Introduction: Both IFN-γ and IL-21, a member of the common-γ chain cytokine family, support induction and expansion of highly-reactive cytotoxic CD8+ T-cells. In addition, IL-21 is a key cytokine for differentiation of alloantigen activated naïve and memory B-cells into antibody producing plasma cells. We questioned whether the frequency of donor-specific IFN-γ and IL-21 can predict kidney transplant rejection, and evaluated these cytokines in a cross-validation study.
Methods: All patients received a living-donor kidney and were treated with basiliximab induction therapy. Maintenance therapy consisted of tacrolimus, MMF and corticosteroids. Corticosteroids were tapered at month 3 post-transplantation and withdrawn thereafter. The training group consisted of PBMC samples from 47 patients obtained at 6 months after transplantation of whom 14 patients developed a late rejection (i.e. after month 6). The independent validation group included pre-transplantation samples of 38 patients of whom 17 patients had an early rejection (<3 months). The frequency of circulating IFN-γ and IL-21 producing cells (pc) to donor cells was determined by Elispot assay. Stimulation with 3rd-party cells, staphylococcal enterotoxin B (SEB) and a cocktail of peptides to influenza, CMV and EBV (ICE) served as positive control for the HLA class II and class I response, respectively.
Results: Remarkably, no relation was found between donor-specific IFN-γ pc frequency and rejection in both groups. Also, no differences were found between rejecting and non-rejecting patients in IFN-γ and IL-21 pc frequencies after third-party, SEB and ICE stimulation. However, significantly higher donor-specific IL-21 pc numbers were found in patients who developed rejection compared to those without rejection in both the training (p=0.020) and validation (p=0.024) group. ROC-curve analysis of donor-specific IL-21 pc frequencies distinguished the development of rejection from non-rejection with a specificity of 88% and 80% in the training and validation group, and a sensitivity of 50% and 73%, respectively. Patients with low IL-21 pc frequencies had a significantly increased rejection free survival rate in both the training (p=0.0008) and validation group (p=0.0005) compared to those with high frequencies. In addition, a positive correlation was found between donor-specific IL-21 pc numbers and serum creatinine concentrations at 12 months (training cohort: rs=0.35, p=0.015) and at 1 and 2 months (validation cohort: rs=0.41, p=0.013 and rs=0.34, p=0.047) post-transplantation. Moreover, patients with pre-transplant anti-HLA antibodies had significantly higher numbers of pre-transplant circulating donor-reactive IL-21 pc than patients without antibodies (rs=0.388, p=0.031).
Conclusion: The frequency of donor-specific IL-21 producing cells is linked to an increased risk of rejection, giving it the potential to be a new biomarker in predicting rejection in different phases of transplantation.