Phenotypic characterization of the human myeloma cell growth fraction

Multiple myeloma (MM) is primarily conceptualized as an accumulative disease. Indeed, most studies evaluating the labeling index (LI) of plasma cells (PCs) revealed that it rarely exceeded 1%. However, as early as 1981, Drewinko et al¹  showed that in vivo myeloma cells were generally nonproliferating, although a small fraction of them could proliferate (growth fraction, GF). Thus, they proposed the concept that this minor GF could give rise to the major nonproliferating fraction.

Myeloma cells either lack or express a weak to intermediate level of CD45.^2,3  However, in previous studies we have shown that CD45 and also CD11a are frequently expressed by a subset of myeloma cells only.³  It has been demonstrated that CD45 expression is highly correlated with the proliferation rate of myeloma cells.^4,5  With regard to normal counterparts, PCs are heterogeneous in terms of CD45 phenotype. Recently, Medina et al⁶  confirmed the association of maturity with decreasing CD45 expression.^7-9  Generation of PCs from B cells, mainly studied in vitro in humans, is a multistep process that involves both proliferation and maturation/differentiation.^10-13  The aim of this study was to evaluate the proliferation of different types of normal PCs in relation to their phenotype, especially to CD45, to understand the biology of polyclonal and monoclonal PC expansions that are reactive plasmacytoses (RP) and MM, respectively.

Forty-nine consecutive patients with MM (25 at diagnosis, 24 at relapse) were included in this study. Bone marrow and blood samples from healthy donors or patients with reactive plasmacytosis, and tonsil samples were obtained and prepared as described.^3,13  Approval was obtained from the Nantes University Hospital Institutional Review Board for these studies. Informed consent was provided according to the Declaration of Helsinki. Antibodies directed against (1) CD11a, CD19, CD45, CD138, Apo2.7 and (2) CD28, CD38, CD56, 5-bromo-2′deoxyuridine (BrdU) were from Beckman Coulter (Miami, FL) and BD Biosciences (San Jose, CA), respectively.

Mononuclear cells (MNCs; 0.5-3 × 10⁶) were stained in a 4-color assay with anti-CD45-fluorescein isothiocyanate (FITC; J33), anti-CD138-phycoerythrin cyanine 5 (PECy5; B-B4), anti-CD38-APC (HB7) and control-PE, anti-CD56-PE (MY31), anti-CD28-allophycocyanin (PE; L293), anti-CD19-PE (J4.119), anti-APO2.7-PE (2.7A6A3), anti-CD11a-PE (25.3) monoclonal anti-bodies (mAbs) as described.¹⁴  To determine CD45-FITC fluorescent staining, MNCs were stained with isotype-FITC, anti-CD38-APC, anti-CD45-PECy5, and anti-CD138-PE mAbs. For intracellular Bcl-2 staining, MNCs were first stained with anti-CD38-APC, anti-CD45-PECy5, and anti-CD138-PE mAbs, then with anti-Bcl-2-FITC (124; Dako, Glostrup, Denmark) or control isotype-FITC mAbs after permeabilization (Intra Prep; Beckman Coulter). For LI, cells were incubated overnight with or without 50 μM BrdU (5-bromo-2′-deoxyuridine; Sigma, St Louis, MO) at 37°C in culture medium, then stained with anti-CD45-APC, anti-CD38-PE, and anti-CD138-PECy5 mAbs (with anti-CD45-APC and anti-CD38-PE mAbs for tonsil and blood PCs), permeabilized, and stained with anti-BrdU as described.¹² 

Data were acquired by means of a Becton Dickinson fluorescence-activated cell sorting FACSCalibur with CellQuest Pro software (BD Biosciences).¹⁴  PCs were identified using a sequential gating strategy (Figure 1).

Statistical analyses were performed using nonparametric Wilcoxon rank sum, Spearman correlation coefficient, and the sign tests.

The percentage of BrdU-positive myeloma cells was low, generally around 1% (n = 49 patients). In all patients, myeloma cells were heterogeneous for CD45 expression and SSC, allowing a delineation of 2 compartments: CD45^brightSSC^high and CD45^int/lowSSC^low. The CD45^brightSSC^high compartment was always present, representing 12% (median) of myeloma cells (illustrated in Figure 1). The second compartment CD45^int/lowSSC^low represented the majority of tumor cells (88%). We observed that the LI of the CD45^brightSSC^high compartment was always greater than that of the CD45^int/lowSSC^low compartment and far above 1% (median, 6.8%; n = 49; Table 1). In general, the CD45^brightSSC^high compartment had a 6.8 higher proliferative level than the CD45^int/lowSSC^low compartment (P < .01). As outlined in Table 1, the LI of the CD45^bright compartment could reach 40% of myeloma cells, which was similar to what we observed in reactive plasmacytoses,¹²  suggesting that in these patients, almost all CD45^bright myeloma cells were proliferating. Of note, the total LI correlated highly with the LI of the CD45^bright compartment (r = 0.81, P < .001) and less with that of the CD45^low compartment (r = 0.44, P < .01).

As observed for myeloma cells, 2 compartments of PCs delineated by CD45 and SSC were found in normal bone marrows (n = 11, illustrated in Figure 1). The CD45^brightSSC^high compartment, which represented 65% of the PCs, was highly proliferative (LI = 18.4% for BM1). The second CD45^lowSSC^low compartment included 35% of the PCs with a much lower LI: 2.6% (BM1 and Table 1). A subset of PCs which lacked CD45 expression was observed in 2 cases.

In Figure 2, we show that PCs from both tonsils and blood, the prototypes of immature PCs, were homogeneously CD45^bright and capable of proliferation (LI > 10%). Reactive PCs turned out to be similar to blood PCs with higher LI.

Since we have previously observed that Bcl-2 inversely correlates with the LI in both reactive PCs and myeloma cells,¹⁶  we looked for a correlation within myeloma and PC subsets (Figures 1, 2). We found that CD45^bright myeloma cells always expressed a lower level of Bcl-2 compared with that of CD45^low/neg myeloma cells (P < .01). In normal bone marrow (BM), a significant increase in Bcl-2 MFIR (mean of fluorescence intensity ratio) was observed in the CD45^low bone marrow PCs in comparison with that of the CD45^bright PCs: 7.4 (4.9-13) versus 4.2 (3.1-9), P < .01. Concerning tonsillar and circulating PCs, we found a highly significant negative correlation between Bcl-2 MFIR and the LI (r = -0.854, P < .001). This inverse correlation in PCs is in agreement with data showing that (1) Bcl-2 negatively controls the proliferation of different human cells types,¹⁷  and (2) Bcl-2 transgenic mice have increased B-cell memory and PC-producing immunoglobulins.¹⁸  Up-regulation of Bcl-2 thus appears to be a critical event during this process of maturation with proliferation decrease and arrest.

We further characterized the phenotype of the CD45^bright myeloma cells using relevant MM markers (CD19, CD20, CD27, CD28, CD56). As illustrated in Figure 3, CD19, CD27, CD28, and CD56 were expressed or not by myeloma cells, and their expression was unrelated to CD45 subsets. However, we found that CD11a overlapped completely with the CD45^bright compartment in all the patients studied (Figure 3), while CD45^low/neg myeloma cells were negative for CD11a. Ahsmann et al¹⁹  have published that lymphocyte function-associated antigen 1 (LFA-1) (CD11a-CD18) expression correlated with tumor growth in MM. LFA-1 is involved in either homotypic or heterotypic interactions in MM. Indeed, human myeloma cells, like stromal cells, express the 3 ligands of LFA-1: intercellular adhesion molecule-1 (ICAM-1) (CD54), ICAM-2 (CD102), and ICAM-3 (CD50).^3,20  These interactions, restricted to the most proliferative compartment, could have some important consequences for tumor behavior. For example, we have previously observed that ICAM-2/LFA-1 interactions were involved in the (negative) control of myeloma cell growth through CD40.²⁰ 

Interleukin 6 (IL-6) has been shown to be a survival and growth factor for human myeloma cells.^21,22  More recently, it has been demonstrated that IL-6 preferentially stimulates CD45^bright myeloma cells to proliferate^23,24  through activation of CD45-associated src kinase.²⁴  Of note, CD45^bright myeloma cells express more IL-6 receptors.²⁵  IL-6 is also a survival and proliferative factor for nonmalignant PCs.^12,13,26  Taken altogether, these data show that CD45^bright myeloma cells, like their normal counterparts (CD45^bright normal PCs), could be the main target of IL-6 to sustain survival and proliferation.

In conclusion, we have found that all patients with MM have a small proliferative compartment of myeloma cells characterized by a bright expression of CD45 and a specific expression of CD11a as well as a low Bcl-2 expression (sensitive to apoptosis). This CD45^brightCD11a^pos population of myeloma cells could constitute the growth fraction as defined in vivo by Drewinko et al¹  more than 20 years ago. Thus, this “to be killed” population could be targeted through CD45- or CD11a-targeted therapies.