Abstract: FR-OR086

The Molecular Phenotype of Polyomavirus Nephropathy and Its Discrimination from T-Cell Mediated Rejection

Session Information

Category: Transplantation

  • 1702 Transplantation: Clinical and Translational


  • Adam, Benjamin A., University of Alberta, Edmonton, Alberta, Canada
  • Nickeleit, Volker, The University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States
  • Randhawa, Parmjeet S., University of Pittsburgh, Pittsburgh, Pennsylvania, United States
  • Regele, Heinz, Medical University of Vienna, Vienna, Austria
  • Mengel, Michael, University of Alberta, Edmonton, Alberta, Canada
  • Wagner, Siegfried, University of Alberta, Edmonton, Alberta, Canada
  • Braesen, Jan H., Medizinische Hochschule Hannover, Hannover, Germany
  • Broecker, Verena, Sahlgrenska University Hospital, Gothenburg, Sweden
  • D'Agati, Vivette D., Columbia University College of Physicians and Surgeons, New York, New York, United States
  • Drachenberg, Cinthia, University of Maryland School of Medicine, Baltimore, Maryland, United States
  • Farkash, Evan A., University of Michigan, Ann Arbor, Michigan, United States
  • Farris, Alton Brad, Emory University, Atlanta, Georgia, United States
  • Geldenhuys, Laurette, Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia, Canada

Improved immunosuppression protocols have reduced the incidence of T-cell mediated rejection (TCMR) but still carry a significant risk of BK polyomavirus nephropathy (BKVN). Despite requiring opposite treatments, BKVN and TCMR often have overlapping clinical and histological presentations. Molecular testing may allow for more precise diagnosis.


NanoString® was used to measure the expression of 800 genes in 40 formalin-fixed paraffin-embedded human samples. The genes included 795 human immune-related genes and 5 polyomavirus (PV) genes (Agnoprotein, LTAg, VP1, VP2, VP3). The samples included native kidney BKVN (n=5), pure TCMR (n=9), SV40 immunohistochemistry-positive carcinoma (tumor BK, n=9), and normal implant kidney biopsies (n=8). Using six additional nephrectomies with mixed BKVN and TCMR, regions showing histologic features of only BKVN (mixed BKVN, n=6) and only TCMR (mixed TCMR, n=3) were isolated with laser capture microdissection. Differential gene expression and diagnostic performance were assessed.


All five PV genes were significantly increased (FDR<0.05) in native BKVN versus pure TMCR but no human genes were differentially expressed (Figure 1). PV gene expression was also significantly higher (versus pure TCMR) in tumor BK (p<0.01), mixed BKVN (p<0.001), and mixed TCMR (p<0.05, except VP1). ROC analysis revealed excellent discrimination between BK-positive (including mixed TCMR) and BK-negative cases (AUC=0.96-1.00). As a 5-gene set, the PV genes demonstrated near-perfect diagnostic performance (AUC=0.99) with improved sensitivity (0.96) over histology (0.87).


These data suggest that PV gene expression is more sensitive than histology and can more precisely discriminate BKVN and TCMR. However, at the molecular level, no significant difference in immune response was identified.