Intra- and Inter-Organ Endothelial Cell Differences Contribute to TA-TMA through HMOX1, Hypoxia and Reactive Oxygen Species Signaling

Introduction Transplant-associated thrombotic microangiopathy (TA-TMA) outcomes are poor in those with multiorgan disease, particularly gastrointestinal (GI) TA-TMA. We hypothesized that differences in endothelial cell responses to circulating soluble factors determine organ injury patterns in TA-TMA. There is no animal model of TA-TMA so we developed a patient-derived in vitro model using primary organ endothelial cells incubated with patient serum collected at TA-TMA diagnosis. Objectives We sought to identify how intra- and inter-organ endothelial cell diversity contribute to TA-TMA. Methods We tested our hypothesis in two ways. First, we performed single cell RNAseq to identify whether specific kidney endothelial cell subtypes are important in TA-TMA. Next, we performed bulk RNAseq on endothelial cells from kidney and GI microvasculature to test how inter-organ endothelial cell diversity contributes to TA-TMA. Sera from 29 hematopoietic stem cell transplant (HSCT) recipients (15 with TA-TMA and 14 without TA-TMA) were used. Results Single cell RNAseq showed 3 distinct kidney endothelial cell clusters, confirming that endothelial cell subtypes exist in our model (Fig. 1A). Endothelial cells cultured with TA-TMA serum were transcriptomically different in cluster 1 compared to cells cultured with control serum. A comparison of TA-TMA vs control serum exposed cells in cluster 1 generated 93 differentially expressed genes (DEGs). The top DEGs are shown in Fig.1B and notably include HMOX1 (padj=6e-159). Pathway differences in complement and coagulation (known to be critical in TA-TMA) were observed as well as novel differences in hypoxia, mesenchymal transition and mTORC1 signaling. UMAP clusters 2 and 3 showed similar cell distribution in TA-TMA and control groups, indicating that TA-TMA related differences occurred in a specific endothelial cell subtype.Bulk RNAseq of GI endothelial cells showed DEG overlap with kidney endothelium but also found organ-specific differences. The top 25 DEGs in GI endothelial cells cultured with TA-TMA sera vs controls are shown in Fig. 2A. HMOX1 (padj=2.3e-5) is a leading DEG and network analysis of all 85 DEGs identified protein-protein interactions between HMOX1, GCLM, TXNRD1 and NQO1 (Fig. 2B). Notably, complement and coagulation are not seen as significant pathways in GI cells (Fig. 2C). In contrast, reactive oxygen species (ROS) signaling was significantly different in both GI endothelium and kidney endothelium, suggesting that complement independent pathways may be important in GI TA-TMA and may contribute to known poor responses to complement blockade. Conclusion These data show that our model is a powerful tool for elucidating organ-specific TA-TMA mechanisms. This model identified targetable complement-independent pathways and endothelial cell subtypes that are being further explored in GI TA-TMA.