AFFILIATIONS
1 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
2 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
3 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
4 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
5 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States
6 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
7 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
8 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
9 Dana-Farber Cancer Institute, Boston, Massachusetts, United States
10 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
11 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
12 Colorado Blood Cancer Institute, Denver, Colorado, United States
13 Oregon Health and Science University, Portland, Oregon, United States
14 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
15 Kyoto University Hospital, Kyoto, Japan
16 Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
17 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States, Massachusetts General Hospital, Boston, Massachusetts, United States
18 Broad Institute of Harvard and MIT, Cambridge, Massachusetts, United States, Dana-Farber Cancer Institute, Boston, Massachusetts, United States, Harvard Medical School, Boston, Massachusetts, United States
ABSTRACT
Introduction: Teclistamab is moving from the treatment of heavily pretreated relapsed/refractory multiple myeloma into earlier disease settings, where deep responses are observed in patients with limited prior treatment exposure and preserved BCMA expression. Nevertheless, recent clinical trials evaluating teclistamab in earlier-stage myeloma have identified a small subset of patients with primary resistance despite the absence of an identifiable antigen-escape mechanism. We hypothesize that primary resistance is driven by spatially organized cellular interactions in the bone marrow microenvironment that limit productive teclistamab-mediated T-cell engagement and tumor clearance.
Methods: We applied Xenium spatial transcriptomics on 20 whole-slide bone marrow FFPE samples from 10 patients with high-risk smoldering multiple myeloma treated with teclistamab (7 responders and 3 nonresponders). Paired samples were collected 1 month before and 4.5 months after teclistamab initiation. A 480-gene panel resolved 13 cell phenotypes across 1.16 million cells. Tissues were represented as cell network graphs to identify recurrent T-cell-anchored motifs and compare observed cell-cell associations with those expected from 1000 permutations of cell labels. Curated gene signatures were used to quantify relevant immune processes. Findings were orthogonally validated using spatial proteomics on 108 regions across an expanded cohort of 27 patients including up to four follow-up timepoints.
Results: Across all patients, baseline bone marrow tumor infiltration consistently ranged from 20-50%, Whole-genome sequencing identified no alterations at the BCMA locus, and baseline tumor BCMA RNA expression was similar. By month 5, all responders achieved CR/sCR and were MRD-negative at 10-6, whereas all nonresponders had stable disease without a decline in soluble BCMA.
Longitudinal analysis of matched samples showed near elimination of tumor cells in responders but persisted in nonresponders, with quantified tumor proportions strongly correlated with pathology whole-slide estimates (R=0.90, p<0.001). T-cell abundance consistently increased post-treatment, whereas increases in macrophages and monocytes were observed in responders with other cell populations remaining stable in nonresponders.
Five recurrent T-cell-anchored spatial motifs were identified. After treatment, T-cell aggregates increased in all patients. Responders also showed depletion of T-tumor and enrichment of T-immune motifs. Spatial gradient analysis expanding from T-tumor motifs showed increasing proximity of T-cell aggregates in nonresponders in contrast to T-immune motifs in responders after treatment. However, inflammatory interferon-responsive signatures were elevated in responders (p<0.001) but remained unchanged in nonresponders. These features suggest that T cells were spatially directed to tumor cells in nonresponders but did not initiate tumor clearance or broader immune activation.
We next evaluated whether T cells preferentially localized near specific cell populations by comparing the frequency of observed neighbors with those expected by chance. In responders, T cells remained preferentially associated with immune cells such as macrophages and monocytes after treatment. In nonresponders, most spatial relationships remained unchanged, but tumor cells became markedly enriched around T cells, indicating persistent T-tumor engagement.
Finally, we examined the link between tumor proximity and T-cell functional state. After treatment, tumor adjacency strongly correlated with increased T-cell dysfunction (R=0.73, p<0.05) and decreased T-cell activation (R=-0.70, p<0.05) in nonresponders, a spatial relationship absent in responders. Consistently, T-cell dysfunction genes (PDCD1, LAG3, HAVCR2, CTLA4, TIGIT, ENTPD1, TOX, IRF4, CXCL13) and their tumor ligands (PD-L1, PD-L2, HLA-DRA, CEACAM1) were elevated in T-tumor motifs after treatment in nonresponders but not in responders.
Conclusions: Primary teclistamab resistance can occur despite limited prior treatment exposure, preserved BCMA expression, and successful T-cell recruitment to tumor. Our findings identify spatially localized T-cell dysfunction, rather than failure of target recognition, as a mechanism limiting tumor elimination. These findings support a key role of the bone marrow microenvironmental state in shaping immunotherapy efficacy and provide new insights for patient stratification and therapeutic intervention.