Single Cell Analysis of Acquired Pomalidomide-Resistance in Multiple Myeloma Cell Lines Reveals Distinct Subclonal Cereblon Mutations and Gene Expression Heterogeneity

Background: Multiple myeloma (MM) is a clonally heterogeneous cancer that recurrently relapses during its clinical course, demonstrating how interplay among subclones can lead to differentially drug-tolerant populations. The determination of onset and mechanisms of drug resistance is of clinical interest, and the improved understanding of this will drive innovative therapeutic approaches.

The IMiD® agents lenalidomide (Len) and pomalidomide (Pom) bind to the E3-ubiquitin ligase receptor, Cereblon (CRBN), leading to ubiquitination and subsequent proteosomal degradation of key transcription factors Ikaros (IKZF1) and Aiolos (IKZF3). As a consequence of this degradation, direct antitumor and immunostimulatory effects are observed. Previous studies have revealed potential mechanisms of resistance to Len and Pom in both MM cell lines and clinical cohorts, including mutations, transcriptional downregulation or alternative splicing of CRBN, or mutations elsewhere in the E3 ligase pathway (CUL4A, DDB1), IKZF1/3 or IRF4. In resistant cell lines and patients, activation of proliferation, cell cycle, IGF-1, MAPK and FOXM1 pathways are reported, in addition to non-activation of interferon signature genes (ISGs).

We hypothesize that rare MM cells/subclones with genetic, epigenetic or proteomic architecture permitting survival against Len or Pom in vitro may exist prior to acquired resistance, but may not be detectable by bulk genomic, transcriptional or proteomic analyses. Our goal is to define such populations using a combination of single cell technologies, initially focusing on acquired Pom-resistance mechanisms and quantifying genetic and transcriptional changes in Cereblon and its mechanistic targets, and to identify features of pre-existing resistant cell populations.

Methods: To investigate the mechanism and timing of onset of acquired resistance, we utilized H929 and MM1S MM cell lines made resistant to Pom by serial dose increments over time, to a final Pom concentration of 10 μM (H929/PR and MM1S/PR respectively). Samples were collected at various timepoints for analysis by genomic, bulk and single cell (10X genomics) RNASeq techniques, and single cell protein quantification of Cereblon, Ikaros and Aiolos by IHC.

Results: Bulk genomic sequencing revealed the H929/PR cell line had gained two CRBN mutations, a frameshift deletion of the glutarimide-binding region present in 10.9% alleles, and a point mutation in the non-glutarimide binding region present in 71% alleles. Concurrent single-cell protein quantification showed a near complete loss of Cereblon expression in the majority of cells, with little change in Ikaros or Aiolos staining. Bulk RNASeq of the H929/PR cell line showed a significant enrichment of cell cycle, E2F and MYC target gene sets. Using single cell transcriptomics, we found high expression of the signature genes of the Pom-resistant population to co-localise with an independently derived FOXM1 target gene set to a cluster of sensitive single cells defined by a gene expression profile identifying cells at the G2M checkpoint.

Sanger sequencing of CRBN in the MM1S/PR cell line showed a de novo 12 bp deletion in intron 5 which is required for spliceosome assembly, resulting in deletion of exon 6 of the mature mRNA transcript. Interestingly, CRBN exon 6 is required for interaction with DDB1, suggesting loss of a functional E3-ligase. Analysis by single cell IHC showed substrate levels similar to sensitive parental cell lines. As in the H929/PR cell line, bulk RNASeq again showed enrichment of E2F and MYC targets, and independently derived Pom-resistant and FOXM1 target gene sets co-localised to the same cluster of MM1S drug-naïve single cells, identified by a G2M checkpoint signature.

Conclusions: A range of CRBN mutations, Cereblon protein loss and enrichment of cell cycle, FOXM1 target and G2M checkpoint-related signatures was found at the onset of Pom resistance. This association with altered cell cycle progression dynamics may lead to, correlate with or result from drug survival. Tracking mutation- and non mutation-associated transcription factor signatures through sequential time points from first drug exposure to full resistance may reveal mechanisms by which functional CRBN loss causes drug resistance.