Human T Lymphocyte Proliferative Response to ...

4 downloads 0 Views 62KB Size Report
allogeneic situation (5). Two in vitro ... detect in vitro in both allogeneic and xenogeneic combinations ..... Mechanisms of alloantigen recognition in transplanta-.
Human T Lymphocyte Proliferative Response to Resting Porcine Endothelial Cells Results from an HLA-Restricted, IL-10-Sensitive, Indirect Presentation Pathway But Also Depends on Endothelial-Specific Costimulatory Factors1 Isabelle Valle´e,2 Jean-Maurice Guillaumin, Gilles Thibault, Yves Gruel, Yvon Lebranchu, Pierre Bardos, and Herve´ Watier3 To investigate the mechanisms of cellular rejection in pig-to-human xenotransplantation, the proliferation of different human purified lymphocyte subpopulations in response to swine leukocyte Ag class II-negative porcine aortic endothelial cells (PAEC) was measured in the presence or absence of human autologous adherent cells (huAPC). CD81 lymphocytes proliferated moderately in the absence of huAPC, and the immune response was slightly increased when huAPC were added. CD561 lymphocytes failed to proliferate in response to PAEC whether huAPC were present or not. CD41 lymphocytes alone did not proliferate in response to PAEC, but a strong proliferative response was observed in the presence of metabolically active huAPC. This response was totally abolished by mAbs directed against HLA class II molecules or by pretreatment of huAPC by human IL-10. Even in the presence of huAPC, CD41 lymphocytes failed to respond to fixed PAEC or to PAEC-lysates, suggesting that PAEC must be viable to support lymphocyte proliferation. Finally, none of the nonendothelial porcine adherent cells tested was able to induce human lymphocyte proliferation, despite the fact that they also provided a large set of xenogeneic peptides. Our results show that the indirect presentation pathway of xenoantigens by huAPC to CD41 lymphocytes is crucial in the response to porcine endothelial cells, and that IL-10 could be of therapeutic interest to prevent human lymphocyte activation by this pathway. Furthermore, we demonstrated that stimulatory signals specifically provided by endothelial cells are also necessary for this huAPC-restricted proliferative response. The Journal of Immunology, 1998, 161: 1652–1658.

T

ransplantation of pig vascularized organs to humans could become a reality in the near future. Promising results in the prevention of hyperacute rejection have indeed been obtained in pig-to-primate models by the use of various therapeutic procedures as well as by genetic engineering (1, 2). Nevertheless the occurrence of an “early delayed” xenogeneic rejection could result in graft endothelial cell activation associated with adhesion and infiltration of the transplant by host mononuclear cells such as monocytes and NK cells (3). Such early delayed xenogeneic rejection may be followed by a more classical cellular rejection involving T cell activation and graft infiltration (4), similar to the allogeneic situation (5). Two in vitro xenogeneic models have been developed to study the mechanisms of a such cellular rejection: mixed lymphocyte

Unite´ Propre de Recherche de l’ Enseignement Supe´rieur-Jeune Equipe 1992 “Interactions Hoˆte-Greffon”, Laboratoire d’Immunologie, Faculte´ de Me´decine, Tours cedex, France Received for publication September 5, 1997. Accepted for publication April 9, 1998. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work is supported by the Fondation Langlois, the Fondation Marcel Me´rieux, and the Conseil Ge´ne´ral d’Indre et Loire. I.V. was supported by Ministe`re de la Recherche et de l’ Enseignement Grant 9318.

reaction (XMLR)4 and mixed lymphocyte endothelial cell culture (XMLEC). The widely used XMLR models have led to the conclusion that the proliferative response of human T lymphocytes to porcine bone marrow-derived cells (peripheral blood lymphocytes, monocytes, and dendritic cells) results from a direct interaction between human TCRs and swine leukocyte Ag (SLA) of both class I and class II (6 –11). The least-studied XMLEC models have been used to study the cellular interactions occurring at the first interface between the graft and the recipient immune system, and especially to examine the stimulatory properties of endothelial cells (12). Thus, our team (13) and others (14, 15) have developed XMLEC models in the pig-to-human combination. Using XMLEC models with SLA class II positive porcine aortic endothelial cells (PAEC) either constitutively (7, 15, 16) or by induction with recombinant porcine IFN-g (14), it has been demonstrated that PAEC induced human CD41 T cell proliferation through a direct presentation pathway, as in XMLR. Nevertheless, we (13, 17) and others (14, 18) have used PAEC devoid of any SLA class II molecules, even after coculture with human (hu) PBMC (13) or with human purified CD81 T cells (14). In these experimental conditions, PAEC do not induce proliferation of purified huCD41 T cells (14), but do induce strong huPBMC proliferation (13, 18). We have previously demonstrated that the depletion of adherent huAPC dramatically decreases the proliferation of huPBMC during such an XMLEC, suggesting an indirect presentation pathway of xenoantigens (13). More recently Dorling et al.

2 Current address: Department of Immunology, Institute for Animal Health, Pirbright, Surrey GU24 ONF, United Kingdom. 3

Address correspondence and reprint requests to Dr. Herve´ Watier, Unite´ Propre de Recherche de l’ Enseignement Supe´rieur-Jeune Equipe 1992 “Interactions Hoˆte-Greffon”, Laboratoire d’Immunologie, Faculte´ de Me´decine, 2 bis boulevard Tonnelle´, 37032 Tours cedex, France.

Copyright © 1998 by The American Association of Immunologists

4 Abbreviations used in this paper: XMLR, xenogeneic mixed lymphocyte reaction; XMLEC, xenogeneic mixed lymphocyte endothelial cell culture; SLA, swine leukocyte Ag; PAEC, porcine aortic endothelial cells; hu, human; rhu-IL-10, human rIL-10.

0022-1767/98/$02.00

The Journal of Immunology (18, 19) also provided evidence of an indirect presentation pathway using immortalized porcine endothelial cell lines and unfractionated huPBMC. These XMLEC models based on SLA class II-negative PAEC thus provide unique opportunities to study the indirect presentation pathway, free from all direct presentation of xenopeptides by porcine endothelial cells. Indeed, the indirect presentation pathway is usually difficult to detect in vitro in both allogeneic and xenogeneic combinations because of the usual powerful CD41 lymphocyte proliferation induced by the direct presentation pathway. In the past few years, the relevance of the indirect pathway has been highlighted in organ allograft and in skin xenograft rejection processes (see Refs. 20 and 21 for reviews). Although the recipient APC play an obligatory and crucial role in induction of lymphocyte proliferation by the indirect pathway, the possible contribution of graft stimulatory cells (such as endothelial cells) remains to be defined. Therefore, this study was undertaken to characterize the indirect presentation pathway in our XMLEC assay. We first tested the ability of purified lymphocyte subpopulations, alone or in the presence of autologous huAPC, to proliferate in response to PAEC. Our results provide evidence of a self restricted indirect presentation pathway of xenopeptides to huCD41 lymphocytes by huAPC. Moreover, we demonstrated that nonendothelial adherent MHC class II negative porcine cells were very poor inducers of human lymphocyte proliferation, emphasizing the fact that porcine endothelial cells are the only “stimulatory” cells, even in a self-restricted indirect presentation pathway.

Materials and Methods Porcine cells PAEC were isolated from miniature swine homozygous for the SLAd haplotype (22) as previously described (17). Briefly, PAEC were harvested after treatment of aortas with collagenase A from Clostridium histolyticum (Boehringer Mannheim, Meylan, France). They were seeded in 25 cm2 gelatin-coated culture flasks in RPMI 1640 medium (Life Technologies, Cergy-Pontoise, France) supplemented with 10% heat-inactivated FCS (Life Technologies), 25 mM sodium bicarbonate (Flow, Les Ulis, France), 2 mM glutamine (Flow), 1 mM sodium pyruvate (Flow), 60 mg/ml tylocin (Life Technologies), 50 IU/ml penicilline (Flow), and 50 mg/ml streptomycin (Flow) (referred to as culture medium). PAEC were subcultured after trypsin-EDTA (Life Technologies) treatment and used for the second to the eighth subcultures. Porcine skin fibroblasts (originating from a SLAd homozygous swine) were obtained after removing adipose tissue, followed by treatment of the dermis with collagenase A from C. histolyticum (Boehringer Mannheim) at 37°C for 30 min. Cells were then collected and pelleted and subsequently cultured in 25-cm2 culture flasks in the same medium as above. The following porcine epithelial cell lines were used: PDH, PK15, and LLC-PK1 (porcine kidney epithelial cells) and ST (porcine testis epithelial cells). PDH and ST were kindly provided by Dr. H. Laude (Institut National de la Recherche Agronomique, Jouy-en-Josas, France) (23). LLCPK1 (CL-101) and PK15 (CCL-33) were obtained at the American Type Culture Collection (ATCC; Manassas, VA). All cell lines were cultured in 75-cm2 culture flasks with the same culture medium as PAEC.

Human cells Human PBMC from healthy human volunteers were isolated by centrifugation (20 min, 800 3 g) of heparinized blood over lymphocyte separation medium (Lymphoprep; Nycomed, Oslo, Norway) and cells were collected from the interface. To prepare T- and NK-enriched lymphocytes, huPBMC were first depleted in adherent cells by two 45-min adhesion cycles on plastic. The nonadherent lymphocytes (huPBL) were then passed through a nylon wool column (Uni-Sorb; Valbiotech, Paris, France) for 60 min at 37°C and gently eluted. After washing in RPMI 1640 medium, these cells were mixed with anti-CD4 or anti-CD8-coated magnetic beads (Dynabeads; Dynal, Oslo, Norway) at a bead:cell ratio of 3:1. After incubation at 4°C with gentle end-over-end mixing for 1 h (anti-CD4) or for 30 min (anti-CD8), the rosetted cells were washed six times in RPMI 1640 medium (Life Technologies) using a magnet to retain the rosettes, and then

1653 resuspended in 100 ml RPMI with 10 ml DetachABead (Dynal) for 45 min at room temperature to free the lymphocytes from the magnetic beads. To isolate human CD561 NK cells, cell populations eluted on a nylon wool column was incubated for 45 min (room temperature, end-over-end mixing) with anti-CD56 mAb Leu19 (Becton Dickinson, Mountain View, CA) at a final concentration of 15 mg/ml. After three washes in PBS, cells were mixed with magnetic-beads coated with sheep anti-mouse IgG Abs (Dynal) at a bead:cell ratio of 1:1 for 30 min at 4°C under end-over-end mixing. The rosetted cells were trapped by the magnet and washed six times before being suspended in RPMI 1640 –10% FCS and incubated overnight (37°C, 5% CO2) to detach the beads. CD191 B lymphocytes were prepared by incubating huPBL with antiCD19-coated magnetic beads (Dynal) for 30 min at 4°C. After six washes, rosetted cells were resuspended in 100 ml RPMI 1640 medium and incubated with 10 ml DetachABead for 30 min at room temperature. All the purified lymphocyte subpopulations were pure (.98%) as determined by FACS analysis, and cell viability was always .98% as determined by trypan blue exclusion test. Human APC were isolated using a procedure described by Freundlich and Avdalovic (24). Briefly, huPBMC were incubated for 45 min at 37°C in a plastic culture flask (Becton Dickinson) precoated with bovine gelatin (Sigma, St-Quentin-Fallavier, France) and autologous plasma. After washing out nonadherent cells, huAPC were collected following incubation with cold EDTA (10 mM in PBS) mixed with an equal volume of RPMI-20% FCS for 15 min at 4°C. This population always contained more than 70% CD141 cells as determined by FACS analysis.

Xenogeneic mixed cultures XMLECs were performed as previously described (13). Briefly, PAEC were seeded in 96-well tissue culture plates (Falcon 3072) to obtain confluent monolayers (3 3 104 cells/well) and then irradiated (30 Gy). When nonendothelial cells were used as stimulating cells instead of PAEC, they were also seeded at 3 3 104 cells/well and irradiated. To obtain PAEC lysates, PAEC (3 3 105 cells/ml) were sonicated with ultrasonic waves (Vibra cell; Bioblock Scientific, Illkirch, France) until no whole cells remained. Samples of 100 ml of this suspension were used in some experiments (corresponding to 3 3 104 cells). Responding cells (huPBMC or purified human lymphocytes) were added in different amounts (0.5 3 105, 1 3 105, or 1.5 3 105 cells/well). The following murine mAbs were used in blocking experiments: L243 (IgG2a, ATCC, HB-55) and L2 (IgG1) (25) directed against monomorphic HLA-DR and HLA-DQ determinants, respectively; 2-27-3 (IgG1) (26) and 74-11-10 (IgG2b; Veterinary Medical Research and Development, Pullman, WA) directed against SLA class I molecules; and W6/32 (IgG2a, Dako, Glostrup, Denmark) directed against HLA-class-I molecules. The mAbs U7.27 (IgG2a; Immunotech, Marseille, France), MOPC-195 (IgG2b; Immunotech), and 679.1 Mc7 (IgG1; Immunotech) were used as isotypic control. In some experiments, CD41 T lymphocytes were cultured for 3 days in the presence of PHA (0.05 mg/ml, Sigma), either alone or on porcine adherent cell monolayers. In all cases, coculture was performed over 6 days (37°C, 5% CO2), the time which was previously determined as optimal for maximal proliferation. Cocultures were examined once a day throughout culture. Cells were pulsed with 1 mCi (3.6 3 104 Bq) tritiated thymidine (Amersham, Little Chalfont, U.K.) for 18 h before they were collected onto filter discs using an automated harvester (Skatron Instrument, Lier, Norway). Tritiated thymidine incorporation was measured with a liquid scintillation b-counter (Tri-Carb 2550 TR/LL, Packard, Rungis, France). Results were expressed in cpm as mean 6 SD of triplicate wells.

Human APC or stimulating cells treatments Human APC or PAEC were fixed using a procedure described by Moreno et al. (27). Cells were washed twice in PBS and incubated with paraformaldehyde (0.1% in PBS) for 5 min at 37°C. The reaction was stopped by adding cold 0.06% glycine. Cells were washed twice in PBS, resuspended in culture medium, and incubated for 1 h at 37°C. Cells were then pelleted and resuspended in medium before being used in XMLECs. Recombinant human IL-10 (rhuIL-10) was kindly provided by Dr. J. Banchereau (Schering-Plough, Dardilly, France). Human APC (5 3 106 cells/ ml) were incubated for 24 h (37°C, 5% CO2) with 0.2 mg/ml (100 U/ml) of rhuIL-10 diluted in culture medium. Human APC were washed three times in RPMI 1640 medium after this incubation to remove rhuIL-10.

Flow cytometry analysis The following murine mAbs were used: MSA-3, an IgG2a directed against a SLA-DR monomorphic determinant (Veterinary Medical Research and Development) (28); SPV-L3, an IgG2a anti-HLA-DQ (Immunotech) used for SLA-DQ recognition (17); PT85A, an IgG2a directed against SLA class

1654

INDIRECT PRESENTATION PATHWAY IN THE PIG-TO-HUMAN XMLEC

FIGURE 1. Proliferation of huCD41 cells to PAEC in the presence of autologous huAPC (A) or huCD191 cells (B). A, Confluent PAEC monolayers and different amounts huAPC were irradiated (30 Gy) before starting coculture with 105 purified huPBMC or huCD41 cells per well. B, Confluent PAEC monolayers and huCD191 lymphocytes were irradiated (30 Gy) before starting coculture with 105 purified huCD41 cells per well. At 18 h before termination of coculture, wells were pulsed with tritiated thymidine (3.6 3 104 Bq/well). Lymphocyte proliferation after 6-day coculture was measured by thymidine uptake quantitated on a liquid scintillation b-counter. Results are expressed in cpm as mean 6 SD of triplicate determination in one experiment representative of five (A) and two (B) experiments.

I determinant (VMRD); and BL4, an IgG2a anti-human CD4 (Immunotech) used as an IgG2a isotypic control. Porcine cell lines (5 3 105 cells) were incubated for 30 min at 4°C with saturating amounts of Abs (ascitic fluids, 1:200 dilution in PBS; purified Abs, 1:5 dilution), then washed twice in cold PBS and incubated for 30 min at 4°C with FITC-conjugated goat anti-mouse IgG F(ab)92-polyclonal Ab (Immunotech). Flow cytometry analysis was performed with a FACStarPlus (Becton Dickinson).

Results

Response of purified CD41 lymphocytes to PAEC needs adherent huAPC Having previously established that APC-depleted human lymphocytes failed to proliferate in response to PAEC (13), we purified human lymphocyte subpopulations and incubated them on PAEC monolayers, either alone or in the presence of autologous adherent huAPC. Highly purified huCD41 lymphocytes (.98% CD41 cells) which were unable to proliferate in the presence of PHA (data not shown) also did not proliferate in the presence of PAEC monolayers (Fig. 1A). By contrast, a strong proliferative response was obtained when irradiated adherent huAPC were coincubated with huCD41 lymphocytes in the assay (Fig. 1A). Paraformaldehyde-fixed huAPCs, used instead of metabolically active huAPC, did not stimulate the proliferation of purified huCD41 lymphocytes (data not shown). The magnitude of the huCD41 lymphocyte response was directly dependent upon the number of viable adherent huAPC added, showing that accessory cells were the re-

FIGURE 2. Proliferation of huCD81 cells (A) or huCD561 NK cells (B) to PAEC in the presence or absence of autologous huAPC. Confluent PAEC monolayers and huAPC were irradiated (30 Gy) before starting coculture with 105 purified huCD81 cells or 105 huCD561 NK cells per well (or 105 huPBMC per well as control). At 18 h before termination of coculture, wells were pulsed with tritiated thymidine (3.6 3 104 Bq/well). Lymphocyte proliferation after 6-day coculture was measured by thymidine uptake quantitated on a liquid scintillation b-counter. Results are expressed in cpm as mean 6 SD of triplicate determinations in one experiment representative of four (A) and two (B) experiments.

strictive cells in the assay (Fig. 1A). On the other hand, because 1% of B cells have a B cell receptor specific to pig antigeneic determinants (29) and are thus putative APC, purified human CD191 lymphocytes (.99% purity) were compared with adherent huAPC in XMLEC experiments. Human CD41 lymphocytes proliferated in the presence of autologous adherent huAPC but completely failed to respond in the presence of autologous B cells (Fig. 1B). On the other hand, the proliferative response of purified huCD81 lymphocytes alone (.98% purity) to PAEC was weak (Fig. 2A). Inhibition of the proliferation (between 61% and 100%) was observed in the presence of the anti-SLA class-I mAbs 2-27-3 or 74-11-10. The anti-HLA class-I mAb, W6/32, did not affect proliferation. The CD81 lymphocyte response was slightly increased in the presence of adherent huAPC (Fig. 2A) but was always lower than that obtained with CD41 cells and adherent huAPC. Furthermore, no proliferation with purified human CD561 NK cells (95% CD561 cells) was observed, even in the presence of adherent huAPC (Fig. 2B). Presentation of xenopeptides is restricted by HLA class II molecules Because the CD41 lymphocyte proliferative response was strictly dependent on the presence of metabolically active autologous adherent huAPC, we determined the mechanisms of xenogeneic Ag presentation. The presence of the anti-HLA-DR blocking mAb L243 strongly inhibited lymphocyte proliferation, down by 85% of the proliferation observed in the presence of the isotypic control (Fig. 3). The anti-HLA-DQ blocking mAb, L2, had a more moderate inhibitory effect, down by 43% of the proliferation obtained in the presence of the isotypic control (Fig. 3). Anti-HLA-DR and

The Journal of Immunology

FIGURE 3. XMLEC with mAbs directed against HLA class II molecules. Human PBMC (105 cells/well) were cocultured on irradiated PAEC monolayers for 6 days in the presence of mAbs: anti-HLA-DR (L243, IgG2a), anti-HLA-DQ (L2, IgG1), IgG2a (isotypic control), or IgG1 (isotypic control). At 18 h before termination of coculture, wells were pulsed with tritiated thymidine (3.6 3 104 Bq/well), and proliferation was measured by thymidine uptake quantitated on a liquid scintillation b-counter. Results are expressed in cpm as mean 6 SD of triplicate determinations. Data are from one experiment representative of two experiments.

anti-HLA-DQ mAbs together totally inhibited the proliferative response by means of an additive effect, demonstrating the availability of a unique indirect presentation pathway. Induction of CD41 lymphocyte proliferation by huAPC in response to PAEC is IL-10 sensitive We had observed in preliminary experiments that rhuIL-10 strongly decreased huPBMC proliferation in response to PAEC. The inhibitory effect was maximum with 100 U/ml of rhuIL-10 (down by 76% of the proliferation without IL-10) and was not further enhanced with higher doses (data not shown). Because huAPC might be one of the targets of IL-10 (30), we tested the ability of rhuIL-10-pretreated huAPC to induce huCD41 lymphocyte proliferation. Human APC were then preincubated with 100 U/ml of rhuIL-10 for 24 h and washed before being added to purified huCD41 lymphocytes in XMLEC experiments. Human CD41 lymphocyte proliferation was strongly decreased when huAPC pretreated with rhuIL-10 were added instead of untreated huAPC (Fig. 4). This dramatic effect first confirmed the ability of IL-10 to interfere with Ag presentation and also strengthened the

FIGURE 4. XMLEC with rhuIL-10-pretreated huAPC. Autologous huAPC were pretreated or not with rhuIL10 (100 U/ml) for 24 h, then extensively washed before adding to XMLEC assay. Human CD41 cells (105 cells/well) and autologous huAPC (3 3 104 cells/well) were cocultured on irradiated PAEC monolayers for 6 days. At 18 h before termination of coculture, wells were pulsed with tritiated thymidine (3.6 3 104 Bq/well), and proliferation was measured by thymidine uptake quantitated on a liquid scintillation b-counter. Results are expressed in cpm as mean 6 SD of triplicate determinations in one experiment representative of four experiments.

1655

FIGURE 5. A, XMLEC was performed with viable PAEC in comparison with paraformaldehyde-fixed-PAEC. Stimulating cells were replaced by PAEC or XMLEC supernatants alone or associated with fixed PAEC. B, PAEC-lysates were used instead of viable PAEC. Lysate A and lysate B correspond to 3 3 104 and 6 3 104 PAEC, respectively. Human PBMC proliferative response was measured after 6-day coculture. At 18 h before termination of coculture, wells were pulsed with tritiated thymidine (3.6 3 104 Bq/well) and proliferation was measured by thymidine uptake quantitated on a liquid scintillation b-counter. Results are expressed in cpm as mean 6 SD of triplicate determinations in one experiment representative of three experiments.

key role of huAPC in the induction of CD41 lymphocyte proliferation in response to PAEC. Human APC-dependent lymphocyte proliferation also requires metabolically active PAEC Attention was then focused on the xenogeneic component of the XMLEC assay to investigate how porcine endothelial cells contribute to lymphocyte proliferation. We first investigated the ability of resting PAEC to supply membrane or soluble stimulatory factors to human lymphocytes. Because huPBMC did not proliferate in response to fixed PAEC or to PAEC supernatant (Fig. 5A), we suspected that a stimulatory soluble factor could be produced by PAEC after their activation during the XMLEC assay. However, huPBMC also failed to proliferate in response to XMLEC-supernatant (Fig. 5A). Moreover, the combination of fixed PAEC with PAEC or XMLEC supernatants was also unable to induce huPBMC proliferation (Fig. 5A). Finally, the ability of PAEC constituents to supply xenopeptides was analyzed. When PAEC lysates were used as stimulants instead of whole viable PAEC, huPBMC totally failed to respond (Fig. 5B). Taken together, these results demonstrate that PAEC must be viable to induce the proliferation of human lymphocytes. Role of PAEC in initiating human lymphocyte proliferation is unique compared with other porcine adherent cells We then determined whether other viable porcine adherent cells could supply xenoantigens to huAPC and induce proliferation of human lymphocytes. Porcine cells originating from different tissues, i.e., PDH, LLCPK-1, PK15 (kidney epithelial cell lines), ST (testis epithelial cell line), and PSF (primary porcine skin fibroblast cells) were used as stimulants at the same cell density as PAEC. All cells expressed the same SLA pattern as PAEC as determined by flow cytometry analysis. Indeed, they all expressed SLA class

1656

INDIRECT PRESENTATION PATHWAY IN THE PIG-TO-HUMAN XMLEC

Table I. Proliferation of human lymphocytes in response to various porcine nonendothelial cells a Renal

huPBMC proliferation (n) PHA-treated huCD41 proliferation (n 5 3)

Testis

Skin Fibroblast

PAEC

LLCPK1

PK15

PDH

ST

PSF

100 100

25.0 6 10 (3) 0

19.8 6 7 (4) 0

2.0 6 1.8 (5) 0

1.9 6 1 (4) 0

22.5 6 11 (5) NT

a The magnitude of the lymphocyte response is expressed as a percentage of the proliferation in relation to PAEC-stimulating cells. n, number of experiments; NT, not tested.

I Ags but did not express any SLA class II molecules (SLA-DR or SLA-DQ) (data not shown). In the different experiments where these nonendothelial cells were compared with PAEC, the most potent stimulating cells were always PAEC. The proliferative response to PAEC was thus deemed to be 100% to provide the reference for interexperiment comparisons (Table I). Human PBMC totally failed to respond to ST and PDH epithelial cell lines and moderately responded (,25% of the mean proliferative response obtained with PAEC) to LLC-PK1 and PK15 epithelial cell lines and skin fibroblasts (Table I). Because all these nonendothelial cell lines probably have approximately the same vast array of xenopeptides to be processed by huAPC and presented to human lymphocytes as PAEC, we hypothesized that the stronger proliferative response of human lymphocytes to PAEC might be due to unique costimulatory properties exhibited by porcine endothelial cells. To bypass the need of huAPC to deliver the signal through TCR/ MHC-peptide interaction, we decided to use PHA-activated huCD41 lymphocytes in coculture experiments, a model described by Pober and coworkers (31). As expected, resting purified huCD41 lymphocytes alone failed to proliferate either in response to nonendothelial cells or to PAEC (data not shown). More interestingly, only PAEC were able to induce the proliferation of PHAactivated huCD41 lymphocytes, whereas none of the nonendothelial cell monolayers were able to do so (Table I). This clear-cut difference demonstrated that PAEC display unique costimulatory properties.

Discussion To study the mechanisms of human PBMC proliferation in response to SLA class II negative porcine endothelial cells (13), we purified human lymphocyte subpopulations and analyzed their proliferative response in the presence or in the absence of autologous adherent huAPC. Human CD561 NK cells completely failed to proliferate in response to PAEC even in the presence of autologous adherent APC. Human CD81 cells alone weakly proliferated in response to PAEC, in accordance with previous reports (14 –16), and was inhibited by anti-SLA class I mAbs (14, 15), confirming a direct presentation pathway. In addition, the CD81 proliferation was enhanced in the presence of huAPC, revealing for the first time CD81 cell stimulation by self APC in the XMLEC assay. The indirect presentation pathway to CD81 lymphocytes, already described in allogeneic cocultures (32, 33), needs further investigation in the xenogeneic situation, especially through cytotoxicity experiments. As expected, because PAEC remained SLA class II negative throughout coculture due to the inability of human IFN-g to crossreact with porcine cells (13, 17), purified huCD41 lymphocytes alone never proliferated in response to PAEC, whereas they responded strongly in the presence of autologous adherent huAPC. The dramatic inhibitory effect of anti-HLA class II mAbs in the XMLEC assay clearly demonstrated a self restricted indirect presentation pathway. When used in the same number as adherent

huAPC, purified huCD191 cells were unable to stimulate a CD41 lymphocyte response, in agreement with the results obtained by Dorling et al. (18) with CD191 lymphocyte-depleted huPBMC. The inability of CD191 B cells to support CD41 cell proliferation could be explained by too low a percentage of B cells expressing B cell receptor recognizing pig Ags. The only effective APCs in our model were thus adherent cells, provided that they were metabolically active. Furthermore, the allogeneic indirect presentation pathway is considered to be resistant to immunosuppressive drugs such as cyclosporine A (34) classically used to fight against cell-mediated rejection. This has revealed the need to develop new immunomodulators to overcome lymphocyte activation through this indirect pathway. Because rhuIL-10-treated adherent APC failed to stimulate the proliferation of huCD41 cells, we believe that IL-10 could be an effective therapeutic tool to overcome CD41 cell activation in the pig-to-human combination. Indeed, IL-10 is a well known inhibitor of the APC function of monocytes (30, 35), by decreasing their membrane HLA-class II expression (36) and B7 costimulatory molecule expression (37). It could thus be speculated that high local expression of IL-10 within the xenograft could contribute to prevention of cellular rejection. In the indirect pathway of allorecognition, the allopeptides originating from the degradation of donor polymorphic MHC molecules fill the groove of responder APC MHC class II molecules (38 – 40). In the xenogeneic combination, polymorphism arises not only from SLA molecules but also, and probably mainly, from each other individual protein, due to wide amino acid sequence dissimilarities between species. As a consequence, any porcine cells (whatever their origin) could provide xenopeptides to huAPC. Unexpectedly, among various SLA class II negative adherent monolayers, only endothelial cells proved to be able to induce a strong proliferation of human lymphocytes. These findings clearly demonstrated that the availability of porcine xenopeptides (originating from various cell lines) in the presence of functional APC is not sufficient to promote human lymphocyte proliferation. These results also indicated that porcine endothelial cells display unique stimulatory properties, thus leading to two hypotheses. The first was that porcine endothelial cells express unique xenoantigeneic peptides. However PAEC lysates (a situation in which all endothelial membranes and internal structures are released into the culture medium) were unable to induce lymphocyte proliferation, ruling out this first hypothesis. The second supposition was that porcine endothelial cells contribute to lymphocyte proliferation by providing costimulatory signals. We observed first that resting PAEC in the form of culture supernatants or fixed cells or both did not provide any stimulatory factors. It could not be excluded that some membrane factors might be altered by paraformaldehyde treatment but the procedure described by Moreno and Lipsky (27) is widely used for this purpose. We also suspected that PAEC could be activated during the XMLEC assay

The Journal of Immunology and then could provide stimulatory signals. However, such factors could not be evidenced in XMLEC supernatants, and membrane factors on activated PAEC could not be evaluated. It is indeed known that some membrane costimulatory molecules are induced on endothelial cells, for example in response to CD40L/CD40 ligation (41, 42). However, all our results led to the conclusion that porcine endothelial cells, human CD41 lymphocytes, and autologous huAPC were involved in a dynamic process, closely interacting in a “me´nage-a`-trois.” In this situation PAEC could act either directly on huAPC activation or on CD41 lymphocytes, or both. To investigate the interaction between endothelial cells and human lymphocytes, we used PHA-activated CD41 lymphocytes, an experimental model that was previously developed in allogeneic coculture to bypass the signal delivered by TCR/MHC interaction and to study the costimulatory properties of the endothelium (31). Because PAEC were able to induce proliferation of PHA-activated human CD41 lymphocytes, we concluded that porcine endothelial cells could provide effective costimulatory signals. This then favored the hypothesis of an interaction between endothelial cells and human lymphocytes in our experimental model with three cells (PAEC, huAPC, CD41 lymphocytes). Moreover the hypothesis of such a specific interaction was strengthened by the fact that none of the nonendothelial cell lines were able to induce proliferation of PHA-activated CD41 lymphocytes. These results led to the logical conclusion that porcine endothelial cells might act together with human lymphocytes and huAPC in a trans-regulated activation mechanism and might deliver a cosignal to human lymphocytes. Trans-costimulatory signals have already been observed in an allogeneic model (43), although it has been reported that cis-regulated activation (MHC class II and costimulatory molecules expressed on the same APC) is more effective than transregulated activation (44, 45). It has been observed in preliminary experiments that an anti-human CD2-blocking mAb inhibited the proliferation of PHA-activated huCD41 lymphocytes in response to PAEC (unpublished data), thus supporting the hypothesis of endothelial trans-costimulatory signals. Involvement of the CD2 pathway has already been demonstrated in XMLEC models using SLA class II positive porcine endothelial cells (14, 15), although the nature of the endothelial porcine ligand of human CD2 is not known. However, in our experimental conditions, the hypothesis of a trans-activation pathway does not exclude the possibility that porcine endothelial cells might also interact directly with huAPC, up-regulating their ability to process and present xenopeptides as well as acquiring all the effective costimulatory molecules to activate human CD41 lymphocytes in a classical cis-regulated manner. Such endothelial costimulatory properties need to be further investigated in this useful model of a pure self restricted indirect presentation pathway.

Acknowledgments We thank Dr. H. Salmon for giving access to the pig herd maintained at the Institut National de la Recherche Agronomique, Nouzilly, France. We also thank Mrs. H. Aget, Mr. P. Louisot, M. S. Mu¨ller, Pr. Le Floch, and Pr. Besnard for their valuable collaboration, and we are grateful to Drs. J. Banchereau, H. Laude, C. Renard, J. Lunney, B. Georges, C. Pourcel, and A. Jestin for their precious gifts.

References 1. Cozzi, E., and D. J. White. 1995. The generation of transgenic pigs as potential organ donors for humans. Nat. Med. 1:964. 2. Parker, W., S. Saadi, S. Lin, Z. Holzknecht, M. Bustos, and J. Platt. 1996. Transplantation of discordant xenografts: a challenge revisited. Immunol. Today 17:373. 3. Bach, F., H. Winkler, C. Ferran, W. Hancock, and S. Robson. 1996. Delayed xenograft rejection. Immunol. Today 17:379.

1657 4. Nowak, R. 1994. Xenotransplants set to resume. Science 266:1148. 5. Fedoseyeva, E. V. 1995. Mechanisms of alloantigen recognition in transplantation. Biochemistry (Moscow) 60:169. 6. Kumagai-Braesch, M., M. Satake, O. Korsgren, A. Andersson, and E. Mo¨ller. 1993. Characterization of cellular human anti-porcine xenoreactivity. Clin. Transplant. 7:273. 7. Kirk, A., R. A. Li, M. S. Kinch, K. A. Abernethy, C. Doyle and, R. Bollinger. 1993. The human antiporcine cellular repertoire: in vitro studies of acquired and innate cellular responsiveness. Transplantation 55:924. 8. Dorling, A., R. Binns, and R. Lechler. 1996. Cellular xenoresponses: although vigorous, direct human T cell anti-pig primary xenoresponses are significantly weaker than equivalent alloresponses. Xenotransplantation 3:149. 9. Satake, M., N. Kawagishi, and E. Moller. 1996. Direct activation of human responder T cells by porcine stimulator cells leads to T cell proliferation and cytotoxic T cell development. Xenotransplantation 3:198. 10. Yamada, K., J. D. Seebach, H. DerSimonian, and D. Sachs. 1996. Human anti-pig T-cell mediated cytotoxicity. Xenotransplantation 3:179. 11. Cunningham, A. C., T. J. Butler, and J. A. Kirby. 1994. Demonstration of direct xenorecognition of porcine cells by human cytotoxic T lymphocytes. Immunology 81:268. 12. Pober, J. S., C. G. Orosz, M. L. Rose, and C. O. Savage. 1996. Can graft endothelial cells initiate a host anti-graft immune response? Transplantation 61:343. 13. Valle´e, I., H. Watier, G. Thibault, H. Salmon, Y. Gruel, Y. Lebranchu, and P. Bardos. 1995. Evidence of noninvolvement of swine MHC class II in the in vitro proliferative response of human lymphocytes to porcine endothelial cells. Transplantation 59:897. 14. Murray, A. G., M. M. Khodadoust, J. S. Pober, and A. L. M. Bothwell. 1994. Porcine aortic endothelial cells activate human T cells: direct presentation of MHC antigens and costimulation by ligands for human CD2 and CD28. Immunity 1:57. 15. Rollins, S. A., S. P. Kennedy, A. J. Chodera, E. A. Elliott, G. B. Zavoico, and L. A. Matis. 1994. Evidence that activation of human T cells by porcine endothelium involves direct recognition of porcine SLA and costimulation by porcine ligands for LFA-1 and CD2. Transplantation 57:1709. 16. Bravery, C. A., P. Batten, M. H. Yacoub, and M. L. Rose. 1995. Direct recognition of SLA- and HLA-like class II antigens on porcine endothelium by human T cells results in T cell activation and releases of interleukin-2. Transplantation 60:1024. 17. Watier, H., I. Valle´e, G. Thibault, A. C. Lalmanach, M. Lacord, Y. Gruel, Y. Lebranchu, H. Salmon, and P. Bardos. 1994. Effect of human inflammatory cytokines on porcine endothelial cell MHC molecule expression: unique role for TNF-a in MHC class-II induction. Transplant. Proc. 26:1152. 18. Dorling, A., R. Binns, and R. I. Lechler. 1996. Cellular xenoresponses: observation of significant primary indirect human T cell anti-pig xenoresponses using co-stimulator-deficient or SLA class II-negative porcine stimulators. Xenotransplantation 3:112. 19. Dorling, A., G. Lombardi, R. Binns, and R. I. Lechler. 1996. Detection of primary direct and indirect human anti-porcine T cell responses using a porcine dendritic cell population. Eur. J. Immunol. 26:1378. 20. Auchincloss, H., and H. Sultan,. 1996. Antigen processing and presentation in transplantation. Curr. Opin. Immunol. 8:681. 21. Czitrom, A. A. 1996. The immune response: the afferent arm. Clin. Orthopaedics 326:1. 22. Sachs, D. H., G. Leight, J. Cone, S. Schwarz, L. Stuart, and S. Rosenberg. 1976. Transplantation in miniature swine. I. Fixation of the major histocompatibility complex. Transplantation 22:559. 23. Laude, H., J.-M. Chapsal, J. Gelfi, S. Labiau, and J. Grosclaude 1986. Antigenic structure of transmissible gastroenteritis virus. I. Properties of monoclonal antibodies directed against virion proteins. J. Gen. Virol. 67:119. 24. Freundlich, B., and N. Avdalovic. 1983. Use of gelatin/plasma coated flasks for isolating human peripheral blood monocytes. J Immunol. Methods 62:31. 25. Fermand, J.-P, A. Chevalier, and J.-C. Brouet. 1986. Characterization of a mAb recognizing a monomorphic determinant of the a-chain of class II DQ antigens. Scand. J. Immunol. 24:313. 26. Ivanoska, D., D. C. Sun, and J. K. Lunney. 1991. Production of monoclonal antibodies reactive with polymorphic and monomorphic determinants of SLA class I gene products. Immunogenetics 33:220. 27. Moreno, J., and P. E. Lipsky. 1986. Differential ability of fixed antigen-presenting cells to stimulate nominal antigen-reactive and alloreactive T4 lymphocytes. J. Immunol. 136:3579. 28. Hammerberg, C., and G. G. Schurig. 1986. Characterization of monoclonal antibodies directed against swine leukocytes. Vet. Immunol. Immunopathol. 11:107. 29. Galili, U., F. Anaraki, A. Thall, C. Hillblack, and M. Radic. 1993. One percent of human circulating B-lymphocytes are capable of producing the natural antiGal antibody. Blood 82:2485. 30. Moore, K. W., A. O’Garra, R. de Waal Malefyt, P. Vieira, and T. R. Mosmann. 1993. Interleukin-10. Annu. Rev. Immunol. 11:165. 31. Hughes, C., C. Savage, and J. Pober. 1990. Endothelial cells augment T cell interleukin 2 production by a contact-dependent mechanism involving CD2/ LFA-3 interaction. J. Exp. Med. 171:1453. 32. Rock, K., L. Rothstein, S. Gamble, and C. Fleischacker. 1993. Characterization of antigen-presenting cells that present exogenous antigens in association with class I molecules. J. Immunol. 150:438. 33. Breurvriesendorp, B. S., and P. Ivanyi. 1993. Self-restricted primary human histocompatibility leukocyte antigen (HLA)-specific cytotoxic lymphocytes-T. Int. Immunol. 5:103.

1658

INDIRECT PRESENTATION PATHWAY IN THE PIG-TO-HUMAN XMLEC

34. Sawyer, G. J., R. Dalchau, and J. W. Fabre. 1993. Indirect T cell allorecognition: a cyclosporine A resistant pathway for T cell help for antibody production to donor MHC antigens. Transplant. Immunol. 1:77. 35. Bromberg, J. S. 1995. IL-10 immunosuppression in transplantation. Curr. Opin. Immunol. 7:639. 36. de Waal Malefyt, R., J. Haanen, H. Spits, M.-G. Roncarolo, A. teVelde, C. Figdor, K. Johnson, R. Kastelein, H. Issel, and J. de Vries. 1991. Interleukin 10 (IL-10) and viral IL-10 strongly reduce antigen-specific human T cell proliferation by diminishing the antigen-presenting capacity of monocytes via downregulation of class II major histocompatibility complex expression. J. Exp. Med. 174:915. 37. Willems, F., A. Marchant, J. P. Delville, C. Gerard, A. Delvaux, T. Velu, M. Deboer, and M Goldman. 1994. Interleukin-10 inhibits B7 and intercellular adhesion molecule-1 expression on human monocytes. Eur. J. Immunol. 24:1007. 38. Liu, Z., Y. K. Sun, Y. P. Xi, A. Maffei, E. Reed, P. Harris, and N. Suciu-Foca. 1993. Contribution of direct and indirect recognition pathways to T-cell alloreactivity. J. Exp. Med. 177:1643. 39. Liu, Z., E. Reed, P. Harris, and N. Suciu-Foca. 1993. Role of anti-HLA immune response in organ transplant rejection. In Rejection and Tolerance. J.-L.

40.

41.

42.

43. 44.

45.

Touraine, J. Traeger, H. Be´tuel, J. M. Dubernard, J.P. Revillard, and C. Dupuy, eds. Kluver, Dordrecht, The Netherlands, p. 111. Warrens, A., G. Lombardi, and R. Lechler. 1994. MHC and alloreactivity: presentation and recognition of major and minor histocompatibility antigens. Transplant. Immunol. 2:103. Karmann, K., C. C. W. Hughes, J. Schechner, W. C. Fanslow, and J. S. Pober. 1995. CD40 on human endothelial cells: inducibility by cytokines and functional regulation of adhesion molecule expression. Proc. Natl. Acad. Sci. USA 92:4342. Yellin, M. J., J. Brett, D. Baum, A. Matsushima, M. Szabolcs, D. Stern, and L. Chess. 1995. Functional interactions of T cells with endothelial cells: the role of CD40L-CD40-mediated signals. J. Exp. Med. 182:1857. Lechler, R. I., and F. Marelli-Berg. 1997. Do endothelial cells fit the paradigm of antigen-presenting cells? Biochem. Soc. Trans. 25:723. Liu, Y., and C. Janeway. 1992. Cells that present both specific ligand and costimulatory activity are the most efficient inducers of clonal expansion of normal CD4 T cells. Proc. Natl. Acad. Sci. USA 89:3845. Wingren, A. G., E. Parra, M. Varga, T. Kalland, H. O. Sjogren, G. Hedlund, and M. Dohlsten. 1995. T cell activation pathways: B7, LFA-3, and ICAM-1 shape unique T cell profiles. Crit. Rev. Immunol. 15:235.