HETEROGENEITY OF T CELL CLONES SPECIFIC FOR A SINGLE INDIRECT ALLOANTIGENIC EPITOPE (I-A^b/H-2K^d54–68) THAT MEDIATE TRANSPLANT REJECTION¹

Abstract

Background. 

One of the complexities of solid organ allograft rejection is the inherent diversity of the specific T cell antigenic epitopes that participate in this response, including the role of direct alloantigen recognition and indirect recognition of donor-derived peptides in recipient antigen-presenting cells.To probe the role of distinct T cell receptor (TCR) avidity differences and the role of cytokine expression patterns of different effector T cells that may participate in allograft rejections, we have identified a dominant allopeptide derived from the H-2K^d molecule, recognized by H-2^b CD4 T cells in the context of syngeneic I-A^b.

Methods. 

To identify a stimulatory peptide derived from the H-2K^d molecule, a panel of synthetic overlapping peptides was screened for immunogenicity and a panel of T cell clones established. These clones were characterized for TCR Vβ usage by mAb staining and/or reverse transcribed-polymerase chain reaction analysis, peptide dose sensitivity as a marker of TCR avidity, cytokine expression phenotype in vitro, and their ability to mediate rejection of a vascularized cardiac allograft after adoptive transfer to immunodeficient mice.

Results. 

The H-2K^d54–68 peptide was identified as a dominant stimulatory peptide by the ability of T cells from C57BL/6 (H-2^b) mice primed by a combination of allogeneic spleen cell injection and mixed peptide immunization to mount an in vitro proliferative response and interferon-γ production by peptide stimulation. Furthermore, direct immunization with synthetic H-2K^d54–68 peptide of normal C57BL/6 mice resulted in accelerated rejection of both skin and cardiac allografts from B10.D2 (H-2^d) mice, but not 3rd party B10.BR (H-2^k) grafts. A panel of 15 distinct CD4⁺ clones specific for H-2K^d54–68 peptide were established and shown to utilize a variety of TCR Vβ and different apparent TCR avidities to H-2K^d54–68 peptide when stimulated in vitro. To characterize these clones further, two clones were chosen based on the difference of avidity to H-2K^d54–68 peptide. The cytokine expression pattern was determined and indirect alloantigen specificity confirmed by analysis of responses to purified peptide and B10.D2 spleen cells using normal H-2^b and I-Aβ chain knockout mice as APC donors. Both of these T cell clones were able to mediate rejection of B10.D2, but not B10.BR hearts, in immunodeficient mice, but the morphological pattern of T cell infiltration was distinct.

Conclusions. 

These results demonstrate the potential importance of fine dissection of the alloantigeneic response to solid organ transplants and provide unique insights into the role of TCR avidity and cytokine expression patterns in different morphological patterns of transplant rejection.

The nature of the specific antigens that stimulate the immune response that leads to rejection of solid organ allografts is quite complex. It has long been recognized that the antigens coded for by the major histocompatibility antigens (MHC) are particularly potent for the induction of allograft rejection, but that multiple other “minor” histocompatibility antigens also exist. The special potency of the MHC is most likely accounted for by the “direct” pathway of alloantigen recognition, in which allo-MHC molecules bind an array of peptides and cross-react with multiple different host T cell receptors (TCRs) resulting in a high frequency, vigorous response. Initially, the dominance of the MHC in the minds of most transplantation immunologists seemed to leave little room for a significant role for the “indirect” pathway in which host MHC molecules present donor-derived peptides. It is now clear that all “minor histocompatibility antigens” are presented via the indirect pathway and rapid rejection can occur between strains of rodents that share the entire MHC region, but differ at multiple minor H-loci. Furthermore, the observation that murine skin transplant rejection can proceed at a normal tempo via a CD4 T cell-dependant pathway to reject an allograft genetically deficient in class II antigens (¹), indicates that the indirect pathway can be quite vigorous. Although it is clear that both indirect and direct pathways of allorecognition exist, the relative quantitative role of these two pathways in actual allograft rejection, and whether these two pathways differ in a qualitative way, remains unclear.

One fundamental problem in addressing this question is the inherent complexity of the potential antigenic entities present in an entire organ and the inability to resolve the responses of T cells to reactive to discrete peptide/MHC epitopes within the mixture of responses that occur during allograft rejection. One approach to the dissection of this complexity is the identification of individual T cell clones that can directly mediate allograft rejection in an isolated system without the potentially confounding effects of bystander T cells. This report describes the initial phase of this approach by the identification of a strong allopeptide specific response and analysis of representative T cell clones specific for a single “indirect” alloantigenic epitope.

MATERIALS AND METHODS

Mice.

C57BL/6J (C57BL/6), B10.D2/nSnJ (B10.D2), B10.BR/SgSnJ (B10.BR), C57BL/6J-B2 m^tm1Unc (B6-β2 m KO), C57BL/6J-Prkdc^scid (B6-SCID), and C57BL/6J-Rag1^{tm1 Mom} (B6-Rag1 KO) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and bred in our animal facility. B6-I-Aβ KO mice were obtained from Dr. Laurie Glimcher and kept in our facility.

Identification of potent allo-peptides.

A series of 15-mer overlapping peptides corresponding hypervariable region of α1 domains of H-2K^d and H-2K^k molecules were synthesized in a 3-amino acid step manner by Chiron Mimotopes, San Diego, CA (Fig. 1). These peptides were dissolved in the solution of 40% acetonitrile and 0.1 M HEPES (pH7.6). To screen for functional allo-peptides, C57BL/6 mice were stimulated in vivo with allogeneic irradiated spleen cells (3×107 B10.D2 or B10.BR spleen cells injected i.p.) and boosted with a mixture of synthetic peptides (approximately 0.4 μg/peptide/mouse) in adjuvant. As a primary screen, popliteal lymph node cells were stimulated in vitro with 20 μg/ml of each synthesized peptide in the presence of 3×105 of irradiated normal C57BL/6 spleen cells in 100 μl of 0.5% mouse serum medium (RPMI1640 medium supplemented with 10 mM HEPES, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 50 μM 2-ME, and 0.5% normal C57BL/6 mouse serum) in triplicate in round bottom 96-well plates. At day 4, cells were pulsed with 1 μCi [3H]-TdR (Amersham, Arlington Heights, IL) and harvested on glass fiber filters 16 hr later. The uptake of [3H]-TdR was measured in a liquid scintillation counter. In some experiments, supernatants from these cultures were harvested at day 4 and the concentration of interferon-γ (IFN-γ) measured by sandwich ELISA using commercially available reagents as described by the manufacturer (Pharmingen, San Diego, CA). A larger quantity of a selected peptide (H-2K^d54–68) was synthesized by the Comprehensive Cancer Center, UAB, Birmingham, AL, for further studies.

Figure 1: Amino acid sequences of synthetic 15-mer overlapping peptides used in this study. Amino acid differences in hypervariable regions of α1 domain of H-2K^b (C57BL/6), H-2K^d (B10.D2), and H-2K^k (B10.BR) molecules are shown. The amino acid sequences identical with H-2K^b sequence are indicated in dots. The synthetic H-2K^d and H-2K^k peptides used in this study are indicated with underlines in the amino acid sequences. Numbers assigned above sequence indicate the position of amino acid in H-2K molecule (⁶).

Surgical procedure.

The surgical technique to place heterotopic cardiac allografts was adapted from the technique of Ono and Lindsey (²). Briefly, the donor aorta was anastomosed in end-to-side fashion to the recipient abdominal aorta and the pulmonary artery was similarly anastomosed to the vena cava. The heart was palpated daily; no detectable beating was presumed to be rejection, which was confirmed by laparotomy (³). Skin grafts were performed using 1-cm2 segments of tail skin transplanted to the flank of the recipient and held in place by a plaster body cast for 6 days. Survival of skin grafts was checked by daily direct observation.

Establishing T cell lines/clones specific for H-2K^d 54–68 peptide.

C57BL/6 mice were immunized with 50 μg of peptide emulsified in complete Freund’s adjuvant (Sigma, St. Louis, MO). One week after immunization, lymphocytes from spleen or draining lymph nodes were stimulated with 20 μg/ml of peptide in 24-well-plates. Ten days later T cells were stimulated with 20 μg/ml of peptide in the presence of 6×106 of irradiated C57BL/6 spleen cells and 50 U/ml of exogenous recombinant mouse interleukin- (IL) 2. Afterward, 1×105 of T cells were stimulated with peptide/APC/IL-2 every 2 weeks in 24-well plates. To establish T cell clones from these T cell lines with a stable growth pattern, limiting dilution was performed. T cells were diluted and distributed at 0.3 cells/well in round-bottom 96-well plates and stimulated with 20 μg/ml of peptide in the presence of 2×105 of splenic adherent cells and 50 U/ml of exogenous IL-2. Seven to 10 days later, expanded T cells were stimulated with peptide/APC/IL-2 in 48-well plates. Afterward cloned T cells were maintained in the same way as established T cell lines.

Flow cytometric analysis.

T cell clones were stained with Fluorescein isothiocyanate-labeled anti-mouse CD3 (F500.A2, J. Allison, University of California, Berkeley, CA), Phycoerythrin-labeled anti-mouse CD4 (GK1.5, F. Fitch, University of Chicago, Chicago, IL) and biotinylated anti-mouse CD8 (53–6.72, ATCC). Streptavidin-Cy-Chrome (Pharmingen, San Diego, CA) was used for the detection of biotinylated antibody. To identify TCR Vβ expressed on T cell clones, cells were stained with the following mAbs against mouse TCR Vβ; anti-Vβ3 (KJ25; P. Marrack and J. Kappler, University of Colorado, Denver, CO), anti-Vβ5.1, 5.2 (MR9–4; O. Kanagawa), anti-Vβ6 (RR4–7; O. Kanagawa, Washington University, St. Louis, MO), anti-Vβ8.1,8.2,8.3 (F23.1; M. Bevan, University of Washington, Seattle, WA), Vβ8.2 (F23.2; M. Bevan), Vβ11 (RR3–15; O. Kanagawa), Vβ 8.1, 8.2 (KJ16–133; P. Marrack and J. Kappler) and Vβ8.3 (1B3.3; Pharmingen). After T cells were stained with these mAbs, cells were analyzed with a FACScan (Becton Dickinson, Mountain View, CA).

5′RACE.

When flow cytometric analysis failed to identify Vβ usage of T cell clones, cDNA encoding rearranged TCR Vβ chain was amplified with 5′RACE kit (Life Technologies, Rockville, MD). After extraction of total RNA using acid phenol and chloroform from 2–5×106 of T cells lysed in 0.5 ml of GITC solution (⁴), cDNA was synthesized using reverse transcriptase from Moloney murine leukemia virus (Superscript, Life Technologies) with Cβ-specific anti-sense primer (5′-TGCTCTCCTTGTAGGCCTGA-3′). Then rearranged TCR Vβ gene was amplified, following the manufacture’s 5′RACE protocol. Sequence of anti-sense primer used for amplification was 5′-GGACCTCCTTGCCATTCACC-3′. The 5′RACE products were cloned into pBluescript plasmid (Stratagene, La Jolla, CA). After identification of TCR Vβ usage by sequencing, T cells were stained with a specific mAb (anti-Vβ 13, MR12–3, Pharmingen) and analyzed with a FACScan to confirm the result of 5′RACE.

Allorecognition by T cell clones.

All cultures were done with resting T cell clones in total volume 100 μl of 10% FCS medium in triplicate in flat bottom 96-well plates. At day 3 cells were pulsed with 1 μCi [3H]-TdR for 16 hr and harvested at day 4. To study dose responses of established T cell clones to H-2K^d peptide, 1×104 of T cell clone cells were stimulated with a series of doses of H-2K^d 54–68 peptide in the presence of 8×105 of irradiated syngeneic C57BL/6 spleen cells. To test whether these clones respond to allo-antigens via the direct pathway, 1×104 of T cell clones were stimulated with 8×105 and 16×105 of irradiated spleen cells from B10.D2 or B10.BR mice without syngeneic APC. To test whether these clones recognize peptide derived in situ from H-2K^d allogeneic cells, 1×104 of T cell clones were stimulated with irradiated spleen cells from B10.D2 or B10.BR mice in the presence of 8×105 of syngeneic APC. To test MHC restriction of allo-recognition by T cell clones, spleen cells from C57BL/6, B6-I-Aβ KO and B10-β2 m KO mice were used as syngeneic APC.

Cytokine assay.

Fourteen days after stimulation of T cell clones with antigen, 1×105 of resting T cells were stimulated with 3×106 of irradiated C57BL/6 spleen cells in 500 μl of 10% FCS medium in 48-well plates to examine cytokine production patterns of selected T cell clones (IKB3/4 and IKB7/5). H-2K^d 54–68 peptide used for stimulation was 30 μg/ml for IKB3/4 and 3 μg/ml for IKB7/5. Supernatants were collected at 24 hr after Ag stimulation, aliquoted and kept at −80°C until ELISA was performed. The amount of IFN-γ, tumor necrosis factor- (TNF) α, IL-2, IL-4, IL-5, and IL-10 in supernatants were determined by sandwich ELISA using commercially available pairs of mAbs (Pharmingen, San Diego, CA).

Adoptive transfer of T cell clones.

The 2.5×106 of IKB3/4 and IKB7/5 T cell clones were stimulated with 30×106 of irradiated C57BL/6 spleen cells and H-2K^d 54–68 peptide in the presence of 50 U/ml of exogenous IL-2 in 5 ml of 10% fetal calf serum (FCS) medium in 6-well plates. Peptide doses used for cell stimulation were 30 μg/ml for IKB3/4 clone and 3 μg/ml for IKB7/5. Two days later cells were collected from wells and dead cells were removed using a gradient medium (Histopaque-1083, Sigma). Cells were washed three times in Hanks’ balanced salt solution (HBSS) and 5×106 of cells were intravenously injected to a B6-SCID or B6-Rag1 KO mouse grafted with a B10.D2 or B10.BR heart 2 weeks earlier. Graft survival was monitored by daily palpation.

Histology.

In selected cases animals were killed and the allograft and lymphoid tissues were snap-frozen in OCT compound (Miles, Elkhart, IN) and immunohistochemical analysis was performed as previously described (³). Briefly, staining with rat anti-mouse CD4 mAb (clone H129.19, PharMingen) was performed using indirect staining with anti-rat IgG biotinylated antibody (Vector Labs, Burlingame, CA) and the ABC-horseradish peroxidase reagent (Vector Labs) with 3,3′-diaminobenzidine (DAB, Sigma) as the chromagen. Eosinophils were stained by detecting the endogenous eosinophil peroxidase (EPO) activity using DAB as chromagen after slides were treated with 10 mM of KCN in phosphate-buffered saline (PBS) (pH 6.0) for 1 min (⁵).

RESULTS

Because the “indirect” pathway of alloantigen recognition is nothing more than the conventional recognition of an immunogenic peptide that happens to be present in an allograft, we reasoned that a dominant indirect allopeptide could be identified by screening for a strong T cell epitope among the peptides coded for by an alloantigenic MHC molecule. Several approaches were taken to screen for such a dominant peptide, including isolation of T cells isolated from mice primed by cardiac or skin allografts (data not shown). The approach that proved successful was to synthesize a series of overlapping peptides corresponding to allogeneic MHC molecules (⁶), directly immunize with alloantigenic spleen cells, boost the response in vivo with a mixture of these peptides, and then screen for in vitro T cell responses. One of the series of such peptides is shown in Figure 1, derived from the α1 domain of the mouse H-2K^d molecule. Figure 2 demonstrates the response of draining lymph node cells 10 days after peptide immunization stimulated with each individual peptide in the presence of self (C57BL/6; H-2K^b) APC. Peptides corresponding to H-2K^d amino acid position 54–68 and 57–71 induced significant T cell proliferation as well as IFN-γ production, although other peptides failed to stimulate a significant response. The H-2K^d_54–68 peptide (QEGPEYWEEQTQRAK) was selected for further study and a large scale synthesis of this peptide was conducted. This peptide differs from the sequence of the H-2K^b molecule in three residues (62R→E, 63E→Q, & 66K→R, see Fig. 1). To confirm that this peptide is a vigorous T cell immunogen, C57BL/6 mice were immunized in the footpads with 50 μg of H-2K^d_54–68 peptide in adjuvant. The draining lymph node cells from these immunized mice generated a vigorous proliferative response with clear peptide dose dependence in standard in vitro culture conditions (data not shown). These data imply the H-2K^d_54–68 peptide is able to bind in the I-A^b class II molecule and stimulate normal H-2^b T cells when immunized in the conventional fashion.

Figure 2: Screening of overlapping peptides by cell proliferation and IFN-γ production assay to identify stimulatory allopeptides. Draining lymph node cells from C57BL/6 mice primed by a combination of allogeneic spleen cell injection and mixed peptide immunization were stimulated with a panel of overlapping H-2K^d peptides in vitro. Cell proliferation and IFN-γ production were assayed. H-2K^d54–68 and H-2K^d57–71 peptides were identified as stimulatory peptides by both cell proliferation and IFN-γ production. Data of proliferation assay are expressed as mean±SEM.

The mere coincidence of a strong peptide immunogen within the sequence of an alloantigen does not demonstrate that this epitope is an important specificity in the rejection of transplanted tissues. To determine whether this epitope is relevant to allograft rejection, C57BL/6 mice were immunized with 50 μg of H-2K^d_54–68 peptide in adjuvant or a control immunization and then subsequently given an allograft from either a B10.D2 (H-2K^d) or B10.BR (H-2K^k) mouse. As shown in Figure 3, both full thickness tail skin grafts and vascularized heterotopic cardiac allografts from B10.D2 mice showed accelerated rejection compared to the control. The rejection of tissue derived from H-2^k mice (that fail to express the specific peptide), was not influenced by the presence of immune T cells of this specificity.

Figure 3: Survivals of cardiac and skin grafts in C57BL/6 mice primed with H-2K^d54–68 peptide. Eight days after immunization of C57BL/6 mice with 50 μg of H-2K^d54–68 peptide in CFA, B10.D2, or B10.BR grafts were transplanted in the mice and survival days were compared to those of control mice immunize with BSA in CFA. A, Survival of cardiac allografts. The survival of B10.D2 grafts in the peptide primed mice (n=8) was significantly shorter than in control mice (n=9) (P <0.001, using Mann-Whitney log rank sum test). However, immunization with H-2K^d54–68 peptide had no effect on the survivals of B10.BR grafts (n=10), compared to those in control mice (n=10). B, Survival of skin allografts. The survival of B10.D2 skin (n=5) in the peptide primed mice was significantly shorter than those in control mice (n=5) (P <0.001, using Mann-Whitney log rank sum test). There is no significant difference in survivals of B10.BR skin grafts between peptide immunized (n=5) and control (n=5) groups.

To study T cells specific for H-2K^d_54–68 peptide further, we immunized C57BL/6 mice with this peptide and established T cell clones specific for this peptide. These T cell clones were characterized for a number of criteria, including the pattern of TCR Vβ usage, the sensitivity to doses of the immunogenic peptide, their ability to recognize naturally processed H-2K^d_54–68 peptide, the pattern of cytokine expression after stimulation, and their ability to mediate allograft rejection. As summarized in Table 1, a total of 15 clones were established from 3 independent mice immunized with the H-2K^d_54–68 peptide, all with a CD4⁺CD8⁻ phenotype determined by flow cytometric analysis. The TCR Vβ usage of each of these clones was determined either by flow cytometric analysis with the available seven different anti-TCR Vβ-specific monoclonal antibodies (mAbs) or by the 5′RACE method using a Cβ constant region PCR primer. In this panel of 15 separate clones, a total of 4 distinct Vβ gene segments were used, although each clone is specific for the same H-2K^d_54–68 peptide. To assess the apparent avidity of these distinct T cell clones for the H-2K^d_54–68 peptide, the proliferative response of each clone to different doses of the peptide in the presence of irradiated syngeneic spleen cells was determined (Fig. 4). The clone designated IKB7/5 demonstrated 10 to 20 times higher sensitivity to the peptide than the other clones, suggesting that the TCR of this clone has an especially high affinity for the H-2K^d_54–68/I-A^b complex.

Table 1: Summary of established T cell clones specific for H-2K^d 54–68 peptide

Figure 4: Dose responses of T cell clones to H-2K^d54–68 peptide. Proliferative response of T cell clones stimulated with H-2K^d54–68 peptide. T cell clones specific for H-2K^d54–68 were stimulated with different doses of the peptide in the presence of syngeneic APC and proliferative response was measured by [3H]-TdR incorporation. In this figure, results of seven representative clones were shown, the remaining eight clones were also tested but showed similar patterns of peptide sensitivity. One clone, IKB7/5, showed the highest sensitivity to the peptide than the other clones. Compared to other clones, IKB7/5 demonstrated approximately 10- to 20-fold higher sensitivity. Data of proliferation assay are expressed as mean±SEM.

To further characterize the response to this allopeptide, we selected the apparently high avidity T cell clone IKB7/5 and a representative clone of apparently lower avidity but good in vitro growth characteristics and a distinct TCR Vβ usage (designated IKB3/4) for more intensive analysis. First these clones were stimulated with allogeneic B10.D2 or B10.BR spleen cells alone to test whether these clones respond to allogeneic cells directly. As shown in Figure 5A, neither clone shows a proliferative response without additional syngeneic APC. To formally demonstrate that these clones respond to both the synthetic H-2K^d_54–68 peptide and naturally processed alloantigen, these clones were stimulated in vitro with irradiated allogeneic cells in the presence syngeneic APC. Figure 6B demonstrates that both clones respond to a mixture of B10.D2 spleen cells and syngeneic C57BL/6 spleen cells (middle panels). Furthermore, by using APC from H-2^b mice with targeted gene deletion of either class I (B6-β2 m KO) or class II (B6-I-Aβ KO) MHC molecules, we demonstrate that the response of these two T cell clones is dependant on a source of H-2^d cells and I-A^b expressing APC. The response is stimulated by class I-deficient APC, but not class II deficient APC (Fig. 6B, top and middle panels). Interestingly, the clone with lower apparent avidity for H-2K^d_54–68/I-A^b (IKB3/4) showed a low level cross-reactive response to syngeneic APC in the presence of B10.BR stimulator cells, although the high avidity clone (IKB7/5) failed to show any detectable response to H-2^k-derived peptides (Fig. 6B, bottom panels). These data formally demonstrate that both of these clones respond to the H-2K^d_54–68/I-A^b complex using alloantigen derived from intact alloantigen expressing cells naturally processed through the MHC class II pathway by syngeneic antigen-presenting cells. The cytokine expression pattern of these two clones was determined after stimulation of the clones with H-2K^d_54–68 peptide and irradiated syngeneic spleen cells, by measuring cytokines in culture supernatants using commercially available ELISA mAbs (Table 2). The low affinity clone IKB3/4 produced large amounts of Th1-type cytokines, IFN-γ and TNF-α, although the high affinity clone IKB7/5 produced an atypical combination of cytokines - IL-2, IL-5, and TNF-α, but marginal amounts of IFN-γ, IL-4, and IL-10.

Figure 5: Specificity of T cell clones to alloantigens. A, Direct response of T cell clones to allogeneic and 3rd party antigens. The 1×104 of T cell clones were stimulated with irradiated allogeneic B10.D2 or 3rd party B10.BR spleen cells without syngeneic APC. Neither IKB3/4 clone (left) nor IKB7/5 clone (right) showed any proliferative response to the stimulator cells. B, Indirect response of T cell clones to allogeneic and 3rd party antigens, and MHC restriction of the response. 1×104 of T cell clones were cultured with H-2K^d 54–68 peptide (top), irradiated B10.D2 (middle), or B10.BR spleen cells (bottom) in the presence of APC form a C57BL/6, B6-β2 m KO or B6-I-Aβ KO mouse. Both IKB3/4 clone (left) and IKB7/5 clone (right) proliferated with peptide stimulation when spleen cells from normal C57BL/6 or B6-β₂m KO mice were used as APC but not responded at all when APC from B6-I-Aβ KO were used (top). When B10.D2 spleen cells were used as antigen source, T cell clones proliferated in a dose-dependent manner with APC from normal and B6-β2 m KO mice, but not B6-I-Aβ KO mice (middle). When B10.BR spleen were used as antigen source, IKB7/5 clone demonstrated no proliferative response at all with all kinds of APC, although IKB3/4 clone proliferated slightly with highest dose (8×105) of spleen cells with APC from normal and B6-β2 m KO mice, but not B6-I-Aβ KO mice (bottom). Data of proliferation assay are expressed as mean±SEM.

Figure 6: Survival of cardiac allografts in immunodeficient mice after adoptive transfer of T cell clones specific for H-2K^d54–68 peptide. The 5×106 of T cell clones stimulated with peptide/APCs for 2 days were adoptively transferred to immunodeficient (B6-SCID and B6-Rag1 KO) mice that had been grafted with B10.D2 or B10.BR hearts 2 weeks earlier. In this figure, survival data of allografts in B6-SCID and B6-Rag1 KO mice are combined because of no significant difference in the results. Both IKB3/4 and IKB7/5 were capable of mediating rapid rejection of allogeneic B10.D2 hearts (10/10 for IKB3/4 and 9/10 for IKB7/5 transfer) but not 3rd party B10.BR hearts (n=6 for IKB3/4 transfer and n=5 for IKB7/5 transfer) (P <0.01, using Mann-Whitney log rank sum test).

Table 2: Cytokine production by T cell clones specific for H-2Kd 54–68 peptide

To determine whether these T cell clones can mediate rejection of cardiac allografts in mice, we transplanted either B10.D2 or B10.BR hearts into immunodeficient hosts (B6-SCID or B6-Rag1 KO mice). After allowing the vascularized allograft to heal in the immunodeficient host for 2 weeks, we then adoptively transferred 5×106 of T cell clones 2 days after in vitro antigen stimulation (Fig. 6). Both T cell clones mediated the rapid rejection of the B10.D2 hearts in the immunodeficient hosts with equivalent rejection times (MST of 5.3 days for IKB3/4 and 5.7 days for IKB7/5), but neither clone caused the rejection of the B10.BR hearts. These data confirm that the specificity for H-2^d-derived peptide demonstrated in vitro also demonstrates the in vivo relevance of this response for allograft rejection. To assess the mechanism of rejection mediated by these two clones, we killed some B10.D2 heart transplanted mice 4 days after T cell clone transfer (just before complete rejection), and examined the histopathological pattern of rejection. Surprisingly, as shown in Figure 7, the density of T cell infiltrate was substantially different for these two clones that both cause allograft rejection with similar tempos: there was a massive clonal infiltrate in the allografts after transfer of the low avidity clone IKB3/4 (Fig. 7B), but only a sparse infiltrate associated with the high avidity clone IKB7/5 (Fig. 7A). In addition, there was a prominent eosinophilic infiltrate in allografts from the IKB7/5 clone transfer just before rejection (Fig. 7C), which may correlate with the production of IL-5 by this T cell clone.

Figure 7: Histological analysis of cardiac allografts in immunodeficient mice with adoptive transfer of T cell clones. Immunohistochemical staining of CD4⁺ T cells and histochemical staining of eosinophils were performed on 4 μm of cryosections of B10.D2 allografts from B6-Rag1 KO mice killed at day 4 after adoptive transfer of IKB3/4 or IKB7/5 T cell clone. A substantial number of CD4⁺ T cells infiltrated to the interstitial myocardium in the cardiac graft with adoptive transfer of IKB3/4 (A), although only a few CD4⁺ T cells were detected in the graft with adoptive transfer of IKB7/5 (B, arrows). Interestingly, IKB7/5 transfer induced a massive eosinophil infiltrate in the allograft (D) although only few eosinophils were seen in the allograft in a B6-Rag1 mouse with IKB3/4 transfer (C, arrow). Original magnification x200, conterstained with hematoxylin.

DISCUSSION

One of the fundamental difficulties for analysis of solid organ transplantation is the inherent complexity of this immune response. Not only are multiple different immune mechanisms involved in transplant rejection, but also the antigenic specificities are extremely complex. One element of this complexity are the differences between the direct recognition of MHC molecules on donor derived APC versus the normal pathway of T cell recognition in which peptides from the nominal antigen are processed and presented by syngeneic APC to the responding T cells. Because the small population of “professional” APC that are present in the donor organ normally turn over from a bone marrow stem cell source, the balance between these two pathways may not be constant over time. The possibility that the relative role of these two pathways may be important in the differences between early acute rejection episodes in the perioperative period and the long-term risk of subsequent immune responses that lead to “chronic rejection” has been suggested by several groups (^7–10).

Although the existence of the indirect pathway of alloantigen recognition is expected on theoretical grounds, the actual role for such responses during in vivo allograft rejection have been difficult to delineate. Several studies have shown that T cells specific for allopeptides presented by syngeneic APC could be identified in lymph nodes of animals that had previously rejected an allograft (^11–16). T cells specific for indirect allopeptides have also been demonstrated in conventional MLR cultures (^17,18) and in patients with chronic renal allograft dysfunction (^19–22) (^23–26). Direct immunization of rodents with selected allopeptides can generate vigorous T cell responses as determined by in vitro proliferation and in vivo DTH responses (²⁷) and transfer of DTH (^28,29). Others have demonstrated that immunization with selected allopeptides could result in accelerated rejection of a subsequent allograft expressing the same epitopes used for prior immunization (^30–34). One report demonstrates that T cell line specific for an indirect allopeptide can directly mediate transplant rejection when transferred to an immunodeficient host (³⁵). Perhaps the most potent evidence indicating a key role of indirect presentation was developed by Auchincloss and associates (¹), who showed that skin transplant rejection, via a CD4 T cell-dependant response, could proceed at a normal tempo in response to donor tissue genetically deficient in MHC class II antigens.

This study was motivated by the idea that identification of a dominant allogeneic peptide in the murine system would allow utilization of the wealth of genetically defined murine strains with targeted gene knockout and transgenes to study transplant rejection by such T cells. To identify such a peptide, we reasoned that an indirect allopeptide worthy of detailed study must be both immunogenic in the recipient strain and sufficiently abundant within an actual allograft to stimulate a vigorous immune response. Immunization with allogeneic spleen cells, followed by boosting with a mixture of synthetic peptides, allowed the identification of a strongly immunogenic peptide derived from the H-2K^d molecule. The epitope identified by this approach can stimulate a strong immune response that can result in allograft rejection as indicated by several distinct results: 1) Animals primed with this peptide reject either skin or cardiac allografts in an accelerated fashion compared with normal rejection of the same tissue in unprimed mice or compared with rejection of a 3rd party allograft that does not express the selected peptide. 2) T cell clones specific for this peptide can directly mediate cardiac allograft rejection in immunodeficient hosts (both SCID and Rag 1 KO mice). The identification of this strong epitope, however, does not indicate that this peptide is the dominant epitope in normal first set allograft rejection in this strain combination. The heterogeneity of the antigenic specificities that drive allograft rejection is underlined by the diversity of TCR Vβ gene segments used by T cells that bind to a single peptide epitope, the range of TCR avidity to a single epitope, the diversity of cytokine gene expression pattern of two clones isolated in parallel that are specific for the same epitope, and the distinct morphological features mediated by two clones specific for the same epitope.

Although the two clones studied by adoptive transfer were selected on the basis of divergent peptide dose sensitivity as a marker of TCR avidity, the fact that they also showed divergent cytokine expression patterns complicates the analysis of the mechanism of rejection mediated by these two clones. These results illustrate that the cytokine expression phenotype of “acute rejection” may not be a discrete entity, but indicates that several different immune mechanisms may result in sufficient tissue injury that overall organ function decompensates resulting in “rejection.” Whether the distinct immunological circumstances of an immune response which develops in the presence of normally immunosuppressive drugs, or focuses primarily on vascular endothelium as in “chronic rejection” represent consistent immunological mechanisms or merely final common pathways of organ injury is unclear. Although the conventional Th1-like cytokine expression phenotype of the IKB3/4 clone is similar to the pattern of cytokines expressed in normal first set allograft rejection (^{3, 36–41}), the density of the T cell infiltrate is almost twice the total CD3+ cell density in this response. Conversely, the low density infiltrate present just prior to complete rejection of the allograft with the IKB7/5 clone is lower than in normal rejection, but perhaps more similar to the presumed density of alloantigen specific cells in normal rejection. Because IL-5 is associated with eosinophil activation in other circumstances (^42,43), eosinophils may contribute to cardiac damage to some extent. Indeed a recent study (⁴⁴) has shown that IL-5 and eosinophils are a key mediator of rejection of class II-disparate skin grafts in mice.

The difficulty of distinguishing the role of TCR avidity from that of cytokine expression pattern using isolated T cell clones points to a general problem with analysis of T cell functional differentiation using panels of T cell clones. Although the fine specificity of TCR recognition characteristics can be elegantly dissected using T cell clones, the in vivo functional behavior of isolated antigen specific cells using panels of such T cell clones is complicated. Most clones show aberrant patterns of adhesion molecule expression and tissue migration patterns compared to normal T cells and the quantitative characteristics of cytokine expression are often unstable after multiple passages in vitro. More critical for the current purpose is the lack of plasticity of the cytokine expression phenotype to differentiation signals of established T cell clones. One approach to circumvent this problem is the production of TCR transgenic mice using TCRs from alloreactive T cell clones that can mediate transplant rejection, so that the characteristics of particular TCR specificities can be analyzed independently of the cytokine expression phenotype. Because naïve TCR transgenic cells can be activated under selective conditions in vitro, the same TCR can be used to produce T cells with disparate cytokine expression patterns and the resultant activity of these selected cells can be examined in an adoptive transfer transplantation model such as that described in this report. The rearranged genomic TCR-α and TCR-β from both the IKB7/5 and IKB3/4 T cell clones with significantly different apparent TCR avidity have been isolated for construction of such transgenic mice.

Acknowledgments.

The authors thank Jimin Li for technical assistance and Drs. Casey Weaver, Mark Benfield, and Bratin Saha for critical review of the manuscript.