Osteoblastic niche supports the growth of quiescent multiple myeloma cells

In this report we used a lipophilic fluorescent marker, PKH, to understand the heterogeneity within multiple myeloma (MM) populations and to determine the effects of niches on the heterogeneity of MM cells.¹  The intensity of PKH staining decreases with each cell division in a linear fashion, resulting in the labeling of quiescent cells within a proliferating population.^2,3 

We allowed the PKH-labeled MM cells to undergo the cell cycle within NOD/SCID mice,, and we tracked the PKH⁺ cells in different niches. In particular, we focused on analyzing the effects of different niches within the bone on MM cells because the bone marrow localization of MM cells is reminiscent of hematopoietic stem cells.⁴  Analysis of 70 NOD/SCID mice revealed that after in vivo cycling, quiescent PKH⁺ cells prefer to reside in the endosteal/osteoblastic regions (OS) of the bone marrow compared with the vascular (VS) regions or spleen. PKH⁺ cells from the osteoblastic niche (PKH⁺/OS) readily formed into colonies in PHA-LCM medium. The PKH⁺ MM cells were also highly proliferative in secondary xenograft assays and were resistant to select drugs compared with the PKH^– MM cells. Our report is the first to show that quiescent MM cells have a preference for niches. Given that bone marrow provides a protective environment for MM cells to proliferate and survive during treatment, it is important to target these interactions to improve the efficacies of the current therapies and to eventually improve patient survival.

Cells (RPMI and NCI, 5 × 10⁶) were stained with PKH67 (final concentration of 2 × 10⁻⁶ M) for 5 minutes. Stained cells were examined for PKH expression using fluorescence-activated cell sorting (FACS) analysis to ensure that >99% of cells were PKH⁺. Details on other methods can be found in the supplemental Methods, available on the Blood Web site. This study was conducted in accordance with the Declaration of Helsinki.

Because the bone marrow localization of MM cells is reminiscent of the localization of hematopoietic stem cells, we compared the frequencies of PKH67⁺ cells in different niches that are important for MM proliferation, such as the OS regions of the bone and the VS regions of the bone. We also analyzed the spleen for PKH engraftment for comparison.

At 24, 48, and 72 hours post-transplant, the recipient mice were sacrificed to analyze the PKH⁺ cells. Similar to the hematopoietic stem cells reported previously,³  the highest amounts of PKH⁺ cells were recovered after 48 hours of in vivo cycling (supplemental Figure 1A). Therefore, we performed the following analyses at 48 hours post-transplantation.

Mice were sacrificed 48 hours post-transplantation with PKH67⁺ cells, and the number of PKH67⁺ cells in each niche was analyzed using FACS. After averaging a total of 70 mice, 11.56% of the PKH67⁺ RPMI MM cells were recovered from the OS niche, 7.65% from the VS niche, and 4.3% from the spleen. For the NCI MM cells, 18.3% of the cells were recovered from the OS niche, 8.9% were recovered from the VS niche, and 3.4% were recovered from the spleen (Figure 1A and supplemental Table 1). Quantitative real-time polymerase chain reaction analyses of the cells from the OS or the VS niches demonstrated that the mature bone marker osteocalcin and the bone morphogenetic proteins BMP2, BMP4, and BMP7 were more highly expressed in the cells from the OS niche than in the cells from the VS niche (supplemental Figure 1B), indicating a clear separation of the bone marrows from the bone lining tissues.

Immunohistochemistry analyses of xenograft organs confirmed the FACS data. Increased engraftment of PKH67⁺ (green) and CD138⁺ (red) cells were found in the OS niches compared with the VS and spleen niches (Figure 1B).

In addition, analyses of unmanipulated bones showed the location of the PKH⁺ cells near the OS regions of the bones. In contrast, the PKH^– CD138⁺ cells were mainly located away from the bone-lining regions. It is also noted that fewer PKH^– CD138⁺ cells were found in the same sections compared with PKH⁺ cells (supplemental Figure 2A-B).

Interestingly, the presence of PKH67⁺ cells that did not express the plasma cell surface marker CD138 were found after 48 hours of in vivo cycling (Figure 1B). These results indicate that most quiescent populations within MM can be either CD138⁺ or CD138^–.

Therefore, we separated the MM cell lines or patient cells into CD138⁺ and CD138^– cells and separately labeled those cells with PKH67 before transplantation into the mice. Although there was some variability, in general the CD138^– cells derived from patients and cell lines retained the PKH staining better than CD138⁺ cells in vivo (Figure 1C), indicating that more CD138^– MM cells remained quiescent in vivo compared with the CD138⁺ cells. The OS regions of the bones were still the preferable niche for both PKH⁺CD138⁺ and PKH⁺CD138^– cells.

To increase OS niche specificity, we removed the monocyte lineages from the patient cells using CD14 antibodies. Slightly increased OS engraftment rates were observed upon transplantation of PKH⁺CD138^– cells compared with the PKH⁺CD138⁺ cells. The OS niche was still a favorable niche for both cell populations (supplemental Table 2). We then compared the phenotypes of the CD138^– MM cells with previously published MM stemlike cells using FACS analysis. The CD138^– MM cells were enriched with CD20 and CD27 cell surface markers, as previously reported⁵  (supplemental Figure 2C-D). V(D)J recombination assays demonstrated the clonal relationship between the CD138⁺ and CD138^– cells (supplemental Figure 2E).

We then measured the clonogenic properties of these cells using the PHA-LCM methylcellulose assay.⁵  Compared with the cells from other niches, the PKH⁺ cells isolated from the OS niche (PKH⁺/OS) formed the most colonies in PHA-LCM medium (Figure 1D and supplemental Table 3). Although some PKH67⁺ cells were found in the spleen (PKH⁺/SP), those cells failed to form colonies. The colonies derived from RPMI or NCI cells contained the same light-chain restriction patterns as the parental cell lines (supplemental Figure 3A). These data suggest that PKH⁺/OS cells are readily equipped to form colonies in vitro, which suggests that the niches that the cells reside in in vivo can influence the cells’ properties.

We then transplanted these cells in secondary recipient mice to observe the ability of the PKH⁺/OS cells to form tumors (supplemental Figure 3B). After 3 months, the organs were isolated from each mouse and were evaluated for tumor engraftment by FACS analyses using human CD138 as a marker. The PKH⁺CD138⁺ cells showed a significantly higher level of engraftment in the secondary mice than the PKH^–CD138⁺ cells (Figure 2A). Interestingly, the PKH⁺ RPMI cells from the OS niches (20%) showed a higher CD138⁺ engraftment rate than the PKH⁺ cells from the VS niche (∼7%) or the spleen (∼5%) (Figure 2B). However, the PKH⁺/OS cells from the NCI cells engrafted to the OS, VS, and spleen at comparable levels, indicating that the homing ability of the NCI MM cells to locate to specific niches was somehow lost during serial transplantation. Immunohistochemistry results also supported the data from FACS analyses (Figure 2C). The increased clonogenic activity or tumorigenic properties of the PKH⁺ cells were not the result of differences in cell viability. Only live cells were sorted based on 7-AAD staining, before transplantation to the secondary xenografts (supplemental Figure 3C), and both the PKH⁺ and PKH^– cells showed similar viability in vitro (supplemental Figure 3D).

Because cellular quiescence or stemlike properties are often linked to drug resistance in some tumors,⁶  we tested the drug susceptibility of the PKH⁺ or PKH^–CD138⁺ cells in vitro using bortezomib, cyclophosphamide, vincristine, rituximab, and prednisone. Regardless of where the cells were isolated from, the PKH⁺ cells were more resistant to the drugs than the PKH^– cell populations (Figure 2D). In summary, our report is the first show that quiescent MM cells preferably reside within OS niches. We discovered that both OS and VS niches in the marrow can support the proliferation of MM cells. However, the OS niches are the primary niches that support stemlike functions of MM cells. Our data also suggest that stemlike functions do not necessarily represent drug resistance, which is more likely governed by cell cycle status rather than the location of these cells, and these findings offer valuable preclinical models for drug testing and screenings for future applications.

The tissue samples were provided by the University of Texas MD Anderson Cancer Center Satellite Multiple Myeloma Tissue Bank. The authors thank Dr Naoki Nakayama (Center for Stem Cell and Regenerative Disease, IMM) for the useful discussion on bone-related markers.

This work was supported by The Anderson Cancer Center SPORE in multiple myeloma (P50CA142509).

Contribution: Z.C. performed and analyzed the experiments; R.Z.O., M.W., and L.K. contributed the clinical samples; and N.M. supervised and planned overall experiments and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Nami McCarty, University of Texas-Health Science Center at Houston, 1835 Pressler St, IMM-630H, Houston, TX 77030; e-mail: nami.mccarty@uth.tmc.edu.