Genomic Profiles of Bone Marrow (BM) Clonal Plasma Cells (PCs) Vs Circulating Tumor Cells (CTCs) and Extramedullary (EM) Plasmacytomas in Multiple Myeloma (MM)

EM disease in MM increases from newly-diagnosed into relapsed patients, and typically predicts for inferior survival. In fact, no treatment seems to be effective in cases with plasma cell (PC) leukemia, which represents the most aggressive form of EM MM. Unfortunately, the mechanisms of extramedullary spread in MM are not well understood, and there is almost no data about the genetic landscape of both forms of EMD (plasmacytomas and peripheral blood CTCs).

Here, we performed whole exome sequencing (WES) to analyze the genomic profiles of highly purified FACS sorted BM and EM clonal PCs from 6 patients with relapsed MM. In 5/6 cases we had all three tissue specific clones, whereas in the remaining case no CTCs were detectable. Depending on the amount of genomic DNA from each clone, whole-genome amplification was performed and in such cases, triplicates of 10 ng of DNA were amplified up to 10 µg using the Illustra GenomiPhi HY DNA amplification kit (GE healthcare). Enrichment of exonic sequences was performed for each library using the SureSelectXT Human All Exon V5+UTRs capture kit (Agilent). To identify somatic mutations we used the mutation caller named Varscan. Variants potentially affecting protein function, including non-synonymous variants, frameshifts in the coding sequence, stop codon-introducing (nonsense) or variants potentially affecting splicing, were analyzed. Only those mutations present in 2/3 libraries analyzed per sample were considered positive.

Overall, a median of 89 (67 - 474) somatic mutations were detected. EM plasmacytomas showed the highest mutation load, followed by CTCs and BM clonal PCs (85 vs 77 vs 75, respectively; P=.07), supporting a higher genomic instability/evolution of EM clones. Strikingly, all 6 cases showed lack of concordance in the mutation profiles of the three tissue related clones; even 2x2 comparisons between BM clonal PCs vs CTCs or plasmacytomas, or between the two forms of EM MM (ie. plasmacytomas vs CTCs) showed lack of 100% concordance in every single patient. Despite high inter-tissue heterogeneity, it should be noted that whenever present, recurrent and potentially actionable mutations in genes such as KRAS (n=3), KDM4A (n=2), KMT2A (n=2), ARID5B (n=2), TRIO (n=2), BRAF (n=1) or CCND1 (n=1) were detected in all three clones, with the exception of one patient in which CTCs lacked a KRAS mutation. Only one and less frequent actionable mutation in the EPHB2 gene (eg. Herceptin) was exclusively noted in the EM plasmacytoma from one patient.

In order to understand the cellular origin of the three tissue related subclones, we then investigated the degree of similarity between each pair of clones at the individual patient level. The pair of clones showing the highest similarity in their mutation profiles were EM plasmacytomas with CTCs (60%, 30% - 94%), followed by BM clonal PCs with CTCs (58%, 35% - 92%), and by BM clonal PCs with EM plasmacytomas (57%, 37% - 91%). Thus, these data suggests that while CTC clones may represent a cellular bridge in between BM and EM plasmacytomas, there is continuous genomic evolution once the different clones have seeded in their respective tissue niches. Accordingly, EM plasmacytomas showed the highest number of specific mutations, followed by BM clonal PCs and CTCs (medians of 9, 4 and 3, respective). We then looked for recurrent mutations specifically present in EM clones, and found that CTCs showed exclusive and recurrent mutations in the CEP152, FSIP2, SYNE1 and TENM1 and ZNF585A genes, whereas EM plasmacytomas showed exclusive and recurrent mutations in the PITRM1 (Metalloprotease 1) gene. Furthermore, ATP7B, MTOR, TBC1D21 and ZNF717 mutations were present in both CTCs and plasmacytomas.

In summary, we compared for the first time the genomic profiles of BM vs EM CTCs and plasmacytoma clones, and showed at the individual patient level, the existence of high mutation heterogeneity in between all three tissue related clones. However, actionable mutations in genes such as KRAS or BRAF were typically shared by BM and EM clones. According to the level of similarity between each clone, CTCs would be the most plausible precursor of EM plasmacytomas. Nevertheless, based on the number of specific mutations present in EM clones, it is likely that EM spreading followed by continuous clonal evolution starts much earlier prior to its clinical detection, which emphasizes the need for sensitive techniques to capture the early stages of EM spreading.