GENOMICS AND BIOMARKERS

L. Li¹, S. Heish¹, J.D. Scandling², T. Sigdel³, M. Sarwal^3
1Pediatric, Stanford University, Stanford/CA/UNITED STATES OF AMERICA, ²Medicine, Stanford University, Palo Alto/CA/UNITED STATES OF AMERICA, ³, Stanford University, stanford/UNITED STATES OF AMERICA

Body: Aim: To improve on the sensitivity and specificity of a peripheral blood transcriptional biomarker genepanel for diagnosis and prediction of operational tolerance in a heterogeneous cohort of 63 adult kidney transplant recipients. Methods: 31 peripheral blood samples were analyzed by oligonucleotide Agilent arrays from 3 different phenotypes from different transplant centers: 16 operational tolerant (TOL) kidney transplant patients without medications for more than 1 year, 10 with chronic graft injury and on triple maintenance immunosuppression (CAN), and 5 healthy donors (HD). Samples were used as a training set to define a minimum gene-set for diagnosis of operational tolerance which were cross validated on an 20 independant samples, including 9 TOL, previously run on cDNA microarrays (PNAS, 2007). Q-PCR was run on 21 genes to build a model for prediction of tolerance by logistic regression and tested on an independent sample set containing 7 TOL patients. Bioinformatics analysis including GeneSpring, AILUN, SAM, PAM, logistic regression and IPA were used. Results: Twenty-one unique genes were identified (FDR<5%) by prediction analysis of microarrays (PAM) and statistical analysis of microarray (SAM) as a minimum gene for TOL. Genes overlapping across cDNA and Agilent platforms, among the 21 set, could back predict 9 TOL samples run on cDNA arrays with 100% sensitivity and specificity. These genes are enriched in immune cell trafficking and 13/21 genes are regulated by TNF, IL6 and IL4. Multinomial logistic regression modeling by Q-PCR for all samples identified 3 genes as a minimal gene set for TOL in the array data-set with 93% sensitivity and PPV, and 90% specificity and NPV. The 3-gene model predicted TOL with 100% sensitivity and NPV, 80% specificity, and 88% PPV, in a blinded independent data-set containing 7 TOL patients. 2 out of the final 3 genes are highly expressed in B cells. Conclusion: A highly regulated minimal gene-set in peripheral blood has been validated across multiple patients groups and transplant centers. This gene-set can be used as a non-invasive monitoring tool for screening patients with stable operational tolerance after kidney transplantation. Serial measurement of expression for this gene-set in peripheral blood opens the door for deliberate immunosuppression minimization after transplantation, and should be further tested for its utility in also monitoring patients on tolerance induction protocols in kidney transplantation.

Disclosure: All authors have declared no conflicts of interest.