Bioinformatic analysis of peripheral blood RNA-sequencing sensitively detects the cause of late graft loss following overt hyperglycemia in pig-to-nonhuman primate islet xenotransplantation

Abstract

Clinical islet transplantation has recently been a promising treatment option for intractable type 1 diabetes patients. Although early graft loss has been well studied and controlled, the mechanisms of late graft loss largely remains obscure. Since long-term islet graft survival had not been achieved in islet xenotransplantation, it has been impossible to explore the mechanism of late islet graft loss. Fortunately, recent advances where consistent long-term survival (≥6 months) of adult porcine islet grafts was achieved in five independent, diabetic nonhuman primates (NHPs) enabled us to investigate on the late graft loss. Regardless of the conventional immune monitoring methods applied in the post-transplant period, the initiation of late graft loss could rarely be detected before the overt graft loss observed via uncontrolled blood glucose level. Thus, we retrospectively analyzed the gene expression profiles in 2 rhesus monkey recipients using peripheral blood RNA-sequencing (RNA-seq) data to find out the potential cause(s) of late graft loss. Bioinformatic analyses showed that highly relevant immunological pathways were activated in the animal which experienced late graft failure. Further connectivity analyses revealed that the activation of T cell signaling pathways was the most prominent, suggesting that T cell-mediated graft rejection could be the cause of the late-phase islet loss. Indeed, the porcine islets in the biopsied monkey liver samples were heavily infiltrated with CD3+ T cells. Furthermore, hypothesis test using a computational experiment reinforced our conclusion. Taken together, we suggest that bioinformatics analyses with peripheral blood RNA-seq could unveil the cause of insidious late islet graft loss.

Long-term control of diabetes in a nonhuman primate by two separate transplantations of porcine adult islets under immunosuppression

“…Consistent with our results, Hering et al reported massive infiltration of CD4 + and CD8 + T cells at the rejected graft sites in a pig‐to‐NHP islet transplantation model 1 . In addition, peripheral blood RNA‐seq analysis previously performed by our group also suggested that the T cell‐signaling pathway was actually activated in the monkeys exhibiting graft rejection 14 . With regard to humoral immune response, moderately increased anti‐Gal IgG and IgM, anti‐WT PEC IgG and IgM, and anti‐donor PBMC IgG and IgM levels were observed before the retransplantations (Figure 5).…”

“…To evaluate the mechanisms underlying the gradual loss of func- In addition, peripheral blood RNA-seq analysis previously performed by our group also suggested that the T cell-signaling pathway was actually activated in the monkeys exhibiting graft rejection. 14 With regard to humoral immune response, moderately increased anti-Gal IgG and IgM, anti-WT PEC IgG and IgM, and anti-donor PBMC IgG and IgM levels were observed before the retransplantations (Figure 5). Furthermore, C4d deposition in the transplanted islets was observed in the first and second graft rejections, indicating that humoral rejection was the likely cause.…”

Protein‐protein interaction networks as miners of biological discovery

“…Moon et al. developed PINTnet that integrates PPI network from STRING, KEGG pathway, and user input RNAseq databases [142]. With the assumption that signals flow through active nodes in pathway interactions, the edges and interaction scores are computed from subnetworks based on closeness centrality and shortest path [141].…”

“…With the assumption that signals flow through active nodes in pathway interactions, the edges and interaction scores are computed from subnetworks based on closeness centrality and shortest path [141]. In a study focused on the mechanism of late graft loss in Type 1 diabetes patients, PINTnet was employed to detect key signaling pathways from peripheral blood RNA‐seq databases [142]. Another framework SPECTRA (Specific Tissue/Tumor Related PPI Networks Analyzer) integrates PPI and gene expression from thirteen public repositories including HPRD, IntAct, BioGRID, MINT etc.…”

“…Condition-specific networks are usually constructed based on the topological features in known PPI networks and other biological information. Moon et al developed PINTnet that integrates PPI network from STRING, KEGG pathway, and user input RNAseq databases [142].…”