AUTHORS
Annabel Laidler, Bethan Hudson-Lund, PhD2, Elise Rees, PhD3, Amatta Mirandari, PhD4, Matthew Cowley5, Trupti Gore6, David Scobie7, Suzanne Byrne8, Gayathri Nageswaran9, Adlina Chan10, Rethina Shritharan11, Louise Ainley, MD12, Jessica Kimber13, Emma Lyon14, Alice Lefever15, Daniel Hughes16, Dylan Jankovic17, Grant Vallance18, Ceri Bygrave, MBBCh19, Dean Smith20, Karthik Ramasamy, MD PhD21, Andreas Tiffeau-Mayer22, Eileen Boyle, MD PhD23, Kwee Yong, Prof24, Benny Chain, PhD25, Lydia Sarah Hui Lee, MBBS, MRCP, MRCPath(Haem), PhD26
AFFILIATIONS
1 University College London, London, United Kingdom
2 University College London, London, United Kingdom
3 University College London, London, United Kingdom
4 University College London, London, United Kingdom
5 University College London, London, United Kingdom
6 University College London, London, United Kingdom
7 University College London, London, United Kingdom
8 University College London, London, United Kingdom
9 University College London, London, United Kingdom
10 University College London, London, United Kingdom
11 University College London, London, United Kingdom
12 University College London, London, United Kingdom
13 University College London, London, United Kingdom
14 University College London, London, United Kingdom
15 University College London, London, United Kingdom
16 University College London, London, United Kingdom
17 University College London, London, United Kingdom
18 University of Oxford, Oxford, United Kingdom
19 University Hospital of Wales, Cardiff, United Kingdom
20 Nottingham City Hospital, Nottingham, United Kingdom
21 University of Oxford, Oxford, United Kingdom
22 University College London, London, United Kingdom
23 University College London, London, United Kingdom
24 University College London, London, United Kingdom
25 University College London, London, United Kingdom
26 University College London, London, United Kingdom
ABSTRACT
Introduction Progression from precursor states to multiple myeloma (MM) is driven by tumour genomics and declining anti-tumour immunity. Immune evasion may result from loss of cytotoxic function in immunosuppressive environments as tumours expand, genomic evolution causing loss of immunogenic epitopes, or both. Tumour-reactive T cells and a myeloma-reactive T-cell transcriptional signature (TFiT) have been described in MM, but their role in precursor disease (smouldering myeloma/SMM) remains unclear. We thus sought to identify tumour-targeting clonotypes using two parallel approaches: (1) studying longitudinal T-cell receptor (TCR) dynamics in SMM (2) identifying tumour-targeting clonotypes using an in vitro reactivity assay.
Methods Serial bone marrow (BM) samples from observational SMM COSMOS study (NCT05047107; progressors/P n=8; non-progressors/NP n=7) underwent bulk TCR sequencing. Healthy donor (HD) repertoires came from COVIDsortium (NCT04318314, n=13; paired, 12m apart). Selecting samples with sufficient sequencing depth, TCRs were filtered (edit distance ≤2) and clustered to group similar CDR3 sequences, with presumptive overlapping antigen specificity. Changes from baseline were tested against null model expectations; significant expanding/contracting clusters identified by Fisher's exact test (FDR<0.05). Baseline and progression samples from two progressors (P2 and P6) underwent whole exome sequencing (WES) of isolated CD138+ tumour cells and single-cell sequencing of CD138− fractions. To characterise tumour-reactive T cells in vitro, we cultured BM T cells for 10 days +/- autologous tumour cells. Cultures were re-stimulated with tumour for 24h, then FACS sorted for activation induced markers (AIM:CD69, OX40, 4-1BB, CD25) before single-cell RNA/TCR or bulk TCR sequencing.
Results Total TCR counts and clonality remained stable (HD=13, NP=5, P=4) and the proportion of dynamic clusters were similar across groups. However, repertoires in progressors were dominated by expanding clusters, unlike HDs or NPs. Linear modelling (adjusting for repertoire size) confirmed progression status as significant predictor of less contracting clusters.
Single-cell RNA-seq in one progressor (P6) revealed that dynamic clusters were mostly CD8 effector memory cells at baseline. However, expanded clusters displayed a predominance of earlier effector memory phenotype with reduced cytotoxicity at progression, unlike contracting clusters. P2 deviated from this trend, showing exclusively contracting clusters and a T cell compartment enriched for terminally differentiated effector-like cells. Clonotypes at baseline with high TFiT score contracted at progression, suggesting a loss of tumour-reactive cells.
Comparing baseline and progression WES of tumour cells we investigated a possible genomic basis for these divergent TCR cluster dynamics. P2 harboured a t(14;20)/MAFB translocation and exhibited marked genomic remodelling with increased mutational burden at progression, whereas P6 exhibited stable disease genomics from baseline to progression.
Ex vivo culture of P2 BM T cells with tumour increased CD8 CD69/4-1BB expression, consistent with HLA class I-mediated activation. Expanded clonotypes (log2FC>2 vs controls) were more activated in tumour conditions (56% AIM-positive) vs. controls (5%). Differential gene expression showed expanded tumour-cultured clonotypes upregulated GZMB, GZMK, IFNG, PRF1 and TIGIT consistent with tumour reactivity, whereas the same clonotypes in control conditions retained TCF7, IL7R and CCR7, showing a less differentiated phenotype (p<0.05). Notably, tumour-reactive clonotypes identified in vitro by expansion decreased in frequency at progression in vivo, reflecting contraction of TFiT-high populations observed longitudinally, with 1 TCR matching between approaches.
Conclusion We report a progression-associated shift toward expanding TCR clusters in SMM. In most patients, expanding populations adopted an early effector memory phenotype with reduced cytotoxic function, consistent with impaired immune control. In contrast, a patient with extensive tumour evolution at progression, had loss of tumour-reactive T cells, supported by contracting TFiT score-high TCR clusters and in vitro reactivity. Our findings suggest progression to MM can occur via distinct immunological routes: functionally impaired expanding T cell populations, or depletion of tumour-reactive clones linked to tumour genomic remodelling.
DISCLOSURES
Annabel Laidler: No relevant conflicts of interest to declare.