An acquired immune-metabolic shift to common chemotherapy confers resistance to immunotherapies in relapsed, high-risk multiple myeloma

Abstract

Abstract Introduction: The prognosis of multiple myeloma (MM) has improved significantly with the advent of targeted agents and immunotherapies. However, around 15% of patients with high-risk (HR) disease benefit less, relapse early, and exhaust treatment options more quickly than standard-risk (SR) patients. These individuals are often on continuous therapy and experience higher treatment-related mortality, particularly from infections, suggesting a link between HR biology, therapy burden, and immune dysfunction. Metabolic factors are key in shaping the tumor microenvironment. In solid cancers, nutrient depletion, such as low glucose or glutamine, impairs T cell function and contributes to immune escape. However, the metabolic landscape of the bone marrow (BM) niche in MM, especially in relation to risk status, remains poorly understood. To address this, we investigated the immune-metabolic composition of the BM microenvironment across risk groups. Methods: We collected BM aspirates from MM patients who were newly diagnosed or had relapsed/become refractory (n=221), and isolated the CD138-positive plasma cells for the risk stratification with the micro-array SKY92 MMProfiler, which is based on gene expression. The matching CD138-negative cell fractions were used for multicolor flow cytometry (n=221) and scRNAseq on sorted T cells (n=20). In addition, we examined the metabolic niche in the BM by liquid chromatography-mass spectrometry for water-soluble metabolites in the BM plasma (n=49). The killing capacity of T cells isolated from untreated and treated SR and HR patients were analyzed in vitro by using the MM cell line MM.1s as the target cell in combination with the bispecific antibody teclistamab in different effector-to-target ratios (n=12). Results: We found profound differences in T-cell composition between HR and SR patients, characterized by lower T-cell numbers (BM: 45% vs 53%, p=0.01) but increased CD8a+/CD4+ T-cell ratios in HR disease (BM: 1.13 vs 0.64, p=0.003), in both BM and peripheral blood. This increase was associated with lines of previous therapy, and further analysis revealed a therapy-induced naïve T-cell depletion in HR patients (16.3% vs 4.3% of CD4+ T cells, p<0.05). This naïve T cell depletion was further confirmed by our scRNAseq, in both CD4+ and CD8+ T cells (59% vs 30 %, p<0.0001 and 32% vs 14%, p=0.0005, respectively). ScRNAseq analysis revealed that pathways associated with T-cell activation and oxidative phosphorylation were upregulated in newly diagnosed and treated SR patients, respectively. This indicates that T-cells are more fit in SR disease. T-cells in HR disease upregulated an apoptosis signature and deregulated genes associated with redox balance. For example, they exhibited decreased TXNIP expression. TXNIP is an inhibitor of thioredoxin, a component of the redox system, found in T-cells. Interestingly, glutathione was enriched in the BM plasma of treated HR patients, likely due to increased ROS production by plasma cells in response to melphalan or proteasome inhibitor treatment. This suggests increased intra- and extracellular redox stress in HR disease. Additionally, essential amino acids necessary for proper T-cell function, such as glutamine and arginine, were depleted in the BM plasma of untreated HR patients, indicating a T-cell suppressive microenvironment. This depletion was associated with a higher proliferative profile of HR MM cells and a higher rate of BM plasma cell infiltration. Together, these findings suggest that the BM niche of HR disease is unfavorable for T cell function. To test T cell functionality, we isolated T cells from the BM of HR and SR patients, both with and without chemotherapy treatment, and performed in vitro killing assays. Strikingly, HR BM-T cells exhibited a significantly reduced cytotoxic activity of HR patients that had been exposed to chemotherapy (p=0.03). Conclusions: Our study reveals an emerging characteristic of HR disease that extends beyond genomic changes in tumor cells. Instead, it is related to profound differences in T-cell frequencies, cytotoxic function, and metabolite levels, which are further exacerbated by chemotherapy. These differences may finally lead to decreased activity of T cell engaging immunotherapies. Overall, our study provides evidence for the use of immunotherapy as front-line therapy, or for the use of T cells from untreated patients in the generation of CAR-T cells, particularly in HR patients.