Circulating Alloreactive T Cells Correlate with Graft Function in Longstanding Renal Transplant Recipients

Abstract

Monitoring for alloreactive memory T cells after organ transplantation may allow individualization of immunosuppression. Two pathways of T cell allorecognition have been implicated in chronic graft dysfunction: Direct (recipient T cells respond to donor peptides presented by donor antigen-presenting cells) and indirect (donor peptides are processed and presented by recipient antigen-presenting cells). Previous studies have assessed these alloresponses only during the first 2 yr after kidney transplantation, so this study correlated the presence of circulating donor-reactive memory/effector T cells, primed by both pathways, in 34 longstanding living-donor renal transplant recipients using the highly sensitive IFN-γ Elispot assay. Remarkably, 59% of patients had directly primed donor-reactive T cells, and their presence correlated directly with serum creatinine (P = 0.001) and inversely with estimated GFR (P = 0.042). Multivariate analysis revealed that hyporesponsiveness of direct, donor-specific T cells was the only variable that significantly correlated with graft function and that antidonor indirect alloreactivity was the only variable that significantly correlated with proteinuria. Interestingly, when both allorecognition pathways were considered together, patients with undetectable direct alloreactivity had better long-term graft function, independent of allosensitization by the indirect pathway. In conclusion, circulating donor-specific alloreactive T cells primed by both pathways are detectable long after transplantation and are associated with graft injury. Assessment of alloreactive memory/effector T cells might be helpful to tailor individual immunosuppression regimens for transplant recipients in the future. The development of interstitial fibrosis and tubular atrophy1 is the main cause of chronic allograft dysfunction and therefore of graft loss after renal transplantation.2,3 Besides nonimmunologic factors, both cellular and humoral alloimmunity contribute to the pathogenesis of chronic allograft damage.4–9 Thus, monitoring donor-specific alloimmune responses at different time points after transplantation is an important challenge to recognize the immunologic state of patients who undergo transplantation. T cells are the key initiators and mediators of alloimmune responses. Alloreactive T cells recognize HLA-mismatched tissue via two different pathways: In the direct pathway, responder T cells recognize intact foreign MHC-peptide complexes on the surface of donor antigen-presenting cells (APC).10–14 In the indirect pathway, T cells recognize donor allopeptides on self-MHC molecules after having been processed and presented by recipient APC.15–17 During the past few years, the relevance of both pathways of antigen allorecognition for long-term graft outcome has controversially been discussed. Priming by the direct pathway has classically been associated with the early posttransplantation period and especially with acute rejection, because professional donor APC are present during the first months only.15,18,19 In contrast, priming by the indirect pathway was suggested to play a main role in the long term; therefore, indirectly primed T cells were considered key mediators for chronic immune-mediated graft injury.20–23 However, recent studies suggest that both pathways may persist and be of relevance for interstitial fibrosis and tubular atrophy. Notably, Herrera et al.24 demonstrated how recipient dendritic cells (DC), when co-cultured with allogeneic DC or endothelial cells (EC), can acquire substantial levels of allogeneic MHC-peptide complexes and subsequently prime T cells by both allorecognition pathways. Hence, early directly primed T cells could be maintained and reactivated by graft cell populations such as graft EC but also by recipient DC at later time points. Previous studies have simultaneously investigated direct and indirect naive and memory T cell responses during the first 2 yr after transplantation only. We wondered whether both pathways would be detectable long after transplantation and correlate with chronic graft dysfunction; therefore, frequencies of IFN-γ–producing donor-reactive T cells primed by both allorecognition pathways were investigated in a group of 34 longstanding living-donor kidney transplant patients. The IFN-γ Elispot assay was used as a highly sensitive functional tool assessing only primed memory/effector T cells. Importantly and different from previous experiences, cytoplasmic membrane protein preparations of donor allogeneic cells were used for Elispot stimulation in the indirect pathway to have the full allogeneic HLA and non-HLA epitope repertoire. RESULTS Clinical Data Patient demographic characteristics and most relevant clinical data are shown in Tables 1 and 2, respectively. In these 34 longstanding (56.2 ± 23.6 mo) living-donor renal transplant recipients, we found that patients who had experienced acute rejection (n = 16) had significantly worse serum creatinine than nonrejectors (n = 18; 2.1 ± 1 versus 1.5 ± 0.4 ml/min; P = 0.007). Patients with late acute rejection episodes (n = 12), depicted as later than month 3 after transplantation, showed significantly worse serum creatinine (2.2 ± 2.2 versus 1.5 ± 0.5 ml/min; P = 0.019) compared with the rest (n = 22). Patients with more than one acute rejection event (n = 10) had a trend to have worse serum creatinine (2.2 ± 1.2 versus 1.6 ± 0.5; P = 0.051) than the rest (n = 24). No relation was observed by univariate analyses between renal function (both serum creatinine and GFR by Modification of Diet in Renal Disease [MDRD]) and number of donor/recipient HLA mismatches, different maintenance immunosuppressive therapy, donor and recipient age, gender, proteinuria, and time after transplantation. Likewise, no clinical variable was associated with presence of proteinuria (data not shown). Elispot Monitoring All 34 patients could successfully be tested by the direct pathway and 33 by the indirect pathway. In all cases, cell viability and test performance were proved by the high number of spots shown after phytohemagglutinin stimulation. A positive direct or indirect alloresponse was defined as >20 spots/3 × 105 stimulated peripheral blood mononuclear cells (PBMC). Twenty (58.8%) of 34 patients had detectable direct Elispot frequencies. A wide range of positive direct responses was observed, ranging from 25 to 380 spots/3 × 105 stimulated PBMC. Regarding the indirect pathway, 20 (60%) of 33 patients had positive Elispot responses that ranged from 20 to 80 spots/3 × 105 stimulated PBMC. Direct pathway Elispot frequencies were approximately three times higher than indirect pathway frequencies (Figure 1). Despite that 14 patients had positive response in both pathways and eight showed no donor-specific alloreactivity by any pathway, no relation was observed between both allorecognition pathways in the univariate analyses (P = 0.28). Direct Pathway of Antigen Allorecognition Twenty (58.8%) patients had a positive donor-specific direct pathway Elispot. Directly primed alloreactive T cell frequencies were positively correlated with serum creatinine and inversely with the GFR-MDRD (Figure 2A); however, no correlation was found with presence of proteinuria (data not shown). Although no relationship was found between direct alloreactivity and history of acute rejection in general, patients with late acute rejection events (n = 12) had significantly higher direct frequencies than the others (n = 22; Figure 2B). Likewise, patients with more than one acute rejection episode (n = 10) had significantly higher direct frequencies (117.5 ± 127.2 versus 37.8 ± 71.2; P = 0.026). Fourteen (41.2%) of 34 patients were donor-specific hyporesponders and had significantly better graft function than non–donor-specific hyporesponder patients (Figure 3). Interestingly, the higher the number of donor/recipient HLA mismatches, the higher donor-specific IFN-γ frequencies were detected (Figure 4). In addition, direct alloreactivity was also positively correlated with positive frequencies against a third party (r = 0.551, P = 0.001). In contrast to antidonor alloresponses, the third-party alloresponses did not correlate with graft function. No association was observed between direct alloresponses and time after transplantation (data not shown). Indirect Pathway of Antigen Allorecognition Twenty (60%) of 33 tested patients had a positive donor-specific indirect pathway Elispot. No correlation was found between positive indirect alloresponses and either serum creatinine or GFR-MDRD. Also, no association was observed between this pathway and any history of acute rejection. Conversely, presence of proteinuria was significantly associated with high donor-specific indirect alloreactivity (Figure 5A). Moreover, the indirect pathway Elispot positively correlated with time after transplantation; that is, the longer the patient had had a transplant, the higher donor-specific alloresponse by this pathway was detected (Figure 5B). In line with the results of the direct pathway, frequencies of donor-specific indirect alloresponses positively correlated with frequencies against third party (r = 0.585, P = 0.001). Nevertheless, third-party responses did not correlate with either proteinuria or graft function. Global Cellular Alloreactivity Fourteen (42%) of 33 patients had positive responses of donor-specific alloreactive memory/effector T cells in both pathways. Eight of (24.2%) 33 had negative antidonor cellular alloreactivity in both pathways. Five (15.1%) patients had only positive direct antidonor alloresponses, and six (18.1%) other patients had only positive indirect antidonor alloreactivity (Figure 6). Interestingly, patients with positive alloreactivity by both pathways (DP+/IP+) had significantly worse graft function than those with undetectable frequencies by both pathways (DP−/IP−) and than those with only indirect frequencies (DP−/IP+) (2.1 ± 1.0 versus 1.1 ± 0.3 [P = 0.004] and 2.1 ± 1.0 versus 1.3 ± 0.3 [P = 0.023], respectively); however, patients with positive frequencies by the direct pathway only (DP+/IP−) did not differ from double-positive patients (DP+/IP+; 2.2 ± 0.4 versus 2.1 ± 1.0, respectively) and, likewise, had also significantly worse serum creatinine compared with DP−/IP+ and DP−/IP− (2.2 ± 0.4 versus 1.2 ± 0.3 [P = 0.005] and 2.2 ± 0.4 versus 1.3 ± 0.3 [P = 0.011], respectively). These data suggest that only the direct pathway is related to creatinine and GFR-MDRD. No relation could be observed regarding the presence of proteinuria and total cellular alloreactivity (data not shown). Circulating Donor-Specific Alloantibodies and T Cell Alloreactivity Only one of 33 screened patients had circulating donor-specific alloantibodies (DSA) after transplantation. This particular patient had a positive indirect but a negative direct pathway Elispot. Multivariate Analyses We studied in a multivariate analysis variables that were associated with renal function and proteinuria in the univariate test, as well as other relevant variables influencing outcome, such as acute rejection, time after transplantation, and HLA mismatch (Table 3). No patient experienced delayed graft function; therefore, it was not included in the analysis. As shown in Table 3, only donor-specific hyporesponsiveness was a significant correlate of serum creatinine, although acute rejection reached P = 0.08. Regarding proteinuria, the antidonor indirect pathway Elispot was the only significant correlate of all variables analyzed. DISCUSSION Progressive renal allograft dysfunction, depicted as a continuum between graft inflammation and fibrosis, appears over time.25 Despite that it has been claimed that immunologic and nonimmunologic factors are engaged in the development of chronic allograft damage, the contribution of the donor-specific alloimmune response is still not routinely assessed nowadays. Thus, searching for biomarkers of alloimmunity could be useful, even late after transplantation, to identify patients with progressive immune-mediated graft injury. Recent studies have questioned whether monitoring the cellular alloresponse during the late posttransplantation period should be based only on indirect alloreactivity or direct alloreactivity should also be assessed.24,26; therefore, we simultaneously measured the donor-specific cellular alloresponse by both direct and indirect pathways of allorecognition as well as the presence of circulating DSA in a group of longstanding living-donor kidney transplant patients. In contrast to a previous study that simultaneously determined direct and indirect naive and memory T cell responses in kidney transplant patients with chronic allograft nephropathy (CAN),21 here we used the IFN-γ Elispot assay to assess only primed effector/memory T cells. Also and different from previous Elispot studies evaluating the indirect pathway,26,27 we used allogeneic fragmented donor cells rather than allogeneic synthetic peptides to cover a broad spectrum of epitopes. The specificity of this method for analyzing indirect alloresponses was validated by using T cell lines for direct/indirect pathway. The advantage of this method compared with the one based on donor HLA-mismatch peptides is that it might reproduce more reliably what is happening in vivo, because the complete donor epitope repertoire including minor mismatches is present. The direct pathway IFN-γ–Elispot frequencies measured in this study (on average 61 spots/300,000 PBMC) were considerably higher compared with previously published frequencies28,29 measured during the first 6 mo after transplantation (on average 20 to 30 spots/300,000 PBMC) as well as those observed by Poggio et al.27 during a median follow-up time of 24 mo after transplantation (33 in patients with CAN and 9 spots/300,000 PBMC in control subjects). Our data suggest a splitting into two groups over time: Those developing donor-specific hyporesponsiveness and those with further increasing direct donor-reactivity, associated with good and poor graft function, respectively. These findings may be in conflict with the current empirical management of immunosuppression—that is, the longer the patient after transplantation, the lower the immunosuppression is given. Although this could be applied to some patients, others might require enough immunosuppression to contain the donor-specific alloimmune response. In addition, indirect allosensitization is related to time after transplantation and to proteinuria. Hence, we can speculate that both direct and indirect pathway frequencies of IFN-γ–producing T cells seems to increase sequentially in some patients after transplantation and is associated with graft injury. Interestingly, the direct pathway Elispot not only could identify patients with worse graft function but also could discriminate patients with the better renal function because they were donor-specific hyporesponders. In fact, this immunologic state was the solely significant variable associated with better renal graft function in the multivariate analysis. As previously pointed out, directly primed memory T cells have classically been viewed as losing their activity long after transplantation, because donor-derived APC leave the graft already during the first weeks. Nevertheless, some controversial data have been shown by various groups. On the one hand, Baker et al.,20 using limiting dilution analyses and a 6-d cytotoxic assay, found in a group of longstanding kidney transplant patients a universal donor-specific hyporesponsiveness by the direct pathway despite presence of CAN in many of them. On the other hand, Poggio et al.,26 using an IFN-γ Elispot, showed significantly higher donor-specific frequencies by the direct pathway in patients with CAN than in control subjects with normal graft function. Reasonable explanations for these differences could be based on the different methods used. We and Poggio's group both worked with a sensitive IFN-γ Elispot in which only memory/effector T cells were evaluated, instead of 6-d proliferative and cytotoxicity assays, assessing naive along with memory T cell responses, used by Baker et al.20 There are two options to explain that directly primed memory/effector T cells are still active for a long time after transplantation. First, despite the absence of professional donor APC, EC or other graft cells would be able to maintain activated circulating memory/effector T cells directly primed early after transplantation by donor APC, because memory T cells need less co-stimulation than naive T cells and therefore are not dependent on professional APC for re-stimulation. Second, as shown in a recent elegant study performed by Lechler et al.,24 recipient DC, when co-cultured with allogeneic DC or EC, are able to acquire substantial levels of donor class I and II MHC-peptide complexes. These recipient DC, having acquired allogeneic MHC, can migrate to and prime T cells by both allorecognition pathways. The that the with time in antidonor frequencies observed by some could be as a of a direct pathway which would to an cell it could be that this pathway of allorecognition could also explain the of memory/effector T cells over time not only in an indirect but also by a direct pathway. Interestingly, the number of HLA mismatches has been associated with worse outcome at the long In fact, in we could a positive correlation between cellular alloreactive patients and the number of HLA mismatches, the of HLA in the long is well that the indirect pathway of allorecognition is already active after transplantation, as previously it seems to increasing relevance with time as donor-derived APC are by recipient In line with a positive correlation between the indirect pathway Elispot and time after transplantation was observed in This could also be by the progressive of graft cells, which would to a of donor molecules by recipient these APC could indirectly memory/effector T cells. was no association between the indirect pathway and graft function; however, in with previous a significant association between indirectly primed memory T cells and proteinuria was that a of this pathway also to the development of chronic immune-mediated graft although graft was not for patients, a relationship between donor-specific alloreactivity and of graft dysfunction was have suggested that the donor-specific humoral alloimmune in particular the to seems to on the indirect pathway. this in the only patient with circulating DSA also had high frequencies of indirectly primed T however, many patients with indirect pathway donor-specific memory T cells did not The findings of study the that both antigen allorecognition pathways are detectable in longstanding kidney transplant patients. In fact, patients with chronic allograft dysfunction showed of cellular as a of both direct and indirectly primed memory T cells. Thus, monitoring of alloreactive effector/memory T cells primed by both pathways with a sensitive in assay such as the IFN-γ Elispot may be helpful for patients with immune-mediated chronic rejection. The of the donor-specific alloimmune state would allow to either for increasing immunosuppression for Patients We a group of 34 longstanding living-donor kidney transplant patients from the at the of the and at the of were functional longstanding living-donor kidney for at 2 yr after transplantation and for patient and donor blood and The of was by and were at both creatinine and a of proteinuria in positive when were used as to chronic allograft cells were from of who had previously been HLA were I and and II HLA with The third party was used for both allorecognition pathways in DSA of alloantibodies in was in the the using either or a cytotoxicity In the of a positive test, donor antigen specificity was of and PBMC were used as Elispot responder cells for both direct and indirect allorecognition pathways. were from blood by as PBMC were subsequently in at for the Direct Pathway of Antigen Allorecognition As cells, T PBMC were used. were after T cells from the donor blood using as by the for cell were performed as for responder PBMC cells were in for later at for the Indirect Pathway of Antigen Allorecognition donor as for the indirect pathway, cytoplasmic membrane protein preparations of donor allogeneic cells were First, PBMC were by as of membrane was performed by these cells to three of 20 at and in a at as previously in No intact cells were found after in all fragmented of HLA I and II from the fragmented cells was by analyses with I and II HLA to HLA The of 2 fragmented PBMC were in different for later Cell the of that the indirect pathway could be in by using fragmented donor a T cell line was by 2 × PBMC from a responder individual previously with with the of × fragmented PBMC from an allogeneic individual HLA and in 2 of complete with with and in a cell and positive using only complete or phytohemagglutinin as were in at and during 6 via analysis was (Figure 3 × 105 of the indirectly primed cells were tested for alloreactivity against 2 of fragmented PBMC from the donor in an IFN-γ Elispot assay are shown in Figure Elispot IFN-γ Elispot assay was performed as previously in were first with IFN-γ and at of the direct pathway, with 3 × 105 responder PBMC in complete and with 3 × 105 donor T PBMC were tested in A third party and a positive and a negative control were also in test the indirect pathway, with 3 × 105 responder PBMC in complete and with 2 × donor fragmented PBMC were in Also, the third-party cells used for the direct pathway and a negative and a positive control were 24 and of for the direct and indirect pathways, were with and was and at was for 2 a were with spots were using a Elispot Cellular are presented as of detected in responder PBMC antigen or donor cells after the response of with responder or donor cells We used >20 spots/3 × 105 responder cells as to a positive of Donor-Specific cellular hyporesponsiveness was defined as a negative donor-specific IFN-γ test but positive IFN-γ All data are ± were compared using the test for the or test for and the or test for correlation was used to the relationship between and analyses were performed to the independent correlation of independent variables with serum creatinine and presence of proteinuria, respectively. All P were and the was defined as a Direct and indirect IFN-γ A wide range of positive responses were observed 20 patients for direct pathway, ranging from 25 to 380 spots/3 × 105 stimulated PBMC. Regarding the indirect pathway, 20 patients had positive Elispot responses that ranged from 20 to 80 spots/3 × 105 stimulated PBMC. Direct pathway Elispot frequencies were approximately three times higher than indirect pathway frequencies (on average ± versus ± P = Direct T cell alloreactivity is related to graft function. Directly primed alloreactive T cell frequencies were positively correlated with serum creatinine and inversely with the (r = 0.551, P = and = P = respectively). Patients who experienced late acute rejection events (n = 12) had significantly higher direct frequencies than the others (n = 22; ± versus ± P = hyporesponsiveness is related to good graft function. hyporesponder patients had significantly better GFR-MDRD than patients ± versus ± ml/min; P = Direct T cell alloreactivity is related to HLA The higher the number of donor/recipient HLA mismatches I and and class II the higher the frequencies of IFN-γ–producing cells were Indirect pathway T cell alloresponse is related to proteinuria and time after transplantation. of proteinuria was significantly associated with high donor-specific indirect alloreactivity ± versus ± spots in patients proteinuria, P = The indirect pathway Elispot positively correlated with time after transplantation (r = P = Global T cell graft function in direct and indirect patients Fourteen (42%) of 33 patients had positive donor-specific alloreactive memory/effector T cell responses primed by both pathways. Eight (24.2%) had negative antidonor alloreactivity by any pathway, (15.1%) had only direct antidonor alloresponses, and six (18.1%) other patients had only indirect antidonor Interestingly, patients with positive alloreactivity by both pathways (DP+/IP+) had significantly worse graft function than those with undetectable frequencies by both pathways (DP−/IP−) and than those with only indirect frequencies 2.1 ± 1.0 versus 1.1 ± 0.3 [P = 0.004] and 2.1 ± 1.0 versus 1.3 ± 0.3 [P = 0.023], respectively); however, patients with positive frequencies by the direct pathway only (DP+/IP−) did not differ from double-positive patients (DP+/IP+; 2.2 ± 0.4 versus 2.1 ± 1.0, respectively) and, likewise, also had significantly worse serum creatinine compared with DP−/IP+ and DP−/IP− (2.2 ± 0.4 versus 1.2 ± 0.3 [P = 0.005] and 2.2 ± 0.4 versus 1.3 ± 0.3 [P = 0.011], respectively). These data suggest that only the direct pathway is related to creatinine and GFR-MDRD. No relation could be observed regarding the presence of proteinuria and total cellular alloreactivity (data not indirectly primed T cell Two of PBMC from an individual were during 6 with the of fragmented PBMC from an allogeneic individual HLA and and with complete cells were from the and by using APC and for and cells, respectively. Elispot Elispot of 3 × 105 indirectly primed T cells from patient A after stimulated with the of 2 fragmented PBMC from patient and with Patient demographic Patient clinical and immunologic and analyses using as dependent variables serum creatinine and proteinuria, was by a from This study was by a from We all patients who in the We are to for in the development of the indirect pathway of antigen Also, we and for and helpful