PD-1/ PD-L1 Expression Is Associated with Tissue Inflammation and BRAF Status in Erdheim-Chester Disease

Rationale: In patients with Erdheim-Chester disease (ECD), accumulation of foamy histiocytes leads to multi-systemic disease with various organs involvement. The fact that BRAF^V600E mutation is found in as much as 70% of ECD tissues, led to the reclassification of ECD as a myeloid neoplasm and has already greatly improved therapy for adults with histiocytoses. Despite these advances, there is still a need to further improve therapy for ECD patients as targeted therapies may cause morbidity and late treatment effects from such regimens and as patients almost systematically relapse when these therapies are stopped. In 2015, Galatica et al reported an increased expression of PD-L1 in 4/4 ECD cases tested all of which were BRAF^V600E mutated. We thus decided to analyze a larger case series of patients as this could represent rationale for addition of immune check-point inhibitors in treatment of multisystemic and/or refractory histiocytoses.

Patients and Methods: We included 36 ECD patients for which BRAF status was determined. Biopsy samples were re-read in all cases. The density of inflammatory cells (lymphocytes and plasma cells) other than histiocytes was evaluated as mild (+), moderate (++), or marked (+++). Immunostaining was performed to detect PD-L1 (QR1Clone) in histiocytes and PD-1 (NAT105 clone) in lymphocytes. PD-L1 was assessed as percentage of positive histiocytes. The positivity of PD-L1 was defined as ≥ 5%. PD-1 immunostaining was evaluated as mild (+), moderate (++), or marked (+++).

Results: Overall, BRAF^V600E was present in 19 patients (52.8%), MAP2K1 in 2 (5.5%) and NRAS in 1 (2.8 %). PD-L1 was positive in 15 patients (41.7%), PD-1 in 23 (63.8%) and both were found in 13 (36.1%). The intensity of inflammation in 21 patients (58.3%) was mild, moderate in 9 (25%) and high in 6 (16.7%). Among the 23 PD-1 positive cases, 15 (65.2%) were mild and 8 (34.8%) moderate. We found a strong association between PD-L1 positivity and intensity of inflammation: 13 PD-L1 positive patients were moderate/marked, vs 2 PD-L1 positive being mild; only 2 PD-L1 negative patients were moderate/marked, vs 19 PD-L1 negative being mild (p<0.001). The same association was seen between PD1 positivity and level of inflammation: 14 PD-1 positive patients had moderate/marked inflammation, vs 9 PD-1 positive patients with mild inflammation; 1 PD-1 negative patient had moderate/marked inflammation, vs 12 PD-1 negative with mild inflammation (p<0.01). We found a negative association between PD-L1 positivity and BRAF^V600E mutation: 4 PD-L1 + patients were mutated vs 11 PD-L1 + being BRAF^V600E Wild Type (WT); 15 PD-L1 - patients were mutated vs 6 PD-L1- being WT (p<0.01). No association was found between PD-1 postivity and BRAF^V600E mutation. 80 % of patients which were PD-L1 - PD-1 + were BRAF^V600E mutated. In the contrary, all patients PD-L1 + PD-1 - were WT.

Conclusions: We found a negative association between PD-L1 positivity and BRAF^V600E mutation. Eleven patients only (30.5%) of ECD patients were PD-L1 and PD-1 negative. The recent success of immune checkpoint blockade therapy in some cancer types combined with the expression of immune checkpoint antigens in ECD samples suggests that such therapies should be investigated for refractory ECD.