Intragraft Selection of the T Cell Receptor Repertoire by Class I MHC Sequences in Tolerant Recipients

Background Allograft tolerance of ACI (RT1a) recipients to WF (RT1u) hearts can be induced by allochimeric class I MHC molecules containing donor-type (RT1Au) immunogenic epitopes displayed on recipient-type (RT1Aa) sequences. Here, we sought the mechanisms by which allochimeric sequences may affect responding T cells through T cell receptor (TCA) repertoire restriction. Methodology/Principal Findings The soluble [α1h u]-RT1.Aa allochimeric molecule was delivered into ACI recipients of WF hearts in the presence of sub-therapeutic dose of cyclosporine (CsA). The TCR Vβ spectrotyping of the splenocytes and cardiac allografts showed that the Vβ gene families were differentially expressed within the TCR repertoire in allochimeric- or high-dose CsA-treated tolerant recipients at day +5 and +7 of post-transplantation. However, at day 30 of post-transplantation the allochimeric molecule-treated rats showed the restriction of TCR repertoire with altered dominant size peaks representing preferential clonal expansion of Vβ7, Vβ11, Vβ13, Vβ 14, and Vβ15 genes. Moreover, we found a positive correlation between the alteration of Vβ profile, restriction of TCR repertoire, and the establishment of allograft tolerance. Conclusions Our findings indicate that presentation of allochimeric MHC class I sequences that partially mimic donor and recipient epitopes may induce unique tolerant state by selecting alloresponsive Vβ genes.


Introduction
The possible mechanisms underlying the development of transplantation tolerance include clonal deletion of alloreactive cells, clonal anergy, cell-mediated suppression, and ''infectious'' tolerance [1]. Studies on the mechanisms responsible for the establishment of transplantation tolerance point to the immune regulation as the major determinant. It was shown previously that the CD4+ T lymphocytes from tolerate hosts inhibit rejection of donor-specific allografts after adoptive transfer into test rat recipients [2] and the examination of tolerant cardiac allografts in both early and late post-transplantation stages indicated that in contrast to the rejecting grafts they were extensively infiltrated by the T cells [3,4]. Following allotransplantation, the T cells can recognize graft MHC antigens in two different ways: via ''direct'' recognition of antigens on donor cells and via ''indirect'' recognition of processed donor antigens presented on recipients APCs in the form of peptides combined with MHC molecules [5]. Numerous studies on rodents and humans demonstrate that the indirect recognition of MHC plays a major role in both acute and chronic allograft rejection [6][7][8].
In response to alloreaction, T cells react to allogeneic MHC molecules by displaying a dominant determinant on donor MHC antigens and are restricted by self-MHC elements [9][10][11]. It was shown previously that the alteration of critical amino acid residues in such immunodominant determinants inhibited T cell proliferative responses, induced T cell anergy and altered host immune responses toward native antigens [12][13][14][15]. Furthermore, analysis of T cell responses to nominal protein or transplantation antigens has shown that constraints in antigen processing and presentation usually limit the T cell responses to one or two dominant determinants [3], and that, in humans, T cells responding to dominant allopeptides had limited TCR-Vb gene usage [10,11]. Thus, indirect allorecognition by T cell possibly regulates specific immunointervention [16][17][18][19] through manipulation of such dominant immunogenic epitopes.
Previously we have mapped an immunogenic determinant for rat class I RT1 u , RT1 l and RT1 a alloantigens to amino acid residues 58-80 on a 1 -helical polymorphic region [20,21]. MHC alloantigens are therefore similar to multideterminant protein antigens that bear localized immunodominant amino acid sequences that trigger immune responses toward the whole protein molecule [22][23][24][25]. Therefore, monitoring the profile of alteration of TCR repertoire will provide a proof of indirect recognition of MHC molecules through TCR restriction.
Alterations in the length distribution of complementarydetermining region 3 (CDR3) 3 of TCR Vb have been observed in vitro and in vivo during allorecognition [5]. Indirect [a 1h u ]-RT1.A a allochimeric recognition may induce functionally unique regulatory T cells with distinctive TCR allospecificities. To test this hypothesis, we have performed a comprehensive analysis of TCR Vb gene usage in parallel with CDR3 spectrotyping of T cells. The CDR3 length distribution (CDR3-LD) measurement of the different Vb genes demonstrated the antigenic diversity recognized by a T cell population. Our experiments provide direct evidence of the ability of indirectly presented allochimeric sequences to select the TCR repertoire and thus provide a powerful tool for manipulating the immune response through the sequences of the indirectly presented antigen. The characteristics of the TCR Vb expression pattern may potentially be used as a novel marker to identify operational regulatory T cells in the recipient of organ allograft.

Induction of Tolerance by [a 1h
u ]-RT1.A a to WF (RT1.A u ) Allograft in ACI (RT1.A a ) Recipient We had constructed an immunogenic [a 1h u ]-RT1.A a molecule by altering hypervariable a 1 -helical region (a.a. 51-90) of ACI (RT1.A a ) to WF (RT1.A u ) sequences [20,26]. The resulted [a 1h u ]-RT1.A a molecule bears both RT1.A u and RT1.A a a 1 -helical epitopes. We have previously demonstrated [20,26] that peritransplant portal venous (p.v.) administration of [a 1h u ]-RT1.A a allochimeric molecules induced tolerance to WF (RT1 u ) heart allografts in ACI (RT1 a ) recipients, when administered in conjunction with sub-therapeutic dose of CsA [31]. This indicates that peri-transplant administration of syngeneic class I sequences flanking allogeneic a 1 -helical immunogenic epitopes, in combination with a brief sub-therapeutic dose of CsA, is able to induce donor-specific transplantation tolerance.

TCR Responses in Long-Term Tolerant Allochimeric Conditioned Recipients
CDR3 spectrotyping shows three to seven visible bands of different size for each Vb gene product, which intensities follow Gaussian distribution. Twenty two samples of TCR Vb profile plots (derived from DNA Analyzer and Genotyper 3.7 software) from CsA-treated long-term tolerant ACI hosts bearing allochimeric [a 1h u ]-RT1.A a class I allografts are shown in Fig. 1A. The TCR Vb profile plots from the allografts of CsA-induced longterm tolerant recipients were similar to the profile plots from the recipients' spleen cells (Fig. 1B). In several profile plots (e.g., Vb3, 8.1, 13, 14, and 19) we observed minor alterations in pick distribution patterns between allograft and spleen cells. Because these alterations were inconsistent between experiments we believe that they represent data output ''noise'' inherent to the applied method of analysis. In summary, it seems that a broad immunosuppressive agent such as CsA is unable to induce a clonal restriction of TCR.
To confirm that restrictions of TCR repertoire (i.e. the oligoclonal expansion of T cells) were unique to the graft site rather than being a general feature of T cells of tolerant recipients, we also performed spectrotyping of the host splenocytes. Figure 2B shows that, indeed, these unique dominant peaks were detected exclusively in heart allografts, and were absent in splenocytes from the same tolerant host. Thus, all restrictions of TCR repertoire observed in our studies were allograft-specific.

Frequency of Oligoclonality of T Cells in Tolerated Grafts
The algorithm for calculating the percentage of alterations of expression profile gene and percentage of expression frequency for each Vb gene were performed as described in the legend to Fig. 3A. We found that the expression frequencies of individual Vb genes varied in a series of twelve grafts from allochimeric proteintreated rats (Fig. 3A). On the other hand, the Pearson correlation analysis indicated that there was no significant difference between the expression frequency of various Vbs in CsA-treated and allochimeric-protein treated rats (Correlation coefficient = 20.00938; P = 0.9669).
Although all twenty two Vb gene products were readily detectable in each individual graft from both CsA and allochimeric protein-treated tolerant rats, the overall Vb usage was restricted only in allochimeric protein-treated rats. As summarized in Figure 3B, the restriction frequencies of individual Vb genes varied in grafts from allochimeric protein-treated tolerant rats: the Vb4, 8.3, 10, 12, 16 were slightly restricted in most of the transplants and Vb7, 11, 13, 14, 15 displayed dominant restricted bands, while Vb1, 3.3, 8.2, 9, 19 were expressed much less frequently ( Fig. 3A and 3B). Interestingly, there was no correlation between the frequency of expression of particular Vb genes and their restriction level (Fig. 3A). This indicates that only selective Vb gene families underwent significant clonal expansion and that the restriction of Vb genes was not related to their expression level.
Interestingly, CDR3 spectrotyping of Vb gene products showed remarkable differences between allochimeric-protein-and CsAtreated rats. Two-sample t-test analysis of the differences of the detection frequency of Vb restricted in total TCR repertoire between CsA and allochimeric protein-induced tolerant rats showed 6% restriction frequency in CsA-treated tolerant allograft versus 36% in allochimeric protein -treated allograft (Fig. 3B). This difference was statistically significant (t = 6.2; P,0.0001).

Dynamic Alteration of T Cell Receptor Repertoire in CsA and Allochimeric Conditioned Long-Term Tolerant Hosts
To analyze the restriction pattern of TCR repertoire, the RNA was harvested at the different post-transplantation time points, at day 5, 7, 30, and .100 from CsA-and allochimeric proteintreated tolerant rats and then subjected to CDR3 spectrotyping (Fig. 4). The restriction of TCR repertoire in allochimeric protein conditioned tolerant recipients did not occur before 30 days of post-transplantation, and dominant CDR3 bands in Vb7, 11, 13, 14, 15 gene products were found more frequently at 100 days posttransplantation (Fig. 4B). In contrast, nearly all Vb gene products were unrestricted throughout the whole post-transplantation period in CsA induced long-term tolerant hosts (Fig. 4A).

Discussion
Our previous data suggested that tolerance might be induced by functional presentation of donor immunogenic epitopes via selfsequences [20,26]. Since processing of allochimeric molecules may alter the repertoire of alloreactive T cells, we applied CDR3 spectrotyping of TCR repertoire to analyze the clonal expansion of T cells in allografted hearts of tolerant recipients. Spectrotyping has been successfully used to analyze T cell responses to some wellcharacterized peptide antigens in vivo [31][32][33]. In current studies, we first determined whether there is any preferential accumulation of T cells with limited Vb expression at the graft site. As the CDR3 spectrotyping can detect T cell clones against a background of polyclonal cells at a frequency of at least 1 in 1000 [34][35][36], we then used this novel technique to identify clonally expanded T cells by comparing CDR3 patterns in CsA and allochimeric proteintreated rats. We found that most of Vb genes were expressed at the similar levels in long-term tolerant rats induced by CsA and allochimeric treatment and clonal expansion occurred only in a limited number of Vb genes and was consistently observed in Vb7, 11, 13, 14, 15. In contrast, no restricted clonal expansion was found in either spleencytes or CsA treated tolerant. The restricted clonal expansion observed in our study may indicate the presence of suppressed alloreactive T cells or preferential activation of putative regulatory T cells. Based on our previous findings, we argue that these cells are the regulatory cells [30,38,40]. Functional relevance of the oligoclonal T cell expansion can be addressed by using Vb-specific mAb to selectively target specific Vb gene products [41]. However, at present, this method is limited because only few anti-Vb antibodies are available. An alternative method to test the functional relevance of T cell oligoclonality is the newly developed technique of TCR-specific DNA vaccination that inactivates targeted T cell clone(s) [42]. Our current study is the first to document a unique pattern of CDR3 spectrotyping in the tolerance pathway induced by allochimeric protein. Thus, our experimental design will allow us to further address the mechanism through which the indirectly presented allochimeric sequence could shape the responding TCR repertoire, which has many important implications for tolerance induction and inhibition of chronic rejection. Identification of the specific T cells involved in these processes may also provide an important tool to design intervening reagents in clinical treatment.
The mechanism of the recognition of allochimeric MHC molecules and T cells remains unclear. It was hypothesized that T cells selected by self-MHC molecules in the thymus could proliferate and mutate their receptors to allow them preferentially to recognize ''altered-self'' MHC molecules [39,43,44]. A restricted oligoclonal rather than a single clonal expansion of the Vb TCR repertoire observed in our study indicates a pattern of hierarchical immunodominance or a cross-reactive response to cryptic self epitopes when regulatory cells are generated in response to self-peptide antigens. This issue can be addressed by dissecting the TCR repertoire using a second allochimeric molecule which shares similar self-RT1.A a sequences to allochimeric molecule [a 1h u ]-RT1.A a , but differs in the substituted immunogenic RT1.A u epitope. Comparison of the TCR repertoire in tolerant hosts following allochimeric treatment with either molecule may identify shared CDR3 expansions indicative of regulatory cell clonality induced by shared self-sequences. Our CDR3 spectrotyping analysis of the TCR Vb genes in allochimeric tolerant rats demonstrated predominant and consistent clonal expansion of the Vb7, 11, 13, 14 and 15 genes. However, our data showed that the alteration Vbs in allochimeric proteintreated tolerant rats would not occur until after 30 days posttransplantation. This intriguing finding may indicate the estab- lishment of a relatively delayed tolerant state. This phenomenon may be explained by a sequential activation of T cells over time with subsequent spreading of the response [37].
Though it has been demonstrated that maintenance of regulatory cells is dependant on the continuous supply of antigenic stimulation by the allograft [45], it has not been possible to differentiate their potential functional diversity in vivo because in most studies, tolerance induction and allograft survival were the endpoints of the analysis [46]. In our studies, we utilized two primary therapies-CsA and allochimeric molecules to achieve distinctive tolerant states, which in turn elicited different responses of TCR repertoire. Such unique TCR restriction model will allow us to examine the functional characteristics and allospecific specificities of regulatory cells based on the inducing therapeutic agent.

Site-Directed Mutagenesis and Production of Mutant Class I MHC Molecules
Mutagenesis primers (purified by PAGE) were obtained from Invitrogen (Carlsbad, CA). The ''QuickChange Multi Site-Directed Mutagenesis Kit'' (Stratagene, San Diego, CA) was used to mutate ACI-RT1 a cDNA as described previously [20,26] and resulting plasmids were sequenced to confirm the presence of the mutations.

Production of Allochimeric Protein Containing Mutant Class I MHC Molecules and Allograft Model
The allochimeric proteins containing [a 1h u ]-RT1.A a sequence were expressed in transfected Buffalo hepatoma cells as described previously [26]. For tolerance induction, allochimeric proteins were administered through the portal vein (1 mg/rat) into ACI recipients of WF hearts at the time of transplantation followed by a 3-day course of oral cyclosporine delivered by gavage feed (CsA, 10 mg/kg/day; day 0-2). Controls included transplantation of allogeneic hearts in the presence or absence of the same dose of CsA. Heterotopic cardiac transplants were placed intra-abdominally [27].

TCR Vb RT-PCR
RNA was extracted from splenocytes and allografted cardiac hearts at post-transplantation day 5, 7, 10, 30, 84, 120 and .120 using the Qiagen RNeasy Mini Kit (QIAGEN, Valencia, CA). In order to remove DNA contamination the RNA samples were digested with DNase I (RNase-free DNase Set, Cat. 79254, QIAGEN) and 25 mg of RNA was reverse transcribed into cDNA RNAs from heart allografts were harvested at different time points after transplantation followed by TCR detection. There were .7 samples for each time-point.(B) (A) There was no TCR restriction in the hearts of CsA treated tolerant recipients throughout the whole period of long-term tolerance (.day 120). (B) There was no TCR restriction in the hearts of allochimeric protein-treated tolerant recipients until 30 days post-transplantation. However, the Vb7, 11, 13, 14 and15 were highly restricted and occurred as the dominant peaks 84 days after transplantation. This restriction was maintained throughout the whole period of long-term tolerance (.day 120), doi:10.1371/journal.pone.0006076.g004 (50 ml) by 500 U of SuperScript III reverse transcriptase and SuperScript First-Strand Synthesis System from RT-PCR kit (Invitrogen, Carlsbad, CA). Rat Vb-specific primers were chosen according to Shirwan et al. [28] with minor modifications [29]. A total of 2 ml of cDNA was amplified using a fixed down-stream primer (Cb1) derived from TCR b chain constant region, and one of the 22 different Vb-specific upstream primer. Amplification reactions were conducted in a GeneAmp PCR System 9700. The reactions started with a 10 min denaturation step at 94uC, followed by 35 cycles of 30 sec at 94uC, 30 sec at 55uC, and 1 min at 72uC, and ended with an elongation step of 7 min at 72uC.

Spectrotyping of TCR Vb CDR3
A 6-FAM-labeled internal primer within the Vb PCR products (Cb2 derived from b chain constant region) was synthesized by Applied Biosystems (Foster, CA). 7.5 ml of each Vb-specific PCR product was subjected to 15 cycles of run-off reactions in a final volume of 10 ml, consisting of 0.25 rM of fluorochrome (FAM)labeled internal primer Cb2, 0.2 mM of each dNTP, 2 mM MgCl 2 , and 0.2 U AmpliTaq Gold (Perkin Elmer, Foster City, CA) in 16 PCR Gold buffer. Amplification was conducted in a GeneAmp PCR System 9700 (Perkin Elmer), with cycle conditions as follows: denaturation at 95uC for 10 min, 15 cycles of 30 sec at 95uC, 30 sec at 58uC, and 30 sec at 72uC, and elongation at 72uC for 5 min. At the end of the reaction, 5 ml of each spectrotyping reaction was mixed with 10 ml of GeneScan LIZ 500 Size Standard (Applied Biosystems, Foster, CA) in 96-well plate. After the denaturation at 94uC for 5 min followed by incubation on ice for 5 min, the 96-well plate was analyzed using 3700 DNA Analyzer (Applied Biosystems, Foster, CA).

Processing of Gel Images
The initial images from DNA Analyzer system were converted into profile plots and the Vb CDR3 region profiles were visualized using ABI Prism Genotyper 3.7 software. The lanes of each Vb gene product were selected for profile plot. The plots were then copied to Microsoft Powerpoint files. Two-sample t-test was applied to detect the differences of restriction frequencies of TCR repertoires in CsA and allochimeric protein treated rat tolerant models. Pearson correlation coefficient analysis was performed to determine whether the expression frequency of TCR repertoire correlated with its restriction frequency in allochimeric moleculesinduced tolerant recipients. A P value of ,0.05 was used to indicate significance.