- Email: [email protected]
Isolation of murine small bowel intraepithelial lymphocytes
Journal of Immunological Methods, 63 (1983) 35-44
35
Elsevier JIM 02745
Isolation of Murine Small Bowel Intraepithelial Lymphocytes George S. Leventon ~'*, Sulabha S. Kulkarni 2, Marvin L. Meistrich 3, James R. Newland 4 and Axel R. Zander 2 I The University of Texas Health Science Center at Houston, Dental Branch and Graduate School of Biomedical Sciences, Dental Science Institute, P.O. Box 20068, Houston, TX 77225, z Department of Developmental Therapeutics and ~ Department of Experimental Radiotherapy, The University of Texas M.D. Anderson Hospital and Tumor Institute, Houston, TX, and 4 Department of Patholo,gy, The University of Texas Dental Branch, Houston, TX, U.S.A.
(Received 17 September 1982, accepted 30 March 1983)
A method for the preparation of cellular suspensions of the epithelium of murine small bowel for the purpose of isolation and recovery of intraepithelial lymphocytes employing intraluminal hyaluronidase digestion is described. Discontinuous Percoll centrifugation of these suspensions yielded a 76-82% pure population of intraepithelial lymphocytes. Between 1.6 and 4.0x 10 7 viable intraepithelial lymphocytes were obtained per gram of gut. The isolate contained approximately equal numbers of T, B and null lymphocytes. Key words: intraepithelial lymphocytes - - small bowel lymphocytes
Introduction The l y m p h o c y t e population of the gut is partitioned by the basement m e m b r a n e into lamina propria l y m p h o c y t e (LPL) and intraepithelial lymphocyte (IEL) subpopulations. IELs form a heterogeneous population of cells, and in the h u m a n gut were f o u n d to contain approximately equal numbers of T, B, and null lymphocytes (Bartnik et al., 1980). They form part of a migratory population of mucosal m e m b r a n e associated l y m p h o i d cells (Toner and Ferguson, 1971; Glaister, 1973; G u y - G r a n d et al., 1974) undergoing a maturation phase (Ropke and Everett, 1976). T h e y function in i m m u n e responses on and within intestinal mucosa and m a y play a significant role in the genesis of graft versus host disease (Elson et al., 1977; M o w a t and Ferguson, 1981). Gut lymphocytes have been isolated by various techniques from rabbits (Rudzik
* To whom reprint requests should be sent. 0022-1759/83/$03.00 © 1983 Elsevier Science Publishers B.V.
36 and Bienenstock, 1974), guinea pigs (Arnaud-Battandier et al., 1978), man (CIancy and Pucci, 1978; Crofton et al., 1978; Bookman and Bull, 1979: Goodacre et al., 1979; Bartnik et al., 1980), and recently, mice (Tagliabue et al., 1982). Suspensions of pooled IELs and LPLs have been prepared by mechanical disruption of the gut (Rudzik and Bienenstock, 1974; Clancy and Pucci, 1978; Goodacre et al., 1979) or by enzymatic digestion (Crofton et al., 1978). Separate IEL and LPL pools have been isolated by sequential procedures with the epithelial layer removed first by gentle mechanical manipulation (Arnaud-Battandier et al., 1978) or chelation (Rudzik and Bienenstock, 1974; Bookman and Bull, 1979; Bartnik et al., 1980) followed by enzymatic digestion of the lamina propria. Epithelial layer cell suspensions have been purified by passage through glass bead columns (Goodacre et al., 1979), CaF~ bead columns (Clancy and Pucci, 1978), or glass bead columns followed by centrifugation through bovine serum albumin (Rudzik and Bienenstock, 1974), or Ficoll-Hypaque flotation (Arnaud-Battandier et al.. 1978), and recently by passage through nylon wool followed by centrifugation through Percoll (Tagliabue et al., 1982). The present study was designed to develop a method whereby murine small bowel 1ELs could be conveniently and quickly isolated in sufficient numbers, purity, and viability for transplantation into mice, thereby making it possible to study the effects of these cells on recovery of hematopoiesis, lymphopoiesis, and repopulation of gut-associated lymphoid tissue. Specifically, the goal was to isolate sufficient numbers of viable IELs for simultaneous transplantation of 50 mice with 2 × 105 cells per animal, for a total of 1 × 10v lymphocytes.
Materials and Methods
Six to eight-week-old male BDF~ mice, fed standard laboratory chow and acidified water ad libitum, were used in all experiments. The method described below used intraluminal hyaluronidase digestion to remove the epithelial layer. The animals were killed by cervical dislocation, and the entire small bowel was removed intact. The average length and weight of the small bowel were 36 cm and 1.1 g respectively. The intestine was placed on phosphate buffered saline (PBS) soaked paper toweling, and the proximal end was cannulated with a blunt 27-gauge needle clamped into place by a hemostat. The bowel was flushed with 10-15 ml of PBS. Gentle peristaltic massage was applied to remove mucus and chyme. A hyaluronidase-containing solution was then instilled. When the digestion solution had displaced the PBS wash, the distal end was clamped shut and the gut slowly distended with a total volume of 4.5 ml of fluid. In some early experiments the gut was left flaccid. The digestion solution consisted of 4700 U USP of bovine testis hyaluronidase (U.S. Biochemical Corp., 38593) and 100 U each of penicillin and streptomycin per ml in Eagle's minimum essential medium (MEM) supplemented with amino acids and vitamins, 20% fetal bovine serum, and 0.1% glucose. During the digestion, the gut was bathed in PBS and incubated at 37°C for 20, 30, 60, or 90 rain. After incubation, the distal end was opened and the contents (epithelial cells and IELs)
37 harvested by gentle milking and a 2-3 ml PBS wash. Upon harvesting, the cell suspensions were vigorously agitated for several seconds, and then immediately diluted with centrifugation medium (Percoll) chilled at 0-4°C. Discontinuous Percoll gradients (30%, 40% and 60% or 30%, 40%, and 70%) were constructed in nitrocellulose tubes (Beckman 302237). The sample was placed within the 30% Percoll layer which contained PBS, 5% fetal bovine serum, 0.1% glucose, and 80 ~ g / m l of crude DNase (bovine pancreas; Sigma Biochemical Corp., D0876 (DN)). The composition of the other layers was similar except for the deletion of the crude DNase. Quantitatively the 30% layer was 25 ml, and the others 6 ml. Gradients were spun for 25 min at 628 × gay in a 0 - 4 ° C Sorval RC5B centrifuge using an HB-4 swinging bucket rotor at slow acceleration and the brake off at deceleration below 1000 rpm. Gradient samples were collected in a syringe by puncturing the side of the centrifuge tube with a 26-gauge needle after applying a seal of stopcock grease. Cellular viability was determined by exclusion of 0.4% trypan blue. Cells were enumerated in a hemocytometer under phase contrast microscopy. All tissue for light microscopic examination was fixed in formalin, paraffin embedded, sectioned at 6 /~m intervals and stained with hematoxylin and eosin. Specimens for electron microscopy were fixed in chilled 2% glutaraldehyde and processed in epon. Cell suspensions were converted into cell pellets by centrifugation for processing. Thick sections were cut at 1/~m intervals and stained with methylene blue, and thin sections for transmission electron microscopy were stained with lead citrate and uranyl acetate. The IEL isolate was examined for Thy 1.2, Ly 1, Ly 2 and IgA + M + G surface antigens by direct immunofluorescence. Thymus and peripheral blood cell suspensions were preincubated for 1 h at 37°C in the digest solution with or without hyaluronidase to determine if the enzyme treatment altered the expression of the above markers. Cells were washed in MEM, pelleted and resuspended to 2.0 x l0 T cells per ml. Fifty microliters of the cell suspension were mixed with 50 #1 of FITC conjugated antisera in a flat bottom 0.4 ml multiwell plastic plate (Becton Dickinson, Falcon 3070) and kept on ice for 45 rain. The antisera used were murine monoclonal anti-Thy 1.2, Ly l, or Ly 2 (Becton Dickinson, 1333, 1343, and 1353) diluted in MEM to 1 : 50, 1 : 10 and 1 : 50 respectively, or goat anti-mouse IgA + M + G (Cappel Laboratories, 1211-0231) diluted 1:50 in MEM. After standing, the cells were pelleted by centrifugation at 350 x g for 5 rain, the supernatant removed by aspiration, the cells washed in 0.15 ml MEM twice and finally resuspended in 1-2 drops of glycerol. A portion of a drop of the final suspension was placed on a clean slide with a wooden applicator, cover-slipped, and examined under an ultraviolet microscope. Between 100 and 200 cells were counted and scored for the percentage fluorescence. Three different IEL isolates were examined with slides of each antigen stain prepared in duplicate.
38 Results
Preparation of cell suspensions Instillation of the hyaluronidase solution without distention of the gut produced erratic tissue separation of the epithelium from the lamina propria (Fig. IA). The separation efficiency was greatly improved by distention of the gut with the digestion solution. Hyaluronidase digestion with distention of the gut followed by mechanical milking and a PBS flush removed almost all of the tissue above the basement membrane, leaving the villus cores and Peyer's patches intact (Fig. 1B and C). Epithelial tissue harvested after 20 or 30 min digestion yielded large sheets of cells, whereas 60 and 90 min digestion produced many free cells and small clusters of cells (Fig. 2A). The suspensions were devoid of fibroblast and red blood cell contamination. These results indicated that the lamina propria remained intact. The average total yield per animal was 4.1 × 10 ~ free cells and 1.9 x 10 ~ viable cells (Table I).
Purification of intraepithelial lyrnpho~ytes An average of 2.1 x 108 total cells were loaded within the 30% layer of the Percoll gradients (Table II). During the preliminary experiments the densest layer of Percoll was 60%, whereas in all subsequent experiments it was 70%. Bands formed at the top of these gradients, and at the 30-40% and 40-70% interfaces. The band at the top of the gradients contained dead epithelial cells and debris, the band at the 30-40% interface consisted of epithelial cells, and the 40-70% interface band contained
39
Fig. 1. Efficiency of removal of the epithelial layer of the gut after intraluminal hyaluronidase digestion under flaccid (A) and distended (B and C) conditions. Note the preservation of the integrity of the Peyer's patch in (C).
40
• 0 ~
-
~
Fig. 2. Cell suspension before centrifugation (A) and post-centrifugation lymphocyte-rich 40 70% Percoll interface (B). Arrows in (A) point to intraepithelial lymphocytes.
41 TABLE 1 SUMMARY OF SINGLE CELL SUSPENSION YIELDS Mean ± one standard deviation, 4 experiments. Total yield Viable yield % of yield viable
4.1 ± 2.2 x l0 ~ 1.9± 1.3 X 108 4 8 ± 17%
TABLE II SUMMARY OF D I S C O N T I N U O U S PERCOLL C E N T R I F U G A T I O N EXPERIMENTS Mean ± one standard deviation, 4 experiments.
Cells loaded 30-40% band 40-70% band
Total cells
Viable cells
% of viable cells loaded 100% 4.9_+2.3% 19.9 ± 2.5%
2.1 ± 0.6 x 108
8.0 ± 0.8 x 107
4 . 9 ± 2 . 0 X l0 6
3.8±2.0×
1.7 _+0.3 x 107
1.6 ± 0.3 x 107
10 6
Fig. 3. Transmission electron microscopy of the lymphocyte-rich 40-70% Percoll interface. Cells 'a' and ' b ' which might be misinterpreted as macrophages and plasma cells are goblet cells cut on transverse section as indicated in the inset of the in situ goblet cell.
42 TABLE Ill INTRAEPITHELIAL LYMPHOCYTE SURFACE ANTIGENS Mean ± 1 standard deviation, 3 experiments. Thy 1.2 Ly 1 Ly 2 IgA+ M +G Null
35.0 + 3.8% 33.4+_5.8% 32.0 ± 4.9% 29.1 +_1.4~ 36.0 + 4.8c~
lymphocytes. When the densest layer of Percoll was 60%, bacterial and protozoal gut flora contaminated the lymphocyte-rich fraction, which appeared as a pellet, The predominance of lymphocytes in the 40-70% band was confirmed by light microscopy of 1 /xm epon thick sections and by transmission electron microscopy (Figs. 2B and 3). Differential counts of thin sections of the 40-70% band revealed a population of 76-82% lymphocytes. The contaminant cells were identified as brush border and goblet cells cut in varying planes of section (Fig. 3 inset). The mean yield in the 40-70% interface band was 1.7 × 107 cells with 95% viable (Table II). Of the initially loaded viable cells, a mean of 20% were recovered in this layer. The 30 40% interface band contained 5% of the initial viable cell load.
Surface antigens No quantitative or qualitative differences in staining for Thy 1.2, Ly 1, Ly 2, or IgA + M + G were noted between huffy coat or thymus cell preparations preincubated with or without hyaluronidase. The IEL preparations contained an average of 35% T-cells (Thy 1.2), 29.1% B-cells (IgA + M + G) and 36.0% null cells. Of the T-cell subgroup 95.4% were positive for Ly 1 and 91.4% for Ly 2 (Table III).
Discussion
A method has been developed for the rapid preparation of concentrated IEL suspensions from the intact murine small bowel by the instillation, recovery, and centrifugation of small volumes of a hyaluronidase-containing solution. This technique provided excellent access and even distribution of the enzyme solution on the epithelial surfaces, and permitted nearly 100% removal of the tissues external to the basement membrane without disrupting the subjacent connective tissue. Percoll (polyvinylpyrrolidone-coated colloidal silica) was chosen as the centrifugation medium because of its reported low viscosity, negligible effect on solution osmotic pressure, lack of toxicity, and lack of free polymer (Pertoft et al., 1977). Initially, 3 ml of undiluted cell suspensions were layered over 15 ml continuous 20-65% Percoll gradients. (These were spun at 13,000 × gay in a Beckman model L550 centrifuge with an SW27 rotor for 10 min at 0 - 4 ° C . ) This approach produced an overloading effect that created a dense mat of material at the 0-20% interface.
43 Larger (39 ml) discontinuous gradients with the relatively dilute (25 ml) sample contained within the 30% layer did not induce the overloading effect. The final yield of purified viable IELs per gram of murine small bowel was between 1.6 x 10 7 and 4.0 X 10 7 cells. These yields per gram of intact gut were higher than the reported yields obtained in other species. Rudzik and Bienenstock (1974) obtained 3.7 × 10 6 cells (IELs and LPLs) per gram of rabbit ileum, and Goodacre et al. (1979) achieved a yield of 6.6 x 10 6 cells (IELs and LPLs) per gram of human small intestine. Crofton et al. (1978) recovered 1.1 x 108 IELs per gram of human tissue, but the muscularis had been dissected free prior to weighing. Our isolation method demonstrated several advantages. Notably, it was: (1) rapid, taking about 90 rain to complete; (2) convenient, having a minimum of tissue manipulations and using small volumes (4.5 ml) of digestion solution (which was prepared in batch and frozen as single use portions); (3) selective for the epithelial layer, producing suspensions that included almost all of the tissue above the basement membrane while leaving the lamina propria intact; (4) efficient, with an IEL yield per mouse approximately 4-fold higher than that recently reported by Tagliabue et al. (1982). The numbers of T, B, and null lymphocytes isolated by this method were found to be approximately equal. Similar percentages have been reported for murine small bowel intraepithelial T-cell (Tagliabue et al., 1982) and human colon T, B, and null IELs (Bartnik et al., 1980). The high percentage of T-cells bearing both Ly 1 and Ly 2 reflects a relative immaturity of this subgroup of lymphocytes. This suggests that the intraepithelial location may be a site of mucosal membrane-associated T-cell maturation. The technique of isolating small bowel IELs described herein may be useful in studying the mucosal immune system in the murine model. A reliable quick method for obtaining sufficient purified viable IELs for transplantation of 2 x 105 lymphocytes per animal in 50 mice, for a total of 1.0 x 10 7 cells has been achieved. Recent studies (Leventon, manuscript in preparation) have demonstrated the viability of the isolated IELs in a transplantation assay. Mice lethally irradiated were rescued by a minimal dose of bone marrow cells. Recovery of IELs and plasma cells in the small bowel was augmented by the addition of purified IELs prepared by this technique.
Acknowledgements This investigation was supported by Training Grant DEO 7029-05, and Grants CA-06294 and CA-24770 from The National Institutes of Health.
References Arnaud-Battandier, F., B.M. Bundy, M. O'Neill, J. Bienenstockand D. Nelson, 1978, J. hnmunol. 121, 1059. Barmik, W., S.G. Re Mine, M. Chiba, W.R. Thayer and R.G. Shorter, 1980, Gastroenterology78. 976.
44 Bookman, M.A. and D.M. Bull, 1979, Gastroenterology 77, 503. Clancy, R. and A. Pucci, 1978, Gut 19, 273. Crofton, R.W., C. Cochrane and D.B.L. McClelland, 1978, Gut 19, 893. Elson, C.O., R.W. Reilly and I.H, Rosenberg, 1977, Gastroenterology 72, 886. Glaister, J.R., 1973, Int. Arch. Allergy Appl. Immunol. 45, 854. Goodacre, R., R. Davidson, D. Singal and J. Bienenstock, 1979, Gastroenterology 76, 300. Guy-Grand, D., C, Griscelli and P. Vassalli, 1974, Eur. J. lmmunol. 4, 435. Mowat, A.M. and A. Ferguson, 1981, Transplantation 32, 238. Pertoft, H., K. Rubin, L. Kjellen, T.G. kaurent and B. Klingeborn, 1977, Exp. Cell Res. 110, 449. Ropke, C. and N.B. Everett, 1976, Am. J. Anat. 145, 395. Rudzik, O. and J. Bienenstock, 1974, Lab. Invest. 30, 260. Tagliabue, A., A. Befus, D. Clark and J. Bienenstock, 1982, J. Exp. Med. 155, 1785. Toner, P.G. and A. Ferguson, 1971, J. Ultrastruct. Res, 34, 329.
35
Elsevier JIM 02745
Isolation of Murine Small Bowel Intraepithelial Lymphocytes George S. Leventon ~'*, Sulabha S. Kulkarni 2, Marvin L. Meistrich 3, James R. Newland 4 and Axel R. Zander 2 I The University of Texas Health Science Center at Houston, Dental Branch and Graduate School of Biomedical Sciences, Dental Science Institute, P.O. Box 20068, Houston, TX 77225, z Department of Developmental Therapeutics and ~ Department of Experimental Radiotherapy, The University of Texas M.D. Anderson Hospital and Tumor Institute, Houston, TX, and 4 Department of Patholo,gy, The University of Texas Dental Branch, Houston, TX, U.S.A.
(Received 17 September 1982, accepted 30 March 1983)
A method for the preparation of cellular suspensions of the epithelium of murine small bowel for the purpose of isolation and recovery of intraepithelial lymphocytes employing intraluminal hyaluronidase digestion is described. Discontinuous Percoll centrifugation of these suspensions yielded a 76-82% pure population of intraepithelial lymphocytes. Between 1.6 and 4.0x 10 7 viable intraepithelial lymphocytes were obtained per gram of gut. The isolate contained approximately equal numbers of T, B and null lymphocytes. Key words: intraepithelial lymphocytes - - small bowel lymphocytes
Introduction The l y m p h o c y t e population of the gut is partitioned by the basement m e m b r a n e into lamina propria l y m p h o c y t e (LPL) and intraepithelial lymphocyte (IEL) subpopulations. IELs form a heterogeneous population of cells, and in the h u m a n gut were f o u n d to contain approximately equal numbers of T, B, and null lymphocytes (Bartnik et al., 1980). They form part of a migratory population of mucosal m e m b r a n e associated l y m p h o i d cells (Toner and Ferguson, 1971; Glaister, 1973; G u y - G r a n d et al., 1974) undergoing a maturation phase (Ropke and Everett, 1976). T h e y function in i m m u n e responses on and within intestinal mucosa and m a y play a significant role in the genesis of graft versus host disease (Elson et al., 1977; M o w a t and Ferguson, 1981). Gut lymphocytes have been isolated by various techniques from rabbits (Rudzik
* To whom reprint requests should be sent. 0022-1759/83/$03.00 © 1983 Elsevier Science Publishers B.V.
36 and Bienenstock, 1974), guinea pigs (Arnaud-Battandier et al., 1978), man (CIancy and Pucci, 1978; Crofton et al., 1978; Bookman and Bull, 1979: Goodacre et al., 1979; Bartnik et al., 1980), and recently, mice (Tagliabue et al., 1982). Suspensions of pooled IELs and LPLs have been prepared by mechanical disruption of the gut (Rudzik and Bienenstock, 1974; Clancy and Pucci, 1978; Goodacre et al., 1979) or by enzymatic digestion (Crofton et al., 1978). Separate IEL and LPL pools have been isolated by sequential procedures with the epithelial layer removed first by gentle mechanical manipulation (Arnaud-Battandier et al., 1978) or chelation (Rudzik and Bienenstock, 1974; Bookman and Bull, 1979; Bartnik et al., 1980) followed by enzymatic digestion of the lamina propria. Epithelial layer cell suspensions have been purified by passage through glass bead columns (Goodacre et al., 1979), CaF~ bead columns (Clancy and Pucci, 1978), or glass bead columns followed by centrifugation through bovine serum albumin (Rudzik and Bienenstock, 1974), or Ficoll-Hypaque flotation (Arnaud-Battandier et al.. 1978), and recently by passage through nylon wool followed by centrifugation through Percoll (Tagliabue et al., 1982). The present study was designed to develop a method whereby murine small bowel 1ELs could be conveniently and quickly isolated in sufficient numbers, purity, and viability for transplantation into mice, thereby making it possible to study the effects of these cells on recovery of hematopoiesis, lymphopoiesis, and repopulation of gut-associated lymphoid tissue. Specifically, the goal was to isolate sufficient numbers of viable IELs for simultaneous transplantation of 50 mice with 2 × 105 cells per animal, for a total of 1 × 10v lymphocytes.
Materials and Methods
Six to eight-week-old male BDF~ mice, fed standard laboratory chow and acidified water ad libitum, were used in all experiments. The method described below used intraluminal hyaluronidase digestion to remove the epithelial layer. The animals were killed by cervical dislocation, and the entire small bowel was removed intact. The average length and weight of the small bowel were 36 cm and 1.1 g respectively. The intestine was placed on phosphate buffered saline (PBS) soaked paper toweling, and the proximal end was cannulated with a blunt 27-gauge needle clamped into place by a hemostat. The bowel was flushed with 10-15 ml of PBS. Gentle peristaltic massage was applied to remove mucus and chyme. A hyaluronidase-containing solution was then instilled. When the digestion solution had displaced the PBS wash, the distal end was clamped shut and the gut slowly distended with a total volume of 4.5 ml of fluid. In some early experiments the gut was left flaccid. The digestion solution consisted of 4700 U USP of bovine testis hyaluronidase (U.S. Biochemical Corp., 38593) and 100 U each of penicillin and streptomycin per ml in Eagle's minimum essential medium (MEM) supplemented with amino acids and vitamins, 20% fetal bovine serum, and 0.1% glucose. During the digestion, the gut was bathed in PBS and incubated at 37°C for 20, 30, 60, or 90 rain. After incubation, the distal end was opened and the contents (epithelial cells and IELs)
37 harvested by gentle milking and a 2-3 ml PBS wash. Upon harvesting, the cell suspensions were vigorously agitated for several seconds, and then immediately diluted with centrifugation medium (Percoll) chilled at 0-4°C. Discontinuous Percoll gradients (30%, 40% and 60% or 30%, 40%, and 70%) were constructed in nitrocellulose tubes (Beckman 302237). The sample was placed within the 30% Percoll layer which contained PBS, 5% fetal bovine serum, 0.1% glucose, and 80 ~ g / m l of crude DNase (bovine pancreas; Sigma Biochemical Corp., D0876 (DN)). The composition of the other layers was similar except for the deletion of the crude DNase. Quantitatively the 30% layer was 25 ml, and the others 6 ml. Gradients were spun for 25 min at 628 × gay in a 0 - 4 ° C Sorval RC5B centrifuge using an HB-4 swinging bucket rotor at slow acceleration and the brake off at deceleration below 1000 rpm. Gradient samples were collected in a syringe by puncturing the side of the centrifuge tube with a 26-gauge needle after applying a seal of stopcock grease. Cellular viability was determined by exclusion of 0.4% trypan blue. Cells were enumerated in a hemocytometer under phase contrast microscopy. All tissue for light microscopic examination was fixed in formalin, paraffin embedded, sectioned at 6 /~m intervals and stained with hematoxylin and eosin. Specimens for electron microscopy were fixed in chilled 2% glutaraldehyde and processed in epon. Cell suspensions were converted into cell pellets by centrifugation for processing. Thick sections were cut at 1/~m intervals and stained with methylene blue, and thin sections for transmission electron microscopy were stained with lead citrate and uranyl acetate. The IEL isolate was examined for Thy 1.2, Ly 1, Ly 2 and IgA + M + G surface antigens by direct immunofluorescence. Thymus and peripheral blood cell suspensions were preincubated for 1 h at 37°C in the digest solution with or without hyaluronidase to determine if the enzyme treatment altered the expression of the above markers. Cells were washed in MEM, pelleted and resuspended to 2.0 x l0 T cells per ml. Fifty microliters of the cell suspension were mixed with 50 #1 of FITC conjugated antisera in a flat bottom 0.4 ml multiwell plastic plate (Becton Dickinson, Falcon 3070) and kept on ice for 45 rain. The antisera used were murine monoclonal anti-Thy 1.2, Ly l, or Ly 2 (Becton Dickinson, 1333, 1343, and 1353) diluted in MEM to 1 : 50, 1 : 10 and 1 : 50 respectively, or goat anti-mouse IgA + M + G (Cappel Laboratories, 1211-0231) diluted 1:50 in MEM. After standing, the cells were pelleted by centrifugation at 350 x g for 5 rain, the supernatant removed by aspiration, the cells washed in 0.15 ml MEM twice and finally resuspended in 1-2 drops of glycerol. A portion of a drop of the final suspension was placed on a clean slide with a wooden applicator, cover-slipped, and examined under an ultraviolet microscope. Between 100 and 200 cells were counted and scored for the percentage fluorescence. Three different IEL isolates were examined with slides of each antigen stain prepared in duplicate.
38 Results
Preparation of cell suspensions Instillation of the hyaluronidase solution without distention of the gut produced erratic tissue separation of the epithelium from the lamina propria (Fig. IA). The separation efficiency was greatly improved by distention of the gut with the digestion solution. Hyaluronidase digestion with distention of the gut followed by mechanical milking and a PBS flush removed almost all of the tissue above the basement membrane, leaving the villus cores and Peyer's patches intact (Fig. 1B and C). Epithelial tissue harvested after 20 or 30 min digestion yielded large sheets of cells, whereas 60 and 90 min digestion produced many free cells and small clusters of cells (Fig. 2A). The suspensions were devoid of fibroblast and red blood cell contamination. These results indicated that the lamina propria remained intact. The average total yield per animal was 4.1 × 10 ~ free cells and 1.9 x 10 ~ viable cells (Table I).
Purification of intraepithelial lyrnpho~ytes An average of 2.1 x 108 total cells were loaded within the 30% layer of the Percoll gradients (Table II). During the preliminary experiments the densest layer of Percoll was 60%, whereas in all subsequent experiments it was 70%. Bands formed at the top of these gradients, and at the 30-40% and 40-70% interfaces. The band at the top of the gradients contained dead epithelial cells and debris, the band at the 30-40% interface consisted of epithelial cells, and the 40-70% interface band contained
39
Fig. 1. Efficiency of removal of the epithelial layer of the gut after intraluminal hyaluronidase digestion under flaccid (A) and distended (B and C) conditions. Note the preservation of the integrity of the Peyer's patch in (C).
40
• 0 ~
-
~
Fig. 2. Cell suspension before centrifugation (A) and post-centrifugation lymphocyte-rich 40 70% Percoll interface (B). Arrows in (A) point to intraepithelial lymphocytes.
41 TABLE 1 SUMMARY OF SINGLE CELL SUSPENSION YIELDS Mean ± one standard deviation, 4 experiments. Total yield Viable yield % of yield viable
4.1 ± 2.2 x l0 ~ 1.9± 1.3 X 108 4 8 ± 17%
TABLE II SUMMARY OF D I S C O N T I N U O U S PERCOLL C E N T R I F U G A T I O N EXPERIMENTS Mean ± one standard deviation, 4 experiments.
Cells loaded 30-40% band 40-70% band
Total cells
Viable cells
% of viable cells loaded 100% 4.9_+2.3% 19.9 ± 2.5%
2.1 ± 0.6 x 108
8.0 ± 0.8 x 107
4 . 9 ± 2 . 0 X l0 6
3.8±2.0×
1.7 _+0.3 x 107
1.6 ± 0.3 x 107
10 6
Fig. 3. Transmission electron microscopy of the lymphocyte-rich 40-70% Percoll interface. Cells 'a' and ' b ' which might be misinterpreted as macrophages and plasma cells are goblet cells cut on transverse section as indicated in the inset of the in situ goblet cell.
42 TABLE Ill INTRAEPITHELIAL LYMPHOCYTE SURFACE ANTIGENS Mean ± 1 standard deviation, 3 experiments. Thy 1.2 Ly 1 Ly 2 IgA+ M +G Null
35.0 + 3.8% 33.4+_5.8% 32.0 ± 4.9% 29.1 +_1.4~ 36.0 + 4.8c~
lymphocytes. When the densest layer of Percoll was 60%, bacterial and protozoal gut flora contaminated the lymphocyte-rich fraction, which appeared as a pellet, The predominance of lymphocytes in the 40-70% band was confirmed by light microscopy of 1 /xm epon thick sections and by transmission electron microscopy (Figs. 2B and 3). Differential counts of thin sections of the 40-70% band revealed a population of 76-82% lymphocytes. The contaminant cells were identified as brush border and goblet cells cut in varying planes of section (Fig. 3 inset). The mean yield in the 40-70% interface band was 1.7 × 107 cells with 95% viable (Table II). Of the initially loaded viable cells, a mean of 20% were recovered in this layer. The 30 40% interface band contained 5% of the initial viable cell load.
Surface antigens No quantitative or qualitative differences in staining for Thy 1.2, Ly 1, Ly 2, or IgA + M + G were noted between huffy coat or thymus cell preparations preincubated with or without hyaluronidase. The IEL preparations contained an average of 35% T-cells (Thy 1.2), 29.1% B-cells (IgA + M + G) and 36.0% null cells. Of the T-cell subgroup 95.4% were positive for Ly 1 and 91.4% for Ly 2 (Table III).
Discussion
A method has been developed for the rapid preparation of concentrated IEL suspensions from the intact murine small bowel by the instillation, recovery, and centrifugation of small volumes of a hyaluronidase-containing solution. This technique provided excellent access and even distribution of the enzyme solution on the epithelial surfaces, and permitted nearly 100% removal of the tissues external to the basement membrane without disrupting the subjacent connective tissue. Percoll (polyvinylpyrrolidone-coated colloidal silica) was chosen as the centrifugation medium because of its reported low viscosity, negligible effect on solution osmotic pressure, lack of toxicity, and lack of free polymer (Pertoft et al., 1977). Initially, 3 ml of undiluted cell suspensions were layered over 15 ml continuous 20-65% Percoll gradients. (These were spun at 13,000 × gay in a Beckman model L550 centrifuge with an SW27 rotor for 10 min at 0 - 4 ° C . ) This approach produced an overloading effect that created a dense mat of material at the 0-20% interface.
43 Larger (39 ml) discontinuous gradients with the relatively dilute (25 ml) sample contained within the 30% layer did not induce the overloading effect. The final yield of purified viable IELs per gram of murine small bowel was between 1.6 x 10 7 and 4.0 X 10 7 cells. These yields per gram of intact gut were higher than the reported yields obtained in other species. Rudzik and Bienenstock (1974) obtained 3.7 × 10 6 cells (IELs and LPLs) per gram of rabbit ileum, and Goodacre et al. (1979) achieved a yield of 6.6 x 10 6 cells (IELs and LPLs) per gram of human small intestine. Crofton et al. (1978) recovered 1.1 x 108 IELs per gram of human tissue, but the muscularis had been dissected free prior to weighing. Our isolation method demonstrated several advantages. Notably, it was: (1) rapid, taking about 90 rain to complete; (2) convenient, having a minimum of tissue manipulations and using small volumes (4.5 ml) of digestion solution (which was prepared in batch and frozen as single use portions); (3) selective for the epithelial layer, producing suspensions that included almost all of the tissue above the basement membrane while leaving the lamina propria intact; (4) efficient, with an IEL yield per mouse approximately 4-fold higher than that recently reported by Tagliabue et al. (1982). The numbers of T, B, and null lymphocytes isolated by this method were found to be approximately equal. Similar percentages have been reported for murine small bowel intraepithelial T-cell (Tagliabue et al., 1982) and human colon T, B, and null IELs (Bartnik et al., 1980). The high percentage of T-cells bearing both Ly 1 and Ly 2 reflects a relative immaturity of this subgroup of lymphocytes. This suggests that the intraepithelial location may be a site of mucosal membrane-associated T-cell maturation. The technique of isolating small bowel IELs described herein may be useful in studying the mucosal immune system in the murine model. A reliable quick method for obtaining sufficient purified viable IELs for transplantation of 2 x 105 lymphocytes per animal in 50 mice, for a total of 1.0 x 10 7 cells has been achieved. Recent studies (Leventon, manuscript in preparation) have demonstrated the viability of the isolated IELs in a transplantation assay. Mice lethally irradiated were rescued by a minimal dose of bone marrow cells. Recovery of IELs and plasma cells in the small bowel was augmented by the addition of purified IELs prepared by this technique.
Acknowledgements This investigation was supported by Training Grant DEO 7029-05, and Grants CA-06294 and CA-24770 from The National Institutes of Health.
References Arnaud-Battandier, F., B.M. Bundy, M. O'Neill, J. Bienenstockand D. Nelson, 1978, J. hnmunol. 121, 1059. Barmik, W., S.G. Re Mine, M. Chiba, W.R. Thayer and R.G. Shorter, 1980, Gastroenterology78. 976.
44 Bookman, M.A. and D.M. Bull, 1979, Gastroenterology 77, 503. Clancy, R. and A. Pucci, 1978, Gut 19, 273. Crofton, R.W., C. Cochrane and D.B.L. McClelland, 1978, Gut 19, 893. Elson, C.O., R.W. Reilly and I.H, Rosenberg, 1977, Gastroenterology 72, 886. Glaister, J.R., 1973, Int. Arch. Allergy Appl. Immunol. 45, 854. Goodacre, R., R. Davidson, D. Singal and J. Bienenstock, 1979, Gastroenterology 76, 300. Guy-Grand, D., C, Griscelli and P. Vassalli, 1974, Eur. J. lmmunol. 4, 435. Mowat, A.M. and A. Ferguson, 1981, Transplantation 32, 238. Pertoft, H., K. Rubin, L. Kjellen, T.G. kaurent and B. Klingeborn, 1977, Exp. Cell Res. 110, 449. Ropke, C. and N.B. Everett, 1976, Am. J. Anat. 145, 395. Rudzik, O. and J. Bienenstock, 1974, Lab. Invest. 30, 260. Tagliabue, A., A. Befus, D. Clark and J. Bienenstock, 1982, J. Exp. Med. 155, 1785. Toner, P.G. and A. Ferguson, 1971, J. Ultrastruct. Res, 34, 329.