Invariant natural killer T (iNKT) cells recognize lipid antigens in the context of CD1D and have been shown to effectively kill CD1D-expressing multiple myeloma (MM) cells. It has been reported that iNKT cells in MM patients are deficient and dysfunctional and that CD1D is variably expressed among MM patients and is completely lost in MM cell lines. Recent clinical study suggests that targeting iNKT cells may prevent progression of asymptomatic MM (AMM) to active disease (Richter at al. Blood 2013). All-Trans Retinoic Acid (ATRA) has been used for the treatment of MM but its modes of action have not been completely elucidated. The aims of this study were to shed light on the mechanism of action of ATRA and to identify ATRA target genes in MM. ATRA at clinically relevant concentrations (1-2 μM, 4 days) had minimal growth inhibitory effect on the majority of MM cell lines (10±13% and 14±16%, respectively; n=10). Gene expression profile (GEP) of MM cell lines and primary cases (n=13) treated with ATRA (1 μM, 2 days) revealed upregulation of CD1D (381±204 vs. 5692±1675 mean±SEM GEP signal in CONT and ATRA groups, respectively; p<0.006). Flow cytometry analysis for cell surface CD1D in 21 MM cases treated in vitro with ATRA revealed significant CD1D induction in 11 cases, intermediate induction in 2 cases and no induction in 8 cases. Clinically, MM cells from patients who received ATRA as part of their salvage therapy also had induced CD1D gene expression compared to levels prior to ATRA therapy. Analysis of abnormal plasma cells based on a single 8-color markers flow cytometry showed higher CD1D coefficient of variation (CV) in AMM cases molecularly classified as high-risk (n=5) than low-risk cases (n=9, p<0.003), while mean florescence intensity (MFI) was insignificantly lower in high-risk AMM cases. In vitro ATRA treatment of BM cells from high-risk AMM patient resulted in increased CD1D MFI by 2 folds and reduced CV in abnormal plasma cell restricted population. To test the functional consequences of CD1D upregulation in MM cells, iNKT cells were purified from healthy donors’ peripheral blood mononuclear cells using anti-Vα24Jα18 immunomagnetic beads and expanded in vitro via stimulation with irradiated autologous α-GalCer-loaded mature dendritic cells (DCs) in the presence of IL-2. ATRA treated (1 μM, 2 days) compared to vehicle treated MM cells (n=4) were more sensitive to iNKT cell-mediated lysis in 5 hour 51chromium-release assays or bioluminescence viability assays (e.g. 21% vs. 75% lysis of H929 cells, p<0.0001, effector: target ratio 2.5:1). ATRA-induced MM cell lysis by iNKT cells was markedly blocked by CD1D neutralizing antibody (1-40 μg/ml) indicating that the enhanced cell lysis was mediated through CD1D. Further GEP analysis identified RARRES3 (retinoic acid receptor responder protein 3) as an additional top ATRA target gene in MM cells. RARRES3 gene expression was commonly but variably induced by ATRA whereas MM cells expressing high baseline RARRES3 were resistant to ATRA-induced CD1D gene and protein expression. Expression levels of CD1D and RARRES3 were inversely correlated in MM cells from newly diagnosed patients (n=449, r=-0.4, p<0.0001). Among molecularly classified groups, the MF subtype, known to be associated with poor survival, had the lowest CD1D/highest RARRES3 expression, whereas the favorable CD2 subtype had a reversed pattern. We conclude that CD1D and RARRES3 are ATRA target genes in MM cells and that stimulation of cell surface CD1D expression on MM cells enhances sensitivity to iNKT cell-mediated lysis.


van Rhee:Senesco: PI Other.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution