In June 2021, the First World Consensus Conference on Pancreas Transplantation provided evidence-based guidelines, offering directions for clinical practice after pancreas transplantation (PT) [1, 2]. The primary message emphasized that both simultaneous pancreas/kidney transplantation (SPK) and pancreas transplantation alone (PTA) lead to improved quality of life [3, 4] and long-term patient survival [5, 6] compared to other medical interventions. Experts conclude that according to empirical evidence, preformed donor-specific anti-HLA antibodies (DSA) with MFI level <5000 in recipients with negative T and B cell flow cytometric crossmatches should not be a prohibitive factor for pancreas transplantation, based on the fact that evidence regarding negative impact of pretransplant DSA on transplant outcomes is lacking [2]. Contrarily, detection of dnDSA has been associated with worse outcomes, including graft rejection and failure [7, 8], Hence, rigorous post-transplant surveillance is recommended. In the setting of SPK, the relevance of HLA mismatching is a matter of debate as it did not translate into improved overall graft outcome, even though associated with reduced development of dnDSA and reduced graft rejection [9, 10]. Concerning immunosuppression, tacrolimus and mycophenolate, compared to ciclosporin and azathioprine, showed superior immunological results, i.e., reduced risk of developing dnDSA [11, 12]. Early tapering of corticosteroids was found suitable for a specific subset of pancreas transplant recipients, demonstrating viability without concomitant compromise in outcomes but improved metabolic parameters in the long term [13, 14].

Our study aims to assess the impact of the Predicted Indirectly ReCognizable HLA Epitopes (PIRCHE-II) scores for the first time in predicting the development of dnDSA and graft outcomes in simultaneous pancreas/kidney transplantation recipients. The PIRCHE-II score is an established algorithm to calculate HLA epitope mismatches for certain HLA antigen mismatches. It estimates the number of indirectly recognizable, donor HLA-derived T cell epitopes and predicts T cell-related immune responses against the donor HLA-derived peptides. Moreover, the PIRCHE-II score has demonstrated the ability to predict the incidence of dnDSA in kidney transplantation (KT) independently and was associated with kidney allograft survival in a cohort of kidney transplantation [15, 16].

In our study, we attempted to address the following questions: 1) What is the incidence of dnDSA among SPKTRs vs. KTRs? 2) What risk factors are associated with the development of dnDSA among SPKTRs at 1-year post-transplantation and in the long-term? 3) What risk factors are associated with the development of dnDSA among the whole cohort of SPKTRs/KTRs at 1-year post-transplantation and in the long term?

A well-established correlation has been suggested in kidney transplantation between a higher number of HLA epitope mismatches [17] and an increased risk of developing dnDSA associated with AMR and allograft loss [18–21]. Lachmann et al. revealed in their paper a strong correlation between the total PIRCHE-II score (considering HLA-locus A, HLA-locus B, HLA-locus C, HLA-locus DR, HLA-locus DQ) and an increased risk of development dnDSA, primarily directed against HLA-DQ, followed by HLA-DR, HLA-A, and HLA-B mismatches. This was confirmed in subsequent studies [22, 23].

In contrast to the investigations carried out in a kidney transplantation cohort, the PIRCHE-II scores’ prognostic value in predicting dnDSA development and graft outcomes following other solid organ transplantations is not well studied. Particularly in pancreas transplantation, data regarding the relevance of PIRCHE-II scores is scarce. Based on suggested risk assessment according to the recently published First World Consensus Conference on pancreas transplantation [1, 2], less importance has been attributed to HLA mismatching and preformed DSA. In this context, our study aims to evaluate risk factors for developing dnDSA among SPKTRs. To the best of our knowledge, this study represents the first effort to investigate the influence of the PIRCHE-II scores, adjusted for both HLA class I and II, and HLA locus-specific, on the development of dnDSA in a cohort of SPKTRs.

Firstly, our results indicate that, despite SPKTRs having fewer preformed DSA and lower median total PIRCHE-II scores, there was a higher incidence of dnDSA against HLA class I and II in the first post-transplant year compared to KTRs. Notably, dnDSA were predominantly directed against the HLA-locus DQ, consistent with previous studies [24]. Several factors contribute to this finding. The pancreas allograft is a highly immunogenic organ, and its beta cells can prompt a strong alloimmune response, contributing to a higher incidence of dnDSA development. Pancreatic inflammation and injury, common in the early post-transplant period, can further activate alloreactivity, leading to the development of dnDSA as the immune system reacts to the inflamed or injured pancreatic tissue. Although SPKTRs receive more intense induction immunosuppressive therapy, rapid steroid withdrawal might not be suitable for all SPKTRs and allow DSA development. Given the observed differences in alloreactivity during the early post-transplant period, it is particularly crucial to study the impact of PIRCHE-II scores on dnDSA formation in SPKTRs. To reduce the potential risk of overestimation of dnDSA with an MFI cut-off of 500, all dnDSA were analyzed individually by a specialist in transplantation immunology in a blinded fashion. Here, 1) analyzing the pattern of single-bead reactivity and comparing it to the HLA typing of the donor, 2) investigating for epitope specificity to determine alpha chain binding antibodies in the setting of HLA-DQ and DP, 3) determining unspecific reactivity by comparing the pattern of reactivity to lot-specific reactivity patterns in non-immunized males that are continuously tracked in our transplant laboratory, and 4) incorporating the reactivity to the recipient’s own HLA antigens was applied to reduce overestimation.

Secondly, regarding the risk factors associated with the development of dnDSA in SPKTRs, the presence of preformed DSA and younger donor age were independently associated with an increased risk. Interestingly, total PIRCHE-II scores, PIRCHE-II scores per HLA class, and PIRCHE-II scores per HLA locus were not independently associated with an increased risk for the development of dnDSA. However, PIRCHE-II scores for HLA class II, particularly HLA-locus DQ, may predict the development of dnDSA against HLA class II, although the sample size in our analysis was not sufficient to show this association independently upon multivariate analysis. This finding aligns with the observation that dnDSA against HLA-locus DQ exhibited the highest incidence among all HLA loci. Conversely, the lack of association for HLA class I is likely attributable to the low incidence of dnDSA against HLA class I and the small sample size of our cohort. Nonetheless, our results highlight other factors associated with the development of dnDSA that should be considered in future studies when evaluating the predictive and additive value of PIRCHE-II scores.

Notably, preformed DSA increased the risk of dnDSA development among SPKTRs but not KTRs. Factors associated with SPKT, preformed DSA, and the combination of both may likely explain this elevated risk of dnDSA development. Factors associated with SPKT include 1) the transplantation of two organs, which presents more allo-antigens compared to kidney transplantation alone, and 2) the pancreas as a highly vascularized organ, that may provoke a stronger alloimmune response compared to the kidney alone. Factors associated with preformed DSA include 1) the presence of an already activated immune system with presence of memory B cells, that may get stimulated by the increased antigen load in SPKT, and 2) a high number of shared HLA-epitopes, that contribute to the potential of HLA antibodies cross-reacting with other HLA antigens [25]. Our results suggest, that the combination of SPKT and preformed DSA is decisive for the increased risk of dnDSA development. 1) Preformed DSA may precipitate subclinical or clinical TCMR and AMR, particularly in the pancreas allograft, which can further induce an inflammatory microenvironment that may stimulate antigen presentation and immune activation, leading to dnDSA development. 2) The rapid steroid withdrawal in SPKTRs may also be critical in cases with preformed DSA, facilitating this inflammatory microenvironment that may allow the formation of dnDSA. Organs from younger donors tend to have higher immunogenicity due to increased expression of HLA antigens and costimulatory molecules. Additionally, the presence of more active dendritic cells and other antigen-presenting cells, coupled with increased cellular proliferation, can enhance antigen exposure and the recipient’s immune activation, thereby increasing the risk of DSA formation.

Thirdly, when considering the entire cohort of SPKTRs/KTRs, simultaneous pancreas/kidney transplantation did not independently increase the risk of developing dnDSA. However, our data demonstrated an association between HLA epitope mismatching and dnDSA development, consistent with the literature [15]. Unlike previous studies, we observed the most pronounced association between PIRCHE-II scores for HLA-locus DQ and the development of dnDSA against HLA-class II. Ladowski et al. reported similar results but primarily focused on PIRCHE-II for HLA class II, especially HLA-DQB mismatches [24]. The concept of HLA epitope mismatch load and the impact of dnDSA is most clearly shown for HLA-locus DQ. It remains unknown whether the number of HLA epitope mismatches or the increased likelihood of more immunogenic HLA epitopes contributes to this increased risk [26]. Current evidence suggest HLA-DQ combinations that are more immunogenic than others [27].

To our knowledge, we are among the first to include PIRCHE-II for both HLA-DQB and HLA-DR in the SPKTRs population and demonstrate their role in predicting dnDSA formation against both HLA class I and II. Additionally, Chaigne et al. demonstrated in a cohort of pancreas recipients that the formation of anti-HLA class I antibodies was unrelated to PIRCHE-II scores. In contrast, the development of anti-HLA class II antibodies was influenced by PIRCHE-II scores [28]. We did not observe an association between PIRCHE-II scores for HLA-locus C, which may explain the lack of significance for total PIRCHE-II scores in our cohort. This finding aligns with previous studies, such as those by Lachmann et al., who also considered HLA-A, B, C, DR, and DQ when calculating the PIRCHE-II score [15]. Thus, our observation highlights the lack of association with the PIRCHE-II score for HLA-C, which can potentially contribute to misleading interpretations of total PIRCHE-II scores. Moreover, we observed an increased risk of dnDSA with the use of ciclosporin and lower donor age [29]. These findings from our multivariate analysis are significant because the two most cited studies, by Lachmann et al. [15] and Unterrainer et al [30], are based on data from patients primarily under ciclosporin-based immunosuppression and with incomplete recipient and donor typing. These earlier studies may not fully reflect the current state of transplantation practices.

Our study possesses several strengths. First, it stands as one of the first analyses focusing on the development of dnDSA in SPKTRs compared to KTRs. Second, we included a well-characterized cohort of SPKTRs spanning over a decade, adhering to a standardized immunosuppressive protocol without the use of ciclosporin, and maintained close clinical and immunological post-transplant monitoring, thus providing high data density. Third, our study explored, for the first time in SPKTRs, the total PIRCHE-II scores, PIRCHE-II scores per HLA-class, and PIRCHE-II scores per HLA-locus. Yet, there are also limitations warranting consideration. Most importantly, the retrospective study design, small sample size and the single-center bias concerning allograft allocation, immunosuppressive strategy based on a steroid-free immunosuppression regimen in SPKTRs. Our study also relied on imputed high-resolution HLA alleles for the PIRCHE-II calculation, which could potentially influence our results due to errors in the imputation. The population in our study was predominantly Caucasian and a recent study has suggested that the potential difference in PIRCHE-II score association with dnDSA development in this setting would be minimal [31].

In summary, SPKTRs exhibit a higher incidence of de novo dnDSA in the first year post-transplantation, which is not linked to an increased HLA-epitope mismatch load. The correlation with preformed DSA indicates a higher immunologic risk, particularly under a steroid-free regimen, favoring dnDSA development. Over the long term, a high HLA-epitope mismatch load for HLA locus DQ is similarly crucial for dnDSA development in both SPKTRs and KTRs. The lack of association between the total PIRCHE-II score, PIRCHE-II scores for HLA classes, and other HLA loci suggests that these biomarkers should at the moment not be used for risk stratification post-transplantation.