ASH 2016: MRD Burden Following ASCT in Multiple Myeloma

Saad Z. Usmani, MD
Levine Cancer Institute
Charlotte, NC, USA

Peripheral Immune Profile and Minimal Residual Disease (MRD) Burden Following Autologous Stem Cell Transplantation (ASCT) in Multiple Myeloma (MM) 


The immune system plays an essential role in both promoting and inhibiting the growth of MM. Loss of anti-myeloma immunity involves altered activation and polarization of effector cells as well as immunosuppressive cytokines. Immune dysfunction associated with MM can be partially or totally reversed when patients achieve clinical remission. Detection of MRD has become an achievable goal in MM, and is a powerful predictor of progression-free survival and overall survival. We hypothesized that the immune profile of MRDpos patients is distinct from MRDneg patients.

Patients and Methods:

We studied peripheral NK, NK-T and T cell distribution/activation, and measured bone marrow MRD status by flow cytometry in 30 newly diagnosed MM patients 60+ days after ASCT. Patients were divided in 2 groups based on κ/λ ratio and MM-plasma cell (MM-PC) distribution: MRDneg (n=6), with κ/λ ratio ≤1.8 and MM-PC≤15 per million or MRDpos (n=24), with MM-PC>15 per million. Plasma was also collected from 17 of these patients (5 MRDneg and 12 MRDpos) 60+ days after ASCT to quantify inflammatory cytokines, chemokines and growth factors by multiplex protein assay. Linear regression was used to determine association between each tested variable (25 by flow cytometry and 27 by multiplex protein assay) and MRD status. Unsupervised hierarchical cluster analysis was then applied to post-ASCT samples with selected variables that were differentially expressed between MRDneg and MRDpos patients (p<0.1) using complete linkage and Euclidean distances.


More than a third of immune variables tested by flow cytometry (9/25) were significantly differential between MRDpos and MRDneg patients (p<0.1). Most significantly, MRDpos patients had fewer circulating CD56high NK cells than MRDneg (p=0.02). While NK cells had a propensity to be highly activated (90% NKG2D+) among MRDpos patients, they displayed an impaired killing phenotype with decreased KIR3DL1 expression compared with their MRDneg counterpart (p=0.01). NK-T and T cell distributions were not influenced by MRD status after ASCT, but phenotypically NK-T cells in MRDpos patients exhibited higher expression of NKG2D (p=0.05), but lacked NKG2A and Tim3 expression, indicative of greater mobilization. Plasma concentrations of all 27 soluble immune analytes tested were higher in MRDneg compared with MRDpos patients. Specifically, 2 pro-inflammatory cytokines [IL-12p70, IL-17] and 3 pro-angiogenic growth factors [FGF, PDGF, and VEGF] were highly correlated with MRD status (p<0.1). Unsupervised hierarchical clustering using all differentially expressed cellular and soluble immune markers showed better separation of MRDneg and MRDpos subgroups than using cellular or soluble immune markers separately (heat-map below).


A comprehensive analysis of 52 immune variables assessing cytotoxic cell distribution, mobilization, killing potential, inflammatory status, immune polarization chemotaxis and angiogenesis revealed distinct differences in peripheral immunity between MRDpos and MRDneg MM patients after ASCT. Fewer circulating NK cells exhibiting a loss of killing potential characterized immune dysfunction among MRDpos patients. On the other hand, MRDneg patients more frequently displayed an inflammatory and pro-angiogenic cytokine profile. Taken together, these observations suggest that MRD represents a state of immune equilibrium where detectable or undetectable myeloma cells remain under surveillance by cellular and soluble immune mediators. As a next step, we are validating our observation in a larger prospective cohort by assessing peripheral immune profile as a predictor for MRD negativity.