Background. Despite the transformative impact of BCMA-directed cellular immunotherapies on multiple myeloma (MM), a substantial proportion of patients fail to achieve durable responses. Pre-existing T-cell exhaustion in the bone marrow (BM) niche is increasingly recognised as a contributing factor. We assembled a literature-derived T-cell exhaustion gene

Methods. A 20-gene exhaustion signature was constructed from published scRNA-seq studies of the MM BM compartment. Prognostic performance was evaluated by Kaplan– Meier analysis in a pooled cohort of 1,152 MM patients (GSE24080 and GSE4204) via the KM Plotter platform. The 15 upregulated and 5 downregulated signature genes were queried against 1,165,466 perturbational profiles on clue.io (Touchstone v1.1.1.43). Candidate compounds were prioritised by normalised connectivity score (NCS ≤ −1.50) and filtered for clinical availability.

Results. The composite 20-gene score was associated with significantly improved overall survival (OS) in the pooled cohort (HR 0.76, 95% CI 0.60–0.97, P = 0.025, n = 1,152). Singlegene analyses were heterogeneous: high PDCD1 predicted inferior OS (HR 2.49, P < 10⁻¹⁶), whereas high TOX, CD274, and IRF4 paradoxically associated with better outcome, reflecting immune infiltration signals in bulk data. LINCS L1000 identified leukotriene receptor antagonists (NCS −1.85), glucocorticoid receptor agonists (NCS −1.83), statins (NCS −1.78), HDAC inhibitors (NCS −1.72), and mTOR inhibitors (NCS −1.68) as the top reversal classes.

Conclusions. A bulk T-cell exhaustion score carries prognostic information in MM independent of conventional risk factors. Montelukast — oral, FDA-approved, and entirely untested in MM — emerges as a biologically coherent and clinically tractable candidate for microenvironment conditioning prior to BCMA-directed immunotherapy.