P-226: Study of T cell receptor (TCR) dynamics in smouldering myeloma (SMM) using TCR sequence clustering to infer contribution of the immune microenvironment to progression to multiple myeloma (MM)

Introduction: The drivers of progression from SMM to MM are not fully explained by tumour-intrinsic factors. We hypothesise that T cells maintain disease equilibrium in SMM and that a reduction in tumour-targeting clones contributes to progression. We analysed longitudinal samples from SMM patients to study T cell clonal dynamics.

Methods: Bulk TCR sequencing was performed on cryopreserved peripheral blood (PB) and bone marrow (BM) mononuclear cells. Simpson’s normalised reciprocal index was calculated and then structurally similar TCRs were grouped by TCRdist3. Clusters with significant abundance changes between timepoints were identified.

Results: Patient samples from the COSMOS observational SMM study (progressors=4, non-progressors=3) at successive timepoints were analysed (median samples per patient=5, median time from first to last=14m, range=7-28). We calculated Simpson’s index for each timepoint (per patient), seeking global changes in TCR repertoire diversity at progression. We found no consistent trends in progressors vs non-progressors.

We hypothesise that clustering TCRs with similar sequences would better reveal changes in tumour directed clonotypes. Similarity in TCR sequences implies shared antigen specificity. We identified a median of 1316 (range=37-9508) unique clusters per patient. Focussing our analysis on the tumour niche, we selected patients with good quality (reads >1500) BM samples (2 progressors, 1 non-progressor), and identified clusters that significantly changed in abundance from baseline to progression/final timepoint (ie. deviated from calculated normal distribution); median significant clusters per patient=126 (range=62-1017). We calculated fold change in these clusters over time. No significant change in cluster abundance was observed in the non-progressor. In contrast, both progressors displayed significantly reduced abundance in several TCR clusters at progression (p< 0.05, n=23, n=8) but no clusters increased in abundance at progression. This indicates loss of structurally similar TCRs (hypothesised to bind the same tumour antigen) at progression, thus a loss of tumour-specific T cell immunity which could contribute to a loss of disease control and progression.

We analysed these TCR clusters in corresponding PB samples. The reduction in BM cluster proportion was mirrored in PB in one progressor, but not in the other, indicating variability in the BM/PB TCR repertoire between patients and favouring the importance of tumour niche immunity to disease control.

Conclusions: Progression to MM may involve alterations in tumour specific T cell immunity, not revealed by studying global TCR diversity. Our analysis employing TCR sequence clustering to infer tumour directed clonotypes identified a subset of TCR clusters that reduced in abundance at progression, consistent with loss of immune control. This approach will aid our understanding of how T cell immunity alters at progression, forming the basis for immune interception strategies.