Multi-omics analysis of a pig-to-human decedent kidney xenotransplant

Abstract

Organ shortage remains a major challenge in transplantation, and gene-edited pig organs offer a promising solution^1-3. Despite gene editing, the immune reactions following xenotransplantation can still cause transplant failure⁴. To understand the immunological response of a pig-to-human kidney xenotransplantation, we conducted large-scale multi-omics profiling of the xenograft and the host's blood over a 61-day procedure in a brain-dead human (decedent) recipient. Blood plasmablasts, natural killer cells and dendritic cells increased between postoperative day (POD) 10 and 28, concordant with an expansion of IgG and IgA B cell clonotypes and subsequent biopsy-confirmed antibody-mediated rejection (AMR) at POD33. Human T cell frequencies increased from POD14 and peaked between POD33 and POD49 in the blood and xenograft, which coincided with T cell receptor diversification, expansion of a restricted TRBV2 and TRBJ1 clonotype and histological evidence of combined AMR and cell-mediated rejection at POD49. At POD33, the most abundant human immune population in the graft was CXCL9⁺ macrophages, which aligned with interferon-γ-driven inflammation and a T helper 1-type immune response. There was also evidence of interactions between activated pig-resident macrophages and infiltrating human immune cells. Xenograft tissue showed pro-fibrotic tubular and interstitial injury marked by S100A6 (ref. ⁵), SPP1 (also known as osteopontin)⁶ and COLEC11 (ref. ⁷) expression at POD21-POD33. Proteomic profiling revealed activation of human and pig complement, with a decreased human component after AMR therapy, in which complement was inhibited. Collectively, these data delineate the molecular orchestration of human immune responses to a porcine kidney and reveal potential immunomodulatory targets for improving xenograft survival.

Extended Data Fig. 1:. Further dissection of human macrophages, NK cells, B cells and dendritic cells.

A-B. Dimensionality reduction graph (A) and percentage across all cells by timepoint (B) for human macrophage, monocytes and dendritic subtypes found in the 5.1k panel ST data colored by cell-type (11205 cells). C-D. Dimensionality reduction graph (C) and percentage across all cells by timepoint (D) for B cells, plasma cells, mast cells and neutrophils found in the 5.1k panel ST data (1652 cells). E-G. Marker gene expression of the populations in the 5.1k panel ST data. H. Percentage levels for additional cell-type populations not shown in the main figure. I-J. Cell-type markers (I) and associated percentage (J) of B, plasma, and dendritic cell populations found in the PBMC scRNAseq data. K. Percentage of human macrophages, pDC and NK cells found in the 478 panel ST data. L-M. pDC activation marker genes and CXCL9 – CXCL11 expression in pDC-cDC1 populations in 5.1k ST data (L) and PBMC scRNAseq data (M).

Extended Data Fig. 2:. Early response from human macrophages and B cells as evidenced by ST and snRNAseq in addition to BCR-seq.

A. Expression of immunoglobulin genes in PBMC bulk RNAseq. B. CDR3 length (nt) distribution in bulk BCRseq. C. Somatic hypermutation frequency in bulk BCRseq divided by isotype. D. Shared overlapping clonotypes across bulk BCRseq postoperative timepoints. E. Tracking top BCR IgH clones in bulk BCRseq postoperative time course. F-I. 478 panel ST data highlighting human subpopulations of macrophages and NK cells expressing CXCL9, CXCL10, and CXCL11 as shown in a dimension reduction map (F), distribution of these markers across those subpopulations (G) and across time in the NK - MP populations (H) and percentage distribution of these subpopulations across time (I). J-K. Similar populations of interest in snRNAseq data, as shown in expression of markers corresponding to the 5.1k panel data subtypes markers (J) and their distribution over time (K). L. Expression of human CXCL9, CXCL10, CXCL11 observed in the tissue bulk RNAseq.

Extended Data Fig. 3.. Further human T cell response dissection.

A-C. Human T cell subtypes found in 5.1k ST data (2595 cells), as shown via dimension reduction plot (A), percentage across all cells in each timepoint (B), and marker genes defining these cells (C). D. Cell-type markers of PBMC NK and T cells populations. E. Proportion of proliferative human NK cells found in the 5.1k ST data. F-G. Proportions (F) and marker gene expression (G) of human CD8T and CD4T from the 478 panel ST data. H. Distribution of T and NK subtypes found in the PBMC scRNAseq data, as percentage of all cells across timepoints (for subtypes not shown in Fig. 4). (H). I. Percentage of human Dividing and T cells found in snRNAseq data. J. Expression levels of CD8T markers (CD8A, CD8B, top) and Tregs markers (RTKN2, CTLA4, bottom), from PBMC bulk RNAseq. Treatment with rATG is annotated with red marks. K. Flow cytometry T cells distributions, CD8+ T cells (middle) and CD4+ T cells (bottom), separated by central memory and effector memory cells.

Extended Data Fig. 4.. Distribution of marker genes delineating resident and circulating T and NK cells.

A. NK markers, 5.1k panel ST data. B. T cell markers, 5.1k panel ST data. C. NK markers, PBMC scRNAseq data. D. T cell markers, PBMC scRNAseq data. E-F. Dimension reduction map showing NK subtypes and circulatory and residency markers in 5.1k ST data (E, 2595 cells) and in 478 ST data (F, 606 cells). G-H. Expression of circulatory and residency markers across NK subtypes in 5.1k ST data (G) and in 478 ST data (H).

Extended Data Fig. 5.. Porcine transcriptional response from tissue and porcine resident immune cells dissection.

A. An inflammatory gene cluster identified by bulk RNAseq longitudinal analysis. Expression over time (left) and top pathway enrichment (right). B. Expression distribution of selected time-point specific marker genes (pig probes) in the 478 ST panel. C-D. Macrophage and T-cell marker expression across cell-types in the 5.1k panel (C) and 478 panel (D) ST data. E. T-cell marker expression across timepoints, shown in pig immune cells in the 5.1k panel (top) and 478 panel (bottom) ST data. F. Top 80 marker genes for POD33 pig immune cells, identified by Wilcoxon rank-sum analysis comparing their transcriptomic profile to all other time points. G-H. Expression of marker genes of interest across timepoints in specific pig cell populations, in snRNAseq samples of the xenograft (G) and in 5.1k ST data (H).

Extended Data Fig. 6:. B-HOT AbMR signature revealed by ST.

A–D: B-HOT AbMR signature expression in the 478-gene ST panel, using genes identified as differentially expressed (DEG post- vs. pre-transplantation) in xenografts from a previous study^, E–G: B-HOT AbMR signature expression in the 5.1k-gene ST panel, using the entire B-HOT panel, regardless of overlap^, results. A. Spatial distribution of the B-HOT AbMR signature in POD33 tissue. B. B-HOT signature enrichment in UMAP, overlaid with cell-type composition. C. AbMR signature enrichment across time points. D. Marker expression of B-HOT AbMR genes. E. Expression of genes (human probes) belonging to the B-HOT AbMR signature across timepoints in all human cells. F. Expression of pig genes whose human orthologues are part of the B-HOT panel, across timepoints in all pig cells. G. Spatial distribution in POD33 of the AbMR gene set score, pig MX1, and pig SERPINE1 expression.

Extended Data Fig. 7:. Neighborhood enrichment analysis of spatial transcriptomics data.

A. Analysis based on the 478-gene ST panel. B. based on the 5.1k-gene ST panel. Each subplot corresponds to a selected reference cell-type (indicated above each heatmap). Heatmaps display neighborhood enrichment z-scores, quantifying the spatial colocalization between the reference cell-type (per subplot) and all other cell-types (columns), across all samples (rows). Enrichment scores were computed using a permutation test (1,000 permutations) via Squidpy. The legend scale ranges from −10 to +10. White stars denote z-score ≥ 1.96 or ≤ −1.96.

Extended Data Fig. 8:. Spatial Niche Profiling in the 478 ST Data.

A. Cell-type annotations from 478-gene ST panel on POD33 biopsy. B. Depiction of spatial niches from neighborhood analyses on cell-types from (A). C. Relative cellular contribution to spatial niches by timepoint from which cells belong. D. Distribution of cells within spatial niches at each timepoint. E. Niche residency of indicated cell-type, represented as frequency of total cells of indicated type. F. Heatmap with clustering of cell-types and spatial niches at POD33.

Extended Data Fig. 9:. Spatial Niche Profiling in the 5.1k ST Data.

A. Cell-type annotations from 5.1k gene panel on POD33 biopsy. B. Depiction of spatial niches from neighborhood analyses on cell-types from (A). C. Relative cellular contribution to spatial niches by timepoint from which cells belong. D. Distribution of cells within spatial niches at each timepoint. E. Niche residency of indicated cell-type, represented as frequency of total cells of indicated type. F. Heatmap with clustering of cell-types and spatial niches at POD33.

Figure 1:. Study overview.

A schematic of the study is shown, indicating the major modalities (top) as well as the tissue and blood timepoints, assays utilized at each respective timepoint, treatments/medications utilized, and clinical microbiology testing performed (bottom) in the 61-day period following the pig-to-human Gal-KO thymokidney xenotransplant. The “CTRL” samples from the 5.1k ST panel data correspond to Contralateral pig kidney (pig-only cells) and Native human kidney (human-only cells), which act as positive and negative control for human cell selection in our analysis. BAL: bronchoalveolar lavage, EBV: Epstein-Barr virus, Gal-KO: α-1,3-galactosyltransferase (GGTA1) knockout, HBV: hepatitis B virus, rATG: rabbit anti-thymocyte globulin. Created in BioRender, by Dowdell, A. (2025) https://BioRender.com/zkz9y0q

Fig. 2:. High-resolution spatial-transcriptional profiling reveals infiltrating human cells in porcine kidneys.

A. Analytical workflow for spatial high-resolution tissue transcriptomic profiling (Xenium); examples are from POD49. B. Separation of pig (orange) and human (blue) nuclei based on transcript species origin (sample POD33 is shown as an example). C. Human/pig gene expression distribution among cells profiled with the 5.1k ST panel, colored by their assigned species (or doublets). The x-axis represents the mean log-norm expression of pig genes and the y-axis the mean log-norm expression of human genes. D. Spatial transcriptomics cell-type labeling of the transplanted tissue H&E image (i.), and corresponding spatial transcriptomics DAPI fluorescence image (ii.) with segmentation-based cells colored by cell-type. E. Human cells infiltrating the renal cortex are shown with their cell-type annotation and associated markers (POD49). F. Glomerulus and infiltrating human cells with cell-type annotation and associated markers (POD49). G-L. Dimension reduction map of cells identified in the xenograft tissue, colored by identified cell-type, for human (G, I, K), and pig (H, J, L) cells, in the 478 Xenium panel (G-H, 3417 and 57760 cells), the 5.1k Xenium panel (I-J, 14771 and 320702 cells) and snRNAseq (K-L, 128889 and 2229 nuclei) modalities. M. Dimension-reduction map of cells identified in recipient PBMC scRNAseq, colored by cell-type. Left: main cell-types (132597 cells), right-up: T and NK cells (22422 cells), right-down: B cells, plasma cells, dendritic cells (4848 cells). Panels A, D, E, F show data from the 478 ST panel. MP: Macrophages, CD8T: CD8+ T cells, CD4T: CD4+ T cells, NK: NK cells, pDC: plasmacytoid dendritic cells, FB: Fibroblasts, EC: Endothelial cells, SMC: Smooth muscle cells, vSMC: vascular smooth muscle cells, fSMC: fibroblast-associated smooth muscle cells, VEC: vascular endothelial cells, cDC: classical dendritic cells, TEM: T effector memory cells, HSC: Hematopoietic stem cells, TCM: T central memory cells, MAIT: Mucosal-Associated Invariant T cells, Treg: Regulatory T cells.

Fig. 3:. Early host immune cell response involving macrophages, NK cells, pDCs, and plasma/B cells.

A. Percentage of human cells identified by spatial transcriptomics in the xenograft tissue (5.1k ST panel) across all identified cells. B. Plasmablast levels measured in the recipient PBMC scRNAseq (top), and B marker expression levels from PBMC bulk RNAseq (bottom). C. Human B cell levels measured by spatial transcriptomics in the xenograft tissue (5.1k ST panel, top) in the recipient PBMC scRNAseq (middle); and marker expression levels identified by PBMC bulk RNAseq (bottom). D. Human NK cell levels measured by spatial transcriptomics in the xenograft tissue (5.1k ST panel, top), in recipient PBMC scRNAseq (middle), and by PBMC bulk RNAseq marker expression (bottom). E. Flow cytometry levels of NK and B cells. F. Human pDC levels measured by spatial transcriptomics in the xenograft tissue (5.1k ST panel, top), in recipient PBMC scRNAseq (middle), and by PBMC bulk RNAseq marker expression (bottom). G. Quantification of unique BCR clones from bulk BCRSeq divided by clonal expansion level. H. Quantification of unique BCR clones from bulk BCRSeq divided by BCR isotype family. I. Percentage of human macrophages (top), CXCL9⁺ macrophages (middle), and Interferon⁺ immune cells (bottom) in the xenograft tissue (5.1k ST panel). J. Gene set enrichment analysis (GO Biological Process) of cell-type markers for CXCL9⁺ macrophages (top) and Interferon⁺ cells (bottom). Bars show adjusted P values. Statistical significance was determined using the hypergeometric test with Benjamini–Hochberg FDR correction. K. Expression of CXCL9, CXCL10, and CXCL11 from the infiltrating human immune cells in the xenograft (from 5.1k ST data). ST: spatial transcriptomics, NK: Natural Killer cells, pDC: plasmacytoid dendritic cells. Therapies (rATG, PLEX, rituximab) are shown at the corresponding time points and next to relevant cell types.

Fig. 4:. Human T-cell response to xenotransplantation.

A-D. Percentage of T cell subpopulations in the PBMC scRNAseq data. This includes CD8 TEM (A), Naive CD8 T (B), CD4 TCM-TEM (C) and Tregs (D). E-H. Percentage of T cell subtypes observed in the 5.1k panel ST data, including CD8 TEM (E), Naive CD8T (F), CD4T (G) and Tregs (H). I. Percentage of all T cells from blood scRNAseq data. J-L. T-cell subtypes marker expression in blood bulk RNAseq data, for CD8T markers CD8A and CD8B (J), the CD4T marker TSHZ2 (K) and Treg markers RTKN2 and CTLA4 (L). M-P. T-cell levels in flow cytometry data, for all T cells (M). Q. T and NK cell subtypes expression of ITGAE (CD103), CD38 and TNFRSF9 across data modalities. Samples that have less than 5 cells of the population in question are excluded from the graph. R. Quantification of unique TCR clones from bulk TCRseq divided by clonal expansion level S. V+J gene expression levels of TCR Beta clones from bulk TCRseq T. Tracking top TCR Beta clones in bulk TCRseq postoperative time course. TEM: T effector memory cells. TCM: T central memory cells, Treg: Regulatory T cells, UMI: Unique molecular identifier, NK: Natural killer cells. ST: Spatial Transcriptomics. Treatment with rATG is annotated with red and orange marks on the graphs.

Fig. 5:. Transplanted porcine tissue response highlights the damage signaling at POD21 and POD33.

A. Expression of inflammatory related genes (FOS, FOSB, CXCL2) in the tissue xenograft porcine cells (from the 478 ST panel). B. Proportion of COLEC11+ cells across timepoints (5.1k panel). C. Proportion of SPP1+ cells across timepoints (478 panel). D-E. Expression of markers of interest for SPP1+ cells in the 478 panel (D) and COLEC11+ cells in the 5.1k panel (E). F. POD21 biopsy region corresponding to high levels of SPP1+ cells (478 panel). G. POD33 biopsy (H&E and ST) with high levels of human immune cells (478 panel). H. POD33 biopsy highlighting expression of pig and human genes of interest near human cells and COLEC11+ cells (5.1k panel). FB: Fibroblasts, EC: Endothelial cells, MP: Macrophages, NK: Natural Killer cells, pDC: Plasmacytoid Dendritic Cells, CD8T: CD8+ T cells, CD4T: CD4+ T cells, fSMC: fibroblast-associated smooth muscle cells, vSMC: vascular smooth muscle cells, SMC: Smooth muscle cells, ST: spatial transcriptomics.

Figure 6:. Dynamics of selected complement pathway elements protein levels in the serum.

Serum abundance over time for the indicated complement factors. Statistical analysis of protein regulation was based on XT2 Proteograph XT measurements of serum samples. The middle line, cyan-colored, and dashed line bands correspond to the median and 50% and 95% credible intervals of the protein intensity posterior distribution, respectively. Treatment with pegcetacoplan is annotated in the graph.