Swine Xenografts Share Few Predicted Indirectly Recognisable SLA-Derived Epitopes With HLA-Derived Epitopes From Human Kidney Grafts

Abstract

Swine-derived kidneys are a promising alternative organ source for transplantation, but compatibility in the major histocompatibility complex remains an immunological barrier. Furthermore, in repeat transplantations, CD4+ memory T cells can lead to a more rapid immune response against repeated exposure to the same antigens. Several studies have shown that HLA and SLA proteins share overlapping B cell epitopes due to structural or electrostatic similarities, but the role of overlapping T cell epitopes has not been fully explored. This study aims to computationally analyse the potential risk of memory T cell activation in subsequent human-after-swine and swine-after-human transplantation by evaluating shared T cell epitopes between the two graft sources. We show that while HLA and SLA demonstrate striking structural similarities, their linear protein sequences are very distinct, which translates to disparate HLA- and SLA-derived peptidomes and T cell epitopes. By applying the PIRCHE-II Tmem analysis to a simulated panel of recipients receiving repeat transplantations from a human kidney and from a swine xenograft, we observed a median of 1 shared T cell epitope in the cross-species context, compared to a median of 17 shared between two human-derived kidneys. This suggests that a swine xenograft exposes a low risk of T cell memory against a later human donor, and that xenotransplantation may provide an opportunity to receive a graft for highly HLA-sensitised recipients.

Keywords: PIRCHE; SLA; memory; swine; xenotransplantation.

FIGURE 1

Phylogenetic tree of representative HLA and SLA protein sequences. Phylogenetic trees show comparative similarity of representative MHC Class I amino acid sequences. All HLA‐A, ‐B and ‐C sequences (blue boxes and labels) are more similar to each other than any of the SLA sequences, and all SLA ‐1, ‐2, and ‐3 (pink boxes and labels) sequences are more similar to each other than any of the HLA sequences. Phylogenetic trees of select MHC Class II Alpha chain and Beta chain are shown in Figures S1A and S1B, respectively.

FIGURE 2

A representative alignment of an HLA and SLA protein structure. A FATCAT alignment of PDB structures 7EMA (PDB DOI: 10.2210/7ema/pdb, SLA‐1*01:01) and 1QVO (PDB DOI: 10.2210/1qvO/pdb, HLA‐A*11:01) is shown. The protein structures mostly overlap, indicating a high degree of structural homology.

FIGURE 3

Structural MHC similarity for various matching contexts. The “Adjusted Similarity Score” (SS divided by aligned sequence length) from pairwise structure alignments. Density curves are stratified based on if the two crystal structures are the exact same PDB crystal structure (green), derived from the same MHC protein (red), from the same human or swine MHC locus (blue), separate loci from the same species (pink) or are from separate species (yellow). While more similar MHC molecules expectedly show more similar protein structures, in all cases the comparisons fall within overlapping range.

FIGURE 4

Heatmap of average adjusted structural similarity per MHC protein. Mean adjusted structural similarity scores from all structure alignments are stratified by the source MHC allele. We observe generally higher similarity for identical structures (diagonal) and for comparisons of alleles from within the same locus, with lower structural similarity for comparisons across species (darker blue).

FIGURE 5

PIRCHE‐T2 and PIRCHE‐Tmem score density for human‐ & swine‐derived transplantations. (A) Distributions of PIRCHE‐T2 risk scores for the two kidney origin species. Distributions from swine‐derived transplants demonstrate a higher predicted risk than human‐derived transplants. (B) Distributions of PIRCHE‐T2 risk scores across the simulated donor pool for the optimum (lowest and highest median scores) patients. One patient was found with the lowest median PIRCHE‐T2 score (13) and two patients were identified with the same worst‐case median PIRCHE‐T2 score (84). These best‐ and worst‐case patients' corresponding scores from swine‐derived grafts (73, and 242&246, respectively) are shown as dotted lines. The swine‐derived transplant has a much higher predicted risk than all human‐derived grafts for all patients in our simulated panel, exemplified in the optimum patients. (C) Distributions of PIRCHE‐Tmem scores for repeat transplantations. PIRCHE‐Tmem scores are higher in two subsequent human grafts, compared with cross‐species grafts.