Insights into Biological Mechanisms Responsible for Egression of Circulating Tumor Plasma Cells (CTCs) in Multiple Myeloma

Background:

CTCs have become powerful prognostic factor in newly diagnosed multiple myeloma (NDMM). CTCs reflect tumor burden and global capacity for proliferation and trafficking and thus mirror the biology of underlying disease. Nevertheless, there is limited knowledge about the mechanisms responsible for egression of CTCs to peripheral blood (PB), specifically explaining highly variable levels of CTCs across MM population.

Aim:

Investigate biological differences in clonal bone marrow plasma cells (BMPCs) across NDMM patients (pts) with variable levels of CTCs from undetectable/low levels to high levels/plasma cell leukemia (PCL).

Methods:

CTCs were measured in 116 NDMM pts during 3 years at single institution using NGF according to EuroFlow protocol. Pts were divided into 7 groups defined by CTC level using logarithmic (log) ranges as follows: undetectable; 0.0002-0.001% (10^-6); 0.001-0.01% (10^-5); 0.01-0.1% (10^-4); 0.1-1% (10^-3); 1-10% (10^-2) and 10-100% (10^-1). Clonal BMPCs were FACS sorted followed by isolation of DNA/RNA. In order to represent patients with high CTCs a cohort of PCL patients (N=7) was added for genomic analysis. FISH results were available in 113 pts and del 1p, del 17p, t(4;14) and t(14;16) were considered high risk cytogenetic alterations (HRCA). RNA sequencing was performed in 40 pts proportionally representing each log range group. To identify set of key genes for increasing CTC levels we employed a linear regression model on median gene expression values, combined with pairwise comparisons and intersected with high/low CTC analysis. To validate the impact of significant genes, the overall survival (OS) analysis associated with high/low gene expression was performed on CoMMpass dataset.

Results:

The median CTC level was 0.03% and the distribution of CTC frequencies across log groups followed a Gaussian pattern with majority of pts included in the middle groups (10^-5, 10^-4 and 10^-3 logs). Interestingly, MGUS-like phenotype (<95% clonal BMPCs in PC compartment) had 89% of pts with undetectable CTCs, 14.3% in 10^-6 and 12% in 10^-5, but no MGUS-like pattern was observed in pts with 10^-4 or higher CTC level (p=0.005). FISH data showed that HRCA, 1q gain/amp, del 13q and t(11;14) were significantly more frequent in pts with CTCs above median of 0.03% and their frequency positively correlated with level of CTCs. This was translated into higher median CTC numbers in pts with HRCA vs without (0.226 vs 0.023%; p=0.002), 1q gain/amp (0.065 vs 0.008%; p=0.02) and del 13q (0.07 vs 0.01%, p<0.001).

Next, we leveraged this unique cohort with progressively increasing CTC numbers to identify key genes involved in CTC biology. By focusing on deregulated genes between high and low CTC groups and examining genes with the most linear changes across all groups, we identified a set of 41 hub genes (7 genes decreasing and 34 increasing in CTC high groups). These genes' impact on OS was supported by analysis in the CoMMpass dataset, where 73% (30/41) of hub genes were prognostic, following the same direction of change observed in the RNAseq analysis.

Interestingly, the majority of 41 identified hub genes are involved in processes that might enable PCs to become independent on BM microenvironment:

- 11 are linked to the cytoskeleton and adhesion, e.g., RHOC, PECAM1, MAP1B, TAGLN2, AHNAK, FLNA, with latter 4 already shown to be deregulated in PCL (Bruinink 2022, Ramberger 2024);

- 11 play roles in protein synthesis and proliferation e.g., MCM2, DCK, AREG;

- 4 related to metabolism e.g. ENO1 involved in glycolysis and lactate production or MTHFD2, both are suggested as potential drug targets;

- 4 involved in transcriptional regulation e.g., ZHX2, EZH2, BRCA1;

Finally, we observed that more than 24% (10/41) of hub genes were located on chr1, which might be explained by more frequent presence of 1q gain/amp in higher CTC level groups.

Conclusion:

NDMM pts with undetectable or low CTCs levels are enriched of MGUS-like phenotype, while increasing CTC numbers positively correlate with presence of HRCA, 1q gain/amp, del 13q and t(11;14). Using RNA seq we identified 41 key genes associated with gradually increasing levels of CTCs with 73% (30/41) of them confirmed as prognostic using the CoMMpass dataset. These genes including EZH2, BRCA1, ENO1, DCK, PECAM-1, RHOC, FLNA etc. represent promising therapeutic targets and/or provide insights into the molecular mechanisms driving CTC egression.