Abstract
MRD monitoring is emerging as a powerful prognostic biomarker in MM. In addition to its utility to redefine the depth of response achieved after therapy, the MRD clone represents, from a biological point of view, a very small fraction of tumor cells that are truly chemoresistant, potentially quiescent (not producing M-protein), and able to recapitulate the initial tumor burden at relapse. Thus, the MRD clone represents a unique model to understand chemoresistance (in otherwise serological responding patients) and the characteristics of eventual MM clonogenic cells; however, the MRD clone has never been characterized at the biological level. Here, we compared the immunophenotypic and genomic profiles of paired baseline vs MRD clonal PCs from a total of 40 newly-diagnosed elderly MM patients enrolled in the GEM2010 trial: sequential VMP (9 cycles) followed by Rd (9 cycles) vs alternating VMP with Rd (18 cycles). MRD monitoring was performed by multidimensional flow cytometry (MFC) at cycles 9 and 18.
First, we focused on the immunophenotypic expression profiling (iPEP) of patient-paired baseline vs MRD clonal PCs through 23-color MFC, combining backbone markers for the identification of clonal PCs with additional integrins, adhesion, activation and maturation molecules. Our results show that the iPEP from MRD clonal PCs at cycle 9 differed from paired baseline tumor cells by significantly (P<.05) increased MFI of CD11c, CD29, CD44, CD49d CD49e, CD54, and CD138, indicating that among the initial tumor bulk, the few chemoresistant cells were those with stronger expression of integrin and adhesion molecules. This phenomenon was particularly evident in patients randomized to the sequential arm; in fact, only after VMP there was a specific selection of CD117+chemoresistant clonal PCs (P=.03). Moreover, MRD clonal PCs showed significantly increased MFI of CD28 (a pro-survival mediator through dendritic cell interaction) and HLADR.
Using patient-specific aberrant phenotypes, we then sorted clonal PCs (purity ≥97%) by FACS for subsequent genomic studies. Patient-specific paired comparison of baseline vs MRD clonal PCs (n=11) genomic profile was performed by high density Cytoscan750K array. The pattern of copy number abnormalities (CNA; only those with minimum of 25 consecutive imbalanced markers/segment and minimum 100 Kb length were considered) significantly varied from baseline to the chemoresistant MRD PC clone (accounting for whole chromosome, chromosomal arm or interstitial imbalances). Individual patient analysis showed one case in which both PC clones showed exactly the same 25 CNA; a second case in which MRD cells showed +1p36.21, -9p23, -17q11.2 and +17q25.3 in addition to 10 CNAs already present at baseline; 4 cases in which MRD cells maintained initial chromosomal imbalances but lacked specific CNA (range: 1-34) present at diagnosis; and 5 patients in which MRD cells displayed none of the CNA (range: 7-20) detected at baseline. It should be noted that markedly subclonal heterogeneity was detected at diagnosis. Accordingly, in one out of the 5 latter cases baseline but not MRD clonal PCs had 1q+ and 13q- which were further confirmed by FISH (in 85% and 58% of clonal PCs, respectively); this suggests that MRD PCs were present at the baseline state as a minor sublclone with respect to the initial tumor bulk. On the other hand, copy number neutral loss of heterozygosity (larger than 3Mb) was slightly more frequent in MRD as compared to patient paired baseline clonal PCs (13 vs 8, respectively).
Only in 5/40 patients sufficient RNA was extracted from baseline and paired MRD FACS-sorted clonal PCs to perform gene expression profiling (GEP; HumanGene 1.0ST). Although the numbers are small, it should be noted that GEP of baseline vs. patient-paired MRD clonal PCs mostly overlapped, with only 19 genes found to be down-regulated after chemotherapy (SAM Excel add-in with a FDR q-value<10-5). Interestingly, 4 of those genes were connected to the glucocorticoid receptor signaling pathway and receptor gene NR3C1, a mediator of glucocorticoids activity in MM cells.
In summary, this integrated phenotypic and genomic analysis suggests that from the baseline MM tumor bulk, primarily chemoresistant MRD PCs may be those potentially closer to the BM niche (with initial increased expression of integrins and adhesion molecules), as well as a more ancestral subclone (with less CNA) in a significant proportion of patients.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.