Immune Profiles of Myeloma Tumor Microenvironment May Predict Sensitivity or Resistance to Elotuzumab

Background: Elotuzumab (ELO), an immunobiologic therapy targeting SLAMF7/CD319, is effective in the treatment of multiple myeloma (MM) with clinical indications in relapsed/refractory disease in combination with lenalidomide or pomalidomide. ELO binds SLAMF7 on the surface of MM cells to increase immune recognition while also binding SLAMF7 on NK cells resulting in the activation of an effector cell population capable of readily killing via antibody-dependent cellular cytotoxicity (ADCC) mechanisms. ELO efficacy has also been linked to enhanced antibody-dependent cellular phagocytosis (ADCP). Given the dependence of ELO activity on immune elements of the tumor microenvironment (TME), we hypothesized that characterization of the immunologic constituents of the immune TME (iTME) would provide cues capable of predicting clinical efficacy.

Methods: Nineteen patients were enrolled on a single-center clinical trial run at Moffitt Cancer Center comparing clinical activity of ELO in combination with low-dose (10mg) vs high-dose (25mg) lenalidomide and dexamethasone in patients with biochemical relapse while on lenalidomide maintenance after first line therapy. Bone marrow aspirate (BMA) and peripheral blood (PB) samples we collected at baseline, after completion of 2 treatment cycles, and at time of progression. Cells isolated from the BMA were analyzed by multiparameter flow cytometry (MPFC) using 4 panels; 3 panels characterized lymphocytes according to maturation, activation and polarization (T _H1, T _H2, T _H17 or T _Reg) and a 4 ^th panel characterized myeloid elements. Data was acquired on a BD Symphony cytometer and analysis was performed using FlowJo software.

Results: Patient samples were categorized according to initial therapeutic response. Responders (n=6) achieved >PR at best response, Progressors (n=5) were identified as patients whose disease markers increased >25% and Stable Disease (n=8) was identified as demonstrating <25% change in serologic disease markers. Here, we present a comparison of phenotypic profiles from the BM samples acquired at baseline before treatment initiation from patients classified as Responders or Progressors (patients demonstrating stable disease as best response were analyzed separately). Sample analysis by MPFC revealed distinct populations uniquely associated with responses as well as with treatment refractoriness. Principal component analysis (PCA) of the myeloid immune compartment shows 2 clusters representing distinct populations detectable only in Responders, while another large cluster, also expressing a granulocytic phenotype consistent with granulocytic myeloid suppressor cells, is prevalent in the Progressor patients but nearly absent from Responder patients, as might be predicted. Similarly, PCA identified monocyte/macrophage clusters uniquely associated with Responder or Progressor samples. Evaluation of lymphoid populations using 3 lymphoid panels also reveals distinctive clustering patters by PCA highlighting populations differentially associated with Responder or Progressor samples. Notably, while NK cells are present in all patient samples, there is a population of immature NK cells enriched in Responder patients that is largely absent from the Progressor population.

Conclusion: Immune profiling of the myeloid and lymphoid components of the baseline BM iTME in MM prior to treatment with ELO reveals distinct patterns of cellular constitution associated with responsiveness or refractoriness to subsequent treatment. While NK cells are known to be important to the ELO mechanism of action, our data suggests that ELO activity is compromised in the absence of immature NK cells. This work suggests that measurable characteristics of the cellular components of the iTME may provide important cues to help predict ELO therapeutic efficacy. These observations will need to be confirmed in a larger data set.