Antibody‐mediated vascular rejection of kidney allograft: characterization of different kidney allograft rejection phenotypes via histology, C4d deposition and donor‐specific antibodies

According to previous Banff classifications, acute antibody-mediated rejection (AMR) was characterized by a combination of acute renal dysfunction, histological features such as inflammation of the microcirculation and positivity of C4d staining and the presence of anti-human leucocyte antigen (HLA) donor-specific antibodies (DSA) 1. AMR is currently the leading cause of graft loss 2, and more recently it has been shown that various phenotypes of acute AMR exist. The first newly described phenotype is subclinical AMR. Deceased-donor renal transplant recipients at high immunological risk were investigated in a pilot study 3. Patients were transplanted with current negative cross-match and, because they all had a history of high serum levels of DSA, were treated with plasma exchange, anti-CD20 and intravenous immunoglobulin (IVIg). At 12 months post-transplant, 95% of patients experienced good graft survival, renal function was adequate with serum creatinine levels of 142 ± 53 μmol/l, there was minimal proteinuria and a low incidence of humoral rejection 3. However, evaluation of biopsies from these patients revealed a high rate of acute lesions, including glomerulitis and peritubular capillaritis, that persisted up to 3 years post-transplant even though renal function was normal and stable 4. In contrast, there was a steady and significant increase in chronic lesions over this period, such as transplant glomerulopathy, interstitial fibrosis/tubular atrophy 4 and arteriosclerosis 5. Therefore, a new definition of AMR was created that is subclinical and characterized by no or mild renal dysfunction, both acute and chronic histological lesions and the presence of DSA 4. This is important, because subclinical AMR has a strong prognostic impact. Evaluation of biopsies at 1 year from patients without rejection or with cellular subclinical rejection reveals that the prognosis is good; conversely, patients with subclinical AMR have a poor prognosis at 1 year (unpublished data). The second AMR phenotype is C4d-negative AMR. It is recommended that all renal allograft biopsies are stained for C4d and the Banff classification includes a scoring system of C4d staining based on the percentage of peritubular capillary stained with C4d: score 0, negative; score 1, minimal; score 2, focal; and score 3, diffuse 6. Protocol biopsies from patients with preformed DSA were obtained at 3 months and 1 year post-transplant and reviewed 7. Of the biopsies that were C4d-positive, more than 90% had evidence of concomitant microvascular inflammation (MI). This is in contrast to the C4d-negative biopsies, where only 55% showed evidence of microvascular injury. It was shown that having this type of inflammation correlated with the presence of class II DSA. Although class I DSA did not vary between groups, class II DSA levels rose progressively with increasing C4d status at both 3 months and 1 year post-transplant 7. In addition, MI was correlated with worse renal function, as shown by significantly higher serum creatinine levels at last follow-up. Serum creatinine levels rose progressively from 145 ± 7·9, 213 ± 22 and 263 ± 31 μmol/l for summed C4d scores of: (i) persistently negative (score 0); (ii) mild (score 1–3); and (iii) high (score 4–6), respectively (P = 0·006) 7. Considering three distinct humoral states, C4d−/MI−, C4d−/MI+ and C4d+/MI+ [splitting this last group into focal (C4d = 1–2) and diffuse (C4d = 3)], the risk of subsequent chronic AMR increased gradually by 22·2, 40·0, 50·0 and 61·8%, respectively 7. Importantly, the increase begins in C4d−/MI+ patients. These results show a lack of sensitivity of C4d as a measure of injury in patients who, despite having microvascular injury, are C4d-negative. Such cases are at increased risk of subsequent chronic AMR. Therefore, Cd4-negative AMR has been included in the most recent Banff classification 8. The positivity of C4d staining is considered as a means to score severity of rejection. The third AMR phenotype is AMR with vascular lesions. A comparison of patients who had acute biopsy-proven rejection (n = 302) with patients without rejection (n = 1777) identified four distinct patterns of kidney allograft rejection: (i) T cell-mediated vascular rejection [26 patients (9%)]; (ii) antibody-mediated vascular rejection [64 (21%)]; (iii) T cell-mediated rejection without vasculitis [139 (46%)]; and (iv) AMR without vasculitis [73 (24%)] 9. Antibody-mediated vascular rejection is a previously unrecognized phenotype characterized by endarteritis associated with circulating DSA; this phenotype has the poorest graft survival. Initially, vascular lesions were not used to score the severity of rejection. However, this recent data suggest that vascular lesions are the most potent prognostic factor of such rejections and are now included in the latest Banff classification 8. The fourth AMR phenotype is characterized by AMR without anti-HLA antibodies but with DSA of other origin (e.g. vimentin). However, the origin of DSA is not always known. We showed recently that anti-HLA-C DSAs as well as DSAs of unknown origin (in males with no past history of blood transfusion or transplantation) were as deleterious as other anti-HLA DSAs 10,11. In conclusion, acute AMR is not a single entity but rather a series of phenotypes with varying severity and potentially different treatments. Therefore, it is important to give the right diagnosis in order to provide the most effective treatment.