Bone morphogenic protein 6: a member of a novel class of prognostic factors expressed by normal and malignant plasma cells inhibiting proliferation and angiogenesis

Article metrics

Abstract

Pathogenesis of multiple myeloma is associated with an aberrant expression of pro-proliferative, pro-angiogenic and bone-metabolism-modifying factors by malignant plasma cells. Given the frequently long time span from diagnosis of early-stage plasma cell dyscrasias to overt myeloma and the mostly low proliferation rate of malignant plasma cells, we hypothesize these to similarly express a novel class of inhibitory factors of potential prognostic relevance. Bone morphogenic proteins (BMPs) represent possible candidates as they inhibit proliferation, stimulate bone formation and have an effect on the survival of cancer patients. We assessed the expression of BMPs and their receptors by Affymetrix DNA microarrays (n=779) including CD138-purified primary myeloma cell samples (n=635) of previously untreated patients. BMP6 is the only BMP expressed by malignant and normal plasma cells. Its expression is significantly lower in proliferating myeloma cells, myeloma cell lines or plasmablasts. BMP6 significantly inhibits the proliferation of myeloma cell lines, survival of primary myeloma cells and in vitro angiogenesis. A high BMP6 expression in primary myeloma cell samples delineates significantly superior overall survival for patients undergoing high-dose chemotherapy independent of conventional prognostic factors (International Staging System (ISS) stage, β2 microglobulin).

Introduction

Multiple myeloma (MM) is an incurable malignant disease of clonal plasma cells that accumulate in the bone marrow (BM) causing clinical signs and symptoms related to the displacement of normal hematopoiesis, BM neovascularization, formation of osteolytic bone lesions and production of monoclonal protein (Vacca et al., 1994; Kyle and Rajkumar, 2004). Pathogenesis of MM is associated with an aberrant expression of pro-proliferative, pro-angiogenic and bone-metabolism-modifying factors by malignant plasma cells (MM cells, MMCs), or their induced production in the BM microenvironment. Several of these factors have been identified by others and by us (Zhan et al., 2002; Klein et al., 2003; De Vos et al., 2006; Hose et al., 2009a).

Given the frequently long time from first diagnosis of early-stage plasma cell dyscrasias to overt myeloma and the mostly low proliferation rate of MMCs (Witzig et al., 1999), we hypothesize these to express a novel class of inhibitory factors of potential prognostic relevance. Bone morphogenic proteins (BMPs), especially BMP6, represent a possible candidate, as they are expressed in MMCs (Zhan et al., 2002) and inhibit the proliferation of myeloma and memory B cells (Ro et al., 2004; Kersten et al., 2005). It likewise stimulates osteoblast differentiation (Ebisawa et al., 1999), osteoclast development (Wutzl et al., 2006) and bone formation (Cheng et al., 2003). In different cancer entities, BMP6 shows a contradictory effect on patients' survival, whereas in renal cell carcinoma, BMP6 acts as an inhibitor of tumor growth (Kim et al., 2003). In prostate (Hamdy et al., 1997; Haudenschild et al., 2004; Dai et al., 2005) and breast cancer (Clement et al., 1999), BMP6 expression promotes tumor progression and metastasis. BMP6 has been described to be a pro-angiogenic factor in terms of endothelial cell migration and tube formation in vitro (Valdimarsdottir et al., 2002).

Bone morphogenic proteins are members of the transforming growth factor-β superfamily, and exert an effect by binding to two different types of serine/threonine kinase receptors. Three type I receptors bind BMPs: activin-like kinase-2, (Alk-2, ACVR1), activin-like kinase-3 (Alk-3, BMPR1A) and activin-like kinase-6 (Alk-6, BMPR1B). Similarly, three type II receptors have been identified, namely, BMP receptor II (BMPR2), activin type II receptor (ActRII, ACVR2) and activin type IIB receptor (ActRIIB, ACVR2B) (Ebisawa et al., 1999). Both type I and type II receptors are required for signaling (Kawabata et al., 1998). All BMPs use BMPR2, but use different BMP type I receptors. BMP6 preferably binds to ACVR1 (Ro et al., 2004).

Intracellular BMP signals are transduced mainly by small mothers against decapentaplegic (SMAD) proteins. Eight different SMAD proteins have been identified in vertebrates, which can be subclassified into receptor-regulated SMADs, common-partner SMADs and inhibitory SMAD proteins. In vertebrates, BMP signaling acts through SMAD1 and the close homologs SMAD5 and SMAD8. Phosphorylated SMAD1, -5 or -8 proteins form a complex with SMAD4, the only member of the common-partner SMAD proteins, and translocate to the nucleus, where they interact with other transcription factors (Massague, 2000). Alternate BMP signaling pathways include prostanoid generation through cyclooxygenase-2 (Ren et al., 2007) and MAPK (mitogen-activated protein kinase)-dependant activation of p38 or the Ras- and Erk pathway (Nohe et al., 2004; Du et al., 2007). Both pathways have been reported to be present in MMCs (Hoang et al., 2006; Trojan et al., 2006).

Assessing the expression of BMPs, its receptors and members of the signaling-transduction pathway, we found BMP6 to be the only BMP expressed in normal (BMPCs) and malignant plasma cells. In vitro, BMP6 binds to MMCs, blocks their proliferation, induces apoptosis and inhibits angiogenesis. The BMP6 expression is a favorable prognostic factor in two independent cohorts of a total of 168 myeloma patients treated with high-dose chemotherapy and autologous stem cell transplantation. This is validated by an independent cohort of 345 patients treated within the total therapy 2 protocol of the Little Rock (LR) group (Barlogie et al., 2006). The prognostic value of BMP6 expression is independent of classical prognostic parameters, that is, serum β2 microglobulin (B2M) and ISS (International Staging System). Thus, BMP6 exemplifies a novel class of prognostic pro-apoptotic and anti-angiogenic factors expressed by normal as well as by malignant plasma cells.

Results

Expression of BMP6, BMP receptors and downstream SMAD proteins

Expression of BMPs, BMP receptors and members of the downstream signal-transduction chain was evaluated using U133 A+B (Heidelberg/Montpellier-group 1, HM1) and U133 2.0 plus (HM2) Affymetrix microarrays (see Table 1A–C, Supplementary Table S1).

Table 1 Presence of expression of BMP6, BMP receptors and downstream SMADs

BMP6 is the only BMP expressed by normal and malignant plasma cells in both HM1 and HM2 (Table 1A and C, Supplementary Table S2). The mean BMP6 expression is significantly and by several orders of magnitude higher in BMPCs and MMCs than in B-cell precursor cells (MBCs and polyclonal plasmablastic cells, PPCs; P<0.001). Human myeloma cell lines (HMCLs) show a lower expression of BMP6 compared with BMPCs (P<0.001 in HM1, Supplementary Table S1).

Of the BMP receptors, four are expressed in BMPCs and MMCs. BMPR2 is present in most BMPCs and precursors without significant change throughout plasma cell differentiation, as well as in MMCs. ACVR1 is an early plasma cell marker, lacking expression in MBCs (Table 1B, Figure 1a), and ACVR2B is aberrantly expressed in MMCs of 12.5% of patients (Table 1B). No significantly different gene expression could be detected for BMP6 or BMP receptors between MMCs from early-stage (monoclonal gammopathy of unknown significance and MMI according to Durie-Salomon stage) and advanced-stage (MMII and MMIII) patients (Supplementary Table S1D).

Figure 1
figure1

Expression of BMP6, BMP receptors and SMAD proteins and validation of gene expression by flow cytometry and western blotting. (a1–3) Expression of BMP6, ACVR1, BMPR2, BMPR1A, ACVR2A, ACVR2B, SMAD1, SMAD4 and SMAD5 in normal plasma cells (BMPC), memory B cells (MBC), polyclonal plasmablastic cells (PPC), myeloma cells (MMC) and human myeloma cell lines (HMCL) within the Heidelberg/Montpellier group 1 (HM1). For HM2, see Supplementary Figure S1. SMAD8 is expressed neither in HM1 nor in HM2. (b) To validate gene expression data, intracytoplasmatic expression of BMP6 was determined by flow cytometry. Shown are cell lines (b1) XG-10, (b2) XG-11 and (b3) U266; the first shows a consistent absence of BMP6 by gene expression profiling, quantitative real-time PCR and flow cytometry, the other two show a consistent expression. (b4) Expression of BMP6 by an exemplary myeloma patient (pMMC). Light gray line: control without primary antibody; black line: measurement with corresponding primary and secondary antibody. (c) Phosphorylation of downstream SMAD proteins can be observed within 15 min after incubation with BMP6. After pre-incubation with heparin, no SMAD phosphorylation is detectable. Consistent with literature, SMAD-2 is not phosphorylated. Actin was used as loading control.

Of the downstream signaling cascade, SMAD5 is expressed in all plasma cell precursor, plasma cell and 166 of 168 MMC samples (Table 1, Figure 1), whereas SMAD1 is aberrantly expressed in 100 of 233 MMC samples and 39 of 40 HMCL samples. SMAD4 is expressed in 1 of 13 MBC, 4 of 12 PPC, 4 of 14 BMPC and 147 of 233 MMC samples, as well as in 36 of 40 HMCL samples. Thus, transducing SMAD1–SMAD4 or SMAD5–SMAD4 complexes are increasingly present from B-cell precursors over MMCs to cell lines. SMAD8 is not expressed.

Of the populations investigated within the whole BM (WBM), only BMPCs, MMCs and a subfraction of mesenchymal stromal cells (MSCs) express BMP6 (Table 1A and C). The low number of BMPCs and MSCs in the BM of normal donors could explain the lack of a detectable BMP6 expression. In myelomatous WBM, expression of BMP6 correlates significantly (rs=0.45, P=0.001) with the percentage of plasma cell infiltration.

Validation of gene expression

By quantitative real-time PCR (qRT–PCR), BMP6 is expressed in 9 of 10 HMCLs (absent in XG-10) and in 10 of 10 primary MMC samples, consistent with results by Presence-Absence calls with Negative Probesets (PANP)/gene expression profiling (GEP) (see Materials and methods). BMP6 receptors (BMPR2 and ACVR1) are expressed in all HMCLs and primary MMC samples investigated. The BMP6 expression measured by qRT–PCR and GEP correlates well for HMCLs (rs=0.78, P=0.009) and MMCs (rs=0.56, P=0.05).

By flow cytometry, an intracellular BMP6 expression can be detected in 9 of 10 HMCLs and 3 of 3 primary MMC samples. In agreement with PANP/GEP and qRT–PCR, BMP6 expression is absent in XG-10. Exemplary data are shown in Figure 1b.

Performing ELISA, amounts of BMP6 around the level of the detection limit could be found in supernatants of HMCLs. For the highly BMP6-expressing HMCLs, XG-11 and U266 (see Figure 1b), BMP6 levels of 47.59 and 57.51 pg/ml could be measured, respectively. In agreement with GEP, qRT–PCR and flow cytometry, no BMP6 was detectable in supernatants of XG-10.

To determine whether BMP6 induces phosphorylation of downstream SMAD proteins, phosphorylated SMAD1, -5 and -8 were investigated by western blotting. After incubation with BMP6, SMAD activation can be detected within 15 min. After pre-incubation with heparin, no SMAD phosphorylation is detectable (Figure 1c, see below).

Biological and clinical correlations of BMP6 expression

The BMP6 expression in MMCs inversely correlates (although weakly) with a gene expression-based proliferation index (HM1 r=−0.45, P<0.001, HM2 r=−0.35, P<0.001). No correlation of the expression of BMP6 or BMP receptors in MMCs with either the presence of chromosomal aberration t(11;14), t(4;14), gain of 11q13, hyperdiploidy (as measured by our copy number score, see below), gain of 1q21, deletion of 17p or deletion of 13q14.3 could be detected. Similarly, no correlation could be detected with the expression of D-type cyclins (CCND1, CCND2 and CCND3) or clinical parameters (including B2M, ISS, Durie–Salmon stage, serum albumin; data not shown). Similarly, we detected no difference of genetic or clinical markers in BMP6high and BMP6low patients.

MMCs bind BMP6 through membrane heparan sulfate proteoglycans

For BMP family members, a binding to heparin/heparan sulfates has been described (Irie et al., 2003). At the same time, we have shown heparan sulfate-binding members of the epidermal growth factor family to bind syndecan-1 (CD138), the only heparan sulfate proteoglycan constantly present on the surface of BMPCs and MMCs (Mahtouk et al., 2006). Therefore, HMCLs were incubated with saturating concentrations of BMP6 and analysed using flow cytometry. We could show BMP6 binding to 10 of 10 HMCLs. Binding in all cases is reduced by incubation of BMP6 with heparin, functioning as a competitor that likely captures BMP6 (Figure 2, exemplary data). Together with the fact that BMP6-induced apoptosis is abrogated by heparin (see below), these data suggest that BMP6 is a heparan sulfate-binding molecule.

Figure 2
figure2

Binding of BMP6 to human myeloma cell lines. Myeloma cell line U266 was incubated with primary anti-BMP6 antibody and secondary PE-labeled antibody (control), pre-incubated with BMP6 or with BMP6 plus heparin, respectively. Light grey line: control; black line: BMP6, primary and secondary antibody±heparin. A strong labeling of U266 was found, whereas binding is reduced by incubation with heparin.

BMP6 inhibits proliferation and induces apoptosis in HMCLs and primary MMCs

BMP6 significantly inhibits the proliferation of all HMCLs investigated in a dose-dependent manner (Figure 3a). The maximum inhibition at 4 μg/ml ranged from 27.9% (RPMI-8226) to 91.1% (OPM-2). For 6 of 10 cell lines, a 50% inhibition (IC50) could be obtained, ranging from 0.08 (XG-11) to 2.15 (LP-1) μg/ml.

Figure 3
figure3figure3figure3

Inhibition of proliferation and induction of apoptosis of myeloma cell lines by BMP6 as well as survival of primary myeloma cells. (a) Inhibition of proliferation of myeloma cell lines by BMP6 in graded concentrations vs medium control measured by 3H-thymidine uptake. The IC50 (in μg/ml) and the maximal inhibition at 4 μg/ml (in %) are for XG-1 NA/57.2%, XG-10 0.6 μg/ml/81.9%, XG-11 0.08 μg/ml/88.6%, XG-13 0.525 μg/ml/74.8%, XG-19 NA/33.7%, OPM-2 0.175 μg/ml/94.3%, RPMI-8226 NA/27.9%, SKMM-2 0.155 μg/ml/86%, U266 NA/30.7% and LP-1 2.125 μg/ml/51.7%. The two bars for each concentration correspond to two independent experiments. (b) Induction of apoptosis by BMP6 as determined by annexin V staining after 8, 24, 48 and 72 h (third row). Apoptosis induction is abrogated by heparin treatment (fourth row), whereas heparin alone did not influence the apoptosis rate (second row). (c) Survival of primary myeloma cells (pMMC) cultured within their bone marrow microenvironment is significantly inhibited by BMP6 as determined by staining with anti-CD138-FITC and propidium iodine. An asterisk (*) indicates a significant decrease between the medium control and the respective BMP6 concentration. (d) Increasing levels of cleaved caspase-3, -8 and -9 can be detected after BMP6 treatment for 48 and 72 h, respectively. This effect is abrogated by pretreatment with heparin. Actin was used as loading control. (e) The proliferation rate of HMCLs can be increased using BMP6 inhibitors noggin and sclerostin. As shown for U266, a highly BMP6 resistant and highly BMP6-expressing HMCL (see above), proliferation is concentration dependently increased by both inhibitors if cells are either co-exposed to exogenous BMP6 or endogenous BMP6-production is inhibited. Therefore, production of BMP6 by myeloma cells contributes to the inhibition of their growth in vitro. An asterisk (*) indicates a significant increase between the medium control and the respective inhibitor concentration (without exogenous BMP6) or between the BMP6 control and the respective inhibitor concentration (with 1 μg/ml BMP6).

Next, OPM-2 and XG-11 cells were cultured for 3 days with or without BMP6. Cell viability and apoptosis were determined by flow cytometric analysis of annexin V binding and propidium iodine (PI) uptake. BMP6 induced apoptosis between 8 h (12.1 vs 6.2% control) and 72 h (38.5 vs 6.7% control). Heparin pretreatment inhibits BMP6-induced apoptosis (Figure 3b, exemplary data), likely explained by the heparin-induced competitive reduction of BMP6 binding to heparan sulfate chains on plasma cells (see above). Heparin alone did not influence the apoptosis rate.

To test whether the survival of primary MMCs is inhibited as well, these were cultured within their BM microenvironment (negative fraction of plasma cell purification) and exposed to BMP6. After 6 days, cell viability was measured by CD138/PI flow cytometry. As shown in Figure 3c, BMP6 significantly inhibited the survival of 3 of 3 primary MMC samples. The maximal inhibition was 90, 93.6 and 94.6%, respectively.

In terms of apoptosis induction, we demonstrated increasing levels of cleaved caspase-3 (effector caspase), -8 and -9 (initiator caspases) after BMP6 treatment for 48 and 72 h, respectively. This effect is abrogated by pretreatment with heparin (Figure 3d, exemplary data).

Next, we tested whether the production of BMP6 by MMCs themselves inhibits their proliferation. We therefore exposed HCMLs to BMP6 inhibitors noggin (Kersten et al., 2006) and sclerostin (Kusu et al., 2003). Both sclerostin and noggin exposure yielded a concentration-dependent increase of proliferation. Exemplary data for U266 (high endogenous BMP6 production, very low sensitivity to exogenous BMP6; see Figures 2b and 3a) are shown in Figure 3e. As a control, we co-exposed HMCLs with exogenous BMP6 (1 μg/ml) and graded concentrations of noggin or sclerostin, showing that indeed noggin and sclerostin significantly and concentration dependently abrogated the BMP6-mediated inhibition of myeloma cell proliferation (Figure 3e). Therefore, the production of BMP6 by myeloma cells contributes to the inhibition of their growth in vitro.

Inhibition of in vitro tubule formation by BMP6

The angiogenic potential of BMP6 was investigated in the AngioKit assay with graded concentrations of BMP6. BMP6 significantly inhibited in vitro tubule formation with a strong inhibition already observed at 0.032 μg/ml (Figure 4, P=0.04 and P=0.001 in two independent experiments) compared with medium control. The inhibition was as efficient as that provided by suramin, a usual tubule formation inhibitor.

Figure 4
figure4

Inhibition of in vitro induction of angiogenesis by BMP6. Inhibition of endothelial cell growth by BMP6 in the AngioKit model. Immunostaining with monoclonal anti-CD31 antibody. (a) Medium control (RPMI-1640), positive control (vascular endothelial growth factor), negative control (suramin) and BMP6 in concentrations of 4, 0.8, 0.16 and 0.032 μg/ml, respectively. Original magnification, × 40. (b) Boxplot summarizing CD31 ELISA results. BMP6 significantly inhibits tubule formation down to the level of the negative control.

Prognostic value of BMP6, BMP receptors and SMAD expression

In a Cox model, as a single continuous variable, BMP6 expression is significantly predictive for overall survival (OS) in HM1 (P=0.02), HM2 (P=0.01) and LR data (P=0.005). Event-free survival (EFS) is only predictive in LR data (P=0.004). In a Cox model tested with B2M (as a continuous variable), BMP6 expression appears as an independent prognostic factor for OS in HM1 (BMP6 expression: P=0.02, B2M: P=0.7), HM2 (BMP6 expression: P=0.03, B2M: P=0.02) and LR data (BMP6 expression: P=0.01, B2M: P<0.001). The same holds true if the BMP6 expression is tested with ISS in HM1 (BMP6 expression: P=0.04, ISS: P=0.9), HM2 (BMP6 expression: P=0.03, ISS: P=0.03) and LR data (BMP6 expression: P=0.02, ISS: P<0.001). In LR data, BMP6 remains an independent prognostic factor for EFS if tested with B2M (BMP6 expression (P=0.02), B2M (P<0.001)) or ISS (BMP6 expression: P=0.01, ISS: P<0.001). Similarly, patients with BMP6high-expressing MMCs show a significantly better OS compared with patients with a BMP6low expression (n=168, P=0.02, hazard ratio (HR): 0.4, confidence interval (CI): (0.18, 0.87), Figure 5a). No prognostic value could be determined for EFS (P=0.9, P=0.9 and P=0.9, respectively). Similar observations could be made with the patient cohort from the LR group (n=345, OS, P=0.03, HR: 0.67, CI: (0.46, 0.97) and EFS, P=0.15, Figure 5b). Genes coding for BMP receptors or downstream SMAD proteins had no prognostic value. In WBM (n=57), BMP6 expression above the median (WBM-BMP6high) delineated a group with better EFS (n=49, P=0.03, HR: 0.45, CI: (0.21, 0.95)) and a tendency to better OS (P=0.3, Figure 5c).

Figure 5
figure5

Effect of BMP6 expression on event-free survival (EFS) and overall survival (OS). (a) EFS and (b) OS for our patients (HM-group; n=168) and for the Little Rock group (LR-group; n=345). All patients are treated with high-dose chemotherapy and autologous stem cell transplantation. Two groups of patients with high (BMP6high, greater or equal to the median, red curve) and low (BMP6low, below the median, black curve) BMP6 expression. OS is significantly superior for high BMP6 expression (HM1 and HM2: P=0.02, LR-group: P=0.03). (c) EFS and OS for BMP6 expression within the whole bone marrow (WBM; n=57). EFS is significantly superior for the group of patients with BMP6 expression above the median (WBM-BMP6high, P=0.03). This group also shows a tendency to better OS, which does not reach significance (P=0.3), likely because of a low number of events.

Discussion

Expression of BMP6 and its receptors

The BMP6 expression is a characteristic of normal and malignant plasma cells (Table 1A and C): from the populations present within the BM, only MMCs, BMPCs and a subfraction of MSCs express BMP6 (Supplementary Table S2). The latter is in agreement with data from Kersten et al. (2006), who have shown stromal cells to express varying levels of BMP6 mRNA by conventional reverse transcriptase (RT)–PCR. Indeed, the BMP6 expression in MSCs in our data is by several orders of magnitude lower than that in plasma cells as detected by GEP, qRT–PCR and flow cytometry (data not shown). The BMP6 expression by malignant plasma cells is in agreement with data from Zhan et al. (2002). Unlike normal BM in a recent report (Kochanowska et al., 2007) and in our data, BMP6 expression can be detected in 40 of 57 samples of myelomatous BM. The expression of BMP6 within the BM correlates significantly with the percentage of plasma cell infiltration. Regarding the low frequency of MSCs in BM, which further deceases with age (Caplan, 2007), BMP6 expression in myelomatous WBM can be attributed to MMCs. However, in supernatants of HMCLs or in BM sera of myeloma patients, only amounts of BMP6 at the level of the detection limit of ELISA could be found. This could be explained by a significantly lower expression of BMP6 by proliferating cells (see below, Figure 1) and differences in the local concentrations of growth factors. A further possible explanation might be the high turnover of heparan sulfate (syndecan-1)-bound BMP6 (Figure 2).

BMP6 receptors (BMPR2, ACVR1) present in BMPCs and MMCs have previously been detected by RT–PCR in 5 of 5 myeloma cell lines (Ro et al., 2004). Of the members of the SMAD signal-transduction cascade, SMAD5 is expressed in almost all samples during plasma cell development and within MMCs, whereas SMAD1, absent in BMPCs, is expressed at increasing rates from MMCs to HMCLs. SMAD4, the only common-partner SMAD, is expressed in only a minor fraction of BMPCs (21%) but in a high fraction of HMCLs (80%; Table 1B). Thus, transducing SMAD1–SMAD4 or SMAD5–SMAD4 complexes are increasingly present from MBCs over MMCs to HMCLs (Table 1B). This is in agreement with the observation that BMP6 expression develops at the differentiation stage of plasma cells, whereas it is absent in MBCs and only expressed in a small minority of PPC samples. The condition is complicated by the fact that alternative BMP signaling ways through prostanoid generation (Ren et al., 2007) and MAPK (Nohe et al., 2004; Du et al., 2007) have been reported, both pathways being present in MMCs (Hoang et al., 2006; Trojan et al., 2006).

Biological implications

The BMP6 expression is significantly lower in proliferating nonmalignant (PPCs) or malignant (HMCLs) cells and correlates inversely with our gene expression-based proliferation index. In vitro, BMP6 significantly inhibits the proliferation of HMCLs, as well as that of primary MMCs, and induces apoptosis in a time-dependent manner (Figure 3), in agreement with published data (Ro et al., 2004). Autocrine production of BMP6 reduces proliferation of HMCLs, as shown for U266. This effect can be abrogated using noggin or sclerostin as known BMP inhibitors (Figure 3e). At the same time, BMP6 significantly inhibits in vitro tubule formation by human endothelial cells down to the level of the negative control (Figure 4), giving a further possible link with BM angiogenesis and thus prognosis. These were surprising results, as BMP6 has been described as an agonist for endothelial cell migration and tube formation in vitro (Valdimarsdottir et al., 2002) using adenoviral-transfected BMP-receptor-expressing mouse embryonic endothelial cells and bovine aortic endothelial cells. Differences might be explained by species-dependent action of BMP6 and/or receptor expression or a difference in action of BMP6 on embryonic vs adult endothelial cells.

With BMP6, BMPCs express a factor-inhibiting proliferation and angiogenesis, as well as inducing apoptosis and bone formation. We hypothesize this to be a sign for an intense local interaction between BMPCs and the BM microenvironment, for example, important during the settling of plasma cells within their BM niche. We also hypothesize that, as BMP6 expression is absent in MBCs, is expressed in a small minority of PPC samples and appears at the differentiation stage of plasma cells, it might be part of a proliferative block appearing late during the development of terminal differentiated plasma cells. If this is the case, the maintenance of BMP6 expression in the vast majority of MMCs of newly diagnosed patients could be interpreted as an ‘anti-proliferative burden’ in terms of reminiscence to the ‘plasma cell nature’ of MMCs. MMCs in advanced stages would be expected to concomitantly lose BMP6 expression, as observed in HMCLs.

BMP6 and survival

For patients treated with high-dose chemotherapy and autologous stem cell transplantation, the BMP6 expression as a single continuous variable is significantly predictive for OS in the three independent data sets investigated. BMP6 expression remains significantly predictive for OS if tested with classical prognostic parameters, that is, either with B2M or ISS (in all three data sets). In LR data, the BMP6 expression appears as an independent prognostic factor for EFS if tested alone or with either B2M or ISS. Similarly, patients with a BMP6 expression in MMCs above the median (BMP6high) show a superior OS compared with patients with an expression below the median (BMP6low) in our data and in the LR group (Figure 5). In the latter case, no differences in EFS could be observed. Thus, the height of BMP6 expression seemingly has a lower influence on the time to first relapse (disease progression). However, as OS prolongs, BMP6 seems to influence either the time to subsequent relapses or the chemo-sensitivity in relapse. One might speculate that a slower increase in proliferation of BMP6high MMCs during disease progression might reduce the probability of acquiring additional genetic lesions and changing to a more aggressive genotype and phenotype. A possible criticism is that different relapse treatments for BMP6high and BMP6low patient groups might have been used. Although it cannot be evaluated in a sufficiently high number of patients, this is, however, extremely unlikely as (i) the state of BMP6 expression was not known during treatment decisions, (ii) no genetic or biochemical associations could be found for BMP6 expression (which might have driven implicitly different treatment strategies) and (iii) two independent relapse strategies (in Germany/France and the United States, respectively) have been used. It is interesting to note that a higher expression of BMP6 within the WBM (WBM-BMP6high) is associated with significantly superior EFS, despite the correlation of WBM-BMP6 expression and plasma cell infiltration. OS is likewise superior, although not significantly, most likely because of a low number of events (Figure 5c). In contrast to the BMP6 expression within MMCs (above), the BMP6 expression within the WBM assesses the integrated mean value of the BMP6 expression of individual MMCs, together with the percentage of MMC infiltration.

The inhibition of proliferation of all cell lines tested, induction of apoptosis in HMCLs and primary MMCs, as well as inhibition of in vitro angiogenesis give a possible biological explanation for an advantageous effect of BMP6high-expressing MMCs. It is interesting to note that BMP6 expression seems to have either beneficial or detrimental effects on patients' survival in different cancer types. Kim et al. (2003) have shown BMP6 to be present in the kidney and to be a potent growth inhibitor in human renal carcinoma cell lines. In other cancers such as prostate (Hamdy et al., 1997; Dai et al., 2005) or breast cancer (Clement et al., 1999), BMP6 expression promotes tumor progression and development of metastasis.

Conclusion

BMP6 exemplifies a novel class of factors independently prognostic for OS expressed by normal as well as by malignant plasma cells, which inhibit proliferation of MMCs and induction of angiogenesis.

Materials and methods

Patients and healthy donors

Patients presenting with previously untreated MM (n=233) or monoclonal gammopathy of unknown significance (n=12) at the University Hospitals of Heidelberg and Montpellier, as well as 14 healthy normal donors, have been included in this study approved by the ethics committee after obtaining written informed consent. The first 65 patients comprise the HM1 group, the 168 additional patients comprise the independent HM2 group (see below). Patients were diagnosed, staged and response to treatment was assessed according to standard criteria (Durie, 1986; Blade et al., 1998; Greipp et al., 2005). For clinical parameters, see Supplementary Table S3. According or in analogy to the GMMG-HD3 trial (Goldschmidt et al., 2003), 168 patients underwent frontline high-dose chemotherapy with 200 mg/m2 melphalan and autologous stem cell transplantation. Survival data were validated by an independent cohort of 345 patients treated within the total therapy 2 protocol (Barlogie et al., 2006).

Samples

For an overview, see Supplementary Table S4. BM plasma cells were purified as previously published (Hose et al., 2009a). Aliquots of unpurified WBM from patients (n=57) and healthy donors (n=7) were obtained after NH4 lysis (Mahtouk et al., 2007). Alternate aliquots were subjected to FACS sorting (FACSAria, Becton Dickinson, Heidelberg, Germany) in CD3+, CD14+, CD15+ and CD34+ cells. Peripheral CD27+ MBCs were FACS-sorted as described (Moreaux et al., 2005). HMCLs U266, RPMI-8226, LP-1, OPM-2 and SKMM-2 were purchased from DSMZ (Braunschweig, Germany), and XG lines were generated at INSERM U847 (Montpellier, France) as published previously (Zhang et al., 1994). PPCs (Tarte et al., 2002), osteoclasts (Moreaux et al., 2005)) and MSCs (Corre et al., 2007) were generated as published.

Interphase fluorescence in situ hybridization

Analyses were carried out on CD138-purified plasma cells as described (Cremer et al., 2005) using probes for chromosomes 1q21, 9q34, 11q13, 13q14.3, 15q22, 17p13, 19q13, 22q11, and translocations t(4;14)(p16.3;q32.3) and t(11;14)(q13;q32.3) (Kreatech Diagnostics, Berlin, Germany). Ploidy status (excluding gains of 1q21) and clonal/subclonal aberrations (that is, present in 60 vs 20–59% of assessed MMCs) were defined as published (Cremer et al., 2005).

GEP

Gene expression profiling was performed as previously published (Hose et al., 2009b).

In brief, after RNA extraction, labeled cRNA was generated using the small sample labeling protocol vII (Affymetrix, Santa Clara, CA, USA), and was hybridized to U133 A+B GeneChip microarray (Affymetrix) for HM1, and to U133 2.0 plus arrays for HM2, according to the manufacturer's instructions. When different probe sets were available for the same gene, we chose the probe set yielding the maximal variance and the highest signal.

As validation, expression of BMP6 (Hs00233470_m1), BMPR2 (Hs00176148_m1) and Alk-2 (ACVR1, Hs00153836_m1, all Applied Biosystems, Darmstadt, Germany) was assessed by qRT–PCR using the ABI Prism 7700 Sequence Detection System (Applied Biosystems; Mahtouk et al., 2005).

Preparation of cell lysates and western blotting

HMCLs were serum starved for 2 h in SynH medium (ABCell-Bio, Montpellier, France). With or without pretreatment with 4 IU/ml heparin (Leo, Breda, The Netherlands) for 1 h, cells were stimulated with 0.5 μg/ml BMP6 (R&D Systems, Wiesbaden-Nordenstadt, Germany) and cell pellets were resuspended in Laemmli buffer. Proteins were separated using a 10% SDS–polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). Membranes were probed with antibodies against phospho-SMAD2 (Ser465/467) and phospho-SMAD1 (Ser463/465)/SMAD5 (Ser463/465)/SMAD8 (Ser426/428) from Cell Signaling Technology (Beverly, MA, USA). As loading control, membranes were stripped and reprobed with an anti-actin antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA). For caspase detection, HMCLs were exposed to 0.75 μg/ml BMP6 for 48 and 72 h and treated as described above. Membranes were probed with antibodies against cleaved caspase-8 and -9 (Cell Signaling Technology). Detection was performed using ECL (Amersham, Buckinghamshire, UK), according to the manufacturer's instructions.

Intracellular staining for BMP6

Intracellular BMP6 expression of 10 HMCLs and 3 primary MMC samples was measured by flow cytometry. HMCLs were incubated with 3 μg/ml Brefeldin A (eBioscience, San Diego, CA, USA) overnight. Cells were stained according to the manufacturer's instructions (eBioscience). For antibodies used, see below.

Measurement of proliferation by 3H-thymidine

Proliferation of HMCLs was investigated as previously published (De Vos et al., 2001). Cells were cultured in RPMI-1640 containing 10% fetal calf serum and graded concentrations (0.00128, 0.0064, 0.032, 0.16, 0.8 and 4 μg/ml) of BMP6. For XG lines, 2 ng/ml interleukin-6 (R&D Systems) was added. To test for the effect of an endogenous BMP6 production, graded concentrations (0.032, 0.16, 0.8, 4 and 20 μg/ml) of BMP6 inhibitors noggin or sclerostin (both R&D Systems) were added (with or without 1 μg/ml BMP6). Proliferation was evaluated after 54 h of culture: cells were pulsed with 37 kBq of 3H-thymidine for 18 h, harvested and 3H-thymidine uptake was measured.

Apoptosis induction

OPM-2 and XG-11 were cultured in RPMI-1640 containing 10% fetal calf serum and 2 ng/ml interleukin-6 (XG-11) with or without 1 μg/ml BMP6 and 4 IU/ml heparin (Braun, Melsungen, Germany), respectively. After 8, 24, 48 and 72 h, cells were stained for annexin V-FITC and PI according to the manufacturer's instructions (Pharmingen, Heidelberg, Germany).

Flow cytometric analysis of caspase-3 activation

Cells were treated as described above, and stained with the FITC-conjugated active caspase-3 detection kit (Becton Dickinson) according to the manufacturer's instructions. Intracellular fluorescence was determined by flow cytometry.

Survival of primary MMCs

Primary MMCs of three newly diagnosed patients, cultured together with their BM microenvironment, were exposed to 0.00128, 0.0064, 0.032, 0.16, 0.8 and 4 μg/ml BMP6. After 6 days, cell viability was measured by CD138-FITC (IQ products, Groningen, The Netherlands; clone B-A38)/PI (Pharmingen) staining and referred to the medium control (Jourdan et al., 1998). PI of 1 μl at the concentration of 50 μg/ml was used.

In vitro assessment of angiogenesis

The angiogenic potential of BMP6 in concentrations of 0.0064, 0.032, 0.16, 0.8 and 4 μg/ml was investigated in the AngioKit assay (TCS Cellworks, Buckinghamshire, UK) as previously published (Hose et al., 2009a). RPMI-1640, vascular endothelial growth factor (2 ng/ml) and suramin (20 μM) served as medium, positive and negative controls, respectively.

Detection of BMP6 binding to myeloma cells by flow cytometry

Cells were incubated with 1 μg/ml BMP6 for 1 h at 4 °C and washed twice in phosphate-buffered saline before incubation with the corresponding primary (R&D Systems; clone 74219) and secondary antibodies (Dako, Hamburg, Germany). Subsequently, BMP6 was pre-incubated with 4 IU/ml heparin for 1 h at 4 °C. Staining without BMP6 was used as control. Analyses were performed by FACSAria and Infinicyt Software for obtaining overlays (Cytognos, Salamanca, Spain).

BMP6 ELISA

BMP6 levels were measured in culture supernatants of HMCLs (n=9), as well as in BM sera of myeloma patients (n=10) using a human BMP6 ELISA kit (RayBiotech, Norcross, GA, USA).

Statistical analysis

See Supplementary Text S1 and S2.

Conflict of interest

The authors declare no conflict of interest.

References

  1. Barlogie B, Tricot G, Rasmussen E, Anaissie E, van RF, Zangari M et al. (2006). Total therapy 2 without thalidomide in comparison with total therapy 1: role of intensified induction and posttransplantation consolidation therapies. Blood 107: 2633–2638.

  2. Blade J, Samson D, Reece D, Apperley J, Bjorkstrand B, Gahrton G et al. (1998). Criteria for evaluating disease response and progression in patients with multiple myeloma treated by high-dose therapy and haemopoietic stem cell transplantation. Myeloma Subcommittee of the EBMT. European Group for Blood and Marrow Transplant. Br J Haematol 102: 1115–1123.

  3. Caplan AI . (2007). Adult mesenchymal stem cells for tissue engineering versus regenerative medicine. J Cell Physiol 213: 341–347.

  4. Cheng H, Jiang W, Phillips FM, Haydon RC, Peng Y, Zhou L et al. (2003). Osteogenic activity of the fourteen types of human bone morphogenetic proteins (BMPs). J Bone Joint Surg Am 85-A: 1544–1552.

  5. Clement JH, Sanger J, Hoffken K . (1999). Expression of bone morphogenetic protein 6 in normal mammary tissue and breast cancer cell lines and its regulation by epidermal growth factor. Int J Cancer 80: 250–256.

  6. Corre J, Mahtouk K, Attal M, Gadelorge M, Huynh A, Fleury-Cappellesso S et al. (2007). Bone marrow mesenchymal stem cells are abnormal in multiple myeloma. Leukemia 21: 1079–1088.

  7. Cremer FW, Bila J, Buck I, Kartal M, Hose D, Ittrich C et al. (2005). Delineation of distinct subgroups of multiple myeloma and a model for clonal evolution based on interphase cytogenetics. Genes Chromosomes Cancer 44: 194–203.

  8. Dai J, Keller J, Zhang J, Lu Y, Yao Z, Keller ET . (2005). Bone morphogenetic protein-6 promotes osteoblastic prostate cancer bone metastases through a dual mechanism. Cancer Res 65: 8274–8285.

  9. De Vos J, Couderc G, Tarte K, Jourdan M, Requirand G, Delteil MC et al. (2001). Identifying intercellular signaling genes expressed in malignant plasma cells by using complementary DNA arrays. Blood 98: 771–780.

  10. De Vos J, Hose D, Reme T, Tarte K, Moreaux J, Mahtouk K et al. (2006). Microarray-based understanding of normal and malignant plasma cells. Immunol Rev 210: 86–104.

  11. Du J, Yang S, Wang Z, Zhai C, Yuan W, Lei R et al. (2007). Bone morphogenetic protein 6 inhibit stress-induced breast cancer cells apoptosis via both Smad and p38 pathways. J Cell Biochem 103: 1584–1597.

  12. Durie BG . (1986). Staging and kinetics of multiple myeloma. Semin Oncol 13: 300–309.

  13. Ebisawa T, Tada K, Kitajima I, Tojo K, Sampath TK, Kawabata M et al. (1999). Characterization of bone morphogenetic protein-6 signaling pathways in osteoblast differentiation. J Cell Sci 112: 3519–3527.

  14. Goldschmidt H, Sonneveld P, Cremer FW, van der HB, Westveer P, Breitkreutz I et al. (2003). Joint HOVON-50/GMMG-HD3 randomized trial on the effect of thalidomide as part of a high-dose therapy regimen and as maintenance treatment for newly diagnosed myeloma patients. Ann Hematol 82: 654–659.

  15. Greipp PR, San Miguel J, Durie BG, Crowley JJ, Barlogie B, Blade J et al. (2005). International staging system for multiple myeloma. J Clin Oncol 23: 3412–3420.

  16. Hamdy FC, Autzen P, Robinson MC, Horne CH, Neal DE, Robson CN . (1997). Immunolocalization and messenger RNA expression of bone morphogenetic protein-6 in human benign and malignant prostatic tissue. Cancer Res 57: 4427–4431.

  17. Haudenschild DR, Palmer SM, Moseley TA, You Z, Reddi AH . (2004). Bone morphogenetic protein (BMP)-6 signaling and BMP antagonist noggin in prostate cancer. Cancer Res 64: 8276–8284.

  18. Hoang B, Zhu L, Shi Y, Frost P, Yan H, Sharma S et al. (2006). Oncogenic RAS mutations in myeloma cells selectively induce cox-2 expression, which participates in enhanced adhesion to fibronectin and chemoresistance. Blood 107: 4484–4490.

  19. Hose D, Moreaux J, Meissner T, Seckinger A, Goldschmidt H, Benner A et al. (2009a). Induction of angiogenesis by normal and malignant plasma cells. Blood 114: 128–143.

  20. Hose D, Reme T, Meissner T, Moreaux J, Seckinger A, Lewis J et al. (2009b). Inhibition of aurora kinases for tailored risk-adapted treatment of multiple myeloma. Blood 113: 4331–4340.

  21. Irie A, Habuchi H, Kimata K, Sanai Y . (2003). Heparan sulfate is required for bone morphogenetic protein-7 signaling. Biochem Biophys Res Commun 308: 858–865.

  22. Jourdan M, Ferlin M, Legouffe E, Horvathova M, Liautard J, Rossi JF et al. (1998). The myeloma cell antigen syndecan-1 is lost by apoptotic myeloma cells. Br J Haematol 100: 637–646.

  23. Kawabata M, Imamura T, Miyazono K . (1998). Signal transduction by bone morphogenetic proteins. Cytokine Growth Factor Rev 9: 49–61.

  24. Kersten C, Dosen G, Myklebust JH, Sivertsen EA, Hystad ME, Smeland EB et al. (2006). BMP-6 inhibits human bone marrow B lymphopoiesis—upregulation of Id1 and Id3. Exp Hematol 34: 72–81.

  25. Kersten C, Sivertsen EA, Hystad ME, Forfang L, Smeland EB, Myklebust JH . (2005). BMP-6 inhibits growth of mature human B cells; induction of Smad phosphorylation and upregulation of Id1. BMC Immunol 6: 9.

  26. Kim IY, Lee DH, Lee DK, Kim BC, Kim HT, Leach FS et al. (2003). Decreased expression of bone morphogenetic protein (BMP) receptor type II correlates with insensitivity to BMP-6 in human renal cell carcinoma cells. Clin Cancer Res 9: 6046–6051.

  27. Klein B, Tarte K, Jourdan M, Mathouk K, Moreaux J, Jourdan E et al. (2003). Survival and proliferation factors of normal and malignant plasma cells. Int J Hematol 78: 106–113.

  28. Kochanowska I, Chaberek S, Wojtowicz A, Marczynski B, Wlodarski K, Dytko M et al. (2007). Expression of genes for bone morphogenetic proteins BMP-2, BMP-4 and BMP-6 in various parts of the human skeleton. BMC Musculoskelet Disord 8: 128.

  29. Kusu N, Laurikkala J, Imanishi M, Usui H, Konishi M, Miyake A et al. (2003). Sclerostin is a novel secreted osteoclast-derived bone morphogenetic protein antagonist with unique ligand specificity. J Biol Chem 278: 24113–24117.

  30. Kyle RA, Rajkumar SV . (2004). Multiple myeloma. N Engl J Med 351: 1860–1873.

  31. Mahtouk K, Cremer FW, Reme T, Jourdan M, Baudard M, Moreaux J et al. (2006). Heparan sulphate proteoglycans are essential for the myeloma cell growth activity of EGF-family ligands in multiple myeloma. Oncogene 25: 7180–7191.

  32. Mahtouk K, Hose D, Raynaud P, Hundemer M, Jourdan M, Jourdan E et al. (2007). Heparanase influences expression and shedding of syndecan-1, and its expression by the bone marrow environment is a bad prognostic factor in multiple myeloma. Blood 109: 4914–4923.

  33. Mahtouk K, Hose D, Reme T, De Vos J, Jourdan M, Moreaux J et al. (2005). Expression of EGF-family receptors and amphiregulin in multiple myeloma. Amphiregulin is a growth factor for myeloma cells. Oncogene 24: 3512–3524.

  34. Massague J . (2000). How cells read TGF-beta signals. Nat Rev Mol Cell Biol 1: 169–178.

  35. Moreaux J, Cremer FW, Reme T, Raab M, Mahtouk K, Kaukel P et al. (2005). The level of TACI gene expression in myeloma cells is associated with a signature of microenvironment dependence versus a plasmablastic signature. Blood 106: 1021–1030.

  36. Nohe A, Keating E, Knaus P, Petersen NO . (2004). Signal transduction of bone morphogenetic protein receptors. Cell Signal 16: 291–299.

  37. Ren R, Charles PC, Zhang C, Wu Y, Wang H, Patterson C . (2007). Gene expression profiles identify a role for cyclooxygenase 2-dependent prostanoid generation in BMP6-induced angiogenic responses. Blood 109: 2847–2853.

  38. Ro TB, Holt RU, Brenne AT, Hjorth-Hansen H, Waage A, Hjertner O et al. (2004). Bone morphogenetic protein-5, -6 and -7 inhibit growth and induce apoptosis in human myeloma cells. Oncogene 23: 3024–3032.

  39. Tarte K, De Vos J, Thykjaer T, Zhan F, Fiol G, Costes V et al. (2002). Generation of polyclonal plasmablasts from peripheral blood B cells: a normal counterpart of malignant plasmablasts. Blood 100: 1113–1122.

  40. Trojan A, Tinguely M, Vallet S, Seifert B, Jenni B, Zippelius A et al. (2006). Clinical significance of cyclooxygenase-2 (COX-2) in multiple myeloma. Swiss Med Wkly 136: 400–403.

  41. Vacca A, Ribatti D, Roncali L, Ranieri G, Serio G, Silvestris F et al. (1994). Bone marrow angiogenesis and progression in multiple myeloma. Br J Haematol 87: 503–508.

  42. Valdimarsdottir G, Goumans MJ, Rosendahl A, Brugman M, Itoh S, Lebrin F et al. (2002). Stimulation of Id1 expression by bone morphogenetic protein is sufficient and necessary for bone morphogenetic protein-induced activation of endothelial cells. Circulation 106: 2263–2270.

  43. Witzig TE, Timm M, Larson D, Therneau T, Greipp PR . (1999). Measurement of apoptosis and proliferation of bone marrow plasma cells in patients with plasma cell proliferative disorders. Br J Haematol 104: 131–137.

  44. Wutzl A, Brozek W, Lernbass I, Rauner M, Hofbauer G, Schopper C et al. (2006). Bone morphogenetic proteins 5 and 6 stimulate osteoclast generation. J Biomed Mater Res A 77: 75–83.

  45. Zhan F, Hardin J, Kordsmeier B, Bumm K, Zheng M, Tian E et al. (2002). Global gene expression profiling of multiple myeloma, monoclonal gammopathy of undetermined significance, and normal bone marrow plasma cells. Blood 99: 1745–1757.

  46. Zhang XG, Gaillard JP, Robillard N, Lu ZY, Gu ZJ, Jourdan M et al. (1994). Reproducible obtaining of human myeloma cell lines as a model for tumor stem cell study in human multiple myeloma. Blood 83: 3654–3663.

Download references

Acknowledgements

We thank Katrin Heimlich, Maria Dörner and Gabriele Hoock for technical assistance. This work was supported in part by grants from the Hopp-Foundation (Germany), the University of Heidelberg (Germany), the National Centre for Tumor Diseases (Heidelberg, Germany), the Tumorzentrum Heidelberg/Mannheim (Germany), the European Myeloma Stem Cell Network (MSCNET) funded within the 6th framework program of the European Community, Novartis Pharma GmbH (Nuremberg, Germany), the Ligue Nationale Contre Le Cancer (équipe labellisée) (Paris, France). It is also part of a national program called ‘Carte d’Identité des Tumeurs' (CIT) funded by the Ligue Nationale Contre le Cancer (France).

Author information

Correspondence to D Hose.

Additional information

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • gene expression profiling
  • multiple myeloma
  • angiogenesis
  • proliferation
  • survival

Further reading