- Email: [email protected]
Human intestinal epithelial cell lines produce factor(s) that inhibit CD3-mediated T-lymphocyte proliferation
Immunology Letters 58 (1997) 159 – 165
Human intestinal epithelial cell lines produce factor(s) that inhibit CD3-mediated T-lymphocyte proliferation Andreas D. Christ a, Sean P. Colgan a, Steven P. Balk b, Richard S. Blumberg a,* a
Gastroenterology Di6ision, Thorn Biomedical Research Building, Room 1310, Brigham and Women’s Hospital, Har6ard Medical School, 75 Francis Street, Boston, MA 02115, USA b Beth Israel Hospital, Har6ard Medical School, Boston, MA 02115, USA Received 6 January 1997; received in revised form 4 April 1997; accepted 22 April 1997
Abstract Peripheral blood T lymphocytes (PBT) proliferate more to anti-CD3 stimulation than to anti-CD2 stimulation. On the other hand, fresh, but not cultivated, intestinal intraepithelial lymphocytes (iIEL) exhibit a lower response to CD3 stimulation in comparison to CD2. The goal of this study was to show that the anti-CD3 T-cell response depends on the microenvironment and is independent of the origin of the lymphocytes. Cultured T-cell lines were stimulated with either an anti-CD3 mAb or an anti-CD2 mAb. Either conditioned supernatant from intestinal epithelial cell (IEC) lines or non- conditioned medium (negative control) was added. After 2 days cytokine production and proliferation were measured. Conditioned supernatant decreased the proliferative response of small and large bowel iIEL compared to controls (P = 0.04). In the same experiments, the cytokine production was non-significantly decreased. Immortalized iIEL, that are not regularly stimulated by their CD3 pathway, showed a similar decrease in proliferation (P B0.001) and cytokine production (P = 0.01) when incubated with conditioned supernatant. Similar results were also obtained with a non-immortalized and an immortalized PBT line (P B 0.001). In a small bowel iIEL cell line, that exhibited a significant response to anti-CD2 stimulation, the proliferative response to anti-CD2 stimulation was preserved. Active conditioned supernatant could be generated from three independent IEC lines and a liver derived epithelial cell line, but not from a non-epithelial control cell line or two extraintestinal epithelial cell lines. We conclude that supernatants of cultured IEC contain soluble factor(s) that cause cultured iIEL and extraintestinal lymphocytes to behave like fresh iIEL. These results, therefore, support and extend the studies of others which suggest that the intestinal microenvironment mucosalizes lymphocytes. © 1997 Elsevier Science B.V. Keywords: Human intestinal epithelial cell lines; T-lymphocyte proliferation
1. Introduction Lymphocytes which reside in the intestinal epithelium, intestinal intraepithelial lymphocytes (iIEL), are a unique population of cells which likely represent a general component of the mucosal immune system of most mammalian species [1]. iIEL are located below the tight junctions of the intestinal epithelium in close contact to the basal and lateral membrane of the intes* Corresponding author. Tel: + 1 617 7326917; fax: + 1 617 7305807; e-mail: [email protected] 0165-2478/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 4 7 8 ( 9 7 ) 0 0 0 8 1 - 3
tinal epithelial cell. In humans, iIEL are predominantly T cells expressing CD8 and a limited array of TCR-ab, and to a lesser extent TCR-gd, indicating that they recognize a restricted range of antigens in the context of a MHC class I-related molecule [2]. Although the MHC class I restriction element has not yet been defined for iIEL, it is commonly accepted that the intracellular signals mediated by ligation of the TCR/CD3 complex are diminished. Freshly isolated iIEL respond 20 times less to PHA-P, Con A and anti-CD3 -sepharose than peripheral blood T lymphocytes as measured in a [3H]Thymidine proliferation assay [3]. This is specific for
160
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
the TCR/CD3 complex since the response to anti-CD2 stimulation [3], and corollarily sheep red blood cells [4], is preserved. Similar functional characteristics have been described for lamina propria lymphocytes (LPL) [5 – 7]. It is not known, however, whether this represents the diminished response of the CD4 and/or CD8 fraction of LPL. To an extent, these characteristics are also shared by murine iIEL, where iIEL are cytolytic but have a low proliferative response to anti-CD3 stimulation as compared to other T lymphocytes [8]. Although the function of iIEL is unknown, iIEL influence functional aspects of adjacent intestinal epithelial cells and vice versa. When intestinal epithelial cell monolayers are reconstituted with mucosal-derived lymphocytes in vitro, barrier function and ion transport properties are attenuated by soluble factors released by co-culture [9]. iIEL bearing the gd TCR also promote intestinal epithelial cell growth by secretion of keratinocyte growth factor [10]. On the other hand, intestinal epithelial cells are known to secrete factors that influence T cells, e.g., IL-8 [11], a known chemoattractant for T lymphocytes [12]. It is likely, therefore, that a bi-directional crosstalk exists between intestinal epithelial cells and iIEL. Since we have previously shown that the composition of the TCR/CD3 complex of the iIEL is similar to peripheral blood T cells [13], which exhibit normal response to TCR/CD3 complex mediated stimulation, it is possible that this functional phenotype of human iIEL is related, at least in part, to soluble mediators released by intestinal epithelial cells. As shown here, cultured intestinal epithelial cells are capable of secreting factor(s) that inhibit proliferation of iIEL in response to antiCD3 stimulation. Moreover, the modulatory effects of the factor(s) secreted by intestinal epithelial cells is also capable of influencing the CD3-mediated response of peripheral blood lymphocytes.
MD), 2 mM L-glutamine (Mediatech), 10 mM Hepes buffer (Mediatech) and 1% non-essential amino acids (100× ; Mediatech) (complete RPMI). The cells were stimulated every 10–11 days with PHA-P (Murex, Dartford, England) and irradiated (5000 rads) allogeneic peripheral blood mononuclear cells. T lymphocytes were immortalized with Herpes6irus saimiri as follows. Three days after stimulation with PHA-P and allogeneic feeder cells, an equal amount of the following media mixture containing 10% Herpes6irus saimiri supernatant from completely lysed monkey host cells (kindly provided by Dr T. Kupper, Harvard Skin Disease Center, Boston, MA) was added: complete RPMI and CG-medium (Vitromex, Vilshofen, Germany) mixed 1:1, 10% human serum and IL-2 (five times higher concentration than above). Immortalized iIEL, but not PBT, further received 5 u/ml IL-4. Thereafter, immortalized T cells were continuously cultured in this medium without stimulation by PHA-P or allogeneic feeder cells. The human intestinal epithelial cell lines HT29 and Caco-2, the human liver derived cell line HepG2 and the human lung derived cell line A549 (kindly provided by Dr P. Finn, Respiratory Disease Division, Brigham and Women’s Hospital) were grown in complete DMEM (Mediatech) supplemented with fetal calf serum (2% [HT29] and 10%, respectively; Sigma). The T84 cell line was grown in a mixture of Ham’s F12 nutrients mix and DMEM (1:1; Mediatech) supplemented with 6% fetal calf serum. Epithelial cell lines were expanded 1:2 when confluent (once or twice every week), after dissociation with 0.05% trypsin and EDTA (Mediatech). HL-60 cells and the human keratinocyte cell line A431 (kindly provided by Dr T. Kupper, Harvard Skin Disease Research Center) were grown in complete RPMI supplemented with 10% fetal calf serum.
2.2. Generation of conditioned supernatants 2. Material and methods
2.1. Cell lines Peripheral blood T lymphocytes (PBT) and iIEL from small and large bowel were isolated and primary cultures established as described previously [3,14]. For the purposes of this study, the cells were cultured in RPMI (Mediatech, Herndon, MA) with 10% human serum (Sigma, St. Louis, MO), IL-2 (Ajinomoto Company, Kawasaki, Japan) and 5 u/ml IL-4 (Genetics Institute, Cambridge, MA). The IL-2 concentration was the concentration giving maximal proliferative response in a pilot study, as measured by a proliferation assay performed as described below (data not shown). Medium was supplemented with Penicillin/Streptomycin (100 u/ml each; BioWhittaker, Walkersville,
To generate conditioned supernatant, confluent cells were incubated for 48 h with HBSS with calcium and magnesium (Sigma) as indicated. Cells remaining in the supernatant were removed by centrifugation and the remainder filtered through a 0.45-mm cellulose acetate filter (Intermountain Scientific, Kaysville, UT). Filtered supernatants were stored at 4°C. The pH of the conditioned and the unconditioned HBSS was the same in air containing minimal CO2 (pH 7.9–8.0) and in 5% CO2, (pH 7.1). To assess the viability of epithelial cells in HBSS during conditioning, electrical physiological measurements were done. Confluent monolayers of T84 cells were grown on collagen coated permeable supports and maintained until steady-state transepithelial resistance was achieved, as previously described in detail [15].
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
Measurement of transepithelial resistance, voltage, and short circuit current were done as previously described [16], and carried out after monolayers were thoroughly washed with HBSS. HBSS was used in both apical and basolateral baths during all experiments. All experiments were performed in a 37°C room to ensure that epithelial monolayers, solutions, plasticware, etc., were maintained at uniform 37°C temperature.
2.3. Proliferation assay Ninty-six well flat bottom plates were coated with goat anti mouse IgG Fc antibodies (ImmunoPure; Pierce, Rockford, IL) diluted 1:240 in phosphate buffered saline (PBS) at 4°C overnight. The plates were then washed three times with PBS and incubated with OKT3 mAb (IgG2a, purified hybridoma supernatant, diluted 1:10 in PBS), OKT11/2 and OKT11/3 (both IgG2a, ascites: a kind gift of Dr. Ellis Reinherz, Dana Farber Cancer Institute, Boston, MA; mixed 1:1 and diluted 1:100 in PBS), normal mouse serum (1:100) or mouse anti allogeneic mouse mAb (IgG2a, 1:500) at room temperature for 6 h. After washing, T lymphocytes were seeded at 100 000 cells per well in 100 ml complete RPMI containing 10% human serum but without cytokines. The 96 well plates further received 100 ml of the various test substances per well. After 48 h incubation at 37°C, wells were pulsed with 0.5 mCi [3H]-Thymidine. After another incubation of 18 hours, cells were harvested on filter mats (Walac, Turku, Finland) and read in a scintillation counter.
2.4. Cytokine bioassay T lymphocytes were stimulated as described for proliferation assays. After 48 hours incubation, 10 ml supernatant was harvested from each well and added to 20 000 MO7E cells (a kind gift of Dr S. Porcelli, Harvard Medical School, Boston, MA) in 200 ml low bicarbonate DMEM (Specialty media, Lavallette, NJ) supplemented with L-glutamine, non-essential amino acids, Hepes buffer, antibiotics and 10% fetal calf serum. After a three day incubation period, wells were pulsed with [3H]-Thymidine as above and harvested after 4 h. The growth of MO7E cells depends on IL-3 or GM-CSF.
2.5. Statistics Data is presented as mean9 SEM inhibition in percent of proliferation of the control wells (unconditioned HBSS). The error bars represent SEM as well. Various conditions within a given T cell line were compared by two-tailed, unpaired Student’s t-test. Comparisons involving more than one T cell line were performed by analysis of variance. P values of 0.05 or below were considered significant.
161
3. Results
3.1. Effect of T84 conditioned supernatant on anti-CD3 mediated stimulation of intestinal T cell lines In preliminary studies we found that the response of iIEL to anti-CD3 stimulation was restored when the iIEL were removed from the intestinal microenvironment and established as continuous, PHA-P stimulated cell lines. Table 1 shows the response of two iIEL lines from small intestine and large intestine to anti-CD3 stimulation. Both iIEL lines exhibited a strong proliferative response and cytokine production after anti-CD3 stimulation using crosslinked OKT3 mAb in comparison to the response to wells coated with either normal mouse serum or an irrelevant, isotype matched (IgG2a) mouse mAb. However, if the medium during the antiCD3 stimulation of the two iIEL cell lines was supplemented with an equal volume of HBSS conditioned from T84 cells, there was a reduction of proliferation in comparison to an equal volume of nonconditioned HBSS in both T lymphocyte cell lines (Fig. 1 black bars, PB 0.05).The white bars indicate inhibition of the cytokine production of these two cell lines in response to anti-CD3 stimulation by conditioned supernatant. Although there was no significant change in cytokine production there was a trend towards a small decrease. As shown in Fig. 2, a dose response experiment with the small intestinal IEL cell line indicated, that the inhibition of proliferation observed with the conditioned HBSS supernatant was present up to a dilution of 1:8. Since the nonimmortalized iIEL cell lines were maintained by periodic PHA-P stimulation, it was possible Table 1 Cultured iIEL proliferate to anti-CD3 mediated stimulation Small intestinal IEL (cpm) Proliferation OKT3 34 883 94940* Normal mouse serum 43 912 Mouse anti-allogenic 41 9 12 mouse mAb Cytokine production OKT3 10 673 9 1022* Normal mouse serum 31 915 Mouse anti-allogenic mouse mAb
50 921
Large intestinal IEL (cpm)
121 880 9 8209* 252 976 272 9 51 11 469 9 1712* 63 99 63 920
Cultured iIEL from small and large bowel were stimulated by plate bound OKT3 and normal mouse serum and an isotype matched mouse antibody as negative controls. Proliferation was measured by [3H]-Thymidine uptake. Both cell lines show a strong proliferative response and cytokine production after anti-CD3 stimulation (n= 4, *PB0.001 vs both negative controls).
162
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
Fig. 1. Supernatant from T84 cells inhibits proliferation of iIEL. Small and large bowel iIEL were stimulated by OKT3 in the presence of supernatant from T84 cells or unconditioned medium. Proliferation was measured by [3H]-Thymidine uptake and cytokine production by bioassay. Data is shown as percent inhibition by conditioned supernatant. Proliferation of both iIEL cell lines was inhibited by conditioned medium (black bars; n=4, P= 0.04). No significant inhibition of cytokine production was observed (white bars; n =4, P= ns). The absolute values for the control wells in the proliferation assay were 28 973 9 1388 cpm and 122 485 9 3883 cpm and in the cytokine assay 13 178 9 502 cpm and 12 093 9 705 cpm for small and large bowel iIEL, respectively.
that the observed effects of the intestinal epithelial cell supernatants were related to an alteration of the cell lines in response to repeated CD3 - mediated signals. We, therefore, also assessed the response of these two iIEL cell lines after immortalization with H. saimiri, which does not require periodic PHA-P stimulation for maintenance of growth. Fig. 3 shows the inhibition of the proliferative response (black bars) and of the cy-
Fig. 3. Anti-CD3 mediated stimulation of immortalized iIEL cell lines is inhibited by T84 supernatant. H. saimiri immortalized small and large bowel iIEL were stimulated by OKT3 in the presence of supernatant from T84 cells. Data shown is percent inhibition compared to unconditioned medium. Proliferation (black bars; n=4, P B0.001) and cytokine production (white bars; n =4, P = 0.01) of both immortalized T lymphocyte cell lines was inhibited. The absolute values for the control wells in the proliferation assay were 27 768 9 1424 cpm and 33 038 9 773 cpm and in the cytokine assay 40 430 9 1143 cpm and 13 576 +1025 cpm for small and large bowel iIEL, respectively.
tokine production (white bars) of the immortalized small intestinal and large intestinal cell lines to antiCD3 stimulation in the presence of HBSS conditioned by T84 cells as compared to unconditioned HBSS. As observed with the nonimmortalized cells, conditioned HBSS significantly inhibited the proliferative response to anti-CD3 stimulation (PB 0.001). Similarly, a decrease in cytokine production was also observed (P= 0.01).
3.2. The effect of T84 conditioned supernatant is specific for CD3 pathway
Fig. 2. Inhibition of iIEL proliferation by T84 cell supernatant is dose dependent. Conditioned supernatant from T84 cells was diluted with unconditioned HBSS at various ratios. The proliferative response after stimulation with OKT3 was measured and percent inhibition compared to unconditioned medium is shown. Inhibition can be observed up to a dilution of 1:8 (n=4 for each data point, P= 0.05 for 1:8 dilution). The absolute value for unconditioned HBSS was 33 770 91512 cpm.
One iIEL line, the immortalized small intestinal cell line, exhibited a significant response to anti-CD2 stimulation using crosslinked OKT11/2 and OKT11/3 mAb as assessed by proliferation and cytokine production. This allowed for the determination of the specificity of the effects of the T84 supernatant for CD3 -mediated stimulation. As shown in Fig. 4, the response to antiCD2 stimulation was not reduced by conditioned HBSS. The effect of the T84 supernatant did not, therefore, represent a nonspecific, toxic activity. The actual increase in proliferation in response to anti-CD2 stimulation with conditioned supernatant as compared to unconditioned, as seen in Fig. 4, could be observed in several experiments. However, this increase was statistically significant in only a single experiment.
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
163
3.3. Effect of conditioned T84 supernatant on peripheral blood T-cell lines In order to determine whether the effect of the factor(s) contained within the T84 supernatant could be generalized to other T-cell populations, T-cell lines from peripheral blood were also characterized. A nonimmortalized T-cell line and an immortalized T-cell line from peripheral blood were stimulated with platebound anti-CD3 in the presence of conditioned supernatant from the T84 cell line. As shown in Fig. 5, the proliferative response of these two cell lines to antiCD3 stimulation was significantly diminished by conditioned T84 supernatant compared to controls (black bars; PB 0.001). As observed with the iIEL cell lines, a non-significant trend towards a decrease in cytokine production was also observed (white bars). Thus the factor(s) responsible for the diminished response to anti-CD3 stimulation was also effective on extraintestinal T cells.
3.4. Influence of D-Glucose on T lymphocyte proliferation The only nutrient that is present in HBSS is D-glucose raising the possibility that the observations thus far simply reflect a consumption of glucose by the
Fig. 5. Anti-CD3 mediated proliferative response of PBT is inhibited by supernatant from T84 cells. Non-immortalized and H. saimiri immortalized peripheral blood T cells were stimulated by OKT3 in the presence of supernatant from T84 cells. Data is shown as percent inhibition compared to unconditioned medium. Proliferation (black bars) of both T lymphocyte cell lines was inhibited by conditioned medium (n =4, PB 0.001). A trend, also not significant, towards inhibition of cytokine production (white bars) was observed (n =4, P= ns). The absolute values for the control wells in the proliferation assay were 38 828 92665 cpm and 90 626 9 2359 cpm and in the cytokine assay 66 119 9 5467 cpm and 28 028 9 1605 cpm for non-immortalized and immortalized T cell lines, respectively.
intestinal epithelial cells during conditioning. We, therefore, examined this possibility in the following studies. HBSS lacking D-glucose was prepared and a dose response curve to D-glucose was generated. D-glucose was added in concentrations from 0 to 2 g/l (standard HBSS contains 1 g/l D-glucose). No significant effect on proliferation of a small intestinal IEL cell line after CD3 stimulation was observed (data not shown).
3.5. Viability of epithelial cells in HBSS
Fig. 4. CD2 pathway of stimulation is not affected by conditioned HBSS. Proliferation (black bars) and cytokine production (white bars) of a PBT cell line stimulated with plate bound OKT11 (antiCD2 stimulation) and OKT3 (anti-CD3 stimulation), respectively, was measured. Conditioned supernatant from T84 cells significantly reduced proliferation (P B 0.005) and cytokine production (PB0.05) after anti-CD3 mediated stimulation, whereas the response to antiCD2 mediated stimulation is preserved. The absolute values for the control wells in the proliferation assay were 5513 9370 cpm and 27 768 9 1423 cpm and in the cytokine assay 11 125 9 651 cpm and 40 430 9 1143 cpm for CD2 and CD3 mediated stimulation, respectively.
It was next important to determine whether the conditions used to generate supernatants from T84 cells into HBSS were toxic. First, no significant changes in epithelial morphology were observed at the light microscopy level (data not shown). Secondly, such conditions did not significantly influence epithelial barrier function, a sensitive measure of toxicity [15], as assessed by voltage clamp measurement of transepithelial resistance. Compared to incubation in serum containing media for 48 h as described in Section 2, placing T84 cells in HBSS for 48 h did not result in diminished barrier function (transepithelial resistances of 8299 160 vs. 7789 44 V cm2 for HBSS and media, respectively, P= ns). Finally, forskolin (1 mM final concentration) stimulated Cl − secretion, the epithelial ion transport event responsible for fluid movement [17], was not
164
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
diminished by incubation of T84 cells with HBSS for 48 h (measured as short circuit current [15], 26.79 4.06 vs. 23.5 96.3 mA/cm2 for media and HBSS, respectively, P= ns). These data indicate that it is unlikely that incubation of T84 cells in HBSS for 48 hours results in toxicity to the T84 cells.
3.6. The effects on anti-CD3 stimulation are specific for intestinal epithelial cells In order to determine whether these observations were specific for epithelial cells, conditioned supernatants from two other intestinal epithelial cell lines (HT29 and Caco2) were also assessed and compared to a non-epithelial cell line, HL-60, an immature myeloid cell line. The conditioned supernatants from all the intestinal epithelial cell lines reduced the proliferative response of a PBT cell line after CD3 mediated stimula-
Fig. 7. Extraintestinal epithelial cell lines have no effect on the proliferation of T lymphocytes. Large intestinal T lymphocytes were stimulated with OKT3 and the proliferative response measured in a [3H]-Thymidine uptake assay. The proliferative response was modified by adding conditioned media from a liver derived (HepG2), a human keratinocyte cell line (A431), a human lung derived cell line (A549), and from T84 cells as a positive control. Data is shown as percent inhibition compared to unconditioned medium. Only the positive control (T84) and the liver derived cell line (HepG2) had a suppressive effect on the CD3 induced proliferation of the T lymphocytes. The absolute value for the control wells was 7391 9 137 cpm (n = 3).
tion (Fig. 6). However, conditioned supernatant from the non-epithelial cell line, HL-60, exhibited no effect at all dilutions tested. To determine whether the effect of conditioned supernatant of epithelial cell lines was specific for gut derived epithelial cells, epithelial cell lines derived from liver (HepG2), skin (A431) and lung (A549) were compared to T84 cells. Only conditioned supernatant from T84 and HepG2 cells suppressed the proliferative response of a large intestinal T cell line after CD3 mediated stimulation (Fig. 7). Conditioned supernatant from lung and skin derived epithelial cells, however, had no effect.
4. Discussion
Fig. 6. Several intestinal epithelial cell lines produce a factor that inhibits proliferation of T lymphocytes. Peripheral blood T lymphocytes were stimulated with OKT3 and the proliferative response was measured in a [3H]-Thymidine uptake assay. Proliferative response was modified by adding conditioned media from various epithelial cell lines in two dilutions (solid lines, order on left side from top to bottom: T84 , HT29 ", Caco2 ). Data is shown as percent inhibition compared to unconditioned medium. In both dilutions the three epithelial cell lines inhibited anti-CD3 mediated proliferation (PB 0.001). Supernatant from HL-60 (dashed line) had no significant effect at either dilution. The absolute value of the control wells was 25 555 9725 cpm.
Previous studies have shown that freshly isolated iIEL exhibit a diminished response to anti-CD3 stimulation in comparison to PBT [3]. However, when cultivated in vitro, iIEL regain the ability to proliferate after anti-CD3 stimulation. This study shows that this restored proliferative response to anti-CD3 stimulation can, in part, be reversed by supernatants from intestinal epithelial cells. Furthermore, we show that the effects of intestinal epithelial cell supernatants in reducing responses to anti-CD3 mediated stimulation can be extended to PBT. These results are not likely due to either
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
the depletion of nutrients or release of toxic factor(s) in view of our use of a low nutrient base medium for conditioning (HBSS), the observed independence from glucose (the sole nutrient in HBSS) in the conditioned medium and the preserved responses to anti-CD2 mediated stimulation. In a single experiment proliferation was suppressed only by 6.7% whereas in all other experiments (n =12) 24-52% inhibition was observed. This level of inhibition is likely biologically significant. The lowest suppression was observed with the IEL cell line from large bowel which exhibited the highest proliferative response. It could be hypothesized that this very strong response could be only inadequately suppressed by factor(s) from the epithelial cells. These studies corroborate and significantly extend the work of several other groups. Meuer and colleagues have previously shown that supernatant from crude intestinal mucosa inhibits anti-CD3 mediated proliferation of freshly isolated T lymphocytes [18]. This makes it likely that this effect on anti-CD3 stimulation possibly plays a role in maintaining the suppressive tone of the gut in vivo. Our data showing that suppressive factor(s) exist in supernatant from several intestinal epithelial cell lines, that are not present in supernatant from a control non-epithelial cell line or from lung or skin derived epithelial cell lines, suggests that the intestinal epithelium may be the source of this factor in vivo. The studies described here are also consistent with studies described by Ebert and colleagues who showed that conditioned supernatant from the HT29 cell line suppressed the mitogen induced proliferation of PBT [19]. These studies, however, did not examine the effects on different pathways, i.e., CD2 and CD3 mediated stimulation, nor was the effect of supernatant from other intestinal epithelial cell lines or the effect on iIEL measured. We speculate that the TCR/CD3 complex of iIEL and PBT in humans is not structurally different but exhibits functional differences due, at least in part, to the local microenvironment; specifically due to soluble mediators released by intestinal epithelial cells. Defining the factor(s) responsible for these functional characteristics of iIEL will be important in understanding how intestinal tissue limits the promiscuous responses of mucosal lymphocytes to luminal antigens.
.
.
165
Acknowledgements Dr A.D. Christ was supported by partial fellowships from Janggen-Po¨hn Stiftung St. Gallen, Freiwillige Akademische Gesellschaft Basel and Theodor Engelmann Stiftung Basel, Switzerland. Dr S.P. Colgan was supported by NIH grant DK08777, Dr S.P. Balk by NIH grant AI33911, and Dr R.S. Blumberg by NIH grants DK44319 and DK51362.
References [1] G.K. Sim, Adv. Immunol. 58 (1995) 297 – 343. [2] R.S. Blumberg, C.S. Probert, A.D. Christ, H. Kim, J.E. Polischuk, V. Morales, D, Gerdes, A. Chott, S.P, Balk, in: G.N.J, Tytgat, J.F.W.M. Bartelsman, S.J.H. van Deventer (Eds.), Inflammatory Bowel Disease, Kluwer Academic Publishers, Dordrecht, Boston and London, 1995, pp. 266-274. [3] E.C. Ebert, Gastroenterology 97 (1989) 1372 – 1381. [4] E.C. Ebert, A.I. Roberts, R.E. Brolin, K. Raska, Clin. Exp. Immunol. 65 (1986) 148 – 157. [5] U.C. Pirzer, G. Schu¨rmann, S. Post, M. Betzler, S.C. Meuer, Eur. J. Immunol. 20 (1990) 2339 – 2342. [6] L. Qiao, G. Schu¨rmann, M. Betzler, S.C. Meuer, Immunol Res. 10 (1991) 218 – 225. [7] L. Qiao, G. Schu¨rmann, M. Betzler, S.C. Meuer, Gastroenterology 101 (1991) 1529 – 1536. [8] B.C. Sydora, P.F. Mixter, H.R. Holcombe, P. Eghtasady, K. Williams, M.C. Amaral, A. Nel, M. Kronenberg, J. Immunol. 150 (1993) 2179 – 2191. [9] P. Kaoutzani, S.P. Colgan, K.L. Cepek, P. Grant Burkhard, S. Carlson, C. Delp-Archer, M.B. Brenner, J.L. Madara, J. Clin. Invest. 94 (1994) 788 – 796. [10] R. Boismenu, W.L. Havran, Science 266 (1994) 1253 – 1255. [11] C.-C. Schuerer-Maly, L. Eckmann, M.F. Kagnoff, M.T. Falco, F.-E. Maly, Immunology 81 (1994) 85 – 91. [12] C.G. Larsen, A.O. Anderson, E. Appella, J.J. Oppenheim, K. Matsushima, Science 243 (1989) 1464 – 1466. [13] S.P. Balk, J.E. Polischuk, C. Probert, C. Stevens, E. Ebert, J. She, C. Terhorst, R.S. Blumberg, Int. Immunol. 7 (1995) 1237– 1241. [14] S.P. Balk, E.C. Ebert, R.L. Blumenthal, F.V. McDermott, K.W. Wucherpfennig, S.B. Landau, R.S. Blumberg, Science 253 (1991) 1411 – 1415. [15] K. Dharmsathaphorn, J.L. Madara, Methods Enzymol. 192 (1990) 354 – 359. [16] S.P. Colgan, M.B. Resnick, C.A. Parkos, C. Delp-Archer, D. McGuirk, A.E. Bacarra, P.F. Weller, J.L. Madara, J. Immunol. 153 (1994) 2122 – 2129. [17] D.W. Powell, in: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, Raven Press, New York, 1987. [18] L. Qiao, G. Schu¨rmann, F. Autschbach, R. Wallich, S.C. Meuer, Gastroenterology 105 (1993) 814 – 819. [19] E.C. Ebert, A.I. Roberts, D. Devereux, H. Nagase, Cancer Research 50 (1990) 6158 – 6161.
Human intestinal epithelial cell lines produce factor(s) that inhibit CD3-mediated T-lymphocyte proliferation Andreas D. Christ a, Sean P. Colgan a, Steven P. Balk b, Richard S. Blumberg a,* a
Gastroenterology Di6ision, Thorn Biomedical Research Building, Room 1310, Brigham and Women’s Hospital, Har6ard Medical School, 75 Francis Street, Boston, MA 02115, USA b Beth Israel Hospital, Har6ard Medical School, Boston, MA 02115, USA Received 6 January 1997; received in revised form 4 April 1997; accepted 22 April 1997
Abstract Peripheral blood T lymphocytes (PBT) proliferate more to anti-CD3 stimulation than to anti-CD2 stimulation. On the other hand, fresh, but not cultivated, intestinal intraepithelial lymphocytes (iIEL) exhibit a lower response to CD3 stimulation in comparison to CD2. The goal of this study was to show that the anti-CD3 T-cell response depends on the microenvironment and is independent of the origin of the lymphocytes. Cultured T-cell lines were stimulated with either an anti-CD3 mAb or an anti-CD2 mAb. Either conditioned supernatant from intestinal epithelial cell (IEC) lines or non- conditioned medium (negative control) was added. After 2 days cytokine production and proliferation were measured. Conditioned supernatant decreased the proliferative response of small and large bowel iIEL compared to controls (P = 0.04). In the same experiments, the cytokine production was non-significantly decreased. Immortalized iIEL, that are not regularly stimulated by their CD3 pathway, showed a similar decrease in proliferation (P B0.001) and cytokine production (P = 0.01) when incubated with conditioned supernatant. Similar results were also obtained with a non-immortalized and an immortalized PBT line (P B 0.001). In a small bowel iIEL cell line, that exhibited a significant response to anti-CD2 stimulation, the proliferative response to anti-CD2 stimulation was preserved. Active conditioned supernatant could be generated from three independent IEC lines and a liver derived epithelial cell line, but not from a non-epithelial control cell line or two extraintestinal epithelial cell lines. We conclude that supernatants of cultured IEC contain soluble factor(s) that cause cultured iIEL and extraintestinal lymphocytes to behave like fresh iIEL. These results, therefore, support and extend the studies of others which suggest that the intestinal microenvironment mucosalizes lymphocytes. © 1997 Elsevier Science B.V. Keywords: Human intestinal epithelial cell lines; T-lymphocyte proliferation
1. Introduction Lymphocytes which reside in the intestinal epithelium, intestinal intraepithelial lymphocytes (iIEL), are a unique population of cells which likely represent a general component of the mucosal immune system of most mammalian species [1]. iIEL are located below the tight junctions of the intestinal epithelium in close contact to the basal and lateral membrane of the intes* Corresponding author. Tel: + 1 617 7326917; fax: + 1 617 7305807; e-mail: [email protected] 0165-2478/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 2 4 7 8 ( 9 7 ) 0 0 0 8 1 - 3
tinal epithelial cell. In humans, iIEL are predominantly T cells expressing CD8 and a limited array of TCR-ab, and to a lesser extent TCR-gd, indicating that they recognize a restricted range of antigens in the context of a MHC class I-related molecule [2]. Although the MHC class I restriction element has not yet been defined for iIEL, it is commonly accepted that the intracellular signals mediated by ligation of the TCR/CD3 complex are diminished. Freshly isolated iIEL respond 20 times less to PHA-P, Con A and anti-CD3 -sepharose than peripheral blood T lymphocytes as measured in a [3H]Thymidine proliferation assay [3]. This is specific for
160
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
the TCR/CD3 complex since the response to anti-CD2 stimulation [3], and corollarily sheep red blood cells [4], is preserved. Similar functional characteristics have been described for lamina propria lymphocytes (LPL) [5 – 7]. It is not known, however, whether this represents the diminished response of the CD4 and/or CD8 fraction of LPL. To an extent, these characteristics are also shared by murine iIEL, where iIEL are cytolytic but have a low proliferative response to anti-CD3 stimulation as compared to other T lymphocytes [8]. Although the function of iIEL is unknown, iIEL influence functional aspects of adjacent intestinal epithelial cells and vice versa. When intestinal epithelial cell monolayers are reconstituted with mucosal-derived lymphocytes in vitro, barrier function and ion transport properties are attenuated by soluble factors released by co-culture [9]. iIEL bearing the gd TCR also promote intestinal epithelial cell growth by secretion of keratinocyte growth factor [10]. On the other hand, intestinal epithelial cells are known to secrete factors that influence T cells, e.g., IL-8 [11], a known chemoattractant for T lymphocytes [12]. It is likely, therefore, that a bi-directional crosstalk exists between intestinal epithelial cells and iIEL. Since we have previously shown that the composition of the TCR/CD3 complex of the iIEL is similar to peripheral blood T cells [13], which exhibit normal response to TCR/CD3 complex mediated stimulation, it is possible that this functional phenotype of human iIEL is related, at least in part, to soluble mediators released by intestinal epithelial cells. As shown here, cultured intestinal epithelial cells are capable of secreting factor(s) that inhibit proliferation of iIEL in response to antiCD3 stimulation. Moreover, the modulatory effects of the factor(s) secreted by intestinal epithelial cells is also capable of influencing the CD3-mediated response of peripheral blood lymphocytes.
MD), 2 mM L-glutamine (Mediatech), 10 mM Hepes buffer (Mediatech) and 1% non-essential amino acids (100× ; Mediatech) (complete RPMI). The cells were stimulated every 10–11 days with PHA-P (Murex, Dartford, England) and irradiated (5000 rads) allogeneic peripheral blood mononuclear cells. T lymphocytes were immortalized with Herpes6irus saimiri as follows. Three days after stimulation with PHA-P and allogeneic feeder cells, an equal amount of the following media mixture containing 10% Herpes6irus saimiri supernatant from completely lysed monkey host cells (kindly provided by Dr T. Kupper, Harvard Skin Disease Center, Boston, MA) was added: complete RPMI and CG-medium (Vitromex, Vilshofen, Germany) mixed 1:1, 10% human serum and IL-2 (five times higher concentration than above). Immortalized iIEL, but not PBT, further received 5 u/ml IL-4. Thereafter, immortalized T cells were continuously cultured in this medium without stimulation by PHA-P or allogeneic feeder cells. The human intestinal epithelial cell lines HT29 and Caco-2, the human liver derived cell line HepG2 and the human lung derived cell line A549 (kindly provided by Dr P. Finn, Respiratory Disease Division, Brigham and Women’s Hospital) were grown in complete DMEM (Mediatech) supplemented with fetal calf serum (2% [HT29] and 10%, respectively; Sigma). The T84 cell line was grown in a mixture of Ham’s F12 nutrients mix and DMEM (1:1; Mediatech) supplemented with 6% fetal calf serum. Epithelial cell lines were expanded 1:2 when confluent (once or twice every week), after dissociation with 0.05% trypsin and EDTA (Mediatech). HL-60 cells and the human keratinocyte cell line A431 (kindly provided by Dr T. Kupper, Harvard Skin Disease Research Center) were grown in complete RPMI supplemented with 10% fetal calf serum.
2.2. Generation of conditioned supernatants 2. Material and methods
2.1. Cell lines Peripheral blood T lymphocytes (PBT) and iIEL from small and large bowel were isolated and primary cultures established as described previously [3,14]. For the purposes of this study, the cells were cultured in RPMI (Mediatech, Herndon, MA) with 10% human serum (Sigma, St. Louis, MO), IL-2 (Ajinomoto Company, Kawasaki, Japan) and 5 u/ml IL-4 (Genetics Institute, Cambridge, MA). The IL-2 concentration was the concentration giving maximal proliferative response in a pilot study, as measured by a proliferation assay performed as described below (data not shown). Medium was supplemented with Penicillin/Streptomycin (100 u/ml each; BioWhittaker, Walkersville,
To generate conditioned supernatant, confluent cells were incubated for 48 h with HBSS with calcium and magnesium (Sigma) as indicated. Cells remaining in the supernatant were removed by centrifugation and the remainder filtered through a 0.45-mm cellulose acetate filter (Intermountain Scientific, Kaysville, UT). Filtered supernatants were stored at 4°C. The pH of the conditioned and the unconditioned HBSS was the same in air containing minimal CO2 (pH 7.9–8.0) and in 5% CO2, (pH 7.1). To assess the viability of epithelial cells in HBSS during conditioning, electrical physiological measurements were done. Confluent monolayers of T84 cells were grown on collagen coated permeable supports and maintained until steady-state transepithelial resistance was achieved, as previously described in detail [15].
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
Measurement of transepithelial resistance, voltage, and short circuit current were done as previously described [16], and carried out after monolayers were thoroughly washed with HBSS. HBSS was used in both apical and basolateral baths during all experiments. All experiments were performed in a 37°C room to ensure that epithelial monolayers, solutions, plasticware, etc., were maintained at uniform 37°C temperature.
2.3. Proliferation assay Ninty-six well flat bottom plates were coated with goat anti mouse IgG Fc antibodies (ImmunoPure; Pierce, Rockford, IL) diluted 1:240 in phosphate buffered saline (PBS) at 4°C overnight. The plates were then washed three times with PBS and incubated with OKT3 mAb (IgG2a, purified hybridoma supernatant, diluted 1:10 in PBS), OKT11/2 and OKT11/3 (both IgG2a, ascites: a kind gift of Dr. Ellis Reinherz, Dana Farber Cancer Institute, Boston, MA; mixed 1:1 and diluted 1:100 in PBS), normal mouse serum (1:100) or mouse anti allogeneic mouse mAb (IgG2a, 1:500) at room temperature for 6 h. After washing, T lymphocytes were seeded at 100 000 cells per well in 100 ml complete RPMI containing 10% human serum but without cytokines. The 96 well plates further received 100 ml of the various test substances per well. After 48 h incubation at 37°C, wells were pulsed with 0.5 mCi [3H]-Thymidine. After another incubation of 18 hours, cells were harvested on filter mats (Walac, Turku, Finland) and read in a scintillation counter.
2.4. Cytokine bioassay T lymphocytes were stimulated as described for proliferation assays. After 48 hours incubation, 10 ml supernatant was harvested from each well and added to 20 000 MO7E cells (a kind gift of Dr S. Porcelli, Harvard Medical School, Boston, MA) in 200 ml low bicarbonate DMEM (Specialty media, Lavallette, NJ) supplemented with L-glutamine, non-essential amino acids, Hepes buffer, antibiotics and 10% fetal calf serum. After a three day incubation period, wells were pulsed with [3H]-Thymidine as above and harvested after 4 h. The growth of MO7E cells depends on IL-3 or GM-CSF.
2.5. Statistics Data is presented as mean9 SEM inhibition in percent of proliferation of the control wells (unconditioned HBSS). The error bars represent SEM as well. Various conditions within a given T cell line were compared by two-tailed, unpaired Student’s t-test. Comparisons involving more than one T cell line were performed by analysis of variance. P values of 0.05 or below were considered significant.
161
3. Results
3.1. Effect of T84 conditioned supernatant on anti-CD3 mediated stimulation of intestinal T cell lines In preliminary studies we found that the response of iIEL to anti-CD3 stimulation was restored when the iIEL were removed from the intestinal microenvironment and established as continuous, PHA-P stimulated cell lines. Table 1 shows the response of two iIEL lines from small intestine and large intestine to anti-CD3 stimulation. Both iIEL lines exhibited a strong proliferative response and cytokine production after anti-CD3 stimulation using crosslinked OKT3 mAb in comparison to the response to wells coated with either normal mouse serum or an irrelevant, isotype matched (IgG2a) mouse mAb. However, if the medium during the antiCD3 stimulation of the two iIEL cell lines was supplemented with an equal volume of HBSS conditioned from T84 cells, there was a reduction of proliferation in comparison to an equal volume of nonconditioned HBSS in both T lymphocyte cell lines (Fig. 1 black bars, PB 0.05).The white bars indicate inhibition of the cytokine production of these two cell lines in response to anti-CD3 stimulation by conditioned supernatant. Although there was no significant change in cytokine production there was a trend towards a small decrease. As shown in Fig. 2, a dose response experiment with the small intestinal IEL cell line indicated, that the inhibition of proliferation observed with the conditioned HBSS supernatant was present up to a dilution of 1:8. Since the nonimmortalized iIEL cell lines were maintained by periodic PHA-P stimulation, it was possible Table 1 Cultured iIEL proliferate to anti-CD3 mediated stimulation Small intestinal IEL (cpm) Proliferation OKT3 34 883 94940* Normal mouse serum 43 912 Mouse anti-allogenic 41 9 12 mouse mAb Cytokine production OKT3 10 673 9 1022* Normal mouse serum 31 915 Mouse anti-allogenic mouse mAb
50 921
Large intestinal IEL (cpm)
121 880 9 8209* 252 976 272 9 51 11 469 9 1712* 63 99 63 920
Cultured iIEL from small and large bowel were stimulated by plate bound OKT3 and normal mouse serum and an isotype matched mouse antibody as negative controls. Proliferation was measured by [3H]-Thymidine uptake. Both cell lines show a strong proliferative response and cytokine production after anti-CD3 stimulation (n= 4, *PB0.001 vs both negative controls).
162
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
Fig. 1. Supernatant from T84 cells inhibits proliferation of iIEL. Small and large bowel iIEL were stimulated by OKT3 in the presence of supernatant from T84 cells or unconditioned medium. Proliferation was measured by [3H]-Thymidine uptake and cytokine production by bioassay. Data is shown as percent inhibition by conditioned supernatant. Proliferation of both iIEL cell lines was inhibited by conditioned medium (black bars; n=4, P= 0.04). No significant inhibition of cytokine production was observed (white bars; n =4, P= ns). The absolute values for the control wells in the proliferation assay were 28 973 9 1388 cpm and 122 485 9 3883 cpm and in the cytokine assay 13 178 9 502 cpm and 12 093 9 705 cpm for small and large bowel iIEL, respectively.
that the observed effects of the intestinal epithelial cell supernatants were related to an alteration of the cell lines in response to repeated CD3 - mediated signals. We, therefore, also assessed the response of these two iIEL cell lines after immortalization with H. saimiri, which does not require periodic PHA-P stimulation for maintenance of growth. Fig. 3 shows the inhibition of the proliferative response (black bars) and of the cy-
Fig. 3. Anti-CD3 mediated stimulation of immortalized iIEL cell lines is inhibited by T84 supernatant. H. saimiri immortalized small and large bowel iIEL were stimulated by OKT3 in the presence of supernatant from T84 cells. Data shown is percent inhibition compared to unconditioned medium. Proliferation (black bars; n=4, P B0.001) and cytokine production (white bars; n =4, P = 0.01) of both immortalized T lymphocyte cell lines was inhibited. The absolute values for the control wells in the proliferation assay were 27 768 9 1424 cpm and 33 038 9 773 cpm and in the cytokine assay 40 430 9 1143 cpm and 13 576 +1025 cpm for small and large bowel iIEL, respectively.
tokine production (white bars) of the immortalized small intestinal and large intestinal cell lines to antiCD3 stimulation in the presence of HBSS conditioned by T84 cells as compared to unconditioned HBSS. As observed with the nonimmortalized cells, conditioned HBSS significantly inhibited the proliferative response to anti-CD3 stimulation (PB 0.001). Similarly, a decrease in cytokine production was also observed (P= 0.01).
3.2. The effect of T84 conditioned supernatant is specific for CD3 pathway
Fig. 2. Inhibition of iIEL proliferation by T84 cell supernatant is dose dependent. Conditioned supernatant from T84 cells was diluted with unconditioned HBSS at various ratios. The proliferative response after stimulation with OKT3 was measured and percent inhibition compared to unconditioned medium is shown. Inhibition can be observed up to a dilution of 1:8 (n=4 for each data point, P= 0.05 for 1:8 dilution). The absolute value for unconditioned HBSS was 33 770 91512 cpm.
One iIEL line, the immortalized small intestinal cell line, exhibited a significant response to anti-CD2 stimulation using crosslinked OKT11/2 and OKT11/3 mAb as assessed by proliferation and cytokine production. This allowed for the determination of the specificity of the effects of the T84 supernatant for CD3 -mediated stimulation. As shown in Fig. 4, the response to antiCD2 stimulation was not reduced by conditioned HBSS. The effect of the T84 supernatant did not, therefore, represent a nonspecific, toxic activity. The actual increase in proliferation in response to anti-CD2 stimulation with conditioned supernatant as compared to unconditioned, as seen in Fig. 4, could be observed in several experiments. However, this increase was statistically significant in only a single experiment.
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
163
3.3. Effect of conditioned T84 supernatant on peripheral blood T-cell lines In order to determine whether the effect of the factor(s) contained within the T84 supernatant could be generalized to other T-cell populations, T-cell lines from peripheral blood were also characterized. A nonimmortalized T-cell line and an immortalized T-cell line from peripheral blood were stimulated with platebound anti-CD3 in the presence of conditioned supernatant from the T84 cell line. As shown in Fig. 5, the proliferative response of these two cell lines to antiCD3 stimulation was significantly diminished by conditioned T84 supernatant compared to controls (black bars; PB 0.001). As observed with the iIEL cell lines, a non-significant trend towards a decrease in cytokine production was also observed (white bars). Thus the factor(s) responsible for the diminished response to anti-CD3 stimulation was also effective on extraintestinal T cells.
3.4. Influence of D-Glucose on T lymphocyte proliferation The only nutrient that is present in HBSS is D-glucose raising the possibility that the observations thus far simply reflect a consumption of glucose by the
Fig. 5. Anti-CD3 mediated proliferative response of PBT is inhibited by supernatant from T84 cells. Non-immortalized and H. saimiri immortalized peripheral blood T cells were stimulated by OKT3 in the presence of supernatant from T84 cells. Data is shown as percent inhibition compared to unconditioned medium. Proliferation (black bars) of both T lymphocyte cell lines was inhibited by conditioned medium (n =4, PB 0.001). A trend, also not significant, towards inhibition of cytokine production (white bars) was observed (n =4, P= ns). The absolute values for the control wells in the proliferation assay were 38 828 92665 cpm and 90 626 9 2359 cpm and in the cytokine assay 66 119 9 5467 cpm and 28 028 9 1605 cpm for non-immortalized and immortalized T cell lines, respectively.
intestinal epithelial cells during conditioning. We, therefore, examined this possibility in the following studies. HBSS lacking D-glucose was prepared and a dose response curve to D-glucose was generated. D-glucose was added in concentrations from 0 to 2 g/l (standard HBSS contains 1 g/l D-glucose). No significant effect on proliferation of a small intestinal IEL cell line after CD3 stimulation was observed (data not shown).
3.5. Viability of epithelial cells in HBSS
Fig. 4. CD2 pathway of stimulation is not affected by conditioned HBSS. Proliferation (black bars) and cytokine production (white bars) of a PBT cell line stimulated with plate bound OKT11 (antiCD2 stimulation) and OKT3 (anti-CD3 stimulation), respectively, was measured. Conditioned supernatant from T84 cells significantly reduced proliferation (P B 0.005) and cytokine production (PB0.05) after anti-CD3 mediated stimulation, whereas the response to antiCD2 mediated stimulation is preserved. The absolute values for the control wells in the proliferation assay were 5513 9370 cpm and 27 768 9 1423 cpm and in the cytokine assay 11 125 9 651 cpm and 40 430 9 1143 cpm for CD2 and CD3 mediated stimulation, respectively.
It was next important to determine whether the conditions used to generate supernatants from T84 cells into HBSS were toxic. First, no significant changes in epithelial morphology were observed at the light microscopy level (data not shown). Secondly, such conditions did not significantly influence epithelial barrier function, a sensitive measure of toxicity [15], as assessed by voltage clamp measurement of transepithelial resistance. Compared to incubation in serum containing media for 48 h as described in Section 2, placing T84 cells in HBSS for 48 h did not result in diminished barrier function (transepithelial resistances of 8299 160 vs. 7789 44 V cm2 for HBSS and media, respectively, P= ns). Finally, forskolin (1 mM final concentration) stimulated Cl − secretion, the epithelial ion transport event responsible for fluid movement [17], was not
164
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
diminished by incubation of T84 cells with HBSS for 48 h (measured as short circuit current [15], 26.79 4.06 vs. 23.5 96.3 mA/cm2 for media and HBSS, respectively, P= ns). These data indicate that it is unlikely that incubation of T84 cells in HBSS for 48 hours results in toxicity to the T84 cells.
3.6. The effects on anti-CD3 stimulation are specific for intestinal epithelial cells In order to determine whether these observations were specific for epithelial cells, conditioned supernatants from two other intestinal epithelial cell lines (HT29 and Caco2) were also assessed and compared to a non-epithelial cell line, HL-60, an immature myeloid cell line. The conditioned supernatants from all the intestinal epithelial cell lines reduced the proliferative response of a PBT cell line after CD3 mediated stimula-
Fig. 7. Extraintestinal epithelial cell lines have no effect on the proliferation of T lymphocytes. Large intestinal T lymphocytes were stimulated with OKT3 and the proliferative response measured in a [3H]-Thymidine uptake assay. The proliferative response was modified by adding conditioned media from a liver derived (HepG2), a human keratinocyte cell line (A431), a human lung derived cell line (A549), and from T84 cells as a positive control. Data is shown as percent inhibition compared to unconditioned medium. Only the positive control (T84) and the liver derived cell line (HepG2) had a suppressive effect on the CD3 induced proliferation of the T lymphocytes. The absolute value for the control wells was 7391 9 137 cpm (n = 3).
tion (Fig. 6). However, conditioned supernatant from the non-epithelial cell line, HL-60, exhibited no effect at all dilutions tested. To determine whether the effect of conditioned supernatant of epithelial cell lines was specific for gut derived epithelial cells, epithelial cell lines derived from liver (HepG2), skin (A431) and lung (A549) were compared to T84 cells. Only conditioned supernatant from T84 and HepG2 cells suppressed the proliferative response of a large intestinal T cell line after CD3 mediated stimulation (Fig. 7). Conditioned supernatant from lung and skin derived epithelial cells, however, had no effect.
4. Discussion
Fig. 6. Several intestinal epithelial cell lines produce a factor that inhibits proliferation of T lymphocytes. Peripheral blood T lymphocytes were stimulated with OKT3 and the proliferative response was measured in a [3H]-Thymidine uptake assay. Proliferative response was modified by adding conditioned media from various epithelial cell lines in two dilutions (solid lines, order on left side from top to bottom: T84 , HT29 ", Caco2 ). Data is shown as percent inhibition compared to unconditioned medium. In both dilutions the three epithelial cell lines inhibited anti-CD3 mediated proliferation (PB 0.001). Supernatant from HL-60 (dashed line) had no significant effect at either dilution. The absolute value of the control wells was 25 555 9725 cpm.
Previous studies have shown that freshly isolated iIEL exhibit a diminished response to anti-CD3 stimulation in comparison to PBT [3]. However, when cultivated in vitro, iIEL regain the ability to proliferate after anti-CD3 stimulation. This study shows that this restored proliferative response to anti-CD3 stimulation can, in part, be reversed by supernatants from intestinal epithelial cells. Furthermore, we show that the effects of intestinal epithelial cell supernatants in reducing responses to anti-CD3 mediated stimulation can be extended to PBT. These results are not likely due to either
A.D. Christ et al. / Immunology Letters 58 (1997) 159–165
the depletion of nutrients or release of toxic factor(s) in view of our use of a low nutrient base medium for conditioning (HBSS), the observed independence from glucose (the sole nutrient in HBSS) in the conditioned medium and the preserved responses to anti-CD2 mediated stimulation. In a single experiment proliferation was suppressed only by 6.7% whereas in all other experiments (n =12) 24-52% inhibition was observed. This level of inhibition is likely biologically significant. The lowest suppression was observed with the IEL cell line from large bowel which exhibited the highest proliferative response. It could be hypothesized that this very strong response could be only inadequately suppressed by factor(s) from the epithelial cells. These studies corroborate and significantly extend the work of several other groups. Meuer and colleagues have previously shown that supernatant from crude intestinal mucosa inhibits anti-CD3 mediated proliferation of freshly isolated T lymphocytes [18]. This makes it likely that this effect on anti-CD3 stimulation possibly plays a role in maintaining the suppressive tone of the gut in vivo. Our data showing that suppressive factor(s) exist in supernatant from several intestinal epithelial cell lines, that are not present in supernatant from a control non-epithelial cell line or from lung or skin derived epithelial cell lines, suggests that the intestinal epithelium may be the source of this factor in vivo. The studies described here are also consistent with studies described by Ebert and colleagues who showed that conditioned supernatant from the HT29 cell line suppressed the mitogen induced proliferation of PBT [19]. These studies, however, did not examine the effects on different pathways, i.e., CD2 and CD3 mediated stimulation, nor was the effect of supernatant from other intestinal epithelial cell lines or the effect on iIEL measured. We speculate that the TCR/CD3 complex of iIEL and PBT in humans is not structurally different but exhibits functional differences due, at least in part, to the local microenvironment; specifically due to soluble mediators released by intestinal epithelial cells. Defining the factor(s) responsible for these functional characteristics of iIEL will be important in understanding how intestinal tissue limits the promiscuous responses of mucosal lymphocytes to luminal antigens.
.
.
165
Acknowledgements Dr A.D. Christ was supported by partial fellowships from Janggen-Po¨hn Stiftung St. Gallen, Freiwillige Akademische Gesellschaft Basel and Theodor Engelmann Stiftung Basel, Switzerland. Dr S.P. Colgan was supported by NIH grant DK08777, Dr S.P. Balk by NIH grant AI33911, and Dr R.S. Blumberg by NIH grants DK44319 and DK51362.
References [1] G.K. Sim, Adv. Immunol. 58 (1995) 297 – 343. [2] R.S. Blumberg, C.S. Probert, A.D. Christ, H. Kim, J.E. Polischuk, V. Morales, D, Gerdes, A. Chott, S.P, Balk, in: G.N.J, Tytgat, J.F.W.M. Bartelsman, S.J.H. van Deventer (Eds.), Inflammatory Bowel Disease, Kluwer Academic Publishers, Dordrecht, Boston and London, 1995, pp. 266-274. [3] E.C. Ebert, Gastroenterology 97 (1989) 1372 – 1381. [4] E.C. Ebert, A.I. Roberts, R.E. Brolin, K. Raska, Clin. Exp. Immunol. 65 (1986) 148 – 157. [5] U.C. Pirzer, G. Schu¨rmann, S. Post, M. Betzler, S.C. Meuer, Eur. J. Immunol. 20 (1990) 2339 – 2342. [6] L. Qiao, G. Schu¨rmann, M. Betzler, S.C. Meuer, Immunol Res. 10 (1991) 218 – 225. [7] L. Qiao, G. Schu¨rmann, M. Betzler, S.C. Meuer, Gastroenterology 101 (1991) 1529 – 1536. [8] B.C. Sydora, P.F. Mixter, H.R. Holcombe, P. Eghtasady, K. Williams, M.C. Amaral, A. Nel, M. Kronenberg, J. Immunol. 150 (1993) 2179 – 2191. [9] P. Kaoutzani, S.P. Colgan, K.L. Cepek, P. Grant Burkhard, S. Carlson, C. Delp-Archer, M.B. Brenner, J.L. Madara, J. Clin. Invest. 94 (1994) 788 – 796. [10] R. Boismenu, W.L. Havran, Science 266 (1994) 1253 – 1255. [11] C.-C. Schuerer-Maly, L. Eckmann, M.F. Kagnoff, M.T. Falco, F.-E. Maly, Immunology 81 (1994) 85 – 91. [12] C.G. Larsen, A.O. Anderson, E. Appella, J.J. Oppenheim, K. Matsushima, Science 243 (1989) 1464 – 1466. [13] S.P. Balk, J.E. Polischuk, C. Probert, C. Stevens, E. Ebert, J. She, C. Terhorst, R.S. Blumberg, Int. Immunol. 7 (1995) 1237– 1241. [14] S.P. Balk, E.C. Ebert, R.L. Blumenthal, F.V. McDermott, K.W. Wucherpfennig, S.B. Landau, R.S. Blumberg, Science 253 (1991) 1411 – 1415. [15] K. Dharmsathaphorn, J.L. Madara, Methods Enzymol. 192 (1990) 354 – 359. [16] S.P. Colgan, M.B. Resnick, C.A. Parkos, C. Delp-Archer, D. McGuirk, A.E. Bacarra, P.F. Weller, J.L. Madara, J. Immunol. 153 (1994) 2122 – 2129. [17] D.W. Powell, in: L.R. Johnson (Ed.), Physiology of the Gastrointestinal Tract, Raven Press, New York, 1987. [18] L. Qiao, G. Schu¨rmann, F. Autschbach, R. Wallich, S.C. Meuer, Gastroenterology 105 (1993) 814 – 819. [19] E.C. Ebert, A.I. Roberts, D. Devereux, H. Nagase, Cancer Research 50 (1990) 6158 – 6161.