Characterization of the "Immune Evasion" Phenotype of Richter Syndrome and the Implications for Immune-Checkpoint Inhibitor Therapy

Richter syndrome (RS) is the transformation of chronic lymphocytic leukemia (CLL) to a high-grade B-cell lymphoma and is associated with an aggressive clinical course and poor prognosis. Conventional treatment options for RS are generally associated with low response rates and limited durability making this entity an area of significant unmet therapeutic need. Immune checkpoint inhibitor therapy has shown promise in the treatment of some aggressive lymphoma subtypes. In RS, modest benefits have been reported in small phase two trials of anti-PD-1 monotherapy and in combination with ibrutinib, however larger scale studies are lacking (Ding et al Blood, 2017; Jain et al Blood, 2016). We sought to characterise the immune-evasion phenotype of RS focussing on potential genetic biomarkers which may inform the selection of patients who are most likely to benefit from immune-directed therapies.

We first assessed the gene expression of immune-checkpoint molecules given their potential clinical relevance and ability to be targeted by available therapeutic agents. Given immunohistochemical (IHC) assessment of immune-checkpoint molecules is recognized to be associated with high inter-observer variability and there is a high correlation between gene expression of immune-checkpoint molecules and IHC, we performed gene expression quantification using the Nanostring nCounter Human Immunology V2 panel (Nanostring Technologies, USA). Nanostring analysis was performed on samples from 17 patients with histologically confirmed RS (DLBCL subtype) and compared to 73 cases of de novo (non-transformed) DLBCL. Significant differences in the gene expression of checkpoint molecules was observed between RS and DLBCL biopsies, including higher expression of LAG3, PD1 and TIGIT in RS (p=0.0001, logFC 1.9; p=0.0017, logFC 1.1 and p=0.0437, logFC 0.7 vs DLBCL, respectively). PD-L2 and TIM3 gene expression were both significantly lower in RS compared to DLBCL (p = 0.0059, logFC 0.8; p = 0.012, logFC 0.8). PDL1 and CTLA4 gene expression did not significantly differ between RS and DLBCL.

We next assessed the gene expression of T- and NK- cell markers (including CD3, CD4, CD8, FOXP3 and CD56) and the ratios of these markers to malignant B-cells (CD19). We observed no significant difference between RS and DLBCL, consistent with a similar relative quantity of immune cell infiltration between the two entities. Significantly higher gene expression of CD39, a marker of CD8+ T-cell exhaustion, was observed in RS than DLBCL (p = 0.031; logFC 0.5). Additional immune-related genes were next assessed, including those involved in antigen presentation (e.g. B2M, HLA molecules, TAP), immunosuppressive cytokine generation (e.g. ARG1, IDO1) and apoptosis resistance (e.g. FAS) which showed no significant differences in expression between RS and DLBCL.

To assess whether these findings were consistent across other transformed lymphoma subtypes, we compared RS to a cohort of transformed follicular lymphoma (tFL, n=16) and transformed marginal zone lymphoma (tMZL, n=25). LAG3 expression was significantly higher in RS compared to both tFL and tMZL (p=0.0002, logFC 2.7; p=0.019, logFC 1.7). PD1 expression was also significantly higher in RS than tFL but not tMZL (p=0.0045, logFC 1.7; p=0.39, logFC 0.4).

Given the established association of copy number amplifications involving immune checkpoint molecules (e.g. PD-L1/PD-L2 on 9p24.1) representing a potential predictive biomarker of response in other lymphomas, we performed hybridization-based NGS with whole genome copy number assessment to evaluate immune checkpoint gene loci in the three cohorts. No significant focal amplifications were detected in RS samples with overexpressed immune-checkpoint molecules. In contrast, three patients in the DLBCL/transformed cohort had focal copy number amplifications involving PD-L1. No copy number amplification of LAG3 was observed in either RS or DLBCL.

In summary, we have observed significantly increased gene expression of LAG3, PD1 and TIGIT in RS compared to de novo DLBCL. Combined with increased gene expression of the exhausted cytotoxic T-cell marker CD39, these data provide a strong biological rationale for pursuing LAG3 inhibition either alone or in combination with other immune checkpoint blockade to enhance anti-tumour T cell responses in this difficult-to-treat entity.

CG/JL/YK co-first authors