FTO promotes multiple myeloma progression by posttranscriptional activation of HSF1 in an m⁶A-YTHDF2-dependent manner

Abstract

N⁶-methyladenosine (m⁶A), as the most pervasive internal modification of eukaryotic mRNA, plays a crucial role in various cancers, but its role in multiple myeloma (MM) pathogenesis has not yet been investigated. In this study, we revealed significantly decreased m⁶A methylation in plasma cells (PCs) from MM patients and showed that the abnormal m⁶A level resulted mainly from upregulation of the demethylase fat mass and obesity-associated protein (FTO). Gain- and loss-of-function studies demonstrated that FTO plays a tumor-promoting and pro-metastatic role in MM. Combined m⁶A and RNA sequencing (RNA-seq) and subsequent validation and functional studies identified heat shock factor 1 (HSF1) as a functional target of FTO-mediated m⁶A modification. FTO significantly promotes MM cell proliferation, migration, and invasion by targeting HSF1/HSPs in a YTHDF2-dependent manner. FTO inhibition, especially when combined with bortezomib (BTZ) treatment, synergistically inhibited myeloma bone tumor formation and extramedullary spread in NOD-Prkdc^em26Cd52il2rg^em26Cd22/Nju (NCG) mice. We demonstrated the functional importance of m⁶A demethylase FTO in MM progression, especially in promoting extramedullary myeloma (EMM) formation, and proposed the FTO-HSF1/HSP axis as a potential novel therapeutic target in MM.

Graphical abstract

The role of m⁶A in MM pathogenesis has not yet been investigated. Hu and colleagues demonstrate that aberrant upregulation of m⁶A demethylase FTO enhances the stability of HSF1 by YTHDF2-mediated mRNA decay, thereby promoting MM progression, especially in promoting EMM formation.

Introduction

Multiple myeloma (MM) is the second most common hematological malignancy that develops as a consequence of a multistep transformation process.¹ Despite the development of novel therapeutic agents, MM remains a largely incurable and fatal disease.^2,3 In recent decades, tremendous evidence has indicated that both genetic aberrations, including mutations in oncogenes and tumor suppressor genes, and epigenetic mechanisms, including DNA and histone methylation, abnormal microRNA (miRNA), and other noncoding RNA expression, are involved in the pathogenesis of MM.^4,5 miRNAs are well known to play a central role in MM pathogenesis and drug resistance by destroying mRNA or inhibiting its translation.⁶ Several studies in recent years assessed epigenetic inhibitors in clinical trials of MM, showing that drugs targeting epigenetic modifiers (HDAC or EZH2) can sensitize MM cells to antimyeloma treatment.^7,8 Moreover, miRNAs have been reported to reciprocally impact epigenetic modulators in cancers, highlighting the existence of a regulatory circuit between these two regulatory systems.9, 10, 11, 12 A recent study indicated that the epigenetic miRNA axis could provide potential therapeutic targets in MM patients, especially for those with poor prognosis.¹⁰ Collectively, epigenetic deregulation is now considered a major driving force of MM carcinogenesis and development.

N⁶-methyladenosine (m⁶A), methylation at the N⁶ position of adenosine, is a recently emerging epigenetic molecular mechanism. Recognized by different reader proteins, m⁶A methylation can regulates almost all steps of RNA metabolism, including stability, decay, splicing, nuclear export, and translation.^13,14 Among hundreds of posttranscriptional chemical modifications, m⁶A is the most abundant and pervasive internal modification of eukaryotic mRNAs and noncoding RNAs.15, 16, 17 Analogous to DNA methylation, m⁶A modification is also reversible and is regulated by three groups of molecules commonly referred to as m⁶A writers, erasers, and readers. m⁶A writers, consisting of several methyltransferases such as methyltransferase-like 3 (METTL3), METTL14, and Wilms tumor 1-associated protein (WTAP), catalyze m⁶A formation. Fat mass and obesity-associated protein (FTO) and ALKBH5, two m⁶A demethylases (erasers), facilitate m⁶A removal.^13,18 In addition, the biological effects of m⁶A modification require the recognition of m⁶A marks by m⁶A reader proteins that contain a YT521-B homology (YTH) domain or, alternatively, by eukaryotic initiation factor 3 (eIF3).¹⁸ YTHDF2 is the first identified and the most studied m⁶A reader protein that influences mRNA stability.¹⁹ Recent studies have suggested that m⁶A RNA methylation plays a crucial role in highly conserved bioprocesses, such as stem cell renewal, tissue development, circadian clock regulation, and the heat shock response (HSR).20, 21, 22 Subsequently, compelling evidence indicates that m⁶A modulators have important effects on tumor initiation, invasion, and metastasis in several cancers and function as either oncogenes or tumor suppressor genes.^13,23

The first evidence of the biological impacts of m⁶A regulatory genes on hematological malignancy was found in a study of the m⁶A demethylase FTO. In that work, FTO was reported to play a tumor-promoting role in acute myeloid leukemia (AML) cell differentiation, enhance leukemogenesis, and attenuate the efficacy of all-trans retinoic acid (ATRA) in AML cells by negatively regulating ASB2 and RARA by decreasing their degree of m⁶A modification.²⁴ Additionally, FTO has been shown to act as a crucial oncogene in glioblastoma,²⁵ gastric cancer,²⁶ breast cancer,²⁷ and melanoma,²⁸ likely by regulating cancer-specific targets. Most importantly, epidemiological studies have demonstrated that people with single-nucleotide polymorphisms (SNPs) in FTO or with obesity/overweight tend to have a higher risk of developing various cancers,29, 30, 31 including MM. However, the role of FTO in MM occurrence and development has not yet been investigated.

In the present work, we revealed significantly decreased m⁶A methylation in MM patients, evaluated the role of FTO in MM pathogenesis and progression, and investigated the underlying molecular mechanism. Our findings provide the persuasive evidence that upregulated FTO is associated with the development and progression of MM, especially with extramedullary spread. FTO promoted the proliferation, migration, and invasion of MM in vitro and in vivo through m⁶A- and YTHDF2-dependent HSF1/HSP axis activation. We proposed that FTO may act as a novel therapeutic target in MM, especially in extramedullary myeloma (EMM).

Results

The FTO m⁶A demethylase is highly expressed in MM and EMM samples

Compared with healthy donors, FTO expression was at a higher level in MM patients by transcriptome array analysis (Figure S1A). Using a colorimetric assay, we evaluated the m⁶A level in MM patients (n = 35) and EMM patients (n = 20). Decreased RNA m⁶A levels were detected in primary plasma cells (PCs) from MM patients compared with those from normal donors (NDs) (n = 15), and a further decrease in the m⁶A level was found in primary PCs from EMM patients (Figure 1A). In addition, there was no significant difference between the m⁶A level in the bone marrow (BM) of EMM patients and that in MM patients (data not shown). To investigate the key regulator of m⁶A methylation in MM, we systematically analyzed the levels of the seven major modifying enzymes in eight sets of primary PCs from NDs, MM patients, and EMM patients and identified that FTO, the core m⁶A demethylase, was significantly upregulated in PCs from patients with MM compared with NDs and was further upregulated in PCs from EMM patients (Figures 1B and 1C). Next, we validated the findings in an expanded sample cohort and consistently found that the expression of FTO was significantly elevated in primary PCs from MM patients, especially in EMM patients (Figure 1D). These results were further confirmed by IHC staining of the FTO protein in BM biopsy of MM patients and soft tissues of EMM patients (Figure 1E). Additionally, upregulation of FTO was found in human myeloma cell lines (HMCLs) at both the mRNA and protein levels, especially in ARH77 and JJN3 cells (two kinds of PC leukemia (PCL) cell lines) (Figures 1F and 1G). However, other m⁶A-associated proteins showed no increasing or decreasing trend consistent with the m⁶A level change in NDs, MM, and EMM samples (Figures 1B and 1C); these proteins also showed no progressive trend in the PCs of NDs, MM cell lines or PCL cell lines (Figure 1G). These results demonstrated that FTO was upregulated in MM cells, especially in those from EMM samples.

Figure 1.

FTO is upregulated in MM and EMM

(A) m⁶A mRNA levels in NDs (n = 15), MM patients (n = 35), and EMM patients (n = 20) were measured by an m⁶A ELISA. (B) Expression of m⁶A regulatory enzymes was measured by qPCR in CD138 + cells from NDs, MM patients, and EMM patients. (C) Western blot analysis of major m⁶A enzymes in CD138 + cells isolated from MM patients or EMM patients. N indicates NDs, M indicates MM patients, and E indicates EMM patients. (D) FTO upregulation was further validated in an expanded set of samples from 15 NDs, 35 MM patients, and 20 EMM patients by qPCR. (E) Distribution of primary PCs in samples of different groups was assessed by H&E staining; the FTO protein level was assessed by immunohistochemistry. Average optical density (AOD) of different groups is presented in the histogram on the right. (F and G) qPCR analysis of FTO (F) and Western blot analysis of m⁶A regulatory enzymes (G) in primary PCs from NDs and in MM cell lines. ∗P<0.05, ∗∗P<0.01, ∗∗∗P<0.001, ns: no significance.

FTO promotes the proliferation, migration, and invasion and inhibits the apoptosis of MM cells in vitro

To explore the functional roles of FTO in MM cells, we established stable FTO-KD and FTO-OE in two common MM cell lines, RPMI8226 and MM1R, using lentiviral transduction (Figures 2A and 2B). We found that forced expression of FTO in MM cells promoted cell growth/proliferation, migration, and invasion, while decreasing apoptosis and the global m⁶A RNA modification level (Figures 2C–2G and S1G). In contrast, FTO knockdown (FTO-KD) showed the opposite effects in MM cells (Figures 2C–2G and S1G). Similar effects were observed when the expression of FTO was silenced by siRNAs (Figures S1B–S1F).

Figure 2.

FTO promotes MM cell proliferation and migration/invasion and inhibits apoptosis in vitro

(A and B) Stable FTO-OE and FTO-KD in RPMI8226 and MM1R cells by lentiviral sequencing. The transfection efficiency was confirmed at both the mRNA (A) and protein levels (B). (C) Cell proliferation assay in RPMI8226 and MM1R cells with GFP (vector control), GFP-FTO (FTO-OE), shNC (negative control), or shFTO (FTO-KD). (D) Cell migration assay analysis with cells as in (C). (E) Cell invasion assay with cells as in (C). (F) Apoptosis analysis in cells used in (C). (G) The global m⁶A level was detected in the cells used in (C). ∗P<0.05; ∗∗P<0.01; ∗∗∗P<0.001.

To further verify the oncogenic function of FTO in MM, we used MA2, a selective chemical inhibitor of FTO³² (Figure 3A). As shown in Figure 3B, an obvious growth inhibitory effect of MA2 was detected in MM cells treated with doses of 60–100 μM (P < 0.001). In addition, MA2 suppressed the proliferation and migration of RPMI8226 and MM1R cells at a dose of 60 μM (Figures 3C and 3D), while increasing the apoptosis of these cells (Figures 3E and S1H). Furthermore, to assess the influence of MA2 on the effects of bortezomib (BTZ), a first-line chemotherapeutic agent for MM, we treated MM cells with a combination of MA2 (60 μM) and BTZ (3 nM for RPMI8226 cells and 1.5 nM for MM1R cells for 48 h. The results showed that in the presence of BTZ, MA2 further decreased the proliferation, migration, and invasion but increased the apoptosis of MM cells (Figures 3C–3E and S1H), demonstrating that FTO inhibition potentially augments the efficiency of BTZ in MM cells in vitro. Importantly, MA2 treatment significantly increased the m⁶A levels in MM cells (Figure 3F).

Figure 3.

The FTO inhibitor MA2 suppresses MM cell growth and migration and increases apoptosis in vitro

(A) Western blot analysis was used to assess the inhibitory effect of MA2 on FTO. (B) Cell viability analysis of MM cells treated with MA2. (C) RPMI8226 and MM1R cells were treated with MA2, combined with or without BTZ for 48 h. The effect on viability was determined using a CCK-8 assay. (D) Cell migration was examined in cells as in (C). (E) Apoptosis analysis was conducted in cells as in (C). (F) The global m⁶A level was detected in RPMI8226 and MM1R cells treated with MA2 at different concentrations. ∗P<0.05; ∗∗P<0.01; ∗∗∗P<0.001.

m⁶A-seq and RNA-seq assays identify HSF1 as a potential target of FTO

To investigate the molecular mechanism of FTO and identify its downstream targets in MM, we conducted m⁶A-seq in combination with RNA sequencing (RNA-seq) analysis to explore the potential transcriptome-wide m⁶A-modified gene targets of FTO in RPMI8226 cells with or without FTO overexpression (FTO-OE) or FTO-KD. In stable FTO-OE cells, 224 genes overlapped between the differentially expressed genes (DEGs) by RNA-seq and the differentially methylated mRNA by m⁶A-seq analysis, while 183 genes overlapped in cells with stable FTO-KD cells (Figure 4A). Gene ontology analysis revealed that the significantly upregulated overlapping genes in FTO-OE cells and differentially downregulated overlapping genes in FTO-KD cells were consistently enriched in cell motility, cell migration, cell adhesion, and cell activation, suggesting that FTO may play critical roles in MM metastasis (Figures S2A and S2B). Among these genes, heat shock factor 1 (HSF1) has been reported to be a potent metastasis-promoting gene in melanoma.^33,34 Notably, it has also been reported to be frequently upregulated in MM samples and in almost all EMM samples.35, 36, 37 Moreover, m⁶A-seq analysis showed that FTO-KD increased m⁶A enrichment in the coding sequence (CDS) region and 3′-untranslated region (3′-UTR) but had little effect on the 5′-untranslated region (5′-UTR) (Figure 4B). Moreover, FTO-KD altered the total peak distribution and unique peak distribution (Figure 4C). Consistent with the results of previous studies,³⁸ the 20,355 m⁶A peaks identified in this study were significantly enriched in the most common m⁶A motif (RRACH) (Figure 4D). Importantly, our m⁶A-seq analysis identified HSF1 as a direct m⁶A modification target. m⁶A peaks were detected in the CDS region and 3′-UTR of the HSF1 transcript, and the number of peaks was increased upon FTO-KD (Figure 4E). Therefore, we selected HSF1 as a candidate target of FTO-mediated m⁶A modification for the next investigation.

Figure 4.

m⁶A-seq and RNA-seq identified HSF1 as a potential target of FTO-mediated m⁶A modification

(A) The intersection of DEGs identified by RNA-seq and m⁶A methylation genes in cells with stable FTO-OE and FTO-KD. (B) The distribution of m⁶A peaks across the length of mRNAs. The 5′-UTR indicates the 5′-untranslated region, the 3′-UTR indicates the 3′-untranslated region, and the CDS indicates the coding sequence region. (C) The proportion of m⁶A peaks distributed in the indicated regions across the entire set of mRNAs and the appearance of new m⁶A peaks or the loss of existing m⁶A peaks after FTO-KD. (D) The most common m⁶A motifs identified by HOMER in RPMI8226 cells with or without FTO-KD. (E) The m⁶A levels on HSF1 mRNA transcripts in FTO-KD (shFTO) and control (shNC) cells.

HSF1 is a functionally important target gene of FTO in MM

To assess the biological function of HSF1 in MM, we first performed transcriptome array analysis and identified the DEGs in MM patients (Figure S2C). Consistent with the results of previous studies,³⁷ our results confirmed upregulation of the HSF1 transcript in MM (Figure 5A, P = 0.0447) and further upregulation in EMM patients (Figure 5B, P = 0.00026). To further clarify the role of HSF1 in MM, survival analysis was conducted to evaluate the value of HSF1 in the prognosis of MM. Kaplan-Meier survival analysis showed that the elevated HSF1 expression predicted a poor prognosis in MM patients (GSE9782)³⁹ (Figure S2D). In addition, compared with NDs, the levels of HSF1 and its target genes HSP70 and HSP90 were higher in MM and EMM samples and were also proven to be higher in MM cell lines, especially in PCL cell lines, than in NDs (Figures 5C–5E and S2E–S2G). Moreover, we evaluated the levels of FTO and HSF1 in PCs from NDs, MM patients, and EMM patients, followed by the Spearman correlation test. The results showed a positive correlation between FTO and HSF1 expression (P < 0.001) (Figure 5F). Furthermore, forced expression of FTO increased the protein level of HSF1 and its target HSPs (Figures 5G and S2H), while FTO-KD produced the opposite results (Figure 5G). These findings indicated that HSF1 was a crucial target gene of FTO in MM.

Figure 5.

HSF1 has been validated to be a functionally important target gene of FTO in MM

(A) Identification of DEGs between MM and ND samples by transcriptome sequencing. (B) Identification of DEGs between EMM and MM samples by transcriptome sequencing. (C) HSF1 expression was further validated in EMM patients (n = 20) by qPCR. (D) Western blot analysis of HSF1 and its target HSPs in CD138 + cells isolated from MM patients or EMM patients. N indicates NDs; M indicates MM patients; and E indicates EMM patients. (E) Immunohistochemical analysis of HSF1 and its target HSPs in biopsies of ND, MM, and EMM samples. (F) qPCR analysis of FTO and HSF1 expression in the same samples of ND, MM, and EMM patients. The correlation between the expression of FTO and HSF1 was determined by the Spearman correlation test. (G) Immunoblot analysis of FTO, HSF1, and its target HSPs in stable RPMI8226 cells with or without forced FTO-OE and FTO-KD. (H) The expression of FTO and HSF1 was determined by Western blot in RPMI8226 and MM1R cells with or without FTO-OE and/or HSF1 knockdown. (I) Cell proliferation assay in cells as in (H). (J) Cell migration assay in cells as in (H). (K) Cell invasion assay in cells as in (H). (L) Cell apoptosis analysis in cells as in (H). (M) The quantification of cell apoptosis for (L). ∗P<0.05; ∗∗P<0.01; ∗∗∗P<0.001.

Next, we investigated whether HSF1 mediates the function of FTO. Knockdown of HSF1 in FTO-OE cells resulted in decreased expression of HSF1 (Figure 5H) accompanied by significant decreases in cell proliferation (Figure 5I), migration (Figure 5J), and invasion (Figure 5K) and a marked increase in apoptosis (Figures 5L and 5M). These findings suggested that the effects of FTO-OE were partially rescued by knockdown of HSF1 and that HSF1 was a crucial downstream target gene of FTO, which mediated its effects in MM.

FTO enhances HSF1 mRNA stability through an m⁶A-YTHDF2-dependent pathway

Next, the mechanism by which FTO regulates HSF1 expression was investigated. Gene-specific m⁶A IP qPCR³⁸ was used to confirm that FTO influenced HSF1 through m⁶A modification. We found that forced expression and FTO-KD markedly, respectively, decreased and increased the m⁶A modification level of HSF1 mRNA (Figures 6A and 6B). Previous studies showed that the m⁶A reader protein YTHDF2 can target thousands of transcripts, whose half-lives increase upon YTHDF2 knockdown.¹⁹ Consistent with early reports, our study showed that knockdown of YTHDF2 with siRNA (siYTHDF2) obviously rescued HSF1 expression in FTO-KD cells (Figures 6C and 6D). Via RNA stability assays,⁴⁰ we assessed the effect of m⁶A modification on the stability of the HSF1 transcript. The results suggested that the half-life of the HSF1 transcript was prolonged and shortened in MM FTO-OE and FTO-KD cells, respectively (Figure 6E). Moreover, siYTHDF2 transfection increased HSF1 transcript stability in FTO-KD cells (Figure 6F). Consistent with the above results, siYTHDF2 transfection promoted the proliferation of shFTO cells (Figure S2I). Thus, our findings indicated that the m⁶A demethylase FTO upregulated HSF1 expression at least partially through a YTHDF2-dependent RNA decay effect in MM cells.

Figure 6.

FTO regulates HSF1 expression by inhibiting m⁶A/YTHDF2-mediated mRNA decay

(A and B) Gene-specific m⁶A qPCR analysis of m⁶A enrichment of HSF1 mRNA in RPMI8226 cells with GFP, GFP-FTO (A), shNC, or shFTO (B). (C and D) qPCR and Western blot analysis of the mRNA (C) and protein (D) levels of HSF1 in RPMI8226 and MM1R cells with or without FTO-KD, in combination with siRNA knockdown of the m⁶A reader YTHDF2. € qPCR analysis of the mRNA stability of HSF1 in RPMI8226 or MM1R cells with GFP, GFP-FTO, shNC or shFTO, siNC (FTO), or siFTO. (F) qPCR analysis of the mRNA stability of HSF1 in RPMI8226 cells and MM1R cells with or without FTO-KD in combination with or without siRNA knockdown of YTHDF2. ∗P<0.05; ∗∗P<0.01; ∗∗∗P<0.001.

The FTO inhibitor MA2 significantly inhibits MM formation and extramedullary spread in the NCG mouse model of disseminated myeloma

To further study the effect of FTO on the pathogenesis and extramedullary metastasis of MM in vivo, we used MM1R-V-luc cells to establish a xenograft model in NOD-Prkdc^em26Cd52il2rg^em26Cd22/Nju (NCG) mice via tail intravenous injection. Three days after xenotransplantation, mice were randomly divided into four groups and intraperitoneally injected with MA2 (20 mg/kg) or vehicle control combined with or without BTZ at the indicated time point. We monitored MM development and extramedullary lesion formation weekly by bioluminescence imaging. At early stages (at and before week 3 after treatment), compared with the control treatment, both MA2 and BTZ treatment alone obviously decreased the number of MM cells in the BM of the fore or hind limbs/joints, delayed extramedullary tumor formation (with tumors mainly consisting of extraintestinal tumors, abdominal tumors, and tumors in subcutaneous and other soft tissues, Figure S3A), and reduced the size of EMD regions, although these treatments produced limited effects at late stages (3 weeks after treatment) (Figures 7A and 7B). Notably, combination treatment with MA2 and BTZ almost completely eliminated extramedullary tumors and markedly reduced the sizes of intramedullary and extramedullary tumors at all stages (Figures 7A and 7B). The differences in MM cells loaded in BM sites and the extramedullary region were statistically confirmed by bioluminescence intensity analysis (P < 0.05) (Figure 7B). Treatment effectiveness was further confirmed by assessing the tumor burden (Figure 7C) and mouse survival (Figure 7D). Consistently, flow cytometric analysis showed that mice treated with MA2 or BTZ had fewer MM cell (CD138+)-formed tumors in fore or hind limbs' BM than control mice (P < 0.001) (Figures 7E and 7F) and that combination treatment with MA2 and BTZ further markedly suppressed myeloma bone tumor formation (P < 0.01) (Figures 7E and 7F), highlighting that MA2 in combination with BTZ has synergistic cytotoxic effects. Next, we examined the expression of FTO and HSF1 in fluorescence-positive MM cells. Downregulation of FTO and HSF1 expression was found in mice treated with MA2 alone and in combination with BTZ (Figures 7G–7I and S4G). Additionally, MA2-treated mice exhibited lower expression levels of the target HSPs of HSF1 than mice in other groups (Figures 7I and S4G). In contrast, the expression of these proteins in BTZ-treated mice was not obviously changed compared with that in control mice (Figures 7G–7I and S4G). The BM of MA2- and BTZ-treated mice also contained higher levels of Caspase 3 (a marker of apoptosis) and lower expression of CD34 (a marker of angiogenesis), VEGF (a marker of angiogenesis), Matrix metalloproteinase 9 (MMP9, a kind of metalloproteinases), and Collagen Ⅳ (a type of collagen related to tumor progression) (Figure S3B). We repeated the in vivo studies with another MM cell line (RPMI8226-Luc) and obtained similar results (Figures S4A–S4F). Taken together, these results indicated that MA2 significantly suppressed MM formation and metastasis in vivo and that combination therapy with MA2 and BTZ had a stronger inhibitory effect on the progression of both medullary and extramedullary disease in NCG mice than treatment with either agent alone.

Figure 7.

The FTO inhibitor MA2 significantly inhibits MM progression and enhances the sensitivity of BTZ to MM in vivo

(A) In vivo bioluminescent imaging showing tumor initiation, burden, and metastasis in NCG mice bearing MM1R-luc cells treated with DMSO (the control group), MA2, BTZ, or MA2 in combination with BTZ. W indicates weeks after treatment. (B) Quantification of bioluminescence intensity at BM and extramedullary disease of mice as in (A) after treatment. (C) Tumor burden was analyzed by ELISA of mouse plasma human λ chain concentrations. (D) Survival in mice bearing MM1R-luc tumors treated with vehicle control (CON), BTZ (1 mg/kg per mouse, at days 1, 4, 8, and 11), MA2 (20 mg/kg per mouse, days 1–10), and their combination (MA2+BTZ). (E and F) Flow cytometry analysis of CD138 + cell percentages. (G and H) qPCR and Western blot showing the mRNA (G) and protein (H) levels of FTO and HSF1 in different groups of mice as in (D). (I) Immunohistochemical analysis of FTO, HSF1, and its target HSPs in BM from different groups of mice as in (D). ∗P<0.05; ∗∗P<0.01; ∗∗∗P<0.001.

Discussion

Epigenetic modification plays an important role in the pathogenesis of various diseases.⁴¹ m⁶A RNA modification, as a new layer of epigenetic regulation, has recently been demonstrated to be involved in the tumorigenesis of various cancers.^13,23 However, the roles of m⁶A modification in MM pathogenesis are still unclear. Herein, we revealed a tendency toward decreased m⁶A methylation in MM and EMM patients and identified the FTO demethylase as the main factor for the abnormal m⁶A level in MM. Our data demonstrated the biological role of FTO-mediated m⁶A demethylation in MM for the first time and proposed the FTO-HSF1/HSP axis as a potential novel therapeutic target in MM, especially in EMM.

Accumulating evidence has shown that m⁶A modification exerts tumor-promoting or tumor-suppressing functions by regulating cancer-specific genes.⁴² In this study, we systematically analyzed the key regulators of m⁶A methylation in MM and confirmed that FTO was upregulated in primary PCs from MM patients, especially in EMM patients. This finding was similar to a recent study reporting that upregulated FTO promoted both metastasis and invasion in endometrial cancer.⁴³ Moreover, many additional studies have shown that FTO is upregulated and considered to act as a crucial oncogene in various cancers, including leukemia,²⁴ breast cancer,²⁷ and melanoma,²⁸ promoting cancer development and progression. Thus, it is rational to infer that FTO upregulation is potentially involved in MM tumorigenicity and metastasis. ALKBH5, another m⁶A demethylase, was previously reported to function as an oncogene in GBM⁴⁴ and breast cancer,⁴⁵ whereas it likely plays a tumor suppressor role in AML, owing to its frequent copy number loss.⁴⁶ Therefore, deregulation of m⁶A methylation may be heterogeneous in the context of different kinds of cancers. Herein, we found that upregulated FTO promoted the proliferation, migration, and invasion and inhibited the apoptosis of MM cells. Moreover, MA2 markedly suppressed the proliferation and migration of MM cells in vitro and inhibited tumor growth and extramedullary spread in therapeutic xenograft models, demonstrating a promising strategy based on m⁶A modification in the treatment of MM.

To further explore the molecular mechanism by which FTO promotes MM progression, we analyzed m⁶A-seq and RNA-seq profile data and identified HSF1 as a candidate target of FTO-mediated m⁶A demethylation. Furthermore, HSF1 was found to be significantly upregulated in MM samples and cell lines, especially in EMM tissues and PCL cell lines, and it was proven to be positively correlated with that of FTO. HSF1, as the transcriptional regulator of the mammalian HSR, is required for the expression of HSPs.^47,48 HSF1 is a powerful modifier of tumorigenesis, and its aberrant activation in many cancers is strongly associated with tumor metastasis and poor prognosis.⁴⁹ In a recent genomic study, HSF1 was one of the six screened potent metastasis-promoting genes in melanoma.³⁴ Moreover, the expression of HSF1 and its transcriptional target genes is associated with metastasis and poor outcomes in breast,⁵⁰ colon, and lung tumors.⁴⁹ In addition, Heimberger et al.³⁷ found that HSF1 was overexpressed in half of MM samples and in almost all EMM tissues. HSF1 inhibition was proven to effectively suppress the growth of MM cells.³⁶ Indeed, several inhibitors of HSF1 have been reported, but their low potency and lack of specificity limit their potential for clinical application.⁵¹ These inhibitors, while inhibiting HSF1, could also inhibit the transcription factor AP1 as well as the NF-κB, P53, and other signaling pathways.^52,53 To date, no specific HSF1 inhibitor has been developed.⁵³ Therefore, a better understanding of the HSF1 regulatory network is essential to achieve specific targeting of HSF1. Herein, we observed that HSF1 silencing functionally inhibited MM cell proliferation, migration, and invasion, similar to the effect of FTO-KD. Moreover, via gain- and loss-of-function studies, we found that FTO positively regulated HSF1 in MM but negatively regulated the m⁶A level of HSF1 mRNA. Silencing HSF1 partially antagonized the tumor-promoting effects mediated by FTO-OE. Therefore, we demonstrated that HSF1 was a functional downstream target of FTO and that FTO inhibition may provide a novel strategy for improved targeting of HSF1.

Although we demonstrated that FTO modulated HSF1 expression in a manner mediated by m⁶A demethylation in MM, Mauer et al.⁵⁴ reported that N6,2′-O-dimethyladenosine (m⁶A_m) was another substrate of FTO, which promoted the decay of mRNAs containing m⁶A_m. Interestingly, a recent review showed that FTO demethylated m⁶A_m rather than internal m⁶A, thus reducing the stability of m⁶A_m-containing transcripts.¹⁸ However, FTO-mediated m⁶A demethylation has been clearly clarified in AML,²⁴ breast cancer,²⁷ and melanoma,²⁸ and repressing FTO with R-2-hydroxyglutarate (R-2HG) increased m⁶A levels on MYC/CEBPA mRNAs and promoted mRNA decay.⁵⁵ Notably, more recent studies have indicated that m⁶A_m levels in many cell lines are extremely low, accounting for only 3%–5% of m⁶A modifications; thus, the effect of FTO on mRNA stability is exerted mainly through m⁶A, not m⁶A_m.^56,57 Herein, we showed that the levels of over 50% of potential target mRNAs of FTO were positively related to those of FTO in MM cells. FTO-KD decreased the mRNA and protein levels and stability of HSF1 and increased m⁶A enrichment on the HSF1 transcript in the CDS and 3′-UTR regions, but m⁶A_m is reported to often be located near the 5′-UTR.⁵⁷ Thus, our findings demonstrated that HSF1 was regulated by FTO-mediated m⁶A demethylation.

Mechanistically, m⁶A exerts its effect primarily by recruiting m⁶A-binding proteins.¹⁸ YTHDF2, a reader protein, displays a 10- to 50-fold higher affinity for methylated mRNAs than for nonmethylated mRNAs.^58,59 YTHDF2 selectively binds m⁶A sites and mediates the well-documented instability of m⁶A-containing mRNAs by recruiting the CCR4-NOT deadenylase complex.^19,60 In our study, FTO-KD reduced the stability of mRNA transcripts, suggesting that the m⁶A reader YTHDF2 might participate in this process. We further investigated whether YTHDF2 affected the HSF1 mRNA expression level via m⁶A regulation and found that knockdown of YTHDF2 increased the level and stability of HSF1 and promoted cell proliferation, reversing the effect of FTO-KD. These results suggested that m⁶A enrichment led to HSF1 mRNA decay through a YTHDF2-dependent mechanism in MM. Li et al.²⁴ showed that FTO depletion can lead to an increase in the level of the methyltransferase METTL3, and Yang et al.²⁸ showed that METTL3/METTL14 knockdown reversed the effect of FTO-KD. However, whether FTO-regulated demethylation affects m⁶A methyltransferase activity in MM needs to be further assessed.

In addition, a previous study showed that activation of the HSR limits the efficacy of several current myeloma drugs, including BTZ.³⁶ HSR, including upregulation of HSPs, can be activated in BTZ-treated myeloma cells, and this activation causes the development of resistance to chemotherapies.^36,61 The results of additional studies implied that proteasome inhibitor treatment-induced upregulation of HSPs was critically dependent on HSF1.³⁷ Thus, targeted inhibition of FTO can not only suppress MM formation and progression but also increase sensitivity to chemotherapeutic drugs by inhibiting the HSF1/HSP axis, which seems to be a promising therapeutic strategy for MM patients, especially refractory and relapsed MM patients. MA2, an FTO selective inhibitor, was reported to inhibit glioblastoma stem cell growth and self-renewal,²⁵ and Zhou et al.⁶² suggested that MA2 may increase the sensitivity of cervical cancer to chemoradiotherapy. In our work, MA2 promoted the anti-MM efficiency of BTZ in vitro, and combination therapy with MA2 and BTZ significantly delayed the occurrence of MM and almost completely eliminated extramedullary metastasis in the NCG mouse model through targeting the FTO/m⁶A/HSF1/HSPs axis. Notably, no substantial effect of MA2 on the growth of normal cells was detected.²⁵ Therefore, FTO is a promising therapeutic target to suppress the growth and migration/invasion of MM cells, and treatment with specific FTO inhibitors, especially in combination with BTZ, is considered a preferred therapeutic strategy for MM patients, especially those who are at high risk of extramedullary relapse. Interestingly, our study found the FTO expression in MM cells was partially weakened by MA2 (Figure 7I), while a previous study suggested that MA2 could not decrease FTO expression.⁶³ MA2 has been reported to elevate the levels of cellular m⁶A in the mRNA of human cells.^25,32 Recent studies have indicated that, in addition to the roles of m⁶A modifications in mRNAs, m⁶A methylation can also regulate the generation and function of noncoding RNAs, such as miRNAs and lncRNAs.64, 65, 66 Furthermore, noncoding RNAs can regulate m⁶A modifications.^64,67,68 Several recent studies have shown that noncoding RNAs can influence carcinogenesis by regulating FTO expression in various of tumors. miR-96 elevates the expression of FTO through regulation of the AMPKα2/FTO/m⁶A/myc axis in colorectal cancer.⁶⁹ miR-149-3p modulates the adipogenic differentiation of BMSCs by directly targeting FTO.⁷⁰ miR-1266 promotes colorectal cancer progression by targeting FTO.⁷¹ Therefore, we speculated that MA2 might lead to the elevated expression of noncoding RNAs, which can indirectly decrease the expression of FTO. Further investigation will be required to better elucidate the specific mechanism underlying this result.

In summary, our study demonstrated an FTO-dependent m⁶A demethylation module that regulates HSF1 expression and has functional implications in MM progression and metastasis, acting at least partially through YTHDF2-mediated mRNA decay. Therefore, given the functional importance of FTO/HSF1 in the formation and drug response of MM, targeting the FTO-m⁶A-HSF1 axis with selective inhibitors may be a promising therapeutic strategy for MM patients, especially for those with extramedullary disease (Figure 8). As FTO has also been reported to play an oncogenic role in other cancers, our findings in MM may have great significance for cancer pathogenesis and therapy.

Figure 8.

The m⁶A demethylase FTO promotes myeloma progression by regulating HSF1 in an m⁶A-YTHDF2-dependent manner

Elevated FTO induces mRNA demethylation, leading to a decrease in HSF1 m⁶A modification levels in MM and EMM patients. FTO promotes the proliferation, migration, and invasion and inhibits the apoptosis of MM cells. HSF1 is a functional target of FTO-mediated m⁶A demethylation. FTO attenuated HSF1 degradation in a YTHDF2-dependent manner. The FTO/HSF1/HSP axis is considered a potential novel therapeutic target in MM, especially in extramedullary disease.

Materials and methods

Patient samples

Primary PCs obtained from 15 healthy donors and 35 newly diagnosed MM patients were purified from BM aspirates using CD138 magnetic beads (Miltenyi Biotec, Auburn, CA, USA), according to the manufacturer's instructions. The purity of the PCs was assessed by flow cytometric analysis and was found to be >95%. In addition, primary PCs from 20 EMM patients were isolated by tissue homogenization of 20 extramedullary soft tissue masses resulting from hematogenous spread to the skin, brain, oral mucosa, and liver. The diagnoses of all EMM patients were confirmed from the pathological sections. The study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review board of Huazhong University of Science and Technology.

Cell culture, MM cells transfected with siRNA, or lentivirus

Synthetic FTO, HSF1 or YTHDF2 siRNA, and siRNA negative control were purchased from RiboBio (Guangzhou, China). The sequences of siRNAs are shown in Table S1. Lentivirus particles were produced using a third-generation packaging system by GeneChem Company (Shanghai, China). Additional details are provided in the Supplementary Materials and Methods.

Microarray assays of the MM patients and EMM patients

Starting with the previously determined gene expression profiles of 3 MM patients assessed by our teams, we added several samples to expand upon the analysis. The global protein-coding transcript profiles of a total of 4 MM patients and 4 EMM patients were analyzed using Human LncRNA Microarray v. 4.0. One microgram of labeled RNA was used for each sample. Quantile normalization and subsequent data processing were performed with the GeneSpring GX v 12.1 software package (Agilent Technologies). After quantile normalization of the raw data, significantly differentially expressed mRNAs between groups were identified by p value and fold change filtering.

Quantitative real-time PCR (qPCR), Western blot, cell proliferation and viability assays, flow cytometric analysis, and migration and invasion assays

The primers for target genes are shown in Table S2. The FTO inhibitor MA2³² was kindly provided by Professor Yang (Shanghai Institute of Materia Medica, Chinese Academy of Sciences). Additional details are provided in the Supplementary Materials and Methods.

Colorimetric analysis (RNA m⁶A quantification)

Total RNA (200 ng) was isolated from tissue samples and cells using TRIzol. The m⁶A level in the total RNA was measured with an EpiQuik m⁶A RNA Methylation Quantification Kit (Epigentek, Farmingdale, NY, USA), following the manufacturer's instructions. The m⁶A levels were evaluated colorimetrically by measuring the absorbance of each sample at a wavelength of 450 nm and performing calculations on the basis of the standard curve.

MeRIP-seq and RNA-seq, gene-specific m⁶A qPCR

Methylated RNA immunoprecipitation sequencing (MeRIP-seq) was performed according to a previously reported protocol⁷² with some changes. The detailed experimental protocols are provided in the Supplementary Materials and Methods.

mRNA stability assay

MM stable FTO-OE and FTO-KD cells were harvested at 6, 3, and 0 h after treatment with actinomycin D (5 μg/mL). Total RNA was isolated with a TRIzol kit (Takara). After reverse transcription, mRNA levels of target transcripts were measured by qPCR. mRNA half-lives were evaluated via linear regression analysis.

Cells and mice used in the animal model

The human MM cell lines RPMI8226 and MM.1R were stably transfected with a luciferase (Luc) vector (GeneChem, Shanghai, China) for in vivo monitoring of tumor growth. Female NCG mice were purchased from the Nanjing Biomedical Research Institute of Nanjing University. Six- to eight-week-old female mice (6–8 mice per group) were obtained at the beginning of the animal experiment. Mice were housed in a pathogen-free environment, and the experimental procedure and protocols were approved by the Committee on Animal Handling of Huazhong University of Science and Technology.

Animal model and treatment

A total of 3×10⁶ RPMI8226/MM1R-Luc cells were intravenously injected into NCG mice to establish a disseminated human MM xenograft model. The in vivo antitumor effect of the FTO inhibitor MA2 combined with or without the first-line chemotherapeutic agent BTZ was evaluated as follows: 3 days post xenotransplantation, MA2 (20 mg/kg), or vehicle control was injected intraperitoneally (i.p.) daily for 10 days, and BTZ was injected intraperitoneally on days 1, 4, 8, and 11. Mouse serum was collected at specified time points during the treatment, and the tumor burden was monitored by detecting myeloma cell-secreted λ light chains via a Human Lambda ELISA Kit (Bethyl Laboratories, No. E88-116). Tumor development was monitored weekly after treatment with an in vivo imaging system (IVIS, SI Imaging, Lago, and LagoX). Luciferin (150 mg/kg, YEASEN, Shanghai, China) was injected intraperitoneally into the mice. After 10–15 min, the tumor photon flux was measured using an IVIS imaging system with an exposure time of 30–60 s. Mice were observed daily and sacrificed when hindlimb paralysis was detected. Fluorescent signal-positive primary tumor (fore or hindlimb BM) and metastasis samples, including spleen (data not shown), extraintestinal tumor, subcutaneous tissue, abdominal tumor, and other soft tissue, were surgically removed for subsequent analysis. Flow cytometry was used to analyze the population of CD138 + cells in fluorescence-positive tissues, whereas H&E staining and immunohistochemical staining were used to assess the expression of FTO and the proteins encoded by its target genes.

Immunohistochemistry

Paraffin-embedded sections from biopsy of MM and EMM samples were deparaffinized and blocked, and BM from each group of xenograft mice was fixed in formalin. Additionally, these sections were incubated with antibodies directed against FTO, HSF1, HSP70, HSP90, Caspase3, CD34, VEGF, MMP9, and Collagen Ⅳ at 4°C overnight. Then, the sections were treated with polymer antibodies, developed with DAB-Chromogen, and counterstained with hematoxylin.

Availability of data and materials

Microarray data are available in the GEO database under accession numbers GSE146757 and GSE147841. All other data supporting the findings of this study are available from the corresponding author upon reasonable request.

Ethics approval and informed consent statement

All experiments related to patient samples were approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (No. 2020(S091)). All animal procedures were approved by the Committee on Animal Handling of Huazhong University of Science and Technology (No. 2019s-854).

Statistical analysis

All results are presented as the means ± SDs. Differences in FTO and HSF1 levels between independent groups of samples were evaluated by using the Mann-Whitney U test. Correlations between the expression levels of FTO and HSF1 in primary PCs were analyzed by the Spearman correlation test. Comparisons between other groups were analyzed using unpaired Student's t tests. All statistical analyses were carried out with GraphPad Prism 7, and P < 0.05 was considered statistically significant.