TRANSCRIPTOMIC APPROACHES FOR PROGNOSIS AND DIAGNOSIS OF RENAL CHRONIC ALLOGRAFT REJECTION
A. Scherer1, A. Krause2, J.R. Walker4, A. Korn3 and F. Raulf1,5
1Novartis Institutes for BioMedical Research/Transplantation, and 2CDRA/Biostatistics, and 3Clinical Research and Development/Transplantation BU, Novartis Pharma AG, Basel, Switzerland, 4Genomics Institute of the Novartis Research Foundation, San Diego, CA, USA.
Simultaneous genome-wide expression profiling with DNA microarrays has gained great importance for the diagnosis and prognosis of a variety of diseases. Recently, this technology has been introduced to explore the pathology of various nephrological diseases such as chronic allograft nephropathy (CAN) (1, 2). Histologically, CAN may or may not be associated with vasculopathy (CAN-V). CAN or renal chronic allograft rejection (CR) is one of the major causes of late graft loss and is irreversible to date (3). We applied gene expression profiling to total RNA, preamplified using a modified Eberwine protocol (4), from single renal protocol biopsies to gain insight into the pathology of CR, to explore expression profiling as a tool to predict the development of CR early post transplantation, and to diagnose CAN and CAN-V at the molecular level.
The low incidence rate of CAN-V during the first year post-transplantation renders de novo patient cohort sizes in clinical CR trials large and impracticable. The identification of biomarkers, which are predictive of the occurrence of CAN-V within the first year after transplantation, could be a valuable tool for patient preselection in new clinical CR trials and for the optimization of clinical treatment regimens. We have exploited genechip analysis to identify metagenes in renal protocol biopsies that were taken from a cohort of 17 patients all with functional grafts and normal post-operative clinical and histological parameters. Nine of these patients developed CAN-V between months 6 and 12 post transplantation, while the others remained healthy. Applying extensive statistical analyses, a prognostic marker set of 10 genes with differential expression pattern at month 6 was identified. One of the upregulated genes, HoxB7, is detected in human atherosclerotic plaques in areas adjacent to calcification and scattered in media and neointima, and induces smooth muscle cell differentiation. The marker pattern was sufficient to correctly predict the occurrence or non-occurrence of CAN-V in 15 out of 17 patients (88%) using “one-at-a-time” cross-validation (i.e. a patient’s status was predicted on the basis of the remaining patients’ data) (5). We are currently collecting more serial biopsy samples to validate these findings.
In order to identify gene expression signatures that are differentiating CAN from CAN-V, we analyzed single renal protocol biopsies with CAN or CAN-V and compared them to samples from healthy allografts. The analysis is ongoing at the time of abstract submission.
Incorporation of gene expression profiling data from human renal protocol biopsies into the clinical health status assessment will improve both the understanding of the pathogenesis of CAN and CAN-V as well as the prognosis and diagnosis of CR.
Acknowledgements: We are grateful to our collaborators E. Navarro, M. Hueso, and D. Serón (Hospital Bellvitge, Barcelona), as well as to all investigators of the Certican trial RAD B251 from 16 centers for contributing protocol biopsies.
References: (1) M. Eikmans et al., Kidney Int. 2002; 62:1125. (2) M. Stegall et al., Am J Transpl 2002; 2: 913. (3) R.M. Jindal, S Hariharan, Nephron 1999; 83: 13. (4) A. Scherer et al., BioTechniques 2003; 34:546. (5) A. Scherer et al., Transplantation 2003; 75:1323.
Dr. Friedrich Raulf
Novartis Institutes for BioMedical Research / Transplantation
Novartis Pharma AG
Phone: +41 61 324 6880
Fax: +41 61 324 3576