epithelial cells harvested by warm trypsinization and subcultured at 14 days remained viable in culture at 28 days. The cells were epithelial in appearance when observed by light microscopy, stained positive for cytokeratin, and incorporated [3H]thymidine. Similar in vitro growth of enterocytes has now been reported in several studies using a variety of harvesting techniques [3, 4, 9-11, 18, 191. It remains unclear, however, whether long-term culture is possible and if these cultured enterocytes can establish normal structure, function, and proliferative capacity [ 191. The basement membrane of the intestine is a thin continuous extracellular matrix that separates the epithelium from the underlying subepithelial tissues. Its major components, including Type IV collagen, the glycoprotein laminin, heparan sulphate proteoglycan, and other proteins, have been identified by immunofluorescence
and immunoelectron microscopy [20]. The basement membrane has a role in organogenesis and cellular attachment, polarization, and differentiation [ 13, 21-221. Hahn et al. [lo] found that there was a fourfold increase in the number of attached intestinal epithelial cells in cultures when these cells were plated on pepsin-soluble biomatrix compared to plain plastic surfaces. Presumably, cell attachment is the result of the interaction of specific receptors of the epithelial cell plasma membrane with components of the basement membrane. Although the presence of subepithelial tissue may be important in these processes [21, 221, laminin and basement membrane extract alone have been observed to promote cellular differentiation of fetal rat enterocytes, independent of mesenchymal cell influence [13, 141. Both a combination of basement membrane components and laminin alone enhanced epithelial cell growth in vitro in the present study. This observation has been made by others [13, 141. The combination of components was more effective in promoting epithelial growth than laminin alone. This is consistent with our previous observations in an in oivo model of intestinal regeneration in the rabbit and other in vitro studies [12, 13, 151.
NGUYEN,
THOMPSON,
AND
SHARP:
ISOLATED
FIG. 4. Photograph of plain (left), laminin-coated (middle), and Matrigel-coated lial growth in the flasks coated with basement membrane components.
Because Matrigel appeared to promote more growth than laminin in uitro, we chose to infuse this biomolecular preparation with harvested cells into the in uivo pouches. Maintaining viability of transplanted enterocytes in vivo has been extremely difficult. We previously harvested enterocytes with a chelation technique and infused the cells directly into a serosal pouch [ll]. There was no evidence of enterocyte growth. Similarly, in the
TABLE
1
Comparison of Cytokeratin Staining in Serosal Pouches 14 Days after Infusion of Enterocytes
Control serosa n=5 Percent surface with cytokeratin staining Thickness of cell layer stained Intensity of cytokeratin staining Composite score Composite score > 2
26
f12
1.0 +
Cells infused directly in growth media n=6
Cell cultured in vitro and infused n=3
35
13
f 21
f 10**
0.0
1.6 +
0.8
1.1 +
1.4 + 0.5 0.4 + 0.2 O/5
1.8 + 1.0 f
0.9 0.8*
1.5 + 1.1 0.5 + 0.7** O/3
* P < 0.05 vs control. ** P < 0.05 vs direct infusion.
O/6
0.7
ENTEROCYTE
515
GROWTH
(right) flasks demonstrating
greater confluence
of epithe-
present study harvested cells infused into serosal pouches either directly or after in vitro culture failed to increase the appearance of cytokeratin positive cells in the pouch. Vacanti et al. [3] reported that only 13% of animals having intestinal cells seeded in biodegradable polymers and implanted in the omentum showed evidence of sustained growth. Arnaout et al. [4] found that 50% of enterocyte transplants were successful for up to 3 weeks when placed intraperitoneally on collagen-coated dextron microcarriers. Basement membrane components added to the growth media significantly increased the quantity and density of cytokeratin-stained cells in the serosal pouches compared to control serosa and cells in growth media alone. Two thirds of the animals had substantial cytokeratin staining at both 1 and 2 weeks. At 1 week more than 50% of the basement membrane surface examined had cytokeratin positivity and the epithelial layer was routinely several cells deep with moderate staining intensity. These effects were less marked at 2 weeks and had diminished markedly at 4 weeks after infusion. This suggests that primarily differentiated cells were infused and because of limited proliferative capacity did not maintain long-term viability in the pouches. However, the presence of Matrigel did permit initial attachment and proliferation of these cells. The failure of the cells grown in vitro and subsequently seeded in the pouches to increase cytokeratin expression is potentially an important observation. The ability to amplify the harvested cells and perhaps repeat infusion on multiple occasions may be important for the
516
JOURNAL
OF SURGICAL
RESEARCH:
of Cytokeratin
Staining
in Serosal
Pouches
52, NO. 5, MAY
1992
2
TABLE Comparison
VOL.
at Different
Times
after
Infusion
of Cells with
Matrigel
Cells with Matrigel Control serosa n=5 Percent surface with cytokeratin Thickness of cell layer stained Intensity of cytokeratin staining Composite score Composite score > 2
staining
1 Week n=3 52 2.0 2.6 2.7 2/3
26 k 12 1.0 f 0.0 1.4 f 0.5 0.4 k 0.2 O/5
+ 28* f 0.5* f 0.9* f 2.0* (67%)*
2 Weeks n=7 36 2.2 2.3 2.3 5/7
t 20 f 0.8* k LO* zk 1.8* (71%)*
4 Weeks n=3 19 f lS# 1.5 i 0.7* 2.2 f 1.5 0.7 It 0.7# 0/3#
* P < 0.05 vs control serosa. ** P < 0.05 vs 1 and 2 weeks.
clinical application of this technique. Furthermore, Woehrle et al. [8] found that pancreatic islet cells grown in vitro prior to in vivo infusion underwent less vigorous rejection than those infused directly. Unfortunately, it appears that current in vitro techniques for small intestinal epithelial cells support amplification and differentiation rather than self-renewal characteristics [13]. There was insufficient epithelial growth in the present study to assess function. Arnaout et al. [4] found that transplanted enterocytes would replace bilirubin uridine diphosphoglucuronyl transferase activity in deficient Gunn rats. However, other metabolic activities have not been demonstrated and this will be an important aspect of future studies. While we have clearly demonstrated increased cellularity and cytokeratin expression in the pouches we have not shown conclusively that these changes represent growth of the isolated and transplanted enterocytes. Functional studies or other identifying markers will be necessary. We cannot exclude the possibility that infusion of the cells, media, and Matrigel altered the cytokeratin expression of existing mesothelial cells on the serosal surface. Although attachment and initial growth of enterocytes are presumed to have occurred in the serosal pouches of the Matrigel-supplemented animals in this study, enterocyte transplantation will be feasible only if long-term growth, differentiation, and function of intestinal epithelia occur. Several factors may be important in this process. The maturity of the transplanted cell, i.e., stem cells versus villus cells may influence proliferative capacity. Since the interaction between the epithelial cells and viable mesenchymal cells is important for cellular polarization and differentiation, transplantation of mesenchymal cells in conjunction with epithelial cells may be prerequisite for growing functionally absorptive cells [23]. Alternatively or additionally, preparation of the mesenchymal surface, e.g., stimulation of angiogenesis may be necessary. Various trophic factors are important in epithelial proliferation and differentiation. These have been employed in in vitro culture systems and may also be important in attempts to stimulate and
sustain growth of epithelial cells in vivo. The presence of luminal contents may also be important to promote growth and differentiation of cells in the serosal pouches. Thus, placing the serosal pouches in continuity with the gastrointestinal tract may stimulate more growth and differentiation of the transplanted enterocytes. In summary, basement membrane components enhance enterocyte growth in vitro. A combination of components promoted better growth than laminin alone. While growth of enterocytes was not conclusively demonstrated in vivo, basement membrane components did substantially increase the cellularity and cytokeratin expressivity of epithelial cells lining the serosal pouch. This beneficial effect was relatively short lived, however, suggesting that other factors will be needed to permit long-term growth of enterocytes in vivo. ACKNOWLEDGMENTS The authors gratefully acknowledge the technical assistance of Susan K. Meyer, Don Daley, George Pallas, and Sally Mann and the secretarial support of Anita Solon.
REFERENCES 1. 2.
3.
4.
5.
Schraut, W. H. Current status of small bowel transplantation. Gastroenterology 94: 525, 1988. Williams, J. W., McClellan, T., Peters, T. G., Nog, S., Dean, P., Banner, B., Vera, S. R., and Stenz, F. Effect of pretransplant graft irradiation on canine intestinal transplantation. Surg. Gynecol. O&et. 167: 197, 1988. Vacanti, J. P., Morse, M. A., Saltzman, W. M., Domb, A. J., Perez-Atayde, A., and Langer, R. Selective cell transplantation using bioabsorbable artificial polymers as matrices. J. Pediatr. Surg. 23: 3, 1988. Arnaout, W. S., Moscioni, A. D., Felcher, A. H., Barbour, R. L., Brown, L. L., and Demetriou, A. A. Intraperitoneal transplantation of microcarrier attached enterocytes in rats. Am. J. Surg. 157: 85, 1989. Maganto, P., Cienfuegos, J. A., Santamaria, L., Rodriguez, V., Eroles, G., Andres, S., Castillo-Olivares, J. S., and Municio,
NGUYEN,
THOMPSON,
A. M. Auxiliary liver by transplanted cytes. J. Surg. Res. 48: 24, 1990.
AND
frozen-thawed
SHARP:
ISOLATED
hepato-
6.
Arnaout, W. S., Moscioni, A. D., Barbour, R. L., and Demetriou, A. A. Development of bioartificial liver: bilirubin conjugation in Gunn rats. J. Surg. Res. 48: 379, 1990. Gray, D. W. R. Islet isolation and transplantation techniques in the primate. Surg. Gynecol. Obstet. 170: 225, 1990. 8. Woehrle, M., Markmann, J. F., Silvers, W. K., Barker, C. F., and Naji, A. Transplantation of cultured pancreatic islets to BB rats. Surgery 100: 334, 1986. 9. Quaroni, A., Wands, J., Trelstad, R. L., and Isselbacher, K. J. Epithelioid cell cultures from rat small intestine. Characterization by morphologic and immunologic criteria. J. Cell. Biol. 80: 248, 1979. 10. Hahn, U., Cho, A., Schuppan, D., Hahn,E. G., Merker, H. J.,and Riecken, E. 0. Intestinal epithelial cells preferentially attach to a biomatrix derived from human intestinal mucosa. Gut 28(Sl): 153, 1987. 11. Nguyen, B. L. T., and Thompson, J. S. Growing small intestinal neomucosa in serosa lined pouches. Curr. Surg. 47: 269, 1990. 12. Thompson, J. S. Basement membrane components stimulate epithelialization of intestinal defects in vivo. Cell Tissue Kinetics 23: 443, 1990. 13. Hahn, U., Stallmach, A., Hahn, E. G., and Riecken, E. 0. Basement membrane components are potent promoters of rat intestinal epithelial cell differentiation in vitro. Gustroenterology 98: 322, 1990. 14. Olson, A. D., Davidson, M., Psycher, T., and Bienkoski, R. La-
15.
16. 17. 18.
19.
20.
21.
22. 23.
ENTEROCYTE
GROWTH
517
minin is a major factor in spreading and pattern formation of intestinal epithelial cells in culture. Gastroenterology 94: A333, 1988. Kleinman, H. K., McGarvey, M. L., Hassell, J. R., Star, V. L., Cannon, F. B., Laurie, G. W., and Martin, G. R. Basement membrane complexes with biological activity. Biochemistry 25: 312, 1986. Freshney, R. I. Culture of Animal Cells, 2nd ed., pp. 115-118. New York: A. R. Liss, 1987. Tennant, J. R. Evaluation of the trypan blue technique for determination of cell viability. Transplantation 2: 685, 1964. Whitehead, R. H., Brown, A., and Bhathal, P. S. A method for the isolation and culture of human colonic crypts in collagen gels. In Vitro 23(6): 436, 1987. Madara, J. L., Stafford, J., Dharmsathaphorn, K., and Carlson, S. Structural analysis of a human intestinal epithelial cell line. Gastroenterology 92: 1133, 1987. Timpl, R., and Dziadek, M. Structure, development and molecular pathology of basement membranes. Znt. Reu. Cell Biol. 2: 27, 1986. Hahn, U., Schuppan, D., Hahn, E., Merker, H. J., and Riecken, E. 0. Intestinal cells produce basement membrane proteins in vitro. Gut 28(Sl): 143, 1987. Ekblom, P. Developmentally regulated conversion of mesenthyme to epithelium. FASEB J. 3: 2131,1989. Kedinger, M., Simon-Assman, P., and Haffen, K. Growth and differentiation of intestinal endodermal cells in a coculture system. Gut 28(Sl): 237,1987.
