ALKBH5 Promotes Multiple Myeloma Tumorigenicity through inducing m⁶A-demethylation of SAV1 mRNA and Myeloma Stem Cell Phenotype

Culture of MM primary cells and cell lines

Mononuclear cells were isolated from bone marrow samples of newly diagnosed MM patients and healthy volunteers by Ficoll at Union Hospital, Wuhan, China (Haoyang, Tianjin, China). The present study was approved by the institutional ethics committee of Huazhong University of Science and Technology. CD138⁺ plasma cells were selected with CD138⁺ microbeads (Miltenyi, Auburn, CA, USA). RPMI8226, U266, ARH-77 and NCI-H929 MM cell lines were purchased from American Type Culture Collection (ATCC) (ATCC, Manassas, VA, USA) and cultured in complete RPMI 1640 medium (Gibco, Grand Island, NY, USA) supplemented with 10-15% FBS (Natocor, Villa Carlos Paz; Argentina) in 5% CO₂ at 37°C.

Lentivirus transfection

A total of 1×10⁵ cells were placed in the 24-well plate, and then, 20 μl of polybrene and virus solution (GeneChem, Shanghai, China) was added to wells with sufficient mixing, and the cells were cultured in 5% CO2 at 37°C. Twelve to sixteen hours later, the medium was changed, and the cells continued to be cultured under the same conditions. Stable ALKBH5-knockdown and ALKBH5-overexpressing cells were obtained after being screened by puromycin (Sigma, St. Louis, MO, USA).

Transwell assays

Cell migration and invasion ability was analyzed by the Transwell assay.

Migration ability test: 50,000 cells (200µl) were placed in the Transwell chambers (8µm, Corning, USA), and 600µl RPMI-1640 medium containing 25% FBS was added to the lower compartment. 24h later, the cells in the upper chamber were wiped away. Cells on the lower surface of the chambers were fixed and stained and then counted under an optical microscope (magnification, ×400).

Invasive ability test: Matrigel matrix (Corning, NY, USA) was mixed with serun-free medium in a 1:8 ratio. The diluted mixture was applied on the underside of the Transwell chamber at 100 µl/well in a 37°C incubator for 4 h, and the other steps were the same as the migration experiment.

RIP-qPCR

RIP-qPCR was performed as described with minor modifications.²² 5 μg ALKBH5 and IgG control antibodies were used for RIP assay. Co-precipitated RNAs were then determined by RT-qPCR.

ELISA

Myeloma cells were cultured in complete RPMI 1640 medium, 24h later, the supernatant was collected to measure the levels of VEGF using ELISA kit according to the manufacturer's instruction (Multisciences biotech, Hangzhou, China).

Side population cell detection

A side population (SP) assay was conducted using Hoechst staining as previously described.²³ A total of 1×10⁶ MM cells were sustained in 1 mL of RPMI1640 medium supplemented with 0.5 μL of Hoechst dye (Sigma, St Louis, MO, USA) for 90 min at 37℃ either alone or in the presence of 50 mol/L verapamil. Then, the cells were washed with cold PBS to terminate the reaction. The excitation wavelength was set to 355 nm, and the detection wavelengths were 450 and 675 nm. Cells negative for Hoechst red and Hoechst blue stain were considered positive for side population cells.

Flow cytometry

Cells were washed with PBS and then resuspended in binding buffer at a density of 1 × 10⁶ cells/ml. 100 μl of cell suspension was transferred to a 5 ml culture tube, next, 5μl CD138-BV421 and CD34-APC antibody (BD Bioscience, San Diego, CA, USA) were added to the tube. The mixture was vortexed and incubated for 15 min at 4°C in the dark. Finally, the samples were analysed by flow cytometry(BD Bioscience, San Diego, CA, USA).

Stem cell sphere formation assay

ALKBH5- and control RPMI8226 cells were plated in 6-well plates and cultured in prepared stem cell medium. Briefly, stem cell medium was prepared by RPMI 1640 medium supplemented with rh stem cell factor, rh GM-CSF, rh IL-3 and EPO (Biolegend, San Diego, CA, USA) and used immediately after preparation. The cell spheroids larger than 50μm in diameter under the microscope was counted as stem cell spheres. Sphere formation ability was determined by the number of stem cell spheroids that formed 14 days after seeding.

MeRIP-sequencing and gene-specific m⁶A qPCR

The purified RNA samples were cut into 100 bp fragments according to a previous study²². Then, RNA samples were incubated with m⁶A primary antibody (Abcam, Cambridge, UK) and immunoprecipitated with protein A beads (Merck Millipore, Billerica, MA, USA) for 2 h at 4 °C. Captured RNA was washed and purified by an RNA Clean and Concentrator kit (ZYMO Research, Beijing, China). Sequencing was carried out on an Illumina HiSeq 2000 (Illumina, San Diego, CA, USA), and qRT-PCR was used to examine the relative abundance of m⁶A SAV1 RNA. The sequences of the SAV1 primers were 5′-ATGCTGTCCCGAAAGAAAACC-3′ and 5′-AGGCATAAGATTCCGAAGCAGA-3′.

MeRIP-sequencing Data analysis

The exomePeak (Version 3.8) software was used for peak calling. The m6A peaks were annotated using bedtools (Version 2.25.0). The deepTools (version 2.4.1) was used for peak distribution analysis. The differentially m6A peaks were identified by a python script, using fisher test. Sequence motifs enriched in m6A peak regions were identified using Homer (version 4.10). Gene ontology (GO) analysis and Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis for annotated genes were both implemented by KOBAS software (version: 2.1.1).

Statistical analysis

The difference between groups was analysed as by Student's t-test, and multiple groups were analysed when applicable by one-way ANOVA in Prism 7.0 (GraphPad Software, San Diego, CA, USA). Experiments were carried out in triplicate. A P-value <0.05 was considered statistically significant.

Elevated expression of ALKBH5 in MM patients and MM cell lines

We first investigated the expression of ALKBH5 in bone marrow-derived CD138⁺ cells from 8 healthy volunteers (normal control, NC) and 24 newly diagnosed MM patients through qRT-PCR. Then we verified the levels of ALKBH5 in 9 healthy volunteers and 25 newly diagnosed MM patients by western blot. The results showed that the expression of ALKBH5 in the CD138⁺ MM cells was significantly higher than that in the NC cells (Figure 1A and 1B). We also found that the ALKBH5 mRNA level was positively correlated with the MM stage according to the International Staging System (ISS) (Figure 1D). Consistently, the expression of ALKBH5 in myeloma cell lines, including RPMI8226, ARH-77, NCI-H929 and U266 cells, was more abundant than that in the normal controls (P<0.05) (Figure 1C and 1E). Collectively, these data suggested that ALKBH5 was highly expressed in both the MM patients and MM cell lines.

Effect of ALKBH5 on myeloma cell proliferation, apoptosis, neovascularization, migration and invasion in vitro

To characterize the role of ALKBH5 in the pathogenesis of MM, ALKBH5-RNAi, ALKBH5-OE and control lentiviruses containing GFP were transfected into RPMI8226 cells. The transfection efficiency was confirmed by fluorescence microscopy, qRT-PCR and western blot. The expression level of ALKBH5 in the RPMI8226 cells transfected with ALKBH5-RNAi (ALKBH5^- cells) or ALKBH5-OE (ALKBH5⁺ cells) lentivirus were significantly downregulated and upregulated, respectively, compared with the control cells (see Figure S1). Subsequently, we evaluated the proliferation of myeloma cells by CCK-8 assay and found that the proliferation of the ALKBH5^- RPMI8226 cells was prominently inhibited in contrast to the control cells (P<0.0001, Figure 2A). Similarly, the downregulation of ALKBH5 also suppressed the viability of the NCI-H929 cells (P<0.001, Figure 2B). The overexpression of ALKBH5 promoted the proliferation of RPMI8226 and ARH-77 cells (P<0.001, Figure 2A and 2B). Flow cytometry analysis showed that ALKBH5 deficiency increased the apoptosis rate of the RPMI8226 cells and NCI-H929 cells (P<0.05), while upregulation of ALKBH5 did not lead to a significant change (Figure 2C&D and figure S2). What's more, transwell assays were used to examine cell migration and invasion ability in different group. Our results indicated that cell migration and invasion ability were significantly downregulated in the ALKBH5^- group and upregulated in the ALKBH5⁺ group (P<0.05; Figure 2E-2F). Additionally, the vascular endothelial growth factor (VEGF), as a core proangiogenic cytokine, plays an important part in MM progress. So we tested the VEGF level in supernatant of cultured ALKBH5^- cells and ALKBH5⁺ cells through VEGF ELISA kit, respectively. It showed that knockdown of ALKBH5 could inhibit the VEGF secretion ability in myeloma cells (P< 0.001, Figure 2G).

ALKBH5 inhibition exerts anti-myeloma effect in vivo

To verify the effect of ALKBH5 in vivo, we developed 6 xenograft models in NOD/SCIDs through subcutaneous injection of AKBH5^- and control RPMI8226 cells (ALKBH5^- : n=3, control: n=3). As expected, xenografts established with ALKBH5^- myeloma cells grew significantly slowlier than the controls (P<0.001, Figure 3A). The tumour volumes in the mice injected with the ALKBH5^- cells were obviously smaller compared to those in control mice (Figure 3B). Immunohistochemical staining for BCL-2, BAX and caspase-3 in tumour sections exhibited a marked decrease in BCL-2, Ki67 and pAKT and an increase in BAX and caspase-3 activity and expression in the tumours generated with ALKBH5^- cells (Figure 3C). Together, these results indicate that ALKBH5 inhibition has a protective effect on tumour growth in vivo.

Identification of the target genes of ALKBH5 via MeRIP sequencing analysis

Firstly, the m⁶A level in the RNA samples taken from ALKBH5^- MM cells, control MM cells and bone marrow-derived CD138⁺ cells from healthy volunteers (normal controls, NCs) were detected using an EpiQuick m⁶A RNA methylation quantification kit and found that the m⁶A methylation levels in the NCs and ALKBH5^- myeloma cells were notably higher than those in the control MM cells (Figure 4A). To further investigate the variation of m⁶A modification in the transcripts, we mapped the transcriptome-wide m⁶A-methylomes in the ALKBH5^- and control MM cells as well as NCs by MeRIP-seq (methylated RNA immunoprecipitation with next-generation sequencing) as previously reported.²⁴ According to our data, the m⁶A modifications were mainly located at the CDS, 3'UTR or stop codon regions (Figure 4B), and the consensus motif identified by HOMER was “GGACA(U)” (Figure 4C). A total of 10204 m⁶A peaks were detected in the ALKBH5^- cells, which was notably higher than that in the control cells (8044) (Figure 4D). Further analysis identified that 1649 transcripts were only identifid in ALKBH5^- group while not in control group, called “ALKBH5^--specific genes”. And m⁶A modifications were specifically found on these ALKBH5^--specific transcripts. Also, MeRIP-seq analysis showed that a total of 9786 m⁶A peaks were detected for the NC cells. Among them, 2359 transcripts has specific m⁶A peaks (only identifid in NCs group while not in control MM group) and were called “NC-specific genes”. Then, RNA sequencing analysis showed that 24 genes were downregulated in the “ALKBH5^--specific gene” group compared with control MM cells, and 120 genes were downregulated in the “NC-specific gene” group compared with control MM cells, and SAV1 gene was selected among the mutually downregulated genes in both groups (Figure 4E and 4F). SAV1 is an important component of the HIPPO pathway, which is highly conserved in mammals and plays a crucial role in the development of neoplasms and stem cell fate regulation. Therefore, we used SAV1 as the entry point to further explore the variations in other components of the HIPPO pathway.

Knock-down of ALKBH5 suppressed myeloma progression via activating the Hippo pathway in vivo and in vitro

First, we used gene-specific m⁶A qPCR to confirm the effect of ALKBH5 on SAV1 m⁶A level, and the m⁶A level of SAV1 mRNA was found to be increased in cells transfected with ALKBH5-RNAi lentivirus (Figure 5A). Moreover, qRT-PCR and western blot analyses confirmed that SAV1 expression was significantly decreased in the ALKBH5^- group (P<0.05, Figure 5B and 5E). Given that m⁶A modification in mRNA can regulate mRNA stability,⁶ an RNA stability assay was carried out to detect the changes of SAV1 mRNA stability, and the results showed that the SAV1 mRNA half-life was decreased after ALKBH5 was silenced (Figure 5C) and increased after ALKBH5 was overexpressed (Figure 5D).

In conclusion, ALKBH5 deficiency can regulate SAV1 mRNA stability and protein expression through m⁶A demethylation modification. Since SAV1 is the initiating protein of HIPPO pathway, other HIPPO pathway downstream proteins, including LATS1, MST1 and P-YAP (S127), were also detected and showed similar downward trends when ALKBH5 expression was knocked down (P<0.05, Figure 5E). The YAP protein, which was repressed by the activation of the HIPPO pathway, exhibited increased expression in the ALKBH5^- cells (P<0.05, Figure 5E). YAP can be bound by ABL1 in the cell nucleus and then can be phosphorylated on tyrosine 357, which augments the stability of P73, activates the proapoptotic gene and ultimately promotes the apoptosis of MM cells.²⁵ Therefore, we further examined the levels of P73-target genes (P21, PUMA and BAX) and P-YAP (Y357) through western blot analysis. Consistent with our hypothesis, ALKBH5 knockdown substantially enhanced the P-YAP (Y357), P21, PUMA and BAX protein levels (P<0.05, Figure 5E). Accordingly, the levels of proteins associated with apoptosis, including BAX, BCL-2 and cleaved caspase3 as well as pERK and pAKT, showed a significant decline in the ALKBH5^- cells (P<0.05, Figure 5F). Consistently, immunohistochemical staining for SAV1 showed a marked decrease and YAP an increase in the xenograft generated with ALKBH5^- cells compared with control cells (Figure 3C). Additionally, RIP-qPCR further substantiated the direct binding between ALKBH5 and SAV1 mRNA (Figure 5G&I). To further determined whether ALKBH5 modulated SAV1 expression through its enzymatic activity, we used the drug 3-deazaadenosine (DAA) which suppress m6A methylation chemically.²⁶ The results showed that the mRNA expression levels of SAV1 were significantly increased after DAA treatments in MM cells (Figure 5H). And the reduction of SAV1 mRNA caused by ALKBH5 knockdown was rescued by DAA treatment (Figure 5J).

Taken together, ALKBH5 inhibition could increase SAV1 m⁶A levels, suppressed the HIPPO signalling pathway and ultimately activates the transcription factor YAP in myeloma cells, promoted the expression of P21, PUMA and BAX, exerting an anti-myeloma effect in vivo and in vitro.

ALKBH5 promotes myelomagenesis by up-regulating SAV1

To determine the function of SAV1 in MM, rescue experiments were performed to identify whether SAV1 is involved in the process of ALKBH5-mediated myelomagenesis. We transfected sh-SAV1 lentivirus and control lentivirus into the ALKBH5-overexpressing MM cells, marked as oe-ALKBH5+sh-SAV1 group, oe-ALKBH5+shNC group, and oe-NC+shNC group. Western blot results displayed that the inhibition of SAV1 partially neutralized the effect of overexpressed ALKBH5 on both ALKBH5 and SAV1 protein expression in MM RPMI-8226 cells (Figure 6A). Cell viability assays (Figure 6B), VEGF ELISA assays (Figure 6D), cell apoptosis assays (Figure 6F) and cell migration/invasion assays (Figure 6G&H) and showed that SAV1-silencing in ALKBH5 overexpressing MM cells significantly attenuated the ALKBH5-mediated promotion of MM progression. Western blot assays further verified that ALKBH5 regulated proliferation- and apoptosis-related proteins synergized with SAV1 (Figure 6C&E). The in vivo cell derived xenograft (CDX) models also demonstrated that tumors from oe-ALKBH5+sh-NC group had an increase in tumor growth and progression. While ALKBH5-overexpression and SAV1-silencing group showed significantly delayed tumor growth and progression (Figure S4). These results indicate that ALKBH5 promotes myelomagenesis by up-regulating SAV1.

ALKBH5 inhibition reduced the proportion of MM stem cells in vitro

Since the HIPPO pathway is an essential component controlling stem cell self-renewal,²⁷ and ALKBH5 has been suggested to promote the enrichment of breast cancer stem cells,¹² we inferred that ALKBH5 might influence MM stem cell (MMSC) maintenance, proliferation, and survival. The phenotype of the MMSCs remains controversial because of their heterogeneity, and several markers, such as side population (SP) cells, CD138^-/CD34^- cells and stem cell core genes, are most commonly used to identify MMSCs.^{25, 28-30} Therefore, we first detected the percentage of CD138^-/CD34^- cells through flow cytometry. As expected, the ALKBH5^- group had fewer CD138^-/CD34^- cells than the control group (P<0.05, Figure 7A). Furthermore, we also applied Hoechst staining to examine the percentage of SP cells in the RPMI8226 and NCI-H929 MM cell population via flow cytometry. Our results indicated that the percentage of SP cells in both kinds of ALKBH5^- myeloma cells were significantly lower than that in the corresponding control cells (P<0.05, Figure 7B and 7C). In addition, qRT-PCR and western blot were performed to detect the expression of NANOG, SOX2 and OCT4, which are genes associated with stem characteristics. The results showed that the levels of NANOG, SOX2 and OCT4 were notably decreased in the ALKBH5^- cells compared with those in the control cells (P<0.05, Figure 7D). Moreover, we assessed the sphere formation ability of the stem cells in ALKBH5- and control RPMI8226 cell populations and found that the number of stem cell spheres was decreased after ALKBH5 was downregulated (Figure 7E and 7F). Thus, our findings support the hypothesis that ALKBH5 deficiency inhibited the growth of MM cells by decreasing the percentage of MM stem cells.