Genetic Analysis of Distinct Phenotypic Subsets within MM1.S Multiple Myeloma Cell Line Reveals the Pre-Existence of MM.1R-like Glucocorticoid Resistance and a Sub-Clone with an Activating PI3-Kinase Delta Mutation That Is Preferentially Sensitive to the Selective PI3-Kinase Inhibitor, Idelalisib

Background: Understanding the pathology of Multiple Myeloma and the testing of therapeutic options has relied heavily on isogeneic cell lines due to the inability to sustain myeloma plasma cells in long-term in vitro culture. The cell lines MM.1S and MM.1R are well recognised in the field of myeloma research, providing a model of glucocorticoid drug resistance, primarily believed to be through variable expression of the glucocorticoid receptor NR3C1 but here we found no evidence of a genetic basis for this. Here we set out to examine the phenotype, function and genotype of the MM.1S and MM.1R cell lines in order to explore the origins of glucocorticoid drug resistance manifested by MM.1R cells and establish whether exome analysis could identify sub-clones with preferential sensitivity to molecular targeted inhibitors.

Methods: MM.1S and MM.1R cell lines were purchased from ATCC. A 10-colour flow cytometry panel (CD38, CD138, CD19, CD45, CD56, CD49d, CXCR4, MMP-9, Ki-67, IL-6) was analysed on a BD LSR Fortessa flow cytometer and MM.1S subsets were sorted using a FACS Aria III. Telomere length was assessed using Single Telomere Length Analysis (STELA) and drug toxicity assays using Annexin V-FITC/PI staining. Bioinformatics of whole exome sequencing was carried out on the GATK platform and gene list analysis using Enrichr. PI3K isoforms were analysed by quantitative PCR and immunoblotting.

Results: The MM.1S cell line demonstrated bimodal CD38 expression, with a 1.5 log difference in CD38 expression (p<0.0001) between the two populations (termed MM.1S^dim and MM.1S^bright). In contrast the MM.1R cell line was uniformly CD38^bright, with expression a further 0.5 log higher than MM.1S^bright. When cell sorted subsets of MM.1S cells were subjected to increasing concentrations of Dexamethasone, the MM.1S^bright cells had a significantly higher LC₅₀ than the MM.1S^dim cells (62nM v 29nM respectively; p<0.0001). In contrast, these subsets showed no significant difference in sensitivity to bortezomib (p=0.84. Furthermore, MM.1S^bright cells had a shorter doubling time than both MM.1S^dim (p=0.0001) and MM.1R (p=0.048). This was underscored by an increased proportion of MM.1S^bright cells in S-phase coupled with shorter mean telomere length when compared with MM.1S^dim and MM.1R (2.58 Kb v 3.29Kb v 3.2Kb respectively). We next subjected purified MM.1S^bright, MM.1S^dim and MM.1R cells to whole exome sequencing. The common clonal origin of the three cell lines was evident from the analysis but each line possessed unique genetic lesions. For example, MM.1S^bright had a FIP1L1-PDGFRA fusion mutation that was not present in the MM.1S^dim cells. This was associated with increased expression of the p110d isoform in MM.1S^dim cells. We therefore analysed the effects of the PI3Kd inhibitor, Idelalisib, on the two cell lines and showed that MM.1S^dim cells were more sensitive (p=0.003) to the effects of this agent. The specific nature of this response was confirmed by the fact that the pan p13K inhibitor PKI-402, was equipotent in both MM.1S^bright and MM.1S^dim cells (p=0.89).

Conclusion: Analysis of two phenotypically distinct subsets within the MM.1S cell line revealed differences in function and genetics thereby confirming the sub-clonal architecture within this cell line. Intriguingly, our data point to the pre-existence of a dexamethasone resistant sub-clone the MM.1S^bright (CD38⁺) population. The subsequent production of the dexamethasone resistant cell line (MM.1R) allowed us to perform comparative genomics thereby identifying the genetic origins of dexamethasone resistance (selection) in MM.1S^bright cells and to track the subsequent clonal evolution (induction) in the MM.1R cells. Furthermore, we showed the potential for developing bespoke treatment plans based on the identification of cell signalling pathway mutations via genomic sequencing. By selective targeting of of these genetic lesions it may be possible to remove multiple sub-clones thereby diminishing the potential for clonal tiding and the development of drug resistance. In theory this could result in longer time to relapse and ultimately improved overall survival.