Full Spectrum Multi-Parameter Single-Tube Flow Cytometry Analysis for Cross-Sectional Immune-Profiling of Natural Killer Cells in Multiple Myeloma Clinical Samples

Introduction The bone marrow microenvironment plays a key role in multiple myeloma (MM) pathogenesis, as it contributes to disease progression, the expansion of dysfunctional effector cells and tumor immune escape. Natural killer cells (NK) are one of the dysfunctional populations in the microenvironment, and their malfunction originates from a) their high expression of checkpoints and their interaction with cognate ligands and b) their dysregulated expression of activating/inhibiting receptors. However, NK subpopulations vary in function and have not been fully categorized in MM. Full spectrum multi-color flow cytometry (SFC) is an advanced technique compared to traditional flow cytometry, whereby many fluorescent markers are used to evaluate a complex combination of antigens. We report here an SFC-based immune phenotyping of NK cells from different MM patients using a panel of 18 markers utilizing 22 fluorophores. Our assessment of activating and inhibitory NK receptors provides important information about clinical outcomes.

Methods Mononuclear cells were purified from MM patient peripheral blood (PBMC) and matched bone marrow (BM-MNC) at the Dana-Farber Cancer Institute. Normal PBMCs (n = 7) were used as controls. MM patients (2 maintenance; 2 induction; 18 relapsed/refractory) were treated under DFCI clinical protocols: 19-291; 19-379; 19-389; 19-453; 20-207; 20-212; 23-291. SFC data were acquired on a 5-Laser SONY ID7000 Spectral Cell Analyzer and analyzed using FlowJo. SFC panel consisted of: Zombie NIR viability dye; CD33 to exclude myeloid cells, NK identification by CD3 and CD56 (CD3^-/CD56⁺), NK receptors NKG2A, NKG2D, NKp46, DNAM-1, KLRB1/CD161 and NKp44, immune checkpoints TIM3, LAG3 and TIGIT, functional markers CD25, CD16, CD69, HLA-DR, and degranulation marker CD107a.

Results NK cells were identified as CD33^-/CD3^-/CD56⁺. In normal resting PBMCs, the majority of NK cells were CD56^hi (NK%: 14-18% of total CD33^-, n = 7), and the percentage of NKT (CD3⁺/CD56⁺) was very low (0.5-1.76% of CD33^-; p < 0.05, NK versus NKT). This pattern was similar for relapsed/ refractory MM-NK cells (CD56⁺ NK: 10% ± 3%; NKT: 0.5-1.5%; p < 0.05). However, NK cells from patients undergoing induction therapy or in maintenance protocols showed a significant shift in the CD56⁺ population from being majority CD3^- to CD3⁺ (CD3⁺/CD56⁺; 10-15%, n = 5, p < 0.05; CD3^-/CD56⁺ < 2%), indicating therapy-induced activation of NKT cells and a possible recovery of the immune landscape. This result is significant since this can indicate if a patient is undergoing therapy. Both maintenance and induction samples show high CD16 expression, with the emergence of CD56^low/CD16^hi cells, and a significant increase in CD69 expression (CD56⁺/CD69⁺; 10% ± 3%) indicating activation of NK cells. In contrast, relapsed/refractory NK cells showed insignificant CD69 expression (<2%, p < 0.05).

We found varying expression of DNAM-1, NKp46, NKG2D, NKG2A and CD161 in NK cells. In the two maintenance samples, the NK cells were NKG2A^low, CD69⁺, and negative for TIM3, TIGIT and LAG3 (< 1%), indicating active NK cells. In contrast, the refractory samples were double positive for TIM3 and LAG3, and CD3^-/CD56⁺ cells were mostly CD25^-, indicating exhaustion. As a control, we did not see any CD107a expression in the absence of the protein transport inhibitor Monensin. We found no major difference between NK cells from matched PBMCs and BM-MNCs (p > 0.05)

The dimension reduction analysis (FlowJo tSNE) merged all dot plots into a dimension-reduced space and showed comparable signatures for all normal PBMCs, which differed from MM PB- and BM-MNCs. Indeed, within MM, dimensionality reduction showed pattern shifts between relapsed and maintenance or induction samples. Our BM-NK single-cell RNA sequencing data (3 relapsed/refractory MM & 3 normal) also confirmed the differential expression of genes in the flow panel, showing high expression of inhibitory markers and low expression of activation markers in relapsed vs normal samples (adjusted p < 0.05).

Conclusion We report here an extensive phenotyping of MM patient NK cells using SFC that can be used to characterize clinical outcome. We will extend our study to phenotype the entire tumor microenvironment with a comprehensive 42-color panel. Our study provides a framework for future cross-sectional and longitudinal bone marrow phenotyping in clinical trials to evaluate treatment outcomes.