Clonal dynamics of alloreactive T cells in kidney allograft rejection after anti-PD-1 therapy

Article https://doi.org/10.1038/s41467-023-37230-4 Clonal dynamics of alloreactive T cells in kidney allograft rejection after anti-PD-1 therapy Received: 26 May 2022 Check for updates 1234567890():,; 1234567890():,; Accepted: 8 March 2023 Garrett S. Dunlap 1,2, Daniel DiToro2,3, Joel Henderson4, Sujal I. Shah 2,3, Mike Manos5, Mariano Severgnini5, Astrid Weins2,3, Indira Guleria2,3, Patrick A. Ott 2,5,6,7,8, Naoka Murakami 2,9,10 & Deepak A. Rao 1,2,10 Kidney transplant recipients are at particular risk for developing tumors, many of which are now routinely treated with immune checkpoint inhibitors (ICIs); however, ICI therapy can precipitate transplant rejection. Here, we use TCR sequencing to identify and track alloreactive T cells in a patient with melanoma who experienced kidney transplant rejection following PD-1 inhibition. The treatment was associated with a sharp increase in circulating alloreactive CD8+ T cell clones, which display a unique transcriptomic signature and were also detected in the rejected kidney but not at tumor sites. Longitudinal and crosstissue TCR analyses indicate unintended expansion of alloreactive CD8+ T cells induced by ICI therapy for cancer, coinciding with ICI-associated organ rejection. Immune checkpoint inhibitors (ICIs) have become the standard of care therapy for many cancers1. ICIs block the activation of inhibitory receptors (e.g., CTLA-4, PD-1), driving T cell activation and enhancing anti-tumor immunity2. However, ICI therapy is often complicated by immune-related adverse events (irAEs) that result from loss of T cell tolerance3,4. Kidney transplant recipients have a 3- to 10-fold increased risk of cancer post-transplant, but the use of ICI is challenging due to the high risk of precipitating acute allograft rejection5,6. We hypothesized that ICI-induced allograft rejection might occur due to a vigorous expansion of pre-existing, alloreactive memory T cells. Here we leveraged banked tissue and blood samples from a patient with advanced melanoma who experienced allograft rejection shortly after ICI therapy to identify and track alloreactive T cells longitudinally and across different tissues. Results A 77-year-old man with end-stage kidney disease due to chronic glomerulonephritis underwent deceased kidney transplantation, with six out of six HLA mismatches (A, B, and DR loci) and without any anti-HLA antibodies. His post-transplant course had been uncomplicated and without any rejection episodes, and immunosuppression was maintained with sirolimus (2 mg per day), mycophenolate mofetil (500 mg twice a day), and prednisone (5 mg per day). Baseline serum creatinine (Cr) was 1.3 mg/dl [normal range, 0.7–1.3], estimated glomerular filtration ratio (eGFR) was 50 ml per minute per 1.73 m2 of the body surface area [normal range, >60], and urinalysis did not show proteinuria. Ten years after transplantation, he was diagnosed with cutaneous melanoma arising from the nasal dorsum and left cervical lymph node metastases with extracapsular extension, stage IIIC (BRAF/NRAS/ c-kit wildtype). The patient underwent a wide local surgical excision, complete left cervical lymphadenectomy, and radiation therapy to the left neck, yet subsequently developed osseous and distant lymph node metastases (Stage IV) (Supplementary Fig. 1A). Treatment with pembrolizumab was initiated (200 mg every 3 weeks) with no reduction in immunosuppressive medications. After two doses of pembrolizumab, serum Cr increased to 2.98 mg/dl. Acute transplant rejection was suspected, and an allograft kidney biopsy was performed 5 weeks after 1 Division of Rheumatology, Inflammation, and Immunity, Brigham and Women’s Hospital, Boston, MA, USA. 2Harvard Medical School, Boston, MA, USA. Department of Pathology, Brigham and Women’s Hospital, Boston, MA, USA. 4Department of Pathology, Boston Medical Center and Boston University, Boston, MA, USA. 5Center for Immuno-Oncology, Dana-Farber Cancer Institute, Boston, MA, USA. 6Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA. 7Broad Institute of MIT and Harvard, Cambridge, MA, USA. 8Department of Medicine, Brigham and Women’s Hospital, Boston, MA, USA. 9Division of Renal Medicine, Brigham and Women’s Hospital, Boston, MA, USA. 10These authors contributed equally: Naoka Murakami, Deepak A. Rao. e-mail: ; 3 Nature Communications | (2023)14:1549 1 Article https://doi.org/10.1038/s41467-023-37230-4 initiation of pembrolizumab (Fig. 1A). Pembrolizumab was discontinued, the patient received solumedrol (500 mg daily for 3 days) followed by a prednisone taper (1 mg per kilogram per day), and the dose of mycophenolate mofetil was increased (1000 mg twice a day). Serum Cr level initially improved to 2.2 mg/dl but 2 weeks later increased again to 3.1 mg/dl. The patient received another course of solumedrol (500 mg daily for 3 days), followed by prednisone taper, resulting in Cr stabilization between 1.7 and 1.9 mg/dl (Fig. 1A). No irAEs were observed. Restaging PET/CT 12 weeks after initiation of pembrolizumab demonstrated a mixed response of lymph node and bone metastases (Supplementary Fig. 1B, C). Histopathological analysis of the allograft biopsy showed prominent interstitial inflammation (Banff score, i3), moderate glomerulitis (g2), mild tubulitis (t1), and no identified vasculitis (v0) (Fig. 1B, C)7. C4d staining was negative, and the donor-specific antibody titer was 0, indicating an absence of antibody-mediated rejection. Immunofluorescence staining of the allograft biopsy indicated infiltration of both CD4+ and CD8+ T cells, suggesting acute T cell-mediated rejection (Fig. 1D). Given that the recipient’s graft post-transplant course was stable until ICI therapy, we hypothesized that ICI therapy had activated a population of pre-existing alloreactive T cells. To identify these alloreactive T cells in vitro, we performed a mixed-lymphocyte reaction (MLR) of recipient PBMCs and donor splenocytes, which had been previously banked. Recipient T cells proliferating after MLR were flowsorted, and droplet-based single-cell RNA sequencing (scRNA-seq) was used to obtain paired gene expression and T cell receptor (TCR) sequences for the sorted cells (Supplementary Fig. 2A). After dimensionality reduction of the 1505 sorted cells that passed quality control filtering, we identified nine clusters, including populations of CD8+ and CD4+ T cells (Fig. 2A, Supplementary Fig. 2B–F, Supplementary Data 1). Reasoning that the cells with the greatest alloreactive potential would show the highest proliferative capacity, we determined the proliferative signature of each cluster using a previously reported gene set (Supplementary Data 2)8. Two clusters of CD4+ T cells (C3 and C5) and one cluster of CD8+ T cells (C1) had elevated proliferation signatures compared to all other populations (Supplementary Fig. 3A, B). Further, analysis of the expression patterns of cell cycle phase-associated genes suggested that cells con (...truncated)