Novel Method to Detect Minimal Residual Disease in Multiple Myeloma Predicts Recurrence Better Than CD138-Based Models

Introduction: Diagnosis, responses to treatment, minimal residual disease (MRD) and circulating tumor cells (CTC) in multiple myeloma (MM) are all detected by the gold-standard marker CD138 flow cytometry-based method. However, the presence of clonogenic CD138-negative MM cells and hypoxia-driven CD138 downregulation were shown previously. In this study, we found that CD138 is significantly downregulated in MRD and CTC populations in MM, thus we developed a novel two-color flow cytometry-based set of biomarkers (independent of CD138) to detect MRD and CTC in MM.

Methods: We created a MRD mouse model by treating MM-bearing mice with a high dose of bortezomib. We then tested the effect of hypoxia and drug treatment on the expression of different markers, including CD138. Therefore, to detect MM cells we utilized CD38-APC antibody, followed by the exclusion of non-MM CD38-expressing cells such as T cells, monocytes, NK cells, B cells and dendritic cells using (CD3, CD14, CD16, CD19 and CD123)-V450 antibodies, respectively. To verify the ability of the new method to selectively detect MM cells, mononuclear cells from peripheral blood (PB) from healthy patients and MM cell lines (hypoxic and normoxic) were stained by the antibody cocktail and analyzed by flow cytometry. Next, we compared the sensitivity of the traditional CD138 method to the new method in detecting (a) normoxic or hypoxic MM cells spiked into bone marrow (BM) cells in vitro, (b) normoxic or hypoxic MM cells spiked into PB in vitro, (c) MM1s-GFP+ cells in mice with different tumor burden in the BM in vivo, (d) circulating MM1s-GFP+ in the PB in vivo, (e) MRD cells in the BM samples from 16 patients with complete remission (CR) or very good partial response (VGPR), and (F) CTC in 12 progressive MM patients. We further aimed to predict time to progression (TTP) in 16 patients with complete remission based on the detection of MRD in these patients using the new method and compared with flow cytometry-based CD138 or histology.

Results: In vivo, we found that bortezomib-treated MRD cells were hypoxic, compared to a progressive vehicle-treated cells. CD138 expression in these cells was significantly decreased, but CD38 expression was unchanged. In vitro expression of CD138 was decreased due to hypoxia and bortezomib treatment, whereas CD38 expression was unchanged. Furthermore, we developed a new method using an antibody cocktail, where the MM cell population is defined as CD38+/CD3-/CD14-/CD16-/CD19-/CD123- (APC+ and V450-).

In vitro, the new method detected 100% of hypoxic and normoxic MM cells, and less than 0.5% of mononuclear cells from the PB or BM of healthy donors. In contrast, CD138+ cells failed to detect 50% of hypoxic MM cells and 10-25% of normoxic cells.

In vivo, the amount of cells detected by the new method directly correlated with the number of MM1s-GFP+ cells detected in the BM with a range between 0-60%, with a correlation coefficient (slope) of 0.99 and R² of 0.999. The CD138 detected only a fraction of the MM1s-GFP+ population (<10%) even in mice with a high BM tumor burden. In addition, the new method detected close to a 100% of the circulating MM1-GFP+ cells in vivo, while the CD138 marker detected less than 1% of the circulating MM1s-GFP+ cells.

In patients, the new method detected 0.5-8% MRD cells in the BM of 16 patients with CR or VGPR (which were defined as CD138-negative), and 0.1-1.8% CTC in the PB of 12 progressive MM patients. In contrast, CD138 marker detected less than 0.5% of MRD cells in the BM of the CR and VGPR patients, and less than 0.1% of CTCs in the PB of the progressive patients. Furthermore, we found that, while CD138 and histology failed to predict recurrence in CR patients, the new method successfully detected CD138-negative MM population whose prevalence in the BM inversely correlated with TTP in MM patients defined as CR based on CD138 and histology.

Conclusions: We confirmed that CD138 expression is variable on MM cells, and that it is downregulated in MRD and CTC populations, and that it was not effective in detecting these particular populations in MM. Furthermore, we developed a novel two-color flow cytometry-based biomarker-set to detect MM cells independent of CD138. The new methods detected close to a 100% of all MM cells in vitro and in vivo, including MRD and CTC. Moreover, the new method detected a CD138-negative MRD and CTC in MM patients, and the prevalence of this population inversely correlated with TTP in MM patients.