The drug treatments currently available for multiple myeloma patients are dramatic improvements over historical regimens, stopping or slowing cancer growth in 80–90% of patients and leading to complete remission in approximately 40% of patients. Many of the new treatment regimens include “novel agents” in combination with dexamethasone, one of the most effective agents used to treat myeloma. The direct mechanism by which dexamethasone works in myeloma is not well characterized but it is assumed that it activates glucocorticoid receptors which results in gene expression changes that promote apoptosis in lymphoid cells. However, often the disease becomes resistant to dexamethasone, and the mechanism for this resistance is not entirely known.
To study the mechanism of resistance, two isogenic cell lines, MM.1R and MM.1S, were independently created from the parental cell line MM.1 to represent models of resistance and sensitivity, respectively, to dexamethasone. This model system was created by Steve Rosen and colleagues in the 1990s and was recently deposited in ATCC. Previous studies have demonstrated differential expression of the glucocorticoid receptor NR3C1 but have not precisely identified the genetic difference between MM.1R and MM.1S across the whole genome.
To better understand the mechanism behind the differences in drug sensitivity between these isogenic cell lines, we performed extensive characterization of MM.1R and MM.1S. We purchased both lines from ATCC and analyzed each using flow cytometry, CGH, CGH-SNP, mRNA sequencing, and exome sequencing. First, we broadly examined both cell lines, demonstrating a 300,000-fold difference in IC50 of MM.1R to MM.1S after 6 days of dexamethasone treatment. No significant ploidy difference was found between the two lines by flow cytometry analysis.
Our CGH results identified 4 copy number differences unique to MM.1R (chr2:p37.1–37.3 deletion, chr4:q32.3–33 deletion, chr5:31.3 deletion, and chr7:q36.3 amplification), the third of which suggested a possible homozygous deletion within NR3C1. To confirm this deletion, we designed primer sets at ∼1kb intervals spanning the entire NR3C1 gene and performed PCR on MM.1R and MM.1S. Our results indicate the presence of a ∼5–8kb deletion of NR3C1 in MM.1R. Additionally, we analyzed our mRNA sequencing data using TopHat-Fusion and identified an inverted fusion between NR3C1 and ARHGAP26, which we confirmed through PCR amplification and Sanger sequencing.
From mRNA sequencing, we identified 63 genes with differential expression between MM.1R and MM.1S (FPKM > 5 in either cell line and greater than fourfold change between them). These results demonstrate a reduction in expression of NR3C1 caused by the two independent deletions identified by CGH. The gene with the larges fold change was MGST1, which is associated with drug resistance and thus may be associated with dexamethasone resistance in this model system based on its expression profile.
We analyzed our exome sequencing results for high-confidence (called by both SAMtools and GATK) non-synonymous mutations not present in the 1000 Genomes Project and filtered them for expression (FPKM > 5). We identified 218 mutations in MM.1R, 208 mutations which were also expressed in MM.1S and 10 mutations which were not expressed in MM.1S. The 10 genes with these mutations—PDIA5, TCERG1, RANBP9, MMS22L, PHF19, RNMTL1, AURKB, ERN1, GPCPD1, PIGT—present potential additional contributors to dexamethasone resistance. Specifically, for example, overexpression of RANBPM (the protein from RANBP9) results in increased glucocorticoid activity, suggesting that it may work in concert with NR3C1 to mediate the effects of dexamethasone.
Ultimately, our results indicate that, unlike previous assumptions, there are several contributors to dexamethasone resistance in this model system and likely even more in the general patient populations, not just differential expression of NR3C1. Furthermore, we have discovered that this differential expression is due to biallelic inactivation of NR3C1 in MM.1R. Future studies will test the relative contribution of each factor to the differential sensitivity to dexamethasone observed in this model system and a broader understanding of this problem in multiple myeloma.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.