Introduction: The field of transplantation is in critical need of more accurate tools to predict allograft outcomes. A transcriptomic profile serves as a snapshot of the temporary cell state and thus, analyzing pretransplant kidney biopsies can provide detailed information on the early biological processes associated with posttransplant outcomes. Currently, there are no established predictive biomarkers for posttransplant kidney function.

Methods: Gene expression measurements were performed using Affymetrix GeneChip microarrays (HG-U133A 2.0) in the training set (n=174). Patients with low function had an eGFR <45 mL/min/1.73m2 (avg slope = -6.36 ml/min/1.73m2/year) and those with high function had an eGFR ≥45 mL/min/1.73m2 (avg slope= 0.81 ml/min/1.73m2/year) at 24 months. To identify differentially expressed genes (DEGs) and clinical characteristics associated with 24-month graft function, linear models were fit adjusting for surrogate variables. The performance of DEGs and clinical characteristics was assessed using the area under receiver operating characteristic curves (AUROC). A subset of candidate genes was measured in an independent set of pretransplant biopsies (n=96) using qPCR. A risk score equation was derived from the predictive model and performance was compared to KDPI. 

Results: Donor age, race, and BMI correlated with 24-month function (p<0.05). No recipient characteristics were statistically significant between groups. Nearly 700 DEGs were found to be associated with 24-month outcomes (FDR<0.05). Grafts with low 24-month function showed upregulated genes related to immune responses (e.g., CD3D, CD69, CCL5, CXCR4, C1QB, FCER1G, ILR7) and downregulated genes related to metabolic/tubular functions (e.g., ENO1, FH, POGK, SLC22A8, SL4A4, WDR1) (Fig 1A). Cell-type enrichment analyses identified monocytes, dendritic, myeloid, and natural killer cells as the main cell sources for the upregulated genes (Fig 1B). Using a 55 gene model adjusted for donor age, race, and BMI resulted in an AUROC=0.996 (KDPI AUROC=0.705). In the validation set, 13 genes + 3 donor characteristics (G+D) achieved an AUROC=0.821 (KDPI AUROC=0.691) (Fig 1C). Using a probability equation derived from the G+D model to predict low 24-month function, the sensitivity was 88.9% and the specificity was 66.6%. When using KDPI, the sensitivity was 80.6% and the specificity was 53.3% (Fig 1D). 

Conclusion: This study represents the largest high-throughput transcriptomic analysis of pretransplant donor kidneys predicting 24-month outcomes. A predictive risk score was developed, which combines donor age, race, BMI, and donor quality gene markers, and can be calculated prior to transplantation to predict graft function. Differential pretransplant transcriptional profiles between kidneys with low and high function at 24-months were also identified, providing a deeper insight into the early biological processes leading to graft dysfunction.