Single-cell chromatin accessibility landscape of T cell-mediated rejection and BK polyomavirus-associated nephropathy after kidney transplantation.

Kidney transplantation stands as the optimal treatment for patients with end-stage renal disease. However, the post-transplantation period often brings about complications, chief among them being rejection and infection, which frequently result in graft failure and occasionally patient mortality [ 1 ]. Recognized diagnostic standards such as the Banff classification, divide rejection into T-cell mediated and antibody-mediated categories. However, increasing evidence suggests that this binary classification is insufficient, as there are mixed rejection types and other conditions [ 2 ]. The majority of acute rejection episodes, especially in the first year after transplantation, are T cell-mediated [ 3 ]. Acute T cell-mediated rejection (TCMR) was once the major barrier to successful kidney transplantation. With further improvements in induction therapy and maintenance immunosuppression, acute TCMR rates continued to decline [ 4 ]. However, clinical TCMR within the first year remains between 5%-15%, while subclinical TCMR occurs in up to 30% of biopsies [ 5 ]. Besides, it is still one of the leading reasons for graft loss [ 6 ]. Furthermore, delayed diagnosis and treatment of acute TCMR may result in the production of de novo donor-specific antibodies (dnDSAs) and the appearance of antibody-mediated rejection (AMR), the other type of rejection [ 7 ]. This underscores the crucial necessity of accurately diagnosing TCMR to ensure timely treatment. On the contrary, high immunosuppression after kidney transplantation causes infection in the kidney, which leads to nephropathy [ 8 ]. BK polyomavirus (BKPyV)-associated nephropathy (BKPyVAN) is one of the most common nephropathies after kidney transplantation, which occurs in almost 2% of 10%-30% of kidney transplant recipients who suffer BKPyV viremia [ 8 ]. Alarmingly, a significant proportion of patients with BKPyVAN are destined to experience graft failure, ranging from 50% to 80% [ 9 ]. Given the absence of effective antiviral therapies for BKPyV, early diagnosis of BKPyVAN becomes crucial for graft preservation. Although accurately diagnosing both TCMR and BKPyVAN is crucial for effective treatment, this can often be challenging due to patients remaining asymptomatic. Currently, the standard diagnostic approach for both conditions is the percutaneous core needle biopsy [ 10 , 11 ]. However, this method is not without its limitations. Firstly, in biopsy samples where viruria is present instead of viremia, distinguishing the pathological features (Banff histologic scores, immunohistochemical analysis of inflammatory infiltrates, etc. ) between TCMR and BKPyVAN can be particularly challenging, leading to false-negative rates ranging from 10% to 30% in diagnosing these conditions [ 8 ]. Secondly, there is significant inter-observer variability in interpreting allograft biopsy results, which can further complicate the diagnostic process [ 12 , 13 ]. Lastly, repeatedly performing biopsies to capture the dynamics of the anti-allograft repertory is neither safe nor practical. Therefore, there is an urgent need for improved diagnostic strategies that can enhance accuracy and minimize the associated risks. Molecular characterization holds promise as a potential strategy to identify novel markers for accurate diagnosis of kidney disease. However, previous studies have demonstrated that bulk transcriptomic analysis has been unsuccessful in revealing specific markers for TCMR and BKPyVAN. For instance, Mannon et al. observed that the transcriptional profiles of TCMR and BKPyVAN exhibited high similarities [ 14 ]. The rejection-associated genes were also significantly upregulated in BKPyVAN. Other studies showed that intragraft gene expression [ 15 ] and next-generation sequencing [ 16 ] all failed in the differential diagnosis of TCMR and BKPyVAN. The inability to distinguish TCMR and BKPyVAN through bulk transcriptome profiling suggests that a molecular characterization beyond the transcriptome is necessary to decipher the unique molecular signatures of these conditions. Moreover, considering that TCMR and BKPyVAN are pathological processes that involve multiple cell types, including both renal parenchymal cells and immune cells, bulk molecular profiling may not capture the cell-type-specific characteristics of these diseases. Chromatin accessibility refers to the extent to which nuclear macromolecules can physically interact with chromatinized DNA, and are dynamically changed in response to pathological stimuli [ 17 ]. Profiling chromatin accessibility can not only reveal the overall chromatin structure but also indicate upstream regulatory elements that modulate cellular transcription. Therefore, to further characterize the molecular features of TCMR and BKPyVAN, we propose to investigate the chromatin accessibility landscape with the single-cell assay for transposase-accessible chromatin sequencing (scATAC-seq) in both TCMR and BKPyVAN following kidney transplantation.

In the control group, we collected five donated kidney samples from individuals ranging in age from 34 to 60 years, including two men and three women. All donors had preserved kidney function (mean serum creatine (sCr) = 62 ± 14 μmol/L, eGFR = 106 ± 25 mL/min/1.73 m 2 ). Histologic review showed no significant glomerulosclerosis, interstitial fibrosis, or tubular atrophy ( Figure 1A ). For the experimental groups, we included seven cases of kidney grafts diagnosed with TCMR (TCMR group) and six cases with BKPyV nephropathy (BKPyVAN group). All patients in both groups were hospitalized for biopsies due to elevated serum creatine levels, namely indicated biopsy, for the first time. All recipients received triple immunosuppressive strategy including prednisone, mycophenolate mofetil, and calcineurin inhibitors. The dosages differed depending on the time after transplantation. The baseline information of the patients included is shown in Table 1 . For the TCMR group, samples were diagnosed as TCMR IA or IB, according to Banff 2019. There are no significant differences in Banff histologic scores or inflammatory infiltrates between TCMR and BKPyVAN under light microscopy ( Figure 1B,C ). All BKPyVAN cases were SV40 positive, and viral inclusions were present under electron microscope. Representative images of PAS-stained kidneys of patients included in the control group (a), TCMR group (B), and BKPyVAN group (C). And the scheme of the experimental design for scATAC-seq on kidney biopsy samples (D). Baseline information of patients included. TCMR: T-cell mediated rejection; BKPyVAN: BK polyomavirus (BKPyV)-virus-associated nephropathy; NA: not applicable. a Data was shown as median (range). b Data was shown as mean ± standard deviation. We performed scATAC-seq on control, TCMR, and BKPyVAN samples. We filtered the low-quality cell in which the TSS Score is lower than 4 or detected fragments is lower than 1,000, and finally get 3,950 cell in control group, 15,569 cells in TCMR groups and 6314 cells in BKPyVAN groups. We detected major cell types within the kidney based on the gene scores of lineage-specific markers ( Figure 2A,B ), which includes proximal tubules (PT), thick ascending limbs (TAL), distal tubules (DCT), collecting ducts (principal cells (PC), intercalated cells (ICA and ICB)), endothelial cells (ENDO), glomerular cell types (mesangial cells (MES), podocytes (PODO)), fibroblasts (FIB), and leukocytes (B cells, T cells, monocytes (MONO)) ( Figure 2A,B , Figure S1 ). The cell types and proportions identified by scATAC-seq align with previously reported single-cell transcriptional and chromatin accessibility profiling on human kidneys [ 23 ], demonstrating the reproducibility and reliability of our results. The scATAC-seq atlas of kidney in control, TCMR, and BKPyVAN groups. A. The t-SNE plot of cells in control, TCMR and BKPyVAN group color-coded by the cell types as indicated. B. The t-SNE plots of cells in control, TCMR, and BKPyVAN groups color-coded by the imputed gene score of cell-lineage marker genes. IRF8: marker gene for B cells [ 41 ]; GZMA: marker gene for T cells [ 42 ]; C1QB: marker gene for monocytes; SLC12A3: marker gene for distal tubules; ENG: marker gene for endothelial cells; SLC26A7: marker gene for intercalated cells; HEYL: marker gene for mesangial cells and fibroblasts; AQP2: marker gene for principal cells; CRB2: marker gene for Podo; CUBN: marker gene for proximal tubules; SLC12A1: marker gene for hick ascending limbs. C. The t-SNE plot of cells in control, TCMR and BKPyVAN group color-coded by the sample types. The dashed circle indicated TCMR and BKPyVAN -induced immune-related cell populations. D. The relative log 2 fold change of cell population in control, TCMR, and BKPyVAN group. The comparison was performed between BKPyVAN/TCMR and the control group. To evaluate the impact of TCMR and BKPyVAN on cell population, we examined the single-cell data by samples. Intriguingly, both TCMR and BKPyVAN significantly induced cell population related to immune systems, including B cells, T cells, and monocytes ( Figure 2C,D ). This is consistent with previous knowledge that infiltration of inflammatory cells is identified under the situation of rejection and infection. Interestingly, compared to the control group, there is an increased B cell population in the TCMR group in addition to the obvious increase of T cells ( Figure 2D ), suggesting a potential role of B cells in the TCMR. Besides, we found that there is an increase in the T cell population in the BKPyVAN group compared to that of the control ( Figure 2D ), indicating that our analysis captured the T cell activation in the kidney upon BKPyV infection [ 25 ]. We next examined parenchymal cells in our samples. There were minimal differences observed in the proportions of parenchymal cells between groups, except for MES and FIB. An increased proportion of MES and FIB was detected in the TCMR group ( Figure 2D ). The cell population analysis reveals a general picture of the cellular heterogeneity between TCMR and BKPyVAN samples. To further explore the molecular signatures of BKPyVAN and TCMR, we next examined the altered chromatin accessibility landscapes in each cell type captured in BKPyVAN and TCMR biopsies ( Table S1 ). Overall, TCMR induces more pronounced alterations in chromatin accessibility compared to BKPyVAN, as evidenced by a higher number of dysregulated ATAC-seq peaks in TCMR samples ( Figure 3A ). In BKPyVAN samples, chromatin accessibility changes are primarily observed in PT, TAL, and PC cells, while TCMR affects a broader range of cell types, including PT, TAL, DCT, PC, ICA/ICB, and ENDO ( Figure 3A ). Notably, although both BKPyVAN and TCMR induce chromatin accessibility changes in PT, TAL, and PC cells, their effects on these cells are quite different ( Figure 3B ). This suggests that distinct cellular response mechanisms are activated in these cell types in response to TCMR and BKPyVAN. The enriched functional pathways of genes with differential chromatin accessibility across control, TCMR, and BKPyVAN groups in specific cell types. A. The bar plot showed the number of differentially accessible regions between the control and BKPyVAN or TCMR groups. B. The Venn plot showed the overlap number of differentially accessible regions between the control and BKPyVAN or TCMR groups. C. The dot plot showed the enriched KEGG term in PT, TAL, and PC cells. The differential gene score comparison was performed between BKPyVAN and the control group. The pathways enriched in viral infections within PT cells and carcinogenesis within TAL cells are highlighted in green font. D. The track plot showed the ATAC signal of HLA-DQB1 between BKPyVAN and the control group in PT cells. E. The track plot showed the ATAC signal of CCND1 between BKPyVAN and the control group in TAL cells. F. The track plot showed the ATAC signal of HLA-DRB1 between BKPyVAN and the control group in PC cells. G. The dot plot showed the enriched KEGG term in PT, TAL, PC, and ENDO cells. The differential gene score comparison was performed between TCMR and the control group. H. The track plot showed the ATAC signal of HLA-DQB1 between TCMR and the control group in PT cells. I. The dot plot showed the enriched KEGG term in ENDO, TAL, and PT cells. The differential gene score comparison was performed between TCMR and BKPyVAN groups. J. The track plot showed the ATAC signal of PLA2G1B between the TCMR and BKPyVAN groups in ENDO cells. K. The track plot showed the ATAC signal of TNFSF8 between the TCMR and BKPyVAN groups in TAL cells. L. The track plot showed the ATAC signal of HLA-DPB1 between TCMR and BKPyVAN group in PT cells. Next, we investigated the chromatin accessibility changes triggered by the BKPyVAN condition. As BKPyVAN majorly affects PT, TAL, and PC cells ( Figure 3A ), we decided to focus on these populations. To identify the biological pathways perturbed by BKPyVAN, we conducted a KEGG enrichment analysis on the genomic loci exhibiting altered chromatin accessibility in these cells ( Figure 3C ). PT cell is one of the major cells BKPyV prefer to infect [ 26 ], and we confirmed significant alterations of chromatin alterations in PT cells upon BKPyVAN ( Figure 3A ). For instance, we found that the chromatin accessibility of genes associated with viral infectious pathways, such as HLA-DQB1, was increased ( Figure 3C–E ). This is as expected since BKPyVAN refers to a virus infection pathological condition. In addition, metabolism-associated pathways, including porphyrin, retinol, and pentose and glucuronate interconversions, were activated. These results suggest a global metabolic remodeling of the chromatin landscape of PT cells upon BKPyV infection. TAL is responsible for reabsorbing sodium, chloride, and some water from the filtrate, contributing to the concentration of urine in the kidney [ 27 ]. It has been reported that BKPyVAN may cause tumorigenesis in urothelial cells [ 28 ]. We also noticed that TAL exhibited increased chromatin accessibility of carcinogenesis pathway upon BKPyVAN, suggesting that BKPyVAN may induce tumorigenesis in TAL cells as well ( Figure 3C ). Besides, genes associated with cell cycle and cellular senescence pathway, such as cyclin D1 (CCND1), were also activated in TAL ( Figure 3C,E ), indicating that these cells may undergo cell cycle alterations. PC in the collecting ducts of the kidneys plays a crucial role in reabsorbing water and electrolytes such as sodium and chloride, essential for maintaining body fluid and electrolyte balance and influencing urine concentration and dilution [ 29 ]. In PC, we identified substantial changes in chromatin accessibility changes in multiple genes enrichened in inflammation-related pathways, such as viral protein interaction with cytokine and cytokine receptor, allograft rejection, antigen processing and presentation, complement, and coagulation cascades chemokine signaling cytokine-cytokine receptor interaction. indicating that a substantial inflammation reaction is triggered in PC upon BKPyVAN ( Figure 3C ). HLA-DRB1 was one of their representatives ( Figure 3F ). In addition to these inflammation-related pathways, other cellular function-associated pathways, such as phagosome, calcium signaling pathway, and lipid metabolism (biosynthesis of unsaturated fatty acids), was detected altered in PC, indicating that the inflammation response reshapes the cellular state of PC. We next focused on the chromatin accessibility in TCMR samples ( Figure 3G ). The chromatin accessibility changes in PT, TAL, and PC cells were majorly associated with immune pathways. Genes enriched in antigen processing and presentation, cytokine − cytokine receptor interaction pathways, including HLA-DQB1, exhibited an increased chromatin accessibility ( Figure 3H ). Unlike BKPyVAN, TCMR also affects DCT, ICA/ICB, and ENDO cells ( Figure 3A ). Interestingly, we observed that the endothelial cells exhibited changes in cellular proliferation and death related pathways, including cellular senescence, cell cycle, mitophagy, and ferroptosis, providing a potential mechanism underscoring the malfunction of endothelial cells in TCMR samples ( Figure 3G ). Our previous analysis revealed that although TCMR and BKPyVAN affect PT, TAL, and PC cells, their impacts on chromatin accessibility in these cells differ, suggesting distinct cellular responses in the kidney to TCMR and BKPyVAN. To further distinguish the molecular features of TCMR and BKPyVAN, we conducted a detailed examination of the chromatin accessibility differences between TCMR and BKPyVAN samples. Differences of ATAC-seq signals were majorly exhibited in ENDO, TAL, and PT cells ( Figure 3I ). For ENDO, the major chromatin accessibility differences are in the genes associated with metabolism ( Figure 3I ). The genes affected in metabolism, including PLA2G1B ( Figure 3J ), are enriched in both lipid and carbohydrate related pathways, such as linoleic acid, arachidonic acid, alpha-linolenic acid, and primary bile acid biosynthesis, and butanoate. In TAL, cytokine-cytokine receptor interaction pathway genes, including TNFSF8 ( Figure 3K ), were more active in the TCMR group, indicating a more robust inflammation response in TCMR compared to that of BKPyVAN. PT also exhibited an elevated immune response in TCMR groups compared to that of BKPyVAN groups ( Figure 3I ), as evidenced by the increase of chromatin accessibility on genes enriched in antigen processing and presentation pathways (such as HLA-DPB1) in PT of TCMR group ( Figure 3L ). In summary, our data reveals the differences in the chromatin landscape at the single-cell level between TCMR and BKPyVAN and identifies the pathways associated with the differences. The differences in chromatin accessibility observed between TCMR and BKPyVAN suggest the existence of distinct transcriptional regulatory factors in these two conditions. To identify the cell-type-specific transcriptional factors that underlie these chromatin accessibility differences, we conducted a transcription factor motif enrichment analysis specifically in PT and TAL cells ( Figure 4A ), as these two cell types are affected by both TCMR and BKPyVAN, but exhibited pronounced chromatin accessibility variations between TCMR and BKPyVAN samples ( Figure 3A-B ). Dysregulated transcription factors across control, TCMR, and BKPyVAN groups in specific cell types. A. The heatmap showing the top enriched transcription factors in PT and TAL cells across three comparisons: BKPyVAN vs. control, TCMR vs. control, and BKPyVAN vs. TCMR. The left panel provides an overview of transcription factors enriched in chromatin regions with increased accessibility, while the right panel highlights those enriched in chromatin regions with decreased accessibility. B. The dot plot showed the enriched transcription factors in the differential accessible regions of PT and TAL cells. The comparison was performed between the BKPyVAN and control groups. The upregulated group refers to transcription factors enriched in chromatin regions with increased accessibility in BKPyVAN samples, while the downregulated group refers to transcription factors enriched in chromatin regions with decreased accessibility in BKPyVAN samples. C. The dot plot showed the enriched transcription factor in the differential accessible regions of PT and TAL cells. The comparison was performed between the TCMR and control groups. The upregulated group refers to transcription factors enriched in chromatin regions with increased accessibility in TCMR samples, while the downregulated group refers to transcription factors enriched in chromatin regions with decreased accessibility in TCMR samples. D. The dot plot showed the enriched transcription factor in the differential accessible regions of PT and TAL cells. The comparison was performed between BKPyVAN and TCMR groups. The enriched in BKPyVAN group refers to transcription factors enriched in chromatin regions with increased accessibility in BKPyVAN samples, while the enriched in TCMR group refers to transcription factors enriched in chromatin regions with increased accessibility in TCMR samples. Compared to control groups, we identified several enriched transcription factors in the BKPyVAN group in PT and TAL ( Figure 4B ). For instance, NR3C1, Hepatocyte nuclear factor 4 A (HNF4A), and HNF4G were identified as top upregulated regulators in PT cells upon BKPyV infection. These factors regulated PT differentiation and might possess potential roles in response to cell injury in fully mature PT [ 30 ]. The enrichment of these factors in BKPyVAN samples indicated that they might play a role in repairing PT after BKPyV infection. In addition, we identified PAX1, PAX5, and PAX9 were identified as upregulated TFs in both PT and TAL of BKPyVAN samples ( Figure 4B ). Pax genes encode a family of transcription factors, and PAX2/8 has been reported to play key roles during kidney development [ 31 ]. Our analysis indicated that PAX1/5/9 was activated upon BKPyV infection, and it was worth investigating the function of these factors in response to BKPyV infection. We also identified transcription factors that may exhibit downregulated activity with BKPyVAN ( Figure 4B ), including early growth response 1 (EGR1), Wilms’ tumor 1 (WT1), and specificity protein-1 (SP1) in PT cells, as well as EGR1/2 and ZBTB7B in TAL cells. Notably, EGR1 has been reported to mediate renal epithelial cell regeneration and repair after injury [ 32 ]. Our results indicated that EGR1 is dysfunctional in BKPyVAN samples, which may contribute to the damage in PT cells upon BKPyV infection. For TCMR, our study also identified multiple potential dysregulated molecular factors compared to the control ( Figure 4C ). For instance, we identified FOS, JUN, and JUNB, subunits of activator protein 1 (AP-1), as a prominent transcription factor with increased activity in both PT and TAL cells of TCMR. AP-1, known for its regulatory role in numerous immune pathways, may be integral to the immune response observed in TCMR samples [ 33 ]. Among the factors potentially suppressed in TCMR, WT1, ZNF148, EGR1/2/4, SMAD5, and TFAP2D emerged as significant candidates in both PT and TAL cells ( Figure 4C ). Notably, EGR1 was identified as TF with downregulated activity in both BKPyVAN and TCMR samples ( Figure 4B,C ), indicating that EGR1 inhibition was a common consequence of immune dysfunction in the kidney. In addition, our results revealed analogous alterations to those observed in PT cells, indicating comparable molecular perturbations in these two cell types ( Figure 4C ). Finally, to distinguish the unique transcriptional and epigenetic signatures between BKPyVAN and TCMR samples, we conducted a comparative analysis between these two samples. We identified HNF4A and HNF4G as upregulated regulators in the PT cells of the BKPyVAN group. These factors regulate PT differentiation and might possess potential roles in response to cell injury in fully mature PT. It indicated that injuries in PT may more actively be repaired with infection than rejection. In TAL cells, we observed an enrichment of NR3C1, a steroid hormone receptor, in the BKPyVAN group, pointing to unmet steroid hormone demand in TAL cells upon infection [ 34 ].

The local ethics committee of The First Affiliated Hospital of Zhejiang University, School of Medicine approved all human tissue protocols (20240825 A). Kidney tissues were collected from patients undergoing kidney biopsy during nephrectomy due to kidney donation (Control group) or after kidney transplantation. All patients provided informed consent and the study was performed in accordance with the Declaration of Helsinki. From September 2023 to December 2023, The First Affiliated Hospital of Zhejiang University, School of Medicine, totally 18 cases were included. Control group was defined as normal kidneys. The samples were obtained from donor kidneys before nephrectomy. According to the Banff Classification for Renal Allograft Pathology, 2022 [ 18 ], 7 cases involved kidney grafts were diagnosed with TCMR (IA or IB). As for BKPyVAN, the 6 cases involved were limited to Biopsy-Proven Class 1 to Class 2 [ 19 ] or Stage A to Stage B2 [ 20 , 21 ]. Biopsy samples were cut into small pieced and digested as described before [ 22 ]. Nuclear dissociation followed Humphreys et al. [ 23 ]. Nuclei suspensions are incubated in a transposition mix that includes a transposase. Gel beads-in-emulsion (GEMs) are generated by combining barcoded gel beads, transposed nuclei, a master mix, and partitioning oil on a Chromium Chip E. To achieve single nuclei resolution, the nuclei are delivered at a limiting dilution, such that the majority (∼90-99%) of generated GEMs contains no nuclei, while the remainder largely contain a single nucleus. Silane magnetic beads are used to remove leftover biochemical reagents from the post GEMs reaction mixture. Solid Phase Reversible Immobilization (SPRI) beads are used to eliminate unused barcodes from the sample. Single-cell ATAC library was generated following the mentioned 10x Genomics instructions. The Chromium Single Cell ATAC protocol produces Illumina ® -ready sequencing libraries. Cellranger-atac (v 2.0.0) was used to align fragments to the reference genome (GRCh38). The tabix sorted text file containing fragment start and end positions and the corresponding cell barcodes served as input for downstream analysis. ArchR (v 1.0.1) were used to perform downstream analysis [ 24 ]. The cell with a unique fragment of less than 1000 or with a transcription start sites enrichment score (TSS Score) of less than 4 was defined as a low-quality cell and was discarded. The function addIterativeLSI of ArchR was used to perform dimensionality reduction. Latent Semantic Indexing (LSI), an approach from natural language processing, was used in this function. In brief, the term frequency was calculated with normalization for sequencing depth per cell. The resulting matrix were used to calculate the term frequency-inverse document frequency (TF-IDF) matrix. Then, through singular value decomposition (SVD), the important information across all sample were conducted. The function addHarmony in ArhcR were used to correct the batch effect across samples. T-Stocastic Neighbor Embedding (t-SNE) was used to visualize single cells in reduced dimension space. The function addClusters in ArhcR, in which Louvain algorithm was embedded, were used to cluster the cells. Gene score was estimated based on accessibility of gene-encoding regions and regulatory elements. The estimation was achieved by function addGeneScoreMatrix () in ArchR. Then, the marker feature was generated by function getMarkerFeatures () in ArchR with FDR  = 1. Pseudo-bulk replicates were generated using function addGroupCoverages in ArchR for each cluster. Peaks calling was perform by using addReproduciblePeakSet in ArchR. Among those peaks, the marker peaks were identified by function getMarkers in ArchR with FDR  = 1. The marker peaks were then used to perform motif enrichment. The significant enriched transcription factor was generated by using function peakAnnoEnrichment in ArchR with FDR  = 1/-1. Clusterprofiler (v 4.0.5) was used to perform GO function enrichment and KEGG pathway annotation.

The accurate diagnosis of TCMR versus BKPyVAN is pivotal for prompt clinical intervention [ 9 ]. However, previous bulk transcriptomic analyses of TCMR and BKPyVAN samples have failed to capture their distinctive molecular signatures [ 14 ]. To address this, we exploited scATAC-seq to characterize the chromatin landscapes of TCMR and BKPyVAN samples at the single-cell level. Our findings not only elucidate the epigenetic traits of both TCMR and BKPyVAN but also offer insights into the upstream molecular mechanisms that shape the unique epigenetic profiles of each cell type within these conditions. In our analysis of TCMR and BKPyVAN biopsies, we observed differences in cell populations compared to control samples ( Figure 2C ). Specifically, we observed an increased presence of B cells in the TCMR group ( Figure 2D ), which underscores a potential role of B cells in TCMR. While TCMR is traditionally attributed to T cell-mediated mechanisms, emerging evidence highlights the significance of B cells in this specific scenario [ 3 , 35 ]. B cells facilitate T cell activation by serving as antigen-presenting cells, further exacerbating renal damage through this intricate interplay [ 36 ]. Our results provide additional evidence supporting the role of B cells in TCMR. The key role of T cell immunity in the control of BKPyV infection was initially suggested by the indirect evidence of increased incidence of reactivation and clinical disease related to the degree of immune suppression [ 25 ]. Our results showed an increase in T cell population in BKPyVAN samples compared to control samples ( Figure 2D ), which also supports the role of T cells in mediating the immune response upon BKPyV infection. Using scATAC-seq, we have gained insights into the molecular effects of BKPyVAN and TCMR on different kidney cell populations. Our analysis revealed that the primary differences in BK infection are concentrated in PC, PT, and TAL ( Figure 3A ), aligning with previous understandings that BK infection primarily affects the medulla [ 37 ]. BKPyV infection can be categorized into several patterns, including inflammatory, immune reconstitution, autoimmune, and transforming patterns [ 7 ]. Malignant transformation is known as a transforming pattern. The role of BKPyV infection in certain malignancies is still a subject of debate. It is acknowledged that BKPyV-associated tumor lesions are primarily of urothelial origin. However, it is intriguing that we observed carcinogenesis-related genes in the TAL modulated by BKPyV ( Figure 3C ). This finding raises questions about the potential role of BKPyV in carcinogenesis beyond urothelial tissues. Further investigation is warranted to confirm these carcinogenic effects and explore their potential role in BKPyVAN. HLA-DQB1 plays a key role in antigen presentation and immune regulation, with previous studies linking certain alleles to viral susceptibility [ 38 , 39 ]. In our study, we observed increased chromatin accessibility at the HLA-DQB1 locus during BKPyVAN, suggesting its involvement in immune modulation. Given that BK polyomavirus infection is a major complication in allogeneic kidney transplantation, this finding highlights a potential role for HLA-DQB1 in shaping immune responses, influencing viral reactivation, and contributing to alloreactivity under immunosuppression. Further research may provide insights into its clinical implications for transplant recipients. The primary differences in chromatin accessibility between TCMR and BKPyVAN were observed in metabolism-associated genes within endothelial cells (ENDO) ( Figure 3I ). This finding is particularly intriguing, as all TCMR cases in this study were classified as IA or IB TCMR according to Banff 2022 criteria, which lack endothelial changes detectable by pathologists. However, our profiling captured molecular impacts on ENDO that highlight metabolic alterations. Specifically, lipid and carbohydrate-related pathways were shown to be upregulated in ENDO cells of TCMR samples. These metabolic pathways have been reported to be proinflammatory [ 40 ]. It is, therefore, worth investigating the function of these metabolic rewiring in ENDO and inflammation responses in kidneys upon TCMR and BKPyVAN. Cell-type-specific transcription factor activity analysis revealed the potential upstream regulators. As PT and TAL exhibited the most pronounced chromatin accessibility variations between TCMR and BKPyVAN, a transcription factor motif enrichment analysis was conducted in these two cell types. Compared with the control group, there were separately different enriched transcription factors in TCMR and BKPyVAN groups ( Figure 4B–D ). These results shed light on identifying molecular mechanisms regulating TCMR and BKPyVAN. For instance, a unique enrichment of NR3C1 in the TAL cells of the BKPyVAN group suggests an unmet steroid hormone demand in infection. The control group included in this study is normal kidney samples instead of normal allograft ones. The comparison between the pathological group and control group, therefore, reflected general changes after transplantation, including pathological effect and immunosuppressive effect. While including a group with normal allografts would be ideal for dissecting pathological effects, routine biopsies are limited, and most biopsies are indicated. Further exploration of normal kidneys, normal allografts, and pathological kidneys is required to comprehensively clarify how immunosuppressive treatment and pathological conditions, respectively, modify chromatin accessibility. Another limitation of this study is the small sample size and no validation, making the findings can only be considered preliminary at this point. In summary, our study offers a comprehensive chromatin accessibility atlas for TCMR and BKPyVAN kidney biopsies, pinpointing the molecular differences between these two pathological conditions. These results can be further exploited to develop novel diagnostic strategies to distinguish between TCMR and BKPyVAN in clinical settings.