Inflammation in Fibrotic Areas (i-IF/TA) Identifies a T cell-mediated Rejection Component of IF/TA with Poor Kidney Allograft Outcome

Abstract

Addressing the etiological heterogeneity of interstitial fibrosis in kidney allografts represents an important challenge to improve long-term transplant outcomes. We investigated the determinants, clinical and histological phenotype, and outcome of i-IF/TA in a prospective cohort of kidney recipients. We prospectively enrolled 1539 kidney recipients transplanted between 2004 and 2010, with systematic assessment of i-IF/TA using the i-IF/TA Banff score and of tubulitis in atrophic tubules (t-IF/TA), on allograft biopsies performed at 1 year post-transplantation. We considered donor, recipient and transplant baseline characteristics, immunosuppression, infectious diseases (CMV, pyelonephritis, BK virus), presence of donor-specific anti-HLA antibodies (DSAs) and all the biopsy-proven diagnoses (T cell-mediated rejection [TCMR], antibody-mediated rejection [ABMR], recurrence, BK virus-associated nephropathy [BKVAN], calcineurin inhibithor [CNI] toxicity, acute tubular necrosis) recorded within the first year after transplantation. We identified 946 (61%) patients with IF/TA at one year post-transplant, among whom 394 (42%) patients showed i-IFTA, which was associated with t-IF/TA in 309 (78%) patients. Patients with i-IF/TA at 1 year post-transplant had significantly decreased allograft survival at 8 years compared with patients with non-inflammatory IF/TA (81% vs 87%, P=0.002). Independent risk factors for i-IF/TA included: TCMR (OR=2.7, p<0.001), BKVAN (OR=3.3, p=0.007) and HLA B (OR=1.3, p=0.012) and DR (OR=1.2, p=0.044) mismatching, while steroid therapy (OR=0.6, p=0.039), CNI therapy (OR=0.5, p=0.011) and inosine-5'-monophosphate dehydrogenase inhibitor therapy (OR=0.5, p=0.011) had protective effects on i-IF/TA occurrence. Unsupervised hierarchical clustering based on the whole spectrum of Banff elementary lesions assessed at 1 year post-transplant showed that i-IF/TA aggregated in the TCMR cluster (i, t, ti, t-IF/TA and i-IF/TA) and not in the ABMR (g, ptc, C4d, cg) and chronicity clusters (cv, ci, ct, ah). We observed a positive gradient between increasing level of i-IF/TA and t-IF/TA Banff scores and the risk of allograft loss (p<0.001). i-IFTA may reflect a TCMR subcomponent of IF/TA related to under-immunosuppression and is associated with poor kidney transplant outcome compared to non-inflammatory IF/TA.