RAS mutations occur frequently in multiple myeloma (MM), but apart from driving progression, they can also stimulate antitumor effects by activating tumor-suppressive RASSF proteins. Although this family of death effector molecules are often silenced in cancers, functional data about RASSF proteins in MM are lacking. Here, we report that RASSF4 is downregulated during MM progression and correlates with a poor prognosis. Promoter methylation analysis in human cell lines revealed an inverse correlation between RASSF4 mRNA levels and methylation status. Epigenetic modulating agents restored RASSF4 expression. Enforced expression of RASSF4 induced G2-phase cell-cycle arrest and apoptosis in human cell lines, reduced primary MM cell viability, and blocked MM growth in vivo. Mechanistic investigations showed that RASSF4 linked RAS to several pro-death pathways, including those regulated by the kinases MST1, JNK, and p38. By activating MST1 and the JNK/c-Jun pathway, RASSF4 sensitized MM cells to bortezomib. Genetic or pharmacological elevation of RASSF4 levels increased the anti-MM effects of the clinical relevant MEK1/2 inhibitor trametinib. Kinome analysis revealed that this effect was mediated by concomitant activation of the JNK/c-Jun pathway along with inactivation of the MEK/ERK and PI3K/mTOR/Akt pathways. Overall, our findings establish RASSF4 as a tumor-suppressive hub in MM and provide a mechanistic rationale for combining trametinib with HDAC inhibitors or bortezomib to treat patients with tumors exhibiting low RASSF4 expression.

Significance: These findings provide a mechanistic rationale for combining trametinib with HDAC inhibitors or bortezomib in patients with multiple myeloma whose tumors exhibit low RASSF4 expression. Cancer Res; 78(5); 1155–68. ©2017 AACR.

Multiple myeloma (MM) is a plasma cell malignancy that mainly resides in the bone marrow (BM) and remains most often incurable. An important feature of MM is its genetic and subclonal heterogeneity (1). Mutations in RAS represent the most common mutations in cancer (2). In MM, RAS is also frequently mutated (20%–45.9%) and the incidence increases during disease progression (3, 4). In fact, in a refractory setting up to 72% of the patients harbor a mutation in NRAS, KRAS and/or BRAF (5). The MEK/ERK and PI3K/Akt pathway are the 2 main effector pathways of RAS, thus enhancing cell proliferation and survival (6). In general, RAS mutations correlate with a higher tumor burden, sustained survival of MM cells, and a decreased sensitivity toward dexamethasone, doxorubicin, and melphalan (7–9). Because direct targeting of RAS is still impossible, targeting druggable RAS effectors, like MEK/ERK, PI3K/AKT and RAF kinases, is actively being explored (10).

Interestingly, RAS also mediates growth inhibitory effects and apoptosis. These antitumor effects are mainly mediated through the Ras-Association Domain Family, a group of 10 RAS death effector proteins (RASSF1-10; refs. 11, 12). The 6 classical members (RASSF1-6) harbor a C-terminal Ras-association (RA) domain that binds with activated K-RAS. Importantly, these RAS death effectors lack any enzymatic activity acting as scaffolding proteins and linking RAS to pathways associated with cell death, senescence and DNA repair (13). Inactivation of these potent tumor suppressors due to promotor methylation disconnects RAS from its antitumor effectors and is described in various cancers types (14). The best characterized member is RASSF1a and epigenetic silencing of RASSF1a is one of the most common alterations in human cancer (14, 15). RASSF2-6 are also frequently epigenetically silenced in cancer, correlating with a more aggressive and progressive disease (14, 16–18). Currently, no data are available on the expression and role of RASSF proteins in MM.

In the present study, we investigated the expression and prognostic value of the 6 classical RASSF proteins in MM. Our findings identify RASSF4 as the main tumor-suppressive RASSF protein in MM and suggest that loss of RASSF4 favors MM progression. Moreover, we demonstrate that HDACi can restore RASSF4 expression and provide evidence that RASSF4 sensitizes the MM cells to bortezomib and the specific MEK1/2 inhibitor trametinib.

Cell lines

The human myeloma cell lines (HMCL) AMO-1, OPM-2, RPMI-8226, U266, MM1S, L363, and JJN3 were obtained from the ATCC. IL6-dependent cell lines XG-2, XG-7, XG-11, and XG-25 were obtained as previously described (19 ). The identity of the cell lines was regularly checked by short-tandem repeat analysis. Cell lines were regularly tested for Mycoplasma contamination and passaged no more than one month before experiments. Cells were cultured in RPMI-1640 medium, supplemented with 10% FCS (Biochrom AG) and 2 mmol/L l-glutamine. XG-2, XG-7, XG-11, and XG-25 were maintained in the presence of 2 ng/mL recombinant IL6 (R&D Systems), 10% FCS and 2 mmol/L L-glutamine. 293T cells were cultured in DMEM medium supplemented with 10% FCS and supplements (100 U/ml penicillin/streptomycin and 2 mmol/L l-glutamine). Media and supplements were all from Lonza (Basel, Switzerland).

Compounds

Panobinostat, trametinib, and bortezomib were obtained from Selleckchem. Decitabine and quisinostat were kindly provided by Johnson & Johnson. For in vitro studies, all compounds were dissolved in dimethylsulfoxide. For in vivo experiments, decitabine and quisinostat were used as a filter sterilized 10% hydroxypropyl-cyclodextran suspension.

Quantitative real-time PCR

RNA was isolated using the RNeasy kit (Qiagen) and converted to cDNA by the Verso cDNA Synthesis Kit (ThermoFisher Scientific). Gene-specific primer sequences were as follows, human RASSF4: forward (5′-CTC TAT CAA CGG CCA CTT CTA-3′), reverse (5′-GGC CAT CTT CCA CCC TAA AT-3′), mouse RASSF4: forward (5′-CTA TCC CTG GAG AAT GGG TTT C-3′), reverse (5′-TCT GTT CGC ATG GGT GAT AAG-3′), Actin: forward (5′-TCC TCT CCC AAG TCC ACA C-3′), reverse (5′-GCA CGA AGG CTC ATC ATT CA-3′); all from Integrated DNA Technologies. mGAPDH Quantitect primer (QT01658692) was obtained from Qiagen. RAS mutation status was determined using qBiomarker Somatic Mutation PCR Array for human Ras-Raf Pathway (Qiagen) according to the manufacturer's instructions. This array includes 42 DNA sequence mutation assays designed to detect the most frequent, functionally verified, and biologically significant mutations in the RAS–RAF pathway.

Western blot and immunoprecipitation

Western blot was performed as described previously (20 ). For immunoprecipitation, cells were lysed in lysis buffer containing 1 mmol/L PMSF (Sigma). Anti-MST1 antibody was added to 150 μg of sample and incubated overnight at 4°C. Next, lysates were transferred to pre-washed protein A magnetic beads, incubated for 30 minutes at room temperature and washed 3 times with lysis buffer. Finally, the pellet was resuspended in 2x loading buffer containing β-ME and boiled for 5 minutes at 95°C. Antibodies used were: phospho-SAPK/JNK (T138/Y185; #9255), SAPK/JNK (#9258), phospho-C-Jun (S63; #2361), C-Jun (#9165), phospho-p38 MAPK (T180/Y182; #92155), p38-MAPK (#9212), phospho-p53 (Ser392; #9281), p53 (#9282), phospho-histone H2AX (S139/Y142; #5438), p21 (#2947), BIM (#2933), phospho-p44/42 MAPK (ERK1/2; Thr202/Tyr204; #9106), p44/42 MAPK (ERK1/2; #9102), phospho-Akt (S473; #4058), Akt (#9272), alpha-tubulin (#2144), MST1 (#14946) and mouse anti-rabbit IgG conformation specific HRP-conjugated antibody (#5127). All antibodies were purchased from Cell Signaling Technology, except for RASSF4 (NBP189249), which was purchased from Novus Biologicals.

Gene expression data

Expression and survival analysis of (publicly available) gene-expression microarray data were done using Genomicscape (http://genomicscape.com). The Heidelberg-Montpellier (HM) cohort (Database E-MTAB-372) and the University of Arkansas for Medical Sciences (UAMS) TT2 and TT3 cohorts (datasets GSE4581, GSE2658) contain expression data of malignant plasma cells (PC) of newly diagnosed, untreated MM patients (21,23,). We also used Affymetrix data of relapsed MM patients subsequently treated with bortezomib (GSE9782) from the study of Mulligan and colleagues (24) and microarray data from normal BMPC and premalignant PC of monoclonal gammopathy of undetermined significance (MGUS; GSE5900) and HMCL (E-TABM-1088 and E-TABM-937).

Conditional lentiviral overexpression

The pLenti3.3/TR and pLenti6.3/V5-DEST Gateway constructs were purchased from ThermoFisher Scientific. Human RASSF4 entry clone (GC-W0840) was purchased from Genecopoeia. 293T cells were transiently transfected with transfer plasmid, VSV.G encoding plasmid, pMD.G and packaging plasmids pLP1 and pLP2 (ThermoFisher Scientific). Viral titer was determined using the colorimetric reverse transcriptase assay (Roche). Stably transduced cell lines were selected using 1 mg/mL Genetecin or 20 μg/mL Blasticidin S HCI (ThermoFisher Scientific). To induce overexpression 1 μg/mL doxycycline hyclate (Sigma) was added.

Patient samples

BM samples (n = 10) were collected for routine diagnostic or evaluation purposes after patients' written informed consent and in accordance with the Declaration of Helsinki and Institutional Research Board approval from Montpellier University Hospital (DC-2008-417). Mononuclear cells were obtained after Ficoll density gradient centrifugation (Nycomed). Samples were transduced with lentiviral particles harboring the empty vector or RASSF4 and cultured in the presence of recombinant IL6 (2 ng/mL). At day 4, viability and total cell count were assessed and the percentage CD138+ viable plasma cells was determined by flow cytometry.

Mice

Male C57Bl/KaLwRij mice (6- to 8-weeks-old) were purchased from Envigo. The 5T33MM model originated from aging C57BL/KaLwRij mice that spontaneously developed MM and is maintained as previously described (20 ). 5T33MM mice with established disease were randomized and treated for 5 days with 0,2 mg/kg decitabine (daily, n = 5), 1,5 mg/kg quisinostat (every 2 days, n = 4), or a combination of both (n = 4). Next, all mice were sacrificed and tumor cells were purified by depleting CD11b+ cells (95% purity). Female NOD-SCID mice (5- to 6-weeks-old) were purchased from Janvier laboratories (Le Genest-Saint-Isle, France). In a first experiment, mice were injected subcutaneously with pLenti6.3 and pLenti6.3-RASSF4–transduced AMO-1 cells (5 × 106) into respectively the left and right flank (n = 5). In a second survival experiment, mice were injected subcutaneously into the right flank with either pLenti6.3 or pLenti6.3-RASSF4 transduced XG-7 cells (5 × 106; n = 10/group). Following tumor engraftment (average tumor size 125 to 250 mm3) mice were administered doxycycline via the drinking water (2 mg/mL). Tumor volume was calculated as followed: volume (mm3) = (length × width²)/2). When tumor volumes reached 1500 mm³, mice were sacrificed. For these experiments, no randomization and blinding was possible, however, tumor measurements were performed by 2 independent technical assistants. All experimental procedures were approved by the Ethical Committee for Animal Experiments, VUB (LA 1230281, CEP14-281-5).

Methylation analysis

DNA methylation data were generated with the Illumina Infinium Human Methylation 450 Bead Chip array (HM450K, Illumina Inc.) for nine cell lines (AMO1, L363, XG7, XG25, RPMI8226, LP1, JJN3, XG11, and OPM2; Helixio). The microarray raw intensities were preprocessed using the R/Bioconductor package minfi (25). The CpGs overlapping with RASSF4 promoter region were extracted using the R package GenomicFeatures implemented in Bioconductor (26).

Statistical analysis

GraphPad Prism 5.0 software and Genomicscape were used to conduct statistical analyses. To compare 2 groups, we used the Mann–Whitney test. To compare multiple groups the Kruskal–Wallis test or one-way ANOVA test was used. Difference in overall survival was assayed with a log-rank test and survival curves plotted using the Kaplan–Meier method. P values of ≤ 0.05 were considered statistically significant.

Loss of RASSF4 is associated with MM progression and a worse outcome

We first analyzed the expression of the 6 classical RASSF proteins in MM using publicly available gene expression profiling data of 22 healthy donors, 44 MGUS patients, and 345 MM patients from the UAMS TT2 cohort (22 ). This analysis revealed that RASSF2 and RASSF4 expression was significantly reduced in MM cells compared to normal BMPC, whereas RASSF1, RASSF5 and RASSF6 expression was significantly upregulated in MM cells (Fig. 1A). Next, we investigated the relation between RASSF1-6 mRNA expression and disease outcome using 3 cohorts of newly diagnosed MM patients. Patients were divided into RASSF high and low expressers using the Maxstat R package. In all cohorts, high RASSF1 and RASSF4 expression consistently correlated with a superior overall survival, whereas RASSF3, RASSF5 and RASSF6 all had a bad prognostic value (Fig. 1B and Supplementary Fig. S1A–S1D). For RASSF2, however, conflicting results were obtained as high RASSF2 levels were significantly associated with a superior overall survival in the HM cohort, but an inferior overall survival in the UAMS TT2 and TT3 cohort. Together, these data show that RASSF4 is the only member that is consistently downregulated in MM cells compared to normal BMPC, correlating with an unfavorable prognosis. This indicates that RASSF4 could be a tumor suppressor in MM.

Figure 1.

Expression and prognostic value of the classical RASSF family members in MM. A,RASSF1-6 mRNA expression levels in normal BMPC (n = 22), MGUS (n = 44), and MM cells of newly diagnosed patients (n = 345) from the UAMS TT2 cohort. *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with BMPC. B, Prognostic value of RASSF2 and RASSF4 mRNA levels in terms of overall survival in newly diagnosed patients from the HM (n = 206), UAMS TT2 (n = 256), and UAMS TT3 (n = 158) cohort. Maxstat analysis was used to calculate the optimal separation of patients based on a cutoff value.

Figure 1.

Expression and prognostic value of the classical RASSF family members in MM. A,RASSF1-6 mRNA expression levels in normal BMPC (n = 22), MGUS (n = 44), and MM cells of newly diagnosed patients (n = 345) from the UAMS TT2 cohort. *, P < 0.05; **, P < 0.01; ***, P < 0.001 compared with BMPC. B, Prognostic value of RASSF2 and RASSF4 mRNA levels in terms of overall survival in newly diagnosed patients from the HM (n = 206), UAMS TT2 (n = 256), and UAMS TT3 (n = 158) cohort. Maxstat analysis was used to calculate the optimal separation of patients based on a cutoff value.

Close modal

RASSF4 is heterogeneously expressed in human MM cell lines and localized to the nucleus

We next examined basal RASSF4 expression using gene expression profiling data of a large panel of HMCL (Fig. 2A). Similar to primary MM cells, RASSF4 mRNA expression was highly heterogeneous in HMCL. We also confirmed this highly variable RASSF4 expression in a selected panel of HMCL with different molecular characteristics and RAS mutation status on mRNA and protein level (Fig. 2B–D; ref. 19 ). We found a moderate to high expression in the XG-2, XG-11, OPM-2, RPMI-8266 and JJN3 cell lines and low expression in the XG-7, AMO-1 and U266 cells. We also determined KRAS and NRAS protein levels. Similar to RASSF4, expression of both NRAS and KRAS was found to be heterogeneous. Moreover, no clear correlation between RASSF4 levels and KRAS/NRAS expression levels nor RAS mutation status was found (Fig. 2C and D). Immunofluorescent staining for RASSF4 and α-tubulin furthermore revealed that RASSF4 is mainly localized to the nucleus and centrosome (Supplementary Fig. S2).

Figure 2.

RASSF4 expression and RAS status in human MM cell lines. A,RASSF4 mRNA expression using microarray data from a panel of 42 human cell lines. B,RASSF4 expression in 8 selected HMCL using quantitative RT-PCR. Actin was used as reference gene; data represent the mean ± SD of three independent experiments. C, RASSF4, NRAS, and KRAS expression in 8 selected HMCL using Western blot analysis. Tubulin was used as loading control; one experiment representative of three is shown. D, RAS mutation status of selected HMCL. Cell lines are ordered according to the RASSF4 gene expression level.

Figure 2.

RASSF4 expression and RAS status in human MM cell lines. A,RASSF4 mRNA expression using microarray data from a panel of 42 human cell lines. B,RASSF4 expression in 8 selected HMCL using quantitative RT-PCR. Actin was used as reference gene; data represent the mean ± SD of three independent experiments. C, RASSF4, NRAS, and KRAS expression in 8 selected HMCL using Western blot analysis. Tubulin was used as loading control; one experiment representative of three is shown. D, RAS mutation status of selected HMCL. Cell lines are ordered according to the RASSF4 gene expression level.

Close modal

Epigenetic modulating agents restore RASSF4 expression in MM

Because RASSF proteins are often epigenetically silenced in cancer cells, we reasoned that the loss of RASSF4 in MM might also be due to epigenetic silencing (12, 27, 28). Thus, we first performed methylation analysis of the CpG islands in the RASSF4 promotor region in a selected panel of HMCL. As shown in Fig. 3A–C, we found a significant inverse correlation between the RASSF4 expression levels and the methylation status of the CpG island cg02841844 (r = −0.89, P = 0.0013). For the XG-7 and MM1S cells we also investigated the presence of repressive histone marks, including H3K9me3 and H3K27me3, but found no clear enrichment of any specific histone marks. Treatment of HMCL (XG-7, AMO-1, and JJN3) with the DNA methyltransferase inhibitor (DNMTi) decitabine and/or histone deacetylase inhibitor (HDACi) quisinostat significantly enhanced RASSF4 expression (Fig. 3D). This effect was most apparent in the XG-7 and AMO-1 cells with low basal RASSF4 levels. We also validated these findings in vivo by treating 5T33MM mice with established disease with decitabine and/or quisinostat. In line with the in vitro data, quisinostat (alone and in combination with decitabine) significantly enhanced RASSF4 expression in the 5T33MM cells. Together, these data indicate that in MM RASSF4 inactivation in mainly mediated through promotor methylation and epigenetic modulating agents are able to restore RASSF4 expression.

Figure 3.

Epigenetic regulation of RASSF4. A, Methylation status of CpGs overlapping with RASSF4 promotor region. Only CpGs located in a window of 500 pb around the transcription start site are considered. Cell lines are ordered according to the RASSF4 gene expression level. B, Log2 Affymetrix gene expression level of RASSF4. C, Correlation between RASSF4 expression and the methylation level of cg02841844. D, Effect of the DNA methyltransferase inhibitor decitabine (DAC) and/or histone deacetylase inhibitor quisinostat (Q) on RASSF4 mRNA expression. Top, XG-7, AMO-1, and JJN3 cells were treated with 0.5 μmol/L DAC (4 days), 5 nmol/L Q (1 day), and the combination of both, and RASSF4 expression was analyzed using qPCR. Data represent the mean ± SD of three independent experiments. Bottom, 5T33MM-inoculated mice with established disease were treated for 5 days with 0.2 mg/kg DAC (daily, n = 5), 1.5 mg/kg Q (every 2 days, n = 4), or a combination of both (n = 4), after which, tumor cells were isolated and purified by depleting CD11b+ cells. Left, scheme depicting experimental set-up. Right, qPCR showing the effect of in vivo treatment on RASSF4 mRNA expression. Actin and GAPDH were used as reference genes for human and mouse samples, respectively. *, P < 0.05 compared with untreated conditions; $, P < 0.05 compared with both single agents.

Figure 3.

Epigenetic regulation of RASSF4. A, Methylation status of CpGs overlapping with RASSF4 promotor region. Only CpGs located in a window of 500 pb around the transcription start site are considered. Cell lines are ordered according to the RASSF4 gene expression level. B, Log2 Affymetrix gene expression level of RASSF4. C, Correlation between RASSF4 expression and the methylation level of cg02841844. D, Effect of the DNA methyltransferase inhibitor decitabine (DAC) and/or histone deacetylase inhibitor quisinostat (Q) on RASSF4 mRNA expression. Top, XG-7, AMO-1, and JJN3 cells were treated with 0.5 μmol/L DAC (4 days), 5 nmol/L Q (1 day), and the combination of both, and RASSF4 expression was analyzed using qPCR. Data represent the mean ± SD of three independent experiments. Bottom, 5T33MM-inoculated mice with established disease were treated for 5 days with 0.2 mg/kg DAC (daily, n = 5), 1.5 mg/kg Q (every 2 days, n = 4), or a combination of both (n = 4), after which, tumor cells were isolated and purified by depleting CD11b+ cells. Left, scheme depicting experimental set-up. Right, qPCR showing the effect of in vivo treatment on RASSF4 mRNA expression. Actin and GAPDH were used as reference genes for human and mouse samples, respectively. *, P < 0.05 compared with untreated conditions; $, P < 0.05 compared with both single agents.

Close modal

RASSF4 impairs growth and survival of MM cells

To investigate the function of RASSF4 in MM, we then assessed the effect of forced RASSF4 expression on MM cell growth and survival of KRAS mutated XG-7 and AMO-1 cells, NRAS mutated JJN3 cells and wildtype OPM-2 cells using an inducible system (Supplementary Fig. S3A–S3E; refs. 19, 29). Forced RASSF4 expression strongly and significantly increased apoptosis and the percentage cleaved caspase-3–positive cells in all cell lines tested (Fig. 4A and Supplementary Fig. S3F–S3H). In contrast, doxycycline had no effect on empty vector–expressing control cells. This decreased viability was also confirmed using a life death staining (Fig. 4B). Subsequent flow cytometric analysis revealed a significant accumulation of cells in sub-G1 and G2-phase (Fig. 4C). Moreover, immunofluorescent staining for RASSF4 and α-tubulin revealed that mitotic cells express only very low levels of RASSF4 compared with nondividing cells (Supplementary Fig. S2). Finally, RASSF4 strongly reduced the number of viable primary CD138+ MM cells in all but one sample (n = 10, Fig. 4D), whereas the nonmyeloma cells were not significantly affected. Patient characteristics, RASSF4 expression levels and the RAS mutation status are provided in Supplementary Table S1 and Supplementary Fig. S4. No significant correlation was found between sensitivity to forced RASSF4 expression and basal RASSF4 expression (r = 0.185, P = 0.63) nor RAS mutation status. Altogether, our data identify RASSF4 as a new tumor suppressor, negatively influencing MM cell proliferation and survival.

Figure 4.

Effect of RASSF4 overexpression on cell-cycle progression, apoptosis, and bortezomib sensitivity. A, Effect of RASSF4 on apoptosis. Transduced cells were cultured with or without doxycycline (Dox) for indicated time points (days) and the effect on apoptosis was assessed using an Annexin V-APC/7′AAD staining, followed by flow cytometric analysis. The percentage of apoptotic cells are the sum of the percentage of Annexin V (+) and Annexin V (+)/7′AAD (+) cells. Data represent the mean ± SD of four independent experiments. *, P < 0.05; **, P < 0.01. B, Effect of RASSF4 on cell viability. XG-7-TR-6.3 huRASSF4 cells were cultured with Dox for 10 days, after which, a life/death staining was performed. Green, viable cells; red, death cells. Scale bar, 100 μm. One experiment representative of three is shown. C, Effect of RASSF4 on cell-cycle progression. Cell-cycle profiles based on DNA content were obtained after staining with propidium iodide. Bars and error bars are mean ± SD of three experiments. *, P < 0.05 compared with 6.3 Cnt; $, P < 0.05 compared with huRASSF4 Cnt. D, Effect of RASSF4 overexpression on primary human CD138+ MM cells. Mononuclear cells from 10 MM patients were transfected with lentiviral particles harboring the empty vector (6.3) or huRASSF4 for 4 days and the percentage CD138+ viable plasma cells was determined by flow cytometry. #, patient number. E, Effect of forced RASSF4 expression on bortezomib drug sensitivity. Cells were treated with different concentrations of bortezomib (Bz, nmol/L) 3 (XG-7 and AMO-1) or 7 (JJN3) days after treatment with Dox for an additional 3 days. Apoptosis was determined by flow cytometry using an Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. F, Prognostic value of RASSF4 mRNA levels in terms of overall survival in relapsed MM patients from the Mulligan (n = 188) cohort. Maxstat analysis was used to calculate the optimal separation based on a cutoff value.

Figure 4.

Effect of RASSF4 overexpression on cell-cycle progression, apoptosis, and bortezomib sensitivity. A, Effect of RASSF4 on apoptosis. Transduced cells were cultured with or without doxycycline (Dox) for indicated time points (days) and the effect on apoptosis was assessed using an Annexin V-APC/7′AAD staining, followed by flow cytometric analysis. The percentage of apoptotic cells are the sum of the percentage of Annexin V (+) and Annexin V (+)/7′AAD (+) cells. Data represent the mean ± SD of four independent experiments. *, P < 0.05; **, P < 0.01. B, Effect of RASSF4 on cell viability. XG-7-TR-6.3 huRASSF4 cells were cultured with Dox for 10 days, after which, a life/death staining was performed. Green, viable cells; red, death cells. Scale bar, 100 μm. One experiment representative of three is shown. C, Effect of RASSF4 on cell-cycle progression. Cell-cycle profiles based on DNA content were obtained after staining with propidium iodide. Bars and error bars are mean ± SD of three experiments. *, P < 0.05 compared with 6.3 Cnt; $, P < 0.05 compared with huRASSF4 Cnt. D, Effect of RASSF4 overexpression on primary human CD138+ MM cells. Mononuclear cells from 10 MM patients were transfected with lentiviral particles harboring the empty vector (6.3) or huRASSF4 for 4 days and the percentage CD138+ viable plasma cells was determined by flow cytometry. #, patient number. E, Effect of forced RASSF4 expression on bortezomib drug sensitivity. Cells were treated with different concentrations of bortezomib (Bz, nmol/L) 3 (XG-7 and AMO-1) or 7 (JJN3) days after treatment with Dox for an additional 3 days. Apoptosis was determined by flow cytometry using an Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. *, P < 0.05; **, P < 0.01; ***, P < 0.001. F, Prognostic value of RASSF4 mRNA levels in terms of overall survival in relapsed MM patients from the Mulligan (n = 188) cohort. Maxstat analysis was used to calculate the optimal separation based on a cutoff value.

Close modal

RASSF4 sensitizes MM cells to bortezomib

We also investigated whether RASSF4 expression influences MM cell sensitivity toward the standard of care agent bortezomib. RASSF4 overexpression significantly increased sensitivity of MM cells to bortezomib in XG-7 and JJN3 cells (Fig. 4E and Supplementary Fig. S3H). Moreover, we evaluated the prognostic value of RASSF4 in a cohort of patients treated with bortezomib monotherapy after relapse (Fig. 4F). Again, high RASSF4 expressers had a superior overall survival upon bortezomib treatment. Together, these data suggest that strategies enhancing RASSF4 expression could render MM cells more sensitive toward bortezomib.

RASSF4 delays in vivo tumor growth

To confirm the tumor-suppressive role of RASSF4 in MM in vivo, mice were first injected with AMO-1-6.3 (left side) and AMO-1-RASSF4 (right side) cells (Fig. 5A). As shown in Fig. 5B and Supplementary Fig. S5A, the (mean) tumor volume was significantly lower in the RASSF4-overexpressing tumors compared with controls. Moreover, the mean tumor weight of RASSF4 transduced tumors at the moment of sacrificing the mice was also significantly lower than the control group (Fig. 5C). In a second experiment, we then investigate the effect of forced RASSF4 on overall survival. Mice were injected with either XG-7-6.3 or XG-7-RASSF4 cells in the right flank (Fig. 5D). In line with the first experiment, we found a strong and significant delay in tumor growth (Fig. 5E). Moreover, mice injected with RASSF4-overexpressing cells showed a significant increase in overall survival compared with the control group (median survival empty vector, 16 days vs. mean survival huRASSF4, not yet reached when mice were sacrificed at day 49; Fig. 5F). For both experiments, RASSF4 overexpression was confirmed using either Western blot analysis or immunofluorescence (Supplementary Fig. S5B and S5C).

Figure 5.

Effect of RASSF4 expression on tumor development. A–C, Tumor development upon injection of AMO-1–transduced cells. A, Scheme depicting experimental setup. B and C, Effect of RASSF4 expression on tumor volume (B) and weight (C). Mean tumor volumes were plotted against days. Mean tumor weight of the tumors at the moment of sacrificing the mice is shown. D–F, Tumor development upon injection of XG-7–transduced cells. D, Scheme depicting experimental setup. E, Effect of RASSF4 expression on tumor volume. Mean tumor volumes were plotted against days. F, Effect of RASSF4 expression on survival. Mice were sacrificed when tumor volumes reached 1,500 mm³. Difference in overall survival between the two groups was assayed with a log-rank test and survival curves plotted using the Kaplan–Meier method. The experiment was terminated at day 49. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Figure 5.

Effect of RASSF4 expression on tumor development. A–C, Tumor development upon injection of AMO-1–transduced cells. A, Scheme depicting experimental setup. B and C, Effect of RASSF4 expression on tumor volume (B) and weight (C). Mean tumor volumes were plotted against days. Mean tumor weight of the tumors at the moment of sacrificing the mice is shown. D–F, Tumor development upon injection of XG-7–transduced cells. D, Scheme depicting experimental setup. E, Effect of RASSF4 expression on tumor volume. Mean tumor volumes were plotted against days. F, Effect of RASSF4 expression on survival. Mice were sacrificed when tumor volumes reached 1,500 mm³. Difference in overall survival between the two groups was assayed with a log-rank test and survival curves plotted using the Kaplan–Meier method. The experiment was terminated at day 49. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Close modal

RASSF4 interacts with MST1 and modulates JNK/JUN, p53, p38, and MEK/ERK signaling pathways

RASSF proteins are known to link RAS to proapoptotic pathways by interacting with the mammalian sterile 20-like kinases MST1 and MST2 (12, 30, 31). As shown in Fig. 6A, MST1 protein was expressed in all HMCL tested, whereas no MST2 protein was detected. Immunoprecipitation experiments confirmed MST1-RASSF4 binding under basal conditions in MM cells (Fig. 6B). Importantly, RASSF4 overexpression resulted in a strong increase in the amount of RASSF4 bound to MST1 and this especially in the cell lines with low basal RASSF4 expression. Next, we evaluated the effect of forced RASSF4 expression on the major signaling pathways described to be modulated by RASSF proteins (13). We observed a clear increase in phosphorylation of SAPK/JNK, c-Jun, p38 and p53 (Fig. 6C). Moreover, we also observed an increase in c-Jun, p21 and Bim expression levels and phosphorylation of H2AX (gammaH2AX). Unexpectedly, we also observed a clear increase in phosphorylation of ERK 1/2, whereas phosphorylation of AKT 1/2/3 remained largely unaffected (except for the AMO-1 cells). Of interest, increasing evidence indicates that MST1 is activated by caspase-mediated cleavage, resulting in an approximately 35 kDa cleavage fragment. This hyperactive MST1 fragment will translocate from the cytoplasm to the nucleus and activate the JNK pathway, resulting in phosphorylation of histone H2B and H2AX, induction of chromatin condensation and apoptosis (32, 33). Moreover, bortezomib treatment has also been reported to activate MST1 by inducing caspase-mediated cleavage. Thus, we hypothesized that this might also be the case in MM cells and forced RASSF4 expression might increase MM cell bortezomib sensitivity by further increasing the levels of hyperactive MST. As shown in Fig. 6D, bortezomib treatment alone did not result in MST1 cleavage in XG-7 and JJN3 MM cells. Nevertheless, forced RASSF4 expression did activate MST1 and the levels of hyperactive MST1 were further enhanced when cells were cotreated with bortezomib. In addition, cotreatment resulted in a further increase in phosphorylation of c-Jun and H2AX and c-jun expression. RNA-sequencing analysis furthermore revealed 263 genes significantly differentially expressed between control and RASSF4 overexpressing cells, with 72 genes upregulated (including RASSF4 and c-Jun) and 191 genes downregulated (Supplementary Table S2). Pathway enrichment analysis of downregulated genes showed enrichment in pathways involved in protein metabolism, the unfolded protein response and translation (Supplementary Fig. S6A and S6B and Supplementary Table S3). Together, these data support the notion that RASSF4 acts as a scaffold protein in MM, connecting RAS to various proapoptotic pathways.

Figure 6.

Pathways modulated by RASSF4 overexpression. A, Western blot analysis of basal expression levels of MST1. B, Coimmunoprecipitation of MST1, followed by Western blot analysis for MST1 and RASSF4. Cells were cultured with Dox (+) or without (−) for 3 days. Next, samples were immunoprecipitated for MST1 and immunoblotted for RASSF4. C, Western blot analysis of involved signaling pathways and downstream targets. Cell lines transduced with empty vector or RASSF4 were cultured with Dox (+) or without (−) for 5 days, after which, Western blot analysis was performed for indicated proteins. D, Western blot analysis of the underlying mechanisms of the increased bortezomib drug sensitivity upon forced RASSF4 expression. Cells were treated with 1.5 nmol/L bortezomib (Bz) 3 (XG-7) or 7 (JJN3) days after treatment with Dox for 24 hours, after which, Western blot analysis was performed for indicated proteins. For the Western blot analysis, tubulin was used as loading control and one experiment representative of three is shown.

Figure 6.

Pathways modulated by RASSF4 overexpression. A, Western blot analysis of basal expression levels of MST1. B, Coimmunoprecipitation of MST1, followed by Western blot analysis for MST1 and RASSF4. Cells were cultured with Dox (+) or without (−) for 3 days. Next, samples were immunoprecipitated for MST1 and immunoblotted for RASSF4. C, Western blot analysis of involved signaling pathways and downstream targets. Cell lines transduced with empty vector or RASSF4 were cultured with Dox (+) or without (−) for 5 days, after which, Western blot analysis was performed for indicated proteins. D, Western blot analysis of the underlying mechanisms of the increased bortezomib drug sensitivity upon forced RASSF4 expression. Cells were treated with 1.5 nmol/L bortezomib (Bz) 3 (XG-7) or 7 (JJN3) days after treatment with Dox for 24 hours, after which, Western blot analysis was performed for indicated proteins. For the Western blot analysis, tubulin was used as loading control and one experiment representative of three is shown.

Close modal

RASSF4 sensitizes MM cells to the MEK inhibitor trametinib

Specific targeting of the RAS-activated pathways using small-molecule inhibitors has shown promising preclinical results in MM (6, 34, 35). However, (preliminary) results from clinical trials indicate that the therapeutic benefit of these agents is only limited when used as single agents (34, 36). We hypothesized that simultaneously stimulating the antitumoral RAS activity, while targeting it's protumoral activity might potently enhance each other's antimyeloma activity. Thus, we investigated whether forced RASSF4 expression could increase the antimyeloma activity of the clinical relevant specific MEK1/2 inhibitor trametinib (37). As illustrated in Fig. 7A, RASSF4 overexpression indeed significantly enhanced trametinib mediated cytotoxicity in all 3 cell lines. To elucidate the mechanisms involved in this increased sensitivity, we then performed a phospho-kinase array. Supplementary Table S4 provides an overview of all the kinases that were found differentially phosphorylated in RASSF4-overexpressing cells treated with trametinib compared with parental cells treated with trametinib and/or RASSF4-overexpressing cells. The most striking differences observed compared with trametinib alone were the strong increase in phosphorylation of c-Jun and p27, whereas the most apparent difference compared with forced RASSF4 expression was the complete inhibition of ERK pathway activation. In addition, we also observed a further inactivation of several well-known prosurvival pathways for MM cells, including the PI3K/AKT, p53, MSK1, WNK1 and mTOR pathway, compared with both parental cells treated with trametinib and RASSF4-overexpressing cells. We also validated these findings for several kinases, including ERK, AKT, GSK3α/β and P70 S6 kinase, using Western blotting (Fig. 7B). In addition, we observed a strong increase in phosphorylation of c-Jun and c-Jun, p21 and p27 levels.

Figure 7.

Effect of RASSF4 expression on sensitivity to the MEK1/2 inhibitor trametinib. A, Effect of forced RASSF4 expression on trametinib sensitivity. Cells were treated with different concentrations of trametinib (T, nmol/L) 3 (XG-7 and AMO-1) or 7 (JJN3) days after treatment with Dox for an additional 3 days. Apoptosis was determined by flow cytometry using an Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. *, P < 0.05; ***, P < 0.001. B, Western blot analysis of the underlying mechanisms of the increased trametinib drug sensitivity upon forced RASSF4 expression. AMO-1-RF4 cells were treated with 50 nmol/L trametinib (T) 3 days after treatment with Dox for 24 hours, after which, Western blot analysis was performed for indicated proteins. Tubulin was used as loading control and one experiment representative of three is shown. C, Antimyeloma activity of trametinib in combination with panobinostat. XG-7, AMO-1, and JJN3 cells were treated with trametinib and/or panobinostat for 72 hours. Apoptosis was determined by flow cytometry using Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. P values and combination indexes are shown in Supplementary Table S5. D, Scheme depicting how loss of the RAS death effector RASSF4 unleashes the promitogenic activity of RAS in MM and how this can be counteracted by enhancing RASSF4 and blocking MEK1/2.

Figure 7.

Effect of RASSF4 expression on sensitivity to the MEK1/2 inhibitor trametinib. A, Effect of forced RASSF4 expression on trametinib sensitivity. Cells were treated with different concentrations of trametinib (T, nmol/L) 3 (XG-7 and AMO-1) or 7 (JJN3) days after treatment with Dox for an additional 3 days. Apoptosis was determined by flow cytometry using an Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. *, P < 0.05; ***, P < 0.001. B, Western blot analysis of the underlying mechanisms of the increased trametinib drug sensitivity upon forced RASSF4 expression. AMO-1-RF4 cells were treated with 50 nmol/L trametinib (T) 3 days after treatment with Dox for 24 hours, after which, Western blot analysis was performed for indicated proteins. Tubulin was used as loading control and one experiment representative of three is shown. C, Antimyeloma activity of trametinib in combination with panobinostat. XG-7, AMO-1, and JJN3 cells were treated with trametinib and/or panobinostat for 72 hours. Apoptosis was determined by flow cytometry using Annexin V-FITC/7′AAD staining. Bars and error bars are mean ± SD of three independent experiments. P values and combination indexes are shown in Supplementary Table S5. D, Scheme depicting how loss of the RAS death effector RASSF4 unleashes the promitogenic activity of RAS in MM and how this can be counteracted by enhancing RASSF4 and blocking MEK1/2.

Close modal

To increase the translational value of these findings and since we showed that HDACi can upregulate RASSF4 expression, we finally examined the antimyeloma effects of trametinib in combination with quisinostat or panobinostat using 5 HMCL. Co-treatment of the cells with either HDACi significantly and synergistically enhanced trametinib-mediated apoptosis (Fig. 7C and Supplementary Fig. S7A), with combination indexes far below 1 (Supplementary Table S5). Moreover, quantitative RT-PCR confirmed RASSF4 upregulation upon panobinostat and/or trametinib treatment in all cell lines tested (Supplementary Fig. S7B). Taken together, these findings provide the rationale to test the anti-MM activity of trametinib in combination with HDACi in RASSF4low patients associated with a poor prognosis.

In MM, sustained activity of the RAS/Raf/MEK/MAPK pathway leading to enhanced MM cell survival and growth is mainly triggered through interaction with the BM microenvironment and/or genetic alterations, including activating mutations of N-RAS, K-RAS or B-Raf (10). Recent studies have pointed out that RAS also has antitumor activity by binding RASSF proteins. However, epigenetic inactivation or loss of these RAS death effectors occurs frequently in a wide range of cancers and is often linked with a worse disease outcome, especially in RAS-driven cancers (13, 16). The present study identifies RASSF4 as the only RASSF member consistently downregulated during disease progression in MM. Methylation analysis of the RASSF4 promoter in HMCL shows an inverse correlation between RASSF4 expression levels and methylation status and treatment with the HDACi quisinostat and/or DNMTi decitabine restores RASSF4 transcription both in HMCL and the 5T33MM model. Together, our data suggest that RASSF4 inactivation in MM is, like in most cancers, mainly mediated through promotor hypermethylation. Nevertheless, future studies should be conducted to validate these findings in a large patient cohort. Moreover, the observed correlation between RASSF4 inactivation and inferior overall survival indicates that loss of RASSF4 might unleash the promitogenic activity of RAS in MM thereby driving MM progression. Consistent with this, forced RASSF4 expression resulted in a clear G2-phase arrest together with a strong induction of caspase-3–mediated apoptosis and this especially in the cells with low basal RASSF4 levels. In line with the G2-phase arrest, we also observed a clear increase in cyclin-dependent kinase inhibitor p21 levels. Moreover, mitotic cells consistently expressed very low levels of RASSF4. Together, this demonstrates that RASSF4 inhibits growth of MM cells and potentially has a negative role in mitotic progression. RASSF proteins are widely known to influence cell-cycle progression (14). RASSF1A, for example, exerts its antitumor effects mainly by cell-cycle restriction and inducing a G1- or G2-phase arrest (14, 15). In breast cancer, RASSF4 also blocks cell-cycle progression and induces apoptosis. However, forced RASSF4 expression had no effect on lung cancer cells and a pro-growth effect in alveolar rhabdomyosarcoma, suggesting that its role as a tumor suppressor may be cell context dependent (12, 14, 38). Forced RASSF4 expression also significantly reduced tumor growth and prolonged overall survival in xenograft models and strongly reduced the viability of primary MM cells. Importantly, forced RASSF4 expression had no effect on the viability of the non-myeloma cells, suggesting that RASSF4 overexpression mainly targets MM cells. Moreover, no significant correlation was found between sensitivity to forced RASSF4 expression and basal RASSF4 expression nor RAS mutation status. In fact, even patients with a relative high basal RASSF4 expression still responded to the forced expression and no differences in response were observed between wildtype and RAS mutated samples. In line, we also observed no clear difference between wildtype and RAS mutated HMCL. Together, these findings provide evidence that RASSF4 is a negative regulator of MM growth and survival, regardless of the RAS mutation status.

RASSF4 has a 60% structure homology with RASSF2 (12). Moreover, similar to RASSF2, we found that RASSF4 is mainly localized to the nucleus and the centrosome (14). Thus, we hypothesized that the mechanisms of action of RASSF4 might resemble those of RASSF2. Like most RASSF proteins, RASSF2 mainly acts as a scaffolding protein, modulating the activity of proapoptotic effectors. Well-known binding partners for RASSF2, apart from KRAS, are MST1/2 (14, 39). MST1-RASSF2 binding stabilizes RASSF2 and activates MST1 (17). MST1 is a well-known proapoptotic kinase that becomes activated upon caspase-mediated cleavage, resulting in JNK/c-Jun pathway activation, phosphorylation of H2AX and H2B, chromatin condensation and finally apoptosis. In alveolar rhabdomyosarcoma, RASSF4 was shown to interact with MST1 (38, 40). We demonstrated that RASSF4 also binds with and activates MST1 in MM cells. RASSF2 and RASSF6 both activate the JNK pathway leading to apoptosis and increased drug sensitivity, underscoring the importance of the JNK pathway in RASSF-mediated antitumor effects (14, 17, 41). In line, we observed a clear increase in JNK and c-Jun phosphorylation and expression upon RASSF4 overexpression. In addition, RASSF4 overexpression also increased phosphorylation of H2AX and Bim levels. In MM, low Jun levels correlate with a bad prognosis and increased c-Jun expression inhibits proliferation and induces caspase-dependent apoptosis (22). Moreover, the JNK pathway increases Bim phosphorylation and expression in MM cells (42). Thus, together our data suggest that forced RASSF4 expression in MM cells results in increased MST1 binding and activation, leading to JNK/c-Jun pathway activation, chromatin condensation and MM cell death. Surprisingly, p44/42 MAPK was also stimulated upon RASSF4 expression. This is counter-intuitive and in contrast with an earlier study showing that RASSF4 inhibits MAP kinase signaling in oral squamous cell cancer stem-like cells (43). However, recent evidence increasingly demonstrates that activation of Raf/MAPK signaling consists of multiple feedback loops and should be considered more as a complex network rather than a linear pathway. In this regard, RASSF1a and RASSF6 have been shown to modulate Raf-1, B-Raf, and MEK/ERK activity (13, 44). Taken together, these findings demonstrate that RASSF4 serves as a central signaling hub in MM, connecting RAS to various signaling pathways. However, the exact mechanism by which RASSF4 modulates these pathways remains to be further investigated. In this regard, it will be important to further explore the interaction with MST1 by using, for example, RASSF4 mutants that are unable to interact with MST1.

Forced RASSF4 expression also increased sensitivity to bortezomib. Previously, bortezomib has been demonstrated to also activate MST1. In our study, however, bortezomib did not result in MST1 cleavage. This might be explained by the fact that we used much lower concentrations and looked at earlier time points. Nevertheless, bortezomib treatment of RASSF4-overexpressing cells resulted in higher levels of hyperactive MST1 and increased activation of the JNK/c-Jun pathway. In line with these findings, bortezomib sensitivity of MM cells is linked to the activation status of the JNK pathway and c-Jun expression is important in overcoming bortezomib resistance (22). Thus, our data indicate that forced RASSF4 expression enhances MM cell sensitivity to bortezomib by increasing the levels of hyperactive MST1, and hence the activity of the JNK pathway. RASSF4 also had a good prognostic value in a cohort of patients treated with bortezomib after relapse. These findings indicate that RASSF4 could be an interesting biomarker for bortezomib sensitivity. In support of this, we recently identified genes differentially expressed between responders and nonresponders in the bortezomib arm of this cohort and found that RASSF4 is significantly upregulated in the responders (45). Importantly, panobinostat was recently FDA approved for the treatment of relapsed/refractory MM patients (46). Our work suggests that HDACi mediated re-expression of RASSF4 may sensitize patients to bortezomib (possibly by activation of the JNK signaling pathway) and could be one of the mechanisms involved in this improved outcome. However, this remains to be validated in the future.

On the basis of the crucial role of aberrant RAS/Raf/MEK/MAPK signaling in MM progression, MEK1/2 inhibitors are currently under wide clinical investigation (4, 34). Trametinib is a specific MEK1/2 inhibitor that is FDA approved in melanoma patients harboring a BRAF V600E or V600K mutation. In MM, however, clinical studies investigating the effect of trametinib alone and in combination with AKT inhibitors were disappointing or suspended early due to toxicity issues (35, 47). A possible explanation for these disappointing results may lie in the redundant and ubiquitous stimulation of RAS/RAF/ERK pathway in MM (10). Thus, we hypothesized that enhancing expression of the pro-death RAS effector RASSF4 might increase the antimyeloma activity of trametinib. As expected, RASSF4 overexpression did significantly enhance trametinib-mediated cytotoxicity. Our kinome analysis revealed that trametinib treatment of RASSF4-overexpressing cells had a strong negative influence on several well-known prosurvival pathways, including MEK/ERK, PI3K/mTOR, p53, and GSK3. In addition, we demonstrated increased levels of phosphorylated and total c-Jun, p21, and p27. Thus, we provide evidence that forced RASSF4 expression enhances MM cell trametinib sensitivity by activating the JNK/c-Jun pathway and p21 levels, while further reducing several important prosurvival signaling pathways. Moreover, our findings are in line with an earlier report, showing that MAPK-ERK-2 signaling increases shuttling of RASSF2 from nucleus to cytoplasm, thus negatively influencing RASSF2-mediated effects. Treatment with a MAPK-ERK-2 inhibitor was able to retain RASSF2 in the nucleus, thus enhancing its tumor suppressor effects (14, 48). Importantly, coexpression of MAF in RAS mutated cells confers resistance to trametinib by increasing MEK/MAPK signaling (49). In MM, MAF is highly expressed in MM cells harboring a t(14;16) and t(4;14) (50). Here, we show that RASSF4 significantly enhanced trametinib-mediated cytotoxicity both in RAS mutated cell lines coexpressing MAF (XG-7 and JJN-3) or not (AMO-1), suggesting that RASSF4 sensitizes MM cells to MEK inhibition independent of MAF expression levels. Recent work also showed that pan-RAF inhibition is required to fully abolish MEK/ERK phosphorylation and induce strong apoptosis independently of the RAS mutation status (10). We observed that 2 out of the 3 cell lines used in this study, namely MM1s and KMS11, displayed very high RASSF4 mRNA expression. Thus, it is tempting to speculate that this apoptosis might be partially mediated by MST-1/RASSF4 (44).

Finally, we showed that increasing RASSF4 expression in MM cells using fairly low doses of HDACi strongly and highly synergistically enhanced the anti-MM activity of trametinib. In KRAS mutant pancreatic cancer cells targeting MAPK and PI3K has only limited efficacy. However, adding an HDACi to this combination led to a potent induction of apoptosis (51). In addition, in BRAF mutant colorectal cancer, HDACi overcame resistance to MEK inhibitors (52). Consequently, our work suggests that combining HDACi and MEK inhibitors represents a promising new treatment option for MM patients with low RASSF4 expression in association with an adverse outcome (Fig. 7D).

In conclusion, our findings demonstrate that RASSF4 is as a negative regulator of MM growth and survival and loss of RASSF4 promotes MM progression by disconnecting RAS from its proapoptotic pathways. Moreover, we provide evidence that RASSF4 sensitizes the MM cells to bortezomib and the MEK1/2 inhibitor trametinib. Therefore, since HDACi can restore RASSF4 expression in MM cells, we propose clinical evaluation of a MEK inhibitor in combination with bortezomib and HDACi.

C. Heirman is in product development and has ownership interest (including patents) in eTheRNA immunotherapies NV. No potential conflicts of interest were disclosed by the other authors.

Conception and design: E. De Smedt, K. Maes, K. Vanderkerken, E. De Bruyne

Development of methodology: E. De Smedt, K. Maes, H. Lui, A. Cakana, D. Hose, K. Breckpot, E. De Bruyne

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): E. De Smedt, K. Maes, S. Verhulst, A. Maes, D. Hose, L.A. van Grunsven, J. Moreaux, E. De Bruyne

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): E. De Smedt, K. Maes, S. Verhulst, A. Kassambara, A. Maes, N. Robert, A. Cakana, D. Hose, J. Moreaux, E. De Bruyne

Writing, review, and/or revision of the manuscript: E. De Smedt, K. Maes, S. Verhulst, N. Robert, A. Cakana, D. Hose, K. Breckpot, L.A. van Grunsven, K. De Veirman, E. Menu, K. Vanderkerken, J. Moreaux, E. De Bruyne

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): E. De Smedt, C. Heirman, K. Breckpot, E. De Bruyne

Study supervision: K. Vanderkerken, E. De Bruyne

This work was supported by the International Myeloma Foundation (IMF; to E. De Bruyne); KomOpTegenKanker (KOTK; to E. De Bruyne, E. Menu, and K. Vanderkerken); Fonds voor Wetenschappelijk Onderzoek (FWO; to E. De Bruyne); the ARC foundation, the French National Research Agency (ANR) and CRLR (to J. Moreaux) and the German Research Foundation (DFG) and the Federal Ministry of Education and Research (BMBF; to D. Hose). E. De Bruyne, K. Maes, and K. De Veirman are postdoctoral fellows of FWO. We thank Sofie Seghers, Carine Seynaeve, Elsy Vaeremans, and Petra Roman for their expert technical assistance.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Chesi
M
,
Bergsagel
PL
. 
Advances in the pathogenesis and diagnosis of multiple myeloma
.
Int J Lab Hematol
2015
;
37
Suppl 1
:
108
14
.
2.
Kimmelman
AC
. 
Metabolic dependencies in RAS-driven cancers
.
Clin Cancer Res
2015
;
21
:
1828
34
.
3.
Mulligan
G
,
Lichter
DI
,
Di Bacco
A
,
Blakemore
SJ
,
Berger
A
,
Koenig
E
, et al
Mutation of NRAS but not KRAS significantly reduces myeloma sensitivity to single-agent bortezomib therapy
.
Blood
2014
;
123
:
632
9
.
4.
Heuck
CJ
,
Jethava
Y
,
Khan
R
,
van Rhee
F
,
Zangari
M
,
Chavan
S
, et al
Inhibiting MEK in MAPK pathway-activated myeloma
.
Leukemia
2016
;
30
:
976
80
.
5.
Kortum
KM
,
Mai
EK
,
Hanafiah
NH
,
Shi
CX
,
Zhu
YX
,
Bruins
L
, et al
Targeted sequencing of refractory myeloma reveals a high incidence of mutations in CRBN and Ras pathway genes
.
Blood
2016
;
128
:
1226
33
.
6.
Steinbrunn
T
,
Stuhmer
T
,
Sayehli
C
,
Chatterjee
M
,
Einsele
H
,
Bargou
RC
. 
Combined targeting of MEK/MAPK and PI3K/Akt signalling in multiple myeloma
.
Br J Haematol
2012
;
159
:
430
40
.
7.
Chng
WJ
,
Gonzalez-Paz
N
,
Price-Troska
T
,
Jacobus
S
,
Rajkumar
SV
,
Oken
MM
, et al
Clinical and biological significance of RAS mutations in multiple myeloma
.
Leukemia
2008
;
22
:
2280
4
.
8.
Steinbrunn
T
,
Stuhmer
T
,
Gattenlohner
S
,
Rosenwald
A
,
Mottok
A
,
Unzicker
C
, et al
Mutated RAS and constitutively activated Akt delineate distinct oncogenic pathways, which independently contribute to multiple myeloma cell survival
.
Blood
2011
;
117
:
1998
2004
.
9.
Rowley
M
,
Liu
P
,
Van Ness
B
. 
Heterogeneity in therapeutic response of genetically altered myeloma cell lines to interleukin 6, dexamethasone, doxorubicin, and melphalan
.
Blood
2000
;
96
:
3175
80
.
10.
Muller
E
,
Bauer
S
,
Stuhmer
T
,
Mottok
A
,
Scholz
CJ
,
Steinbrunn
T
, et al
Pan-Raf co-operates with PI3K-dependent signalling and critically contributes to myeloma cell survival independently of mutated RAS
.
Leukemia
2017
;
31
:
922
33
.
11.
Cox
AD
,
Der
CJ
. 
The dark side of Ras: regulation of apoptosis
.
Oncogene
2003
;
22
:
8999
9006
.
12.
Eckfeld
K
,
Hesson
L
,
Vos
MD
,
Bieche
I
,
Latif
F
,
Clark
GJ
. 
RASSF4/AD037 is a potential ras effector/tumor suppressor of the RASSF family
.
Cancer Res
2004
;
64
:
8688
93
.
13.
Donninger
H
,
Schmidt
ML
,
Mezzanotte
J
,
Barnoud
T
,
Clark
GJ
. 
Ras signaling through RASSF proteins
.
Semin Cell Dev Biol
2016
;
58
:
86
95
.
14.
Volodko
N
,
Gordon
M
,
Salla
M
,
Ghazaleh
HA
,
Baksh
S
. 
RASSF tumor suppressor gene family: biological functions and regulation
.
FEBS Lett
2014
;
588
:
2671
84
.
15.
Donninger
H
,
Clark
JA
,
Monaghan
MK
,
Schmidt
ML
,
Vos
M
,
Clark
GJ
. 
Cell cycle restriction is more important than apoptosis induction for RASSF1A protein tumor suppression
.
J Biol Chem
2014
;
289
:
31287
95
.
16.
Richter
AM
,
Pfeifer
GP
,
Dammann
RH
. 
The RASSF proteins in cancer; from epigenetic silencing to functional characterization
.
Biochim Biophys Acta
2009
;
1796
:
114
28
.
17.
Song
H
,
Oh
S
,
Oh
HJ
,
Lim
DS
. 
Role of the tumor suppressor RASSF2 in regulation of MST1 kinase activity
.
Biochem Biophys Res Commun
2010
;
391
:
969
73
.
18.
Mezzanotte
JJ
,
Hill
V
,
Schmidt
ML
,
Shinawi
T
,
Tommasi
S
,
Krex
D
, et al
RASSF6 exhibits promoter hypermethylation in metastatic melanoma and inhibits invasion in melanoma cells
.
Epigenetics
2014
;
9
:
1496
503
.
19.
Moreaux
J
,
Klein
B
,
Bataille
R
,
Descamps
G
,
Maiga
S
,
Hose
D
, et al
A high-risk signature for patients with multiple myeloma established from the molecular classification of human myeloma cell lines
.
Haematologica
2011
;
96
:
574
82
.
20.
De Bruyne
E
,
Bos
TJ
,
Schuit
F
,
Van Valckenborgh
E
,
Menu
E
,
Thorrez
L
, et al
IGF-1 suppresses Bim expression in multiple myeloma via epigenetic and posttranslational mechanisms
.
Blood
2010
;
115
:
2430
40
.
21.
Reme
T
,
Hose
D
,
Theillet
C
,
Klein
B
. 
Modeling risk stratification in human cancer
.
Bioinformatics
2013
;
29
:
1149
57
.
22.
Chen
L
,
Wang
S
,
Zhou
Y
,
Wu
X
,
Entin
I
,
Epstein
J
, et al
Identification of early growth response protein 1 (EGR-1) as a novel target for JUN-induced apoptosis in multiple myeloma
.
Blood
2010
;
115
:
61
70
.
23.
Hose
D
,
Reme
T
,
Meissner
T
,
Moreaux
J
,
Seckinger
A
,
Lewis
J
, et al
Inhibition of aurora kinases for tailored risk-adapted treatment of multiple myeloma
.
Blood
2009
;
113
:
4331
40
.
24.
Mulligan
G
,
Mitsiades
C
,
Bryant
B
,
Zhan
F
,
Chng
WJ
,
Roels
S
, et al
Gene expression profiling and correlation with outcome in clinical trials of the proteasome inhibitor bortezomib
.
Blood
2007
;
109
:
3177
88
.
25.
Aryee
MJ
,
Jaffe
AE
,
Corrada-Bravo
H
,
Ladd-Acosta
C
,
Feinberg
AP
,
Hansen
KD
, et al
Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays
.
Bioinformatics
2014
;
30
:
1363
9
.
26.
Lawrence
M
,
Huber
W
,
Pages
H
,
Aboyoun
P
,
Carlson
M
,
Gentleman
R
, et al
Software for computing and annotating genomic ranges
.
PLoS Comput Biol
2013
;
9
:
e1003118
.
27.
Chow
LS
,
Lo
KW
,
Kwong
J
,
Wong
AY
,
Huang
DP
. 
Aberrant methylation of RASSF4/AD037 in nasopharyngeal carcinoma
.
Oncol Rep
2004
;
12
:
781
7
.
28.
Han
Y
,
Dong
Q
,
Hao
J
,
Fu
L
,
Han
X
,
Zheng
X
, et al
RASSF4 is downregulated in nonsmall cell lung cancer and inhibits cancer cell proliferation and invasion
.
Tumour Biol
2016
;
37
:
4865
71
.
29.
Leich
E
,
Weissbach
S
,
Klein
HU
,
Grieb
T
,
Pischimarov
J
,
Stuhmer
T
, et al
Multiple myeloma is affected by multiple and heterogeneous somatic mutations in adhesion- and receptor tyrosine kinase signaling molecules
.
Blood Cancer J
2013
;
3
:
e102
.
30.
Cooper
WN
,
Hesson
LB
,
Matallanas
D
,
Dallol
A
,
von Kriegsheim
A
,
Ward
R
, et al
RASSF2 associates with and stabilizes the proapoptotic kinase MST2
.
Oncogene
2009
;
28
:
2988
98
.
31.
Ikeda
M
,
Kawata
A
,
Nishikawa
M
,
Tateishi
Y
,
Yamaguchi
M
,
Nakagawa
K
, et al
Hippo pathway-dependent and -independent roles of RASSF6
.
Sci Signal
2009
;
2
:
ra59
.
32.
Teraishi
F
,
Guo
W
,
Zhang
L
,
Dong
F
,
Davis
JJ
,
Sasazuki
T
, et al
Activation of sterile20-like kinase 1 in proteasome inhibitor bortezomib-induced apoptosis in oncogenic K-ras-transformed cells
.
Cancer Res
2006
;
66
:
6072
9
.
33.
Ura
S
,
Nishina
H
,
Gotoh
Y
,
Katada
T
. 
Activation of the c-Jun N-terminal kinase pathway by MST1 is essential and sufficient for the induction of chromatin condensation during apoptosis
.
Mol Cell Biol
2007
;
27
:
5514
22
.
34.
Chang-Yew Leow
C
,
Gerondakis
S
,
Spencer
A
. 
MEK inhibitors as a chemotherapeutic intervention in multiple myeloma
.
Blood Cancer J
2013
;
3
:
e105
.
35.
Tolcher
AW
,
Patnaik
A
,
Papadopoulos
KP
,
Rasco
DW
,
Becerra
CR
,
Allred
AJ
, et al
Phase I study of the MEK inhibitor trametinib in combination with the AKT inhibitor afuresertib in patients with solid tumors and multiple myeloma
.
Cancer Chemother Pharmacol
2015
;
75
:
183
9
.
36.
Munugalavadla
V
,
Mariathasan
S
,
Slaga
D
,
Du
C
,
Berry
L
,
Del Rosario
G
, et al
The PI3K inhibitor GDC-0941 combines with existing clinical regimens for superior activity in multiple myeloma
.
Oncogene
2014
;
33
:
316
25
.
37.
Gilmartin
AG
,
Bleam
MR
,
Groy
A
,
Moss
KG
,
Minthorn
EA
,
Kulkarni
SG
, et al
GSK1120212 (JTP-74057) is an inhibitor of MEK activity and activation with favorable pharmacokinetic properties for sustained in vivo pathway inhibition
.
Clin Cancer Res
2011
;
17
:
989
1000
.
38.
Crose
LE
,
Galindo
KA
,
Kephart
JG
,
Chen
C
,
Fitamant
J
,
Bardeesy
N
, et al
Alveolar rhabdomyosarcoma-associated PAX3-FOXO1 promotes tumorigenesis via Hippo pathway suppression
.
J Clin Invest
2014
;
124
:
285
96
.
39.
Schagdarsurengin
U
,
Richter
AM
,
Hornung
J
,
Lange
C
,
Steinmann
K
,
Dammann
RH
. 
Frequent epigenetic inactivation of RASSF2 in thyroid cancer and functional consequences
.
Mol Cancer
2010
;
9
:
264
.
40.
Ura
S
,
Masuyama
N
,
Graves
JD
,
Gotoh
Y
. 
MST1-JNK promotes apoptosis via caspase-dependent and independent pathways
.
Genes Cells
2001
;
6
:
519
30
.
41.
Liang
YY
,
Chen
MY
,
Hua
YJ
,
Chen
S
,
Zheng
LS
,
Cao
X
, et al
Downregulation of Ras association domain family member 6 (RASSF6) underlies the treatment resistance of highly metastatic nasopharyngeal carcinoma cells
.
PLoS One
2014
;
9
:
e100843
.
42.
Lei
K
,
Davis
RJ
. 
JNK phosphorylation of Bim-related members of the Bcl2 family induces Bax-dependent apoptosis
.
Proc Natl Acad Sci U S A
2003
;
100
:
2432
7
.
43.
Michifuri
Y
,
Hirohashi
Y
,
Torigoe
T
,
Miyazaki
A
,
Fujino
J
,
Tamura
Y
, et al
Small proline-rich protein-1B is overexpressed in human oral squamous cell cancer stem-like cells and is related to their growth through activation of MAP kinase signal
.
Biochem Biophys Res Commun
2013
;
439
:
96
102
.
44.
Romano
D
,
Nguyen
LK
,
Matallanas
D
,
Halasz
M
,
Doherty
C
,
Kholodenko
BN
, et al
Protein interaction switches coordinate Raf-1 and MST2/Hippo signalling
.
Nat Cell Biol
2014
;
16
:
673
84
.
45.
Hofman
IJ
,
Van Duin
M
,
De Bruyne
E
,
Fancello
L
,
Mulligan
G
,
Geerdens
E
, et al
RPL5 on 1p22.1 is recurrently deleted in multiple myeloma and its expression is linked to bortezomib response
.
Leukemia
2017
;
31
:
1706
14
.
46.
Tandon
N
,
Ramakrishnan
V
,
Kumar
SK
. 
Clinical use and applications of histone deacetylase inhibitors in multiple myeloma
.
Clin Pharmacol
2016
;
8
:
35
44
.
47.
clinicaltrials.gov
.
U.S. National Institutes of Health
; 
2016
[June 23rd, 2016]; Available from
: https://clinicaltrials.gov/ct2/results?term=Trametinib+and+myeloma&Search=Search.
48.
Kumari
G
,
Mahalingam
S
. 
Extracellular signal-regulated kinase 2 (ERK-2) mediated phosphorylation regulates nucleo-cytoplasmic shuttling and cell growth control of Ras-associated tumor suppressor protein, RASSF2
.
Exp Cell Res
2009
;
315
:
2775
90
.
49.
Qiang
Y-W
. 
The co-occurrence of MAF translocations in RAS mutated multiple myeloma confers resistance to MEK inhibition
.
Blood
2016
;
128
:
1138
.
50.
Qiang
YW
,
Ye
S
,
Chen
Y
,
Buros
AF
,
Edmonson
R
,
Van Rhee
F
, et al
MAF protein mediates innate resistance to proteasome inhibition therapy in multiple myeloma
.
Blood
2016
;
128
:
2919
30
.
51.
Ischenko
I
,
Petrenko
O
,
Hayman
MJ
. 
A MEK/PI3K/HDAC inhibitor combination therapy for KRAS mutant pancreatic cancer cells
.
Oncotarget
2015
;
6
:
15814
27
.
52.
Carson
R
,
Celtikci
B
,
Fenning
C
,
Javadi
A
,
Crawford
N
,
Perez-Carbonell
L
, et al
HDAC inhibition overcomes acute resistance to MEK inhibition in BRAF-mutant colorectal cancer by downregulation of c-FLIPL
.
Clin Cancer Res
2015
;
21
:
3230
40
.

Supplementary data