Next generation sequencing based assessment of the alloreactive T cell receptor repertoire in kidney transplant patients during rejection: a prospective cohort study

After the introduction of highly potent immunosuppressive regimens in combination with induction therapy, a significant reduction of acute rejection episodes and an improvement in short term graft survival have been achieved [1, 2]. However, a number of transplanted organs are persistently lost following an acute rejection episode, and long-term graft function and survival following this kind of immunological event poses even greater problems to patient management [3, 4].

Acute cellular rejection with infiltrating CD8+ and CD4+ T lymphocytes play a central role and account for the larger fraction of acute rejection episodes [5]. Lymphocyte infiltrates in the kidney are the leading finding of T cell mediated rejection and therefore the nature of infiltrating lymphocytes in the transplanted kidney is of special interest even more as little is known about the antigen specificity of the T cell receptors and the molecular properties of these infiltrates.

Likewise, T cells play a crucial role in the initiation of antidonor antibody formation and consecutively antibody mediated rejection (ABMR). T helper cells support B cells after encountering their antigen by direct interaction and cytokine production which leads to proliferation, differentiation and affinity maturation of B cells to memory B cells or long-lived plasma cells secreting high-affinity antibodies [6‐8].

During early T cell development T cell precursors enter the thymus and differentiate to immature T cells. In the thymus, random recombination of germline-encoded TCR variable (V), diversity (D), and joining (J) gene segments aims for the formation of a functional TCR protein and generation of a highly diverse TCR repertoire [9].

Junctional diversification further increases the combinatorial diversity by trimming gene ends or adding nucleotides between the recombining genes and thereby generating hypervariable complementarity-determining regions (CDR1-CDR3). The complementarity-determining region 3 (CDR3), which is the most variable CDR, spans over the V(D) J junction and forms the primary site of antigen contact [9, 10]. The cells expressing a recombined T cell receptor only survive if self MHC (major histocompatibility complex) and a self-peptide presented by thymic antigen presenting cells is recognized with adequate affinity to transmit a survival signal (positive selection). In a second selection process (negative selection) T cells with an operational TCR but an activation above an affinity- or cross-reactivity threshold are send to apoptosis by a death signal as these cells would be more likely to become activated by self-peptides or by the MHC molecule itself and cause autoimmunity [9].

This complex development of the T cell receptor repertoire and selection process results in the generation of a highly diverse repertoire capable of defending an individual against a variety of pathogens and being ideally tolerant to antigens present in the own body. This leads to a large number of unique TCR rearrangements with in theory over 10^13 unique T cell clonotypes representing the immune repertoire [11‐13].

After the generation of the immune repertoire events such as infections, aging or immune modulating therapies cause an alteration which can lead to the expansion or deletion of clones and effect the diversity of the overall repertoire [10, 14, 15].

In solid organ transplantation T cells recognizing a foreign antigen or MHC protein, so called alloreactive T cells, expand and cause a rejection with severe organ damage. Receptor diversity analysis of these alloreactive cells using conventional methods such as PCR-based hybridization methods, flow cytometry of V region usage or TCR spectratyping, is a major technical challenge [16‐18]. Additionally, these techniques have limitations regarding their resolution:

Fluorescence activated cell sorting (FACS) is limited by the availability of antibodies against T cell receptor’s Variable, Diversity and Joining (VDJ) elements and PCR-based length analysis by spectratyping does not allow to discriminate between clones with identical CDR3 lengths [19, 20].

With the development and technical progress of next-generation sequencing (NGS) technologies it is now feasible to sequence millions of receptor clones simultaneously, which provides a much more detailed and precise picture of the immune repertoire [11, 21, 22]. It is possible to obtain several million sequences per sequencing run and only recently the focus has been put on receptor sequencing of the alloreactive lymphocyte population and the changes in abundancy of these clones and repertoire diversity over time [23].

Two studies of TCR repertoire analysis based on NGS data were published in 2009 using two different experimental designs [11, 21]. These designs differ in the starting material used being either genomic DNA (gDNA) or messenger RNA/complementary DNA (mRNA/cDNA)-based approaches [24, 25]. Both strategies have advantages but also shortcomings. Using genomic DNA in contrast to mRNA as a starting template has the advantage of superior quantification as every T cell only possesses one functional TCR chain, depicting in theory a more precise overall T cell receptor repertoire. A major downside of using gDNA is the significant multiplex PCR bias introduced by different primer annealing or amplification efficiencies leading to over/underrepresentation or dropout of those elements by non-uniform amplification. This bias is reduced in the mRNA/cDNA approach by using template switch during cDNA library preparation which forms universal primer annealing sites on both ends of the library and can be further counteracted by the introduction of unique molecular identifiers (UMIs). However, the mRNA/cDNA approach is known to be less specific and conclusions on absolute frequencies of cells is more difficult due to unequal mRNA expression levels in cells.

Morris et al. recently used mixed lymphocyte reactions (MLR, incubating donor antigen presenting cells (APCs) with recipient T cells) to select alloreactive T cell clones prior to transplantation. Using high-throughput sequencing data of the CDR3 a unique fingerprint of the donor-reactive T cell repertoire was identified before transplantation and donor reactive T cells were tracked posttransplant in blood samples [26]. Further studies validated this approach in different solid organ settings, as in small bowel transplantation, also analyzing T cell repertoires in tissue biopsies [27, 28].

We aim to apply a similar approach in this project to get a closer insight into the alloreactive lymphocyte population prior to transplantation and follow these populations over time and analyze associations of repertoire diversities with the histomorphological phenotype of T cell mediated rejection in the kidney transplantation setting. Further we would like to decipher patterns of expansion of these clones prior to rejection and decline in graft function to hopefully foresee rejection episodes before graft damage occurs.

This is a single center prospective cohort trial designed to evaluate the value of T cell receptor repertoire sequencing of T cells in the periphery and in the allograft during a biopsy proven rejection episode compared to absence of histological findings of rejection conducted at the General Hospital of Vienna/Medical University of Vienna, Department of Nephrology.

These alloreactive T cells are defined prior to transplantation via a mixed lymphocyte reaction as shown in previous studies [26, 27].

Fifteen patients with histological proven acute T cell mediated rejection will be compared to 15 patients without histopathological signs of alloimmune response in time matched for-cause biopsies. Only recipients of donor organs evaluated and accepted by the local transplant team in Vienna will be included due to lack of availability of donor peripheral blood mononuclear cells (PBMCs) of donors evaluated by Eurotransplant, an Europe-wide program allocating and exchanging deceased donor organs across-borders [29]. The proposed duration of the trial is 36 months. The sample size of 15 patients experiencing an acute T cell mediated rejection and an accrual phase of 24 months was calculated based on historic data from our center. The study protocol and informed consent were designed following the “Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) 2013 statement” guidelines and the standards of the world health organization research ethics review committee (Additional files 1 and 2) [30, 31].

Adaptive immunity and T cell activation plays a crucial role in rejection of solid organ transplants [40, 41]. Until now the description of T cells responsible for allograft rejection has been elusive mostly due to technical limitations. With the rising of cutting edge techniques as next generation sequencing it has been possible to characterize and track alloreactive T cells potentially responsible for rejection episodes.

In the transplantation setting the questions raised using these novel methods have until now mostly been focusing on elucidating the mechanisms of tolerance induction in the setting of combined kidney and bone marrow transplantation [26]. In the paper of Morris et al., who were one of the first using the technique of TCR sequencing of alloreactive T cells and tracking these over time, showed a decline of these distinct clones in tolerant patients compared to no significant reduction in a patient experiencing a rejection episode.

These findings raise potential questions if a similar pattern of increase of alloreactive clones in the periphery can be seen in kidney transplant patients at a rejection episode. Of additional interest in our study will be if these alloreactive clones expand and can not only be tracked in the periphery but also in the allograft itself during rejection. These questions have partially been raised in two papers published by two different groups [42, 43].

The results were promising although patient numbers with complete follow up in the paper of Alachkar et al. were low and emphasis was not put on alloreactive clones, but more on TCR clone kinetics during rejection [43]. The detection and TCR sequencing of the pre-transplant defined alloreactive T cell clones targeting the allograft and causing rejection episodes may open new opportunities to predict rejections prior to clinical deterioration of graft function. Further the identification of the patients individual alloreactive T cell repertoire can potentially help to decipher antigen specificity of these cells as new techniques show promising results in predicting epitopes targeted by T cells [44, 45].

The antigen-prediction system developed in the paper of Glanville et al. managed to assign TCRs to antigen-specific binding groups and define enriched sequence motifs which were mainly responsible for antigen binding. They used these motifs to design a set of synthetic TCRs not found in the biological samples and predict their specificity to a known antigen. These synthetic TCRs were proven to successfully induce T cell activation upon antigen contact.

By using the TCR repertoire sequencing approach it will generally be possible to identify T cell clonotypes which are expanded compared to the baseline repertoire and decipher changes in overall diversity and clonality. The analysis of the expanded T cell population taking the clinical phenotype of a patient into consideration could provide new insights into several questions. In delicate situations for example as rise in serum creatinine in presence of a coexisting significant viraemia (e.g. BK Virus, Cytomegalovirus), an expansion of a virus specific T cell population compared to alloreactive clones could give important information for clinical decision making, as virus specific T cells have already been isolated and sequenced successfully [42, 46‐50]. Zeng et al. compared the frequency of BK virus and alloreactive T cell clones, defined by flow cytometric measurement of T cell responses to HLA and BK peptide stimulation in vitro, in the periphery and in kidney biopsies of transplant patients during BK nephropathy and rejection [42]. They found that biopsies with BK nephropathy contain more alloreactive than virus specific clones and theorize that the leading element of tissue injury in viral nephropathy is primary mediated by an ‘innocent bystander’ mechanism and secondary T cell migration caused by both anti-viral and anti-HLA immunity.

Dziubianau et al. included next to BK virus also CMV specific T cell clones in their analysis but final conclusions were hampered by the small sample number [46]. Nevertheless, they could prove TCR repertoire sequencing as a reliable and reproducible technique for the detection of T cell mediated pathology.

Another possible use could be the description of T cell plasticity regarding the conversion of pre-transplant non Treg, effector/central memory cells into Tregs, or other effector cells which were defined by their distinct T cell receptor. This has been already partly shown in a paper of Sprangers et al. [27].

Besides T cell conversion also T cell reconstitution after anti-thymoglobulin (ATG) induction has been studied by several research groups [51‐55]. The results showed an imbalanced reconstitution of the entire T cell population and an altered response of alloreactive T cells after ATG induction therapy [51, 56]. But there was no specific emphasis put on the reconstitution of the alloreactive T cell population.

Regarding the Treg subpopulation, also the properties of specific Treg cells defined by their TCR in tolerant patients or patients with good graft function and low amount of immunosuppressant therapy, compared to patients experiencing rejection episodes as well as the influence of immunosuppressants (e.g calcineurin inhibitors), which are known to particularly affect this population, could be of interest [57, 58].

This promising technique shows to have a broad field of application giving answers to diverse questions regarding the T cell repertoire in kidney transplantation.

An analysis of our data repository showed that these 30 samples will be available within the first 2 years after study initiation accounting for a transplant frequency of local evaluated donors of roughly 120 per year at our center.

Only biopsy proven T cell mediated rejections according to current pathology classification, BANFF 2017 or total absence of rejections will be included in our analysis.

A shortcoming of this study will be a technical limitation regarding the definition of alloreactive T cell clones as the MLR mostly represents antigen presentation via the direct pathway and only to a fewer extent the indirect pathway, which seems the leading mechanism behind chronic allograft function. This may be limiting or hamper conclusions made to chronic allograft nephropathy but should not compromise the assertions concluded in the setting of acute T cell mediated rejection in the first months of transplantation.