Background: We have recently demonstrated that an 'immune score' is strongly and independently prognostic in de novo DLBCL treated with R-CHOP immuno-chemotherapy. The score quantifies the relative composition of immune effectors (T cells) and checkpoints (e.g. PD-1 axis molecules and M2 macrophages), as a measure of net anti-tumoral immunity within the TME. It is also known that a diverse TCR repertoire is a hallmark of a robust anti-HIV T cell immune response; conversely in metastatic melanoma treated with anti-PD-1 checkpoint blockade, narrow more clonal TCR repertoires are associated with favorable response. The relationship between the intra-tumoral TCR repertoire and the TME in DLBCL following R-CHOP immuno-chemotherapy is unknown.


High-throughput unbiased TCR β chain sequencing was performed on 116 nodal tissues (101 de novo DLBCL patients treated with R-CHOP with long-term follow-up including 8 EBV+DLBCL; and 15 age/gender matched healthy lymph nodes). Outcomes included measurement of productive uniques (a measure of the number of functional T cells with a distinct TCR rearrangement or 'richness'); entropy (a measure of TCR 'diversity'), 'clonality' (a measure of clonal expansions) and the 'maximal frequency' of the most highly expressed clone within tumor biopsies. Results were compared to digital quantification (by nanoString) of key immune effector and checkpoint genes within the TME, the immune score, malignant cell-of-origin (COO), R-IPI and patient survival.


First we compared the TCR repertoire in lymphomatous and healthy nodes. There was a marked increase in clonality, reduced diversity and high maximal frequency within DLBCL nodes relative to healthy nodal tissue (both p<0.0001), consistent with an abnormally narrow TCR repertoire of antigen-specific T cells. Next, we tested the relationship between TCR and the TME. Notably, there was modest (r=0.3-0.7) but highly significant (all p<0.001) positive correlations between both richness and diversity (but not clonality) with CD3/CD4/CD8 T cells, and a range of immune checkpoints including PD-L1, PD-L2, LAG-3, CSF-1 and TIM-3. These findings are strongly suggestive of an adaptive immune response, in which malignant B cells influence (i.e. 'adapt') the TME in an attempt to counter an effective anti-lymphoma T-cell response that is in part influenced by the breadth of the TCR repertoire. Then we investigated the TCR repertoire in the context of prognosis and overall survival (OS) following R-CHOP. There were no correlations between COO or R-IPI with any TCR parameter. However, the presence of a high maximal frequency in the tumour biopsy was associated with significantly inferior 5 year OS of 59% compared to 81% in patients without a high maximal frequency (p=0.03, Figure 1). As expected, the immune score stratified patients into highly disparate outcomes: high-score 5-year overall survival 96% versus 42% for low-score (p<0.0001). Interestingly, there were significant differences in the TCR repertoire between the two groups. There was a significant increase for both richness and diversity in high immune score lymphoma patients (p=0.015 and p=0.018 respectively). In keeping, clonality was not increased in high-immune score patients. The only samples associated with increased T cell clonality were those patients with very high levels of intratumoral EBV, potentially reflecting the latent viral antigens expressed by this lymphoma. In the group of patients with poor prognosis (5 year OS 59%), defined by high maximal frequency, the immune score stratified two groups with very different outcomes (5 year OS 90% vs. 30%, p=0.003).


These findings indicate the TCR repertoire as a key parameter of the TME that the malignant B cell attempts to narrow. A broad TCR repertoire is associated with a good prognostic immune score (i.e. increased T cells relative to PD-1 axis molecules and M2 macrophages checkpoints) after R-CHOP immunoÐchemotherapy, whereas a more clonal T cell response is associated with significantly inferior outcome.


Hamm:Adaptive Biotech: Employment.

Author notes


Asterisk with author names denotes non-ASH members.