The pattern and phenotype of T-cell infiltration associated with human liver allograff rejection☆

Abstract Although transplant biopsy remains the best means for assessing liver allograft dysfunction, the presence and degree of rejection often is difficult to determine by current histologic criteria. The added diagnostic and prognostic value of examining liver allograft biopsy specimens by immunopathologic methods, such as pheno-typing inflammatory cell infiltrates, has been inconclusive. To examine the value of assessing the phenotype and location of T-cell infiltrates we compared findings in 20 liver transplant biopsy specimens obtained from patients undergoing allograft rejection with those in 20 biopsy specimens from patients with no evidence of rejection. Serial frozen sections from all biopsies were labeled by immunoperoxidase techniques using monoclonal antibodies to identify cells expressing CD3, CD4, CD8, CD45RA, and CD45RO. As expected, the median of average cell infiltrates was higher in the rejecting versus nonrejecting group for each cell phenotype and region. However, statistical comparisons indicated that only some combinations of cell phenotype and location were significantly greater in the rejecting versus nonrejecting groups. The median number of portal CD3+ T cells per high-power field was increased in the rejecting versus nonrejecting groups (15.15 v 5.00 cells/high-power field; P v 1.55 cells/high-power field; P P