METTL1抑制剂STM9005在实体瘤和血液肿瘤临床前模型中显示抗肿瘤活性
Pharmacological inhibition of METTL1 as a therapy against solid and blood cancers
研究团队开发出首个METTL1 tRNA甲基转移酶抑制剂STM9005,在黑色素瘤、胰腺癌、肺癌和AML等多种小鼠肿瘤模型中观察到肿瘤生长受抑、AML模型生存期延长。该化合物在细胞中降低tRNA的m7G修饰,对METTL1的IC50低于13 nM,在49种甲基转移酶中显示出超过10,000倍选择性。作者指出实验中未观察到肿瘤消退,可能与给药时间较短有关。
Main
Over 150 RNA modifications have been described to date, with transfer RNA (tRNA) being one of the most highly modified RNA species5,6. tRNA modifications are critical for the folding, stability and function of tRNA2,7,8. One of the most prevalent tRNA modifications is m7G1,2. It is deposited at position 46 of tRNA (m7G46) by the METTL1–WDR4 heterodimeric protein complex, with METTL1 functioning as the catalytically active component of the complex, and WDR4 having a scaffolding and RNA-binding role2,3,4. The m7G modification on tRNA has been shown to have an important role in stabilizing the tRNA structure, regulating the translation of mRNA and influencing a number of cellular processes, including stress response and oncogenesis9,10,11,12,13. Recent studies have shown that METTL1 may also deposit the m7G modification on mRNAs and microRNA precursors14,15,16, although at a substantially lower rate compared with the m7G levels on the tRNA fraction.
Many studies have implicated the role of the epitranscriptome in the development and regulation of cancer17,18,19. An increasing amount of evidence shows that dysregulation of tRNA modifications and their pathways contributes to the pathogenesis of human disease, including cancer6,18. Notably, upregulation of METTL1 and its catalytic activity has been linked to oncogenic transformation as well as the initiation and maintenance of aggressive haematological and solid malignancies, including acute myeloid leukaemia (AML), melanoma, glioblastoma, liver and lung cancers11,20,21,22. Moreover, the non-catalytic function of METTL1 has recently been linked to the development of sarcomas through its interaction with the multi-tRNA synthetase complex23.
To date, epitranscriptomic drug discovery efforts have largely focused on the targeting of the m6A mRNA modification and its regulators24,25,26,27. Early prototype inhibitors of METTL1 have recently been described and support the feasibility of pharmacologically inhibiting METTL128,29. However, their limited potency and selectivity render them unsuitable as tools to assess the molecular and therapeutic impact of targeting METTL1 catalytic inhibition. Here we present the development and characterization of STM9005, a potent and highly selective first-in-class tRNA methyltransferase inhibitor of METTL1 that demonstrates significant anti-tumour activity across diverse cancer types through the reduction in m7G on tRNA substrates of METTL1.
Characterization of STM9005
We performed a high-throughput screen of 250,000 diverse drug-like compounds. After initial hit triaging, STM4529 was identified as a singleton non-S-adenosyl methionine (non-SAM)-related hit that inhibited METTL1 with a half-maximum inhibitory concentration (IC50) of around 15 μM (Extended Data Fig. 1a,b). Surface plasmon resonance (SPR) analysis showed that STM4529 binding to METTL1 was competitive with the SAM analogue S-adenosyl homocysteine (SAH), indicating its binding to the catalytic site of METTL1 (Extended Data Fig. 1c). After extensive chemical optimization of potency, selectivity and in vitro and in vivo absorption–distribution–metabolism–excretion properties, we identified STM9005 (Fig. 1a and Supplementary Note 1a).
a, The chemical structure of STM9005. b, Representative biochemical activity assay showing inhibition of the METTL1–WDR4 enzyme complex using a concentration range of STM9005 and the less active analogue STM9001. c, Representative METTL1 cellular target engagement of STM9005 and STM9001 was assessed using the InCELL Pulse assay. For b and c, the geometric mean IC50 and s.d. indicated in brackets were derived from n = 13 (STM9005) and n = 3 (STM9001) independent experiments. d, STM9005 selectively inhibits METTL1–WDR4 in a methyltransferase profiling panel of 49 methyltransferases. The asterisk indicates less than 1% activity remaining. Data are mean ± s.d. n = 2 independent samples. e, The reduction in m7G modification on small-fraction RNA after treatment with vehicle, 0.6 or 1.2 µM of STM9005 was measured using RNA MS in the MOLM-13 and OCI-AML2 cell lines. Ratio of m7G to A, C, G and U nucleosides. Data are mean ± s.d. n = 3 independent samples. P values were calculated using two-way analysis of variance (ANOVA) with multiple-comparison correction. f, Crystal structure of METTL1 (carbon atoms in green) in a complex with STM9005 (carbon atoms are shown in purple). The interactions between the compound and the protein or neighbouring water molecules (red spheres) are indicated (yellow dashed lines). PDB: 9RYL. g, The pharmacokinetic (PK) profile of STM9005 in vivo was measured in mouse plasma after a single i.p. dose of either 10 or 30 mg per kg. Data are mean ± s.d. n = 2 (12 h time point for 10 mg per kg cohort) and n = 3 (all other groups) mice. Compound exposure over time (area under the curve, AUC), maximal concentrations (Cmax) and half-life (t1/2) are indicated in the pharmacokinetics summary table. Pharmacokinetics values after oral dosing are shown for comparison. p.o., oral gavage.
STM9005 inhibited recombinant METTL1 enzyme activity with an IC50 < 13 nM, a value restricted by the lower detection limit of the assay (Fig. 1b). To use a more sensitive assay system and to verify binding to METTL1 in an orthogonal system, we performed SPR analysis, which demonstrated direct stoichiometric high-affinity binding of STM9005 to METTL1 with a Kd of 2 nM (Extended Data Fig. 1d). Moreover, using a cellular target engagement assay, we confirmed cell penetration and binding of STM9005 to METTL1 in cells with an IC50 of 13 nM (Fig. 1c). STM9005 was found to be highly specific to METTL1 as it had no effect on the activity of other RNA-methyltransferases (METTL3/14, METTL16, BCDIN3D, TSG1 and NSUN2) when used at comparable concentrations (Extended Data Fig. 1e), and exhibited greater than 10,000-fold selectivity for METTL1–WDR4 when tested against a broad panel of 49 RNA-, DNA- and protein-methyltransferases (Fig. 1d and Supplementary Table 1). Moreover, no significant off-target toxicity liabilities were identified when STM9005 was tested in a broad panel of 97 kinases (KINOMEscan, Eurofins DiscoverX; Supplementary Table 2) and over 70 assays testing a range of receptors and enzymes commonly associated with drug toxicity (SAFETYscan 47, Eurofins DiscoverX; Supplementary Table 3).
To further investigate the cellular activity of STM9005, we applied RNA mass spectrometry (MS) to assess the specific effect of STM9005 on the m7G modification level of tRNA from the previously known METTL1-dependent AML cell lines MOLM-13 and OCI-AML211,30,31. We observed a significant decrease in m7G levels (Fig. 1e), but not in the levels of the m1A and m5C RNA modifications (Extended Data Fig. 1f). To further quantify the cellular effect of STM9005 on m7G-modified tRNA, we developed an antibody-based m7G detection assay (Meso Scale Discovery (MSD)), which enabled us to measure the effect of STM9005 on tRNA-Val. This demonstrated a concentration-dependent decrease in tRNA-Val-m7G after treatment with STM9005 with an IC50 of 34 nM (Extended Data Fig. 1g), confirming potent inhibition of tRNA modification by STM9005 in cells at nanomolar concentrations. Moreover, we confirmed STM9005 acts specifically on the METTL1–m7G tRNA axis, as we observed no significant effect on m7G modification of 18S ribosomal RNA (Extended Data Fig. 1h).
To characterize the binding of STM9005 to METTL1 further, we performed X-ray crystallography, confirming unequivocally that STM9005 binds to the SAM-binding site of METTL1 (Fig. 1f, Extended Data Fig. 1i, Extended Data Table 1 and Supplementary Note 1b). Notably, part of STM9005 binds in a pocket not previously observed in METTL1 crystal structures, where it forms molecular interactions such as the hydrogen bonds between Tyr184 and the central carbonyl moiety of STM9005, and a pi-carbon interaction between Phe159 and the tetrahydroisoquinoline methyl of STM9005, providing the basis for its high affinity and selectivity. In strong support of this, STM9001, a diastereomer of STM9005 in which the tetrahydroisoquinoline methyl is in an opposite orientation, is around 100-fold less active (Fig. 1b,c and Extended Data Fig. 1a,g), providing further evidence of the specificity and binding mode of STM9005 (Extended Data Fig. 1i and Supplementary Table 1).
The observed binding of STM9005 to the catalytic pocket of METTL1 did not impact the RNA or protein levels of METTL1 after in vitro treatment (Extended Data Fig. 1j,k), in contrast to the significant and selective effect observed on the m7G tRNA modification levels (Fig. 1e and Extended Data Fig. 1f,g). Finally, pharmacokinetic profiling of STM9005 in mice after a single intraperitoneal (i.p.) dose of either 10 or 30 mg per kg showed good bioavailability with maximal plasma exposures of 13 and 56 μM, respectively, and a half-life of around 1 h. Importantly, at the 12 h time point, STM9005 plasma levels remain above the METTL1 cellular IC50, indicating METTL1 inhibition in vivo over a period of at least 12 h (Fig. 1g).
Taken together, these data establish STM9005 as a potent, selective and bioavailable catalytic inhibitor of METTL1 that is suitable for both in vitro and in vivo investigations.
Cellular effects of STM9005
To investigate the anti-cancer potential of STM9005 in cellular models, we initially screened the impact of STM9005 treatment in three different cancer cell panels using colorimetric assays which measure metabolic activity (MTS assay). For these experiments, we used cancer cell models that were either previously reported to be sensitive to the genetic inhibition of METTL111,32,33, or their METTL1-dependency status (sensitivity or resistance) was validated here using CRISPR-based genetic disruption of METTL1/Mettl1 (human and mouse, respectively) in competitive co-culture assays (Extended Data Fig. 2a,b). We observed a dose-dependent reduction in the proliferation of human leukaemic cell lines (Fig. 2a), primary mouse AML models (Fig. 2b) and human and mouse solid cancer models (Fig. 2c) from different tissue origins (skin, lung, pancreas, colorectal, oesophageal and soft tissue). Investigating the direct effect of STM9005 on cell growth using a methylene blue assay, we observed highly potent, dose-dependent effects on the fitness and growth of a subset of these cell lines (Fig. 2d and Extended Data Fig. 3a). Additional reprofiling in proliferation assays using ATP quantitation under optimized culture conditions confirmed the anti-proliferative activity of STM9005 at submicromolar concentrations in these cells (Fig. 2e and Extended Data Fig. 3b). By contrast, the diastereomer control compound STM9001 was significantly less potent in these systems, consistent with its reduced activity against the catalytic activity of METTL1, providing strong evidence that the anti-proliferative effects of STM9005 are due to METTL1 catalytic inhibition (Fig. 2e and Extended Data Fig. 3b). There was also a significant dose-dependent negative effect seen in the colony forming potential of MLL-AF9/Flt3ITD/+-driven primary mouse AML cells after ex vivo treatment with STM9005 (Extended Data Fig. 3c). Notably, the cancer models that appeared to be independent of METTL1 genetic loss (Extended Data Fig. 2a,b) were also not significantly impacted after treatment with STM9005 (Fig. 2d). To confirm that the anti-proliferative effects of STM9005 are mediated specifically through METTL1, we assessed its impact in two insensitive (OE33, mPDAC2) and two sensitive (B16F10, mPDAC1) cancer cell lines with a loss of METTL1 or Mettl1 after CRISPR–Cas9 targeting. STM9005 treatment in both sensitive and insensitive cell lines resulted in no additional negative effect (Extended Data Fig. 3d,e) despite the absence of METTL1 protein (Extended Data Figs. 2a,b and 3a), consistent with the small molecule exerting its mechanism of action specifically through inhibition of METTL1’s catalytic activity.
a–c, STM9005 concentration–response curves measuring cell viability (MTS assay) in panels of human haematological cancer (a), primary mouse leukaemia (b) and human and mouse solid cancer cells (c) after 6 days of treatment. Data are mean ± s.e.m. n = 3 independent cultures. IC50 values are shown in parentheses. d, The normalized confluency was measured by quantification of methylene blue staining of sensitive (top) and insensitive (bottom) cells after 9 days of treatment with vehicle or STM9005 at the indicated concentrations. Data are mean ± s.d. n = 4 independent cultures. P values were calculated using one-way ANOVA. Representative images for each well are shown under the corresponding bars. OE, overexpression. e, Representative STM9005 and STM9001 concentration–response curves (CellTiter-Glo) in LoVo and OE21 cells under optimized culture conditions after 7 days of treatment. Geometric mean IC50 ± s.d. (parentheses) were derived from n = 25 (LoVo, STM9005) and n = 8 (OE21, STM9005) and n = 2 (STM9001) independent experiments. f,g, Myeloid differentiation of MOLM-13 (f) or primary mouse MLL-AF9/Flt3ITD/+ AML (g) cells after 6 days of treatment with vehicle or 1, 2.5 or 5 µM STM9005. Left, representative histogram of CD11b–APC (f) or Gr1–APC (g) staining. Data are mean ± s.d. n = 3 independent cultures. P values were calculated using one-way ANOVA. Right, the mean fluorescence intensity (MFI) of CD11b–APC-stained (f) and Gr1–APC-stained (g) cells. h, Cell cycle analysis of A549 (top) or LoVo (bottom) cells after 7 days of treatment with vehicle or 1 or 2 µM STM9005. n = 3 independent cultures. i, Breakdown of h, shown as the proportion in each cell cycle stage of A549 (top) or LoVo (bottom) cells. Data are mean ± s.d. n = 3 independent cultures. P values were calculated using two-way ANOVA with multiple-comparison correction. j,k, Western blot analysis of a cell cycle protein panel in A549 (j) and LoVo (k) cells after 5 day treatment with vehicle or 0.3 or 1 µM STM9005. n = 3 (A549) and n = 2 (LoVo) independent cultures.
In the context of haematological malignancies, pharmacological inhibition of METTL1 using STM9005 caused significant and dose-dependent myeloid differentiation in both human and primary mouse models of AML (Fig. 2f,g and Extended Data Fig. 3f). Moreover, we and others have previously shown that genetic inhibition of METTL1 negatively affected various cancer models by impacting their cell cycle status11,20. To examine whether the observed effect on cell cycle is dependent on the catalytic activity of METTL1 we analysed a panel of solid cancer cell lines with different tissues of origin and mutational background and showed that treatment with STM9005 causes a significant disruption of cell cycle, with a dose-dependent increase in cells in G0/G1 phase and a significantly decreased number of cells in S phase (Fig. 2h,i and Extended Data Fig. 3g,h) as the most common changes across the cell lines tested. Moreover, western blot analysis of cell-cycle-related markers confirmed the observed cell cycle dysregulation seen in the various cancer models tested here (Fig. 2j,k). Moreover, annexin V analysis showed no significant levels of apoptosis after treatment with STM9005 (Extended Data Fig. 3i).
These data confirm the importance of METTL1’s methyltransferase activity in maintaining both haematological and solid cancers, and that pharmacological inhibition of its enzymatic activity is sufficient to elicit anti-cancer cellular effects.
Molecular effects of STM9005 treatment
To assess the molecular changes caused by treatment with STM9005, we analysed its impact across different layers of the transcriptome. We first performed modification-induced misincorporation tRNA sequencing (mim-tRNA-seq) to investigate the m7G modification status on tRNA fractions from B16F10 melanoma tumours and a primary mouse MLL-AF9/Flt3ITD/+-driven AML. Analysis of the mim-tRNA-seq data revealed a highly specific and significant reduction in the m7G46 mark in known m7G-modified tRNAs—a result that was highly similar between both cancer models (Fig. 3a and Supplementary Table 4).
a,b, mim-tRNA-seq analysis of the mean (red) and individual replicate (grey) log2-transformed fold change (log2[FC]) in misincorporation at the m7G46 site of tRNAs (a) and the log2[FC] (grey) in tRNA levels (b) in B16F10 (left) and MLL-AF9/Flt3ITD/+ (right) cells after 72 h treatment with vehicle or 2 µM STM9005. The purple asterisks represent a significant change (a; P < 0.01, unpaired two-tailed Student’s t-tests) or a significant decrease (b, Padj < 0.05, Wald test). Related to Supplementary Tables 4 and 5. n = 3 independent samples. c, Northern blot analysis of m7G-modified tRNAs and non-m7G-modified RNA controls for MLL-AF9/Flt3ITD/+ (left) and A549 (right) cells after 72 h treatment with vehicle or 2 µM STM9005. n = 2 independent experiments. d, Western blot analysis of CDK4, KDM1A and β-actin after treatment for 7 days with vehicle or 1, 2.5 or 5 µM STM9005 in a panel of cancer models. e,f, Rescue of phenotype in B16F10 melanoma cells after overexpression (OE) of Arg-TCT-tRNA (e) and Cdk4 (f) versus an empty vector control. The normalized confluency was measured by quantification of methylene blue staining of B16F10 cells with Arg-TCT-tRNA overexpression (e), Cdk4-overexpression (f) or an empty vector control, after treatment for 6 days with vehicle or 1.25 or 2.5 µM STM9005. Data are mean ± s.d. n = 3 independent cultures. P values were calculated using two-way ANOVA with multiple-comparison correction. Representative images are shown under the corresponding bar. g,h, Rescue of phenotype in A549 (g) or mPDAC1 (h) cells after overexpression of WT METTL1 (A549) and WT Mettl1 (mPDAC1) versus an empty control vector and METTL1MUT (A549) and Mettl1MUT (mPDAC1) overexpression. The normalized confluency was measured by quantifying methylene blue staining of cells overexpressing WT or mutant METTL1 or an empty control vector that were treated with vehicle or 313.5 nM or 625 nM STM9005, after 6 days of treatment. Data are mean ± s.d. n = 3 independent cultures. P values were calculated using two-way ANOVA with multiple-comparison correction. Representative images are shown under the corresponding bar.
Subsequent investigation of differential tRNA expression from the same two cancer models showed selective downregulation of a subset of m7G-modified tRNAs after pharmacological inhibition of METTL1 (Fig. 3b and Supplementary Table 5), which was validated using northern blotting across a panel of m7G-modified tRNA substrates (Fig. 3c and Extended Data Fig. 4a).
We previously showed that the catalytic function of METTL1 drives the translation of key oncogenic proteins, including CDK4 and KDM1A, by modifying and stabilizing a subset of tRNAs11. To explore the dependence of these outcomes on METTL1’s catalytic activity, we performed western blotting using a panel of cancer models from diverse tissues, confirming that CDK4 and KDM1A protein levels were reduced after STM9005 treatment in a dose-dependent manner (Fig. 3d and Extended Data Fig. 4b), while the same was not seen with corresponding mRNA levels (Extended Data Fig. 4c,d). To verify that STM9005’s mechanism of action is mediated by inhibition of the m7G–tRNA axis, we performed rescue-of-phenotype experiments by overexpressing either the established METTL1–m7G tRNA substrate Arg-TCT or the METTL1 biomarker CDK4, a known substrate of Arg-TCT11. Notably, overexpression of Arg-TCT tRNA in both a solid cancer and a leukaemia model (Fig. 3e and Extended Data Fig. 4e–g) led to a partial but significant response to STM9005 and an increase in the viability of the cancer cells. Similarly, overexpression of Cdk4 in both a solid cancer and a leukaemia model (Fig. 3f and Extended Data Fig. 4h–j) mirrored the partial rescue of phenotype seen when using Arg-TCT tRNA overexpression, supporting the conclusion that the cellular effects caused by STM9005 treatment are dependent on the m7G–tRNA molecular pathway, which is maintained by the catalytic activity of METTL1.
To further validate that STM9005 acts specifically by inhibiting the catalytic function of METTL1 we overexpressed wild-type (WT) METTL1 and mutant (MUT) METTL1 carrying two amino acid changes in the catalytic domain (L160A/D164A for human METTL1 and L154A/D157A for mouse Mettl1). Forced expression of WT METTL1/Mettl1 significantly increased cellular resistance to STM9005, whereas viability did not improve in cells overexpressing METTL1MUT/Mettl1MUT (Fig. 3g,h and Extended Data Fig. 4k,l), confirming that STM9005 exerts its effects through the inhibition of the catalytic activity of METTL1. Moreover, the reduction in the levels of known m7G-dependent protein substrates, such as CDK4 and KDM1A, is reversed when WT METTL1 is overexpressed, while the same is not seen with overexpression of METTL1MUT (Extended Data Fig. 4m). This further accentuates that STM9005’s activity is specifically driven by the catalytic inhibition of METTL1.
To assess the wider downstream effects of STM9005 treatment on global translation, we used ribosome sequencing (Ribo-seq). After treatment with STM9005, Ribo-seq and mRNA sequencing (mRNA-seq) changes were moderately correlated (r = 0.6368; Extended Data Fig. 5a and Supplementary Table 6), suggesting that pharmacological inhibition of METTL1 affects gene expression at both the transcription and translational levels. Ribosome-protected RNA fragment (RPF) density analysis showed that, after treatment with STM9005, there was a significant increase in A-site ribosome occupancy at codons associated with m7G-modified tRNAs compared with those associated with non-m7G modified tRNAs (Extended Data Fig. 5b,c and Supplementary Table 7), suggesting ribosome stalling at sites that are dependent on m7G-modified tRNAs for decoding. Importantly, treatment did not reduce the protein levels of METTL1 (Extended Data Fig. 1k), METTL3 and WDR4 (Extended Data Fig. 5d) at any of the concentrations tested, confirming that the effects seen can be attributed to the catalytic inhibition of METTL1, rather than a global translation defect. To further validate our Ribo-seq findings, we compared genes downregulated in our Ribo-seq dataset with a published SILAC-based proteomic dataset from METTL1 knockdown and METTL1(WT)–WDR4 overexpression studies11. Genes downregulated after STM9005 treatment showed significant overlap with those reduced after short-hairpin-RNA-mediated depletion of METTL1 (shMETTL1; P < 1 × 10−5; Extended Data Fig. 5e and Supplementary Table 8) and were likewise enriched among genes upregulated by METTL1(WT)–WDR4 overexpression (P < 1 × 10−5; Extended Data Fig. 5f and Supplementary Table 8), further confirming on-target, catalysis-dependent effects on the METTL1-regulated translatome. This demonstrates that catalytic inhibition of METTL1 using STM9005 selectively alters the m7G-modified tRNA pool and therefore impacts translation, in concordance with our previous findings using genetic downregulation of METTL1 levels11.
To assess the molecular changes caused by treatment with STM9005 on the mRNA transcriptome, we performed mRNA-seq and differential expression analysis using three cancer models (OE21, LoVo and B16F10) treated with either vehicle or STM9005 (Supplementary Tables 9–11). Gene Ontology analysis of differentially expressed genes showed that treatment with STM9005 dysregulates several key oncogenic pathways, including pathways such as those regulated by MYC, PI3K–AKT–mTOR, p53 and, as shown earlier, the cell cycle (Extended Data Fig. 6a and Supplementary Tables 12–14).
Previous studies have associated METTL1 with the transcriptional regulation of MYC and its pathways12,34. However, there is a lack of evidence directly linking the catalytic function of METTL1 to the regulation of MYC. To directly address this, we investigated whether the observed changes in expression of MYC, caused by isolated catalytic inhibition of METTL1, were evident at the MYC protein level. Indeed, western blot analysis confirmed that treatment with STM9005 leads to a dose-dependent decrease in MYC protein levels in two independent cancer models (Extended Data Fig. 6b), while the same reduction was not seen at the mRNA level (Extended Data Fig. 6c), showing that METTL1 regulates MYC protein levels through its catalytic function. To explore the potential of STM9005 treatment impacting cancer growth through the downstream regulation of MYC levels, we performed rescue-of-phenotype experiments by overexpressing MYC in a human and a mouse solid cancer model. Our experiments showed that MYC overexpression increased viability after treatment with STM9005 compared with the empty vector control (Extended Data Fig. 6d–g), indicating that the negative impact of STM9005 on cancer cells could be partially connected to negative regulation of MYC and its pathways.
Finally, we confirmed that, despite the transcript and protein level dysregulation of several cancer-associated pathways after treatment with STM9005, transcript and protein levels of METTL1 remained unaffected (Extended Data Fig. 6h and Extended Data Fig. 1j), highlighting again that the effects observed were due to the catalytic inhibition of METTL1.
Taken together, our mechanistic analysis further confirms the specificity and selectivity of our small-molecule approach and that treatment with STM9005 affects METTL1-mediated m7G modification levels, which in turn causes defects in the translation and gene regulation of cancer-essential biomarkers.
In vivo efficacy of STM9005
Given our strong evidence for potency, specificity and anti-cancer efficacy of STM9005 in vitro, we performed a series of in vivo experiments using both solid cancers and leukaemia models.
Initially, we tested the in vivo activity of STM9005 using three solid tumour models from different tissues (melanoma, pancreatic ductal carcinoma (PDAC) and lung adenocarcinoma). For the first in vivo solid-tumour experiment, we transplanted immunocompetent mice subcutaneously with B16F10 mouse melanoma cells and treated them twice daily with 30 mg per kg STM9005 for 5 days (Extended Data Fig. 7a). This treatment plan led to a significant decrease in tumour size of the STM9005-treated cohort in comparison to the vehicle-treated mice (Fig. 4a,b and Extended Data Fig. 7b). To confirm on-target activity of STM9005 in vivo, we quantified m7G levels on tRNA extracted from melanoma tumours and kidney tissues of control and treated mice. Notably, we detected significantly lower levels of m7G tRNA modification in the tumours (Fig. 4c) and kidneys (Fig. 4d) of animals treated with STM9005 compared with the vehicle-treated cohort, demonstrating catalytic inhibition of METTL1 in vivo. We also confirmed a significant decrease in MYC and, simultaneously, a significant increase in p53 after in vivo treatment of melanoma tumours with STM9005 (Fig. 4e–g), providing strong proof of a mechanism consistent with reduction of cancer proliferation in vivo and complementing our previous in vitro findings (Extended Data Fig. 6a,b).
a–d, Representative image of tumours (a), the tumour volume (at day 11; b) and normalized tRNA-Val m7G levels in the tumours (c) and kidneys (d) from animals transplanted with B16F10 melanoma cells, treated twice daily with vehicle (n = 6 (b and d) and n = 5 (c)) or 30 mg per kg STM9005 (n = 9 (b), n = 5 (c) and n = 8 (d)). e–g, Western blot analysis of MYC, p53 and β-actin (e) and normalized protein levels of MYC (f) and p53 (g) in the tumours of animals transplanted with B16F10 melanoma cells, treated twice daily with vehicle (n = 6 independent animals) or 30 mg per kg STM9005 (n = 6 independent animals), normalized to β-actin. h–j, Representative image of the tumours (h), tumour volume over time (i) and the tumour volume at the end point (day 26; j) from animals transplanted with mPDAC1 cells and treated twice daily with vehicle (n = 5) or 30 mg per kg STM9005 (n = 9). k–n, Representative image of the tumours (k), the tumour volume over time (l), the tumour volume at the end point (day 23; m) and the normalized tRNA-Val m7G levels in the kidneys (n) from animals transplanted with A549 lung adenocarcinoma cells and treated twice daily with vehicle (n = 5 (tumour analysis) and n = 4 (kidney analysis)) or 30 mg per kg STM9005 (n = 5) at 23 days after transplantation. o,p, Kaplan–Meier survival analysis of animals transplanted with primary mouse MLL-AF9/Flt3ITD/+ AML (o) or AML-491 PDX (p) cells treated twice daily with vehicle (n = 5) or 30 mg per kg STM9005 (n = 7). P values were calculated using the log-rank (Mantel–Cox) test. q, The levels of hCD45+ cells in the bone marrow of mice transplanted with AML-491 PDX and treated twice daily with vehicle (n = 4) or 30 mg per kg STM9005 (n = 5), measured 43 days after transplantation. The green line indicates ongoing treatment. Data are mean ± s.d. P values were calculated using unpaired two-tailed Student’s t-tests (b–d, f, g, i, j, l–n and q).
We followed up with two further in vivo experiments using solid cancer models, one using a mouse PDAC35 (mPDAC1) model driven by Kras and a second using the human lung adenocarcinoma model A549, both transplanted subcutaneously into immunocompromised mice and treated with 30 mg per kg STM9005 or vehicle twice daily for 9 days (Extended Data Fig. 7c,d). As with the B16F10 model, treatment with STM9005 significantly impaired the growth of pancreatic and lung adenocarcinoma tumours compared to the control cohorts (Fig. 4h–m) and reduced tRNA m7G levels (Fig. 4n). Importantly, while STM9005 showed strong anti-cancer efficacy in vivo, it did not cause any acute or lasting signs of toxicity, and caused no measurable effects on the body weight of either immunocompromised or immunocompetent mice (Extended Data Fig. 7e–g).
Having established the anti-proliferative potential of STM9005 treatment in solid tumours, we next investigated the effect of pharmacological inhibition of METTL1 on haematological malignancies. First, we transplanted an aggressive primary mouse AML model driven by MLL-AF9/Flt3ITD/+ into NSG-SGM3 mice and treated them with 30 mg per kg STM9005 or vehicle, twice daily for 5 days (Extended Data Fig. 7h). Pharmacological inhibition of METTL1 significantly prolonged the survival of STM9005-treated mice compared with the vehicle-treated controls (Fig. 4o), thereby demonstrating efficacy in both haematological and solid cancer models.
To accentuate these observations in a clinically relevant context, we next used a patient-derived AML xenograft (AML-491 PDX) and treated the mice twice daily with either vehicle or 30 mg per kg STM9005 for 10 days (Extended Data Fig. 7i). The STM9005-treated PDX cohort had a significantly longer lifespan compared with the vehicle-treated mice (Fig. 4p) and illustrated a significantly lower AML burden in the bone marrow, as measured by the presence of hCD45+ cells (Fig. 4q). Finally, analysis of weight and blood count measurements covering the duration of the AML PDX experiment (Extended Data Fig. 7i) showed no significant differences or signs of toxicity between the vehicle- and STM9005-treated cohorts (Extended Data Fig. 7j–o), consistent with our previous profiling illustrating no off-target toxicity (Supplementary Tables 2 and 3).
To investigate the impact of STM9005 on normal cells, we assessed the effects of STM9005 on primary cells ex vivo. STM9005 treatment did not impair the colony-forming potential of WT mouse bone marrow cells (Extended Data Fig. 8a) or primary human CD34+ cord blood cells (Extended Data Fig. 8b). Similarly, treatment of primary human hepatocytes from normal donors had no effect on cell viability, in contrast to doxorubicin-treated positive controls, which showed a significant reduction in viability (Extended Data Fig. 8c).
To assess the long-term effects of STM9005 on normal tissues in vivo, WT C57BL/6J mice were treated with vehicle or 30 mg per kg STM9005 twice daily for 12 days, and samples were collected 30 days after treatment. The organ weights of the spleen, liver and kidneys showed no differences between treatment groups (Extended Data Fig. 8e–g, respectively). Moreover, there was no differential impact observed on bone marrow cellularity (Extended Data Fig. 8h) and colony-forming potential of bone marrow cells (Extended Data Fig. 8i,j). Haematological parameters during the treatment period (day 7 after treatment initiation), including white blood cell counts, red blood cell counts, platelet counts and haemoglobin levels, were unaffected by STM9005 treatment (Extended Data Fig. 8k–n). Furthermore, haematopoietic cell composition in the bone marrow and spleen (Extended Data Fig. 8o,p), and analysis of major haematopoietic stem and progenitor compartments (Extended Data Fig. 8q–v), also revealed no significant differences between groups. Finally, histological assessment of heart, liver, kidney and spleen illustrated no morphological evidence of tissue pathology in either treatment group (Extended Data Fig. 9).
Taken together, our data show that isolated catalytic inhibition of METTL1 using STM9005 is highly efficacious in vivo, leading to detrimental effects on the proliferation of both solid and haematological malignancies, with no measurable signs of toxicity or long-term adverse effects.
Discussion
The therapeutic potential of pharmacologically targeting different epitranscriptomic regulators is largely underexplored. Recently, we and others have demonstrated that small-molecule inhibition of RNA-modifying enzymes is a credible therapeutic avenue. However, most of the preclinical and clinical efforts to date have been focused on the m6A modification on poly(A)+ RNA24,25,26,27. There are numerous studies suggesting that tRNA-modifying enzymes could be therapeutic vulnerabilities in cancer6,29,36. Yet, the true therapeutic potential of targeting a tRNA-modifying enzyme pharmacologically remains unclear. Here we illustrate the development and preclinical characterization of a bona fide first-in-class tRNA methyltransferase inhibitor, STM9005—a potent, specific and bioavailable small molecule that selectively blocks the catalytic activity of the m7G-writer METTL1.
We show that pharmacological inhibition of METTL1 affects a range of aggressive malignancies through distinct cellular and molecular activity. STM9005 treatment selectively decreases the deposition of m7G on a key subgroup of tRNAs previously associated with oncogenic transformation, cancer stemness, maintenance and aggressiveness11,20,37, which in turn impacts the translation of their mRNA/protein substrates. This molecular mechanism of action affects key oncogenic and therapeutically relevant pathways, such as those related to cell cycle regulators, p53 and MYC-associated proteins, consistent with our previous findings11.
Using a specific and potent catalytic inhibitor of METTL1 enabled us to generate multiple unique datasets to examine the direct effects of inhibiting the catalytic activity of the METTL1–WDR4 complex. It also enabled us to investigate the therapeutic potential of METTL1 inhibition, illustrating robust and on-target anti-proliferative efficacy of STM9005 in vivo using a range of aggressive mouse and human cancer models, including PDXs. Although we illustrated robust tumour growth inhibition in vivo, we did not observe tumour regression, potentially due to short periods of STM9005 dosing. These findings pave the way for further investigations, including possible combinatorial treatments of STM9005 with other cancer drugs, ultimately leading to the improvement of outcomes of current anti-cancer treatments.
Overall, this study provides a strong proof of concept of the in vivo activity and therapeutic efficacy of a specific and selective tRNA-methyltransferase inhibitor against cancer, identifying this class of inhibitors as a promising opportunity for the development of therapies against malignancies with a clear unmet medical need.
Methods
Expression and purification of full-length METTL1–WDR4 complex
Sf21 cells in Sf-900 II serum-free medium (Gibco, 11497-013) were co-infected at a multiplicity of infection of 1 for both His-tagged METTL1 and Flag-tagged WDR4 baculovirus preparations. The infected cultures were incubated for 72 h at 110 rpm and 27 °C. Cells were collected by centrifugation at 3,400g for 15 min at 4 °C and the pellets were stored at −80 °C.
The pellets were lysed in 25 mM Tris pH 8.0 (Invitrogen, 15568025), 300 mM NaCl (Merck, S6546), 1 mM tris(2-carboxyethyl)phosphine (TCEP) (Thermo Fisher Scientific, 20490), 5% glycerol (Thermo Fisher Scientific, 327255000), 0.25% CHAPS (Thermo Fisher Scientific, 28300), 1 mM phenylmethylsulfonyl fluoride (PMSF) (Thermo Fisher Scientific, 36978) supplemented with protease inhibitor tablets (Roche, 4693124001) and 10 U ml−1 benzonase (Merck, E8263). Homogenization was achieved by two cycles at 20 kilopounds per square inch using an IKA high pressure homogenizer. Lysates were briefly centrifuged in between cycles (2,000 rpm for 15 min) to remove foam and finally cleared by ultracentrifugation (45 min at 150,000g, 4 °C).
The protein complex was purified sequentially by immobilized metal affinity chromatography (IMAC), desalting, ion-exchange chromatography (IEX) and size-exclusion chromatography (SEC). IMAC purification was performed on the 5 ml HisTrap FF Crude column (Cytiva) at a flow rate of 4 ml min−1. The columns were washed in IMAC buffer A (25 mM Tris pH 8.0, 300 mM NaCl, 1 mM TCEP, 5% glycerol and 20 mM imidazole (Merck, 56749-1KG)) and protein complexes were eluted in a gradient of 0–100% IMAC buffer B (buffer A + 500 mM imidazole) over 15 column volumes (CV). Desalting was performed using a HiPrep 53 ml desalting column (Cytiva) at a flow rate of 5 ml min−1 and buffer containing 25 mM Tris pH 8.0, 50 mM NaCl, 1 mM TCEP and 5% glycerol. Further purification by IEX was by 5 ml HiTrap Q HP anion-exchange chromatography at a flow rate of 4 ml min−1. The columns were washed in IEX buffer A (25 mM Tris pH 8.0, 50 mM NaCl, 1 mM TCEP and 5% glycerol) and protein complexes were eluted in a gradient of 0–100% IEX buffer B (25 mM Tris pH 8.0, 1 M NaCl, 1 mM TCEP and 5% glycerol) for 15 CV and 100% IEX buffer B for 5 CV. Finally, SEC purification was on the Superdex S200 26/60 (320 ml) column (Cytiva) at a flow rate of 2 ml min−1 in buffer containing 25 mM Tris pH 8.0, 300 mM NaCl, 1 mM TCEP, 5% glycerol and 3 μM SAM (Santa Cruz Biotechnology, sc-286738C) before the samples were concentrated through an Amicon-15 10,000 MWCO centrifugal filter (Merck).
In vitro transcription of pre-tRNA-Phe substrate
The METTL1–WDR4 substrate pre-tRNA-Phe2 was generated using the TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific, K0441) according to the manufacturer’s recommendations. DNA templates were purchased from IDT as complementary upper- and lower-strand DNA Ultramer oligonucleotides (Supplementary Table 15) containing the T7 promoter sequence (TAATACGACTCACTATAGGG). IVT products were purified using the RNA Clean & Concentrator-25 Kit (Zymo Research, R1018) according to the manufacturer’s instructions. For denaturation and renaturation, the tRNA was diluted into 1× annealing buffer (10 mM Tris pH 8.0, 50 mM NaCl and 1 mM EDTA (Invitrogen, AM9260G)), heated to 94 °C for 5 min and allowed to slowly cool down to room temperature (Thermomixer Comfort, Eppendorf). The samples were snap-frozen and stored at −80 °C.
METTL1–WDR4 RapidFire MS methyltransferase assay
The enzymatic assay was established to determine IC50 values for the inhibition of RNA methyltransferase activity. The enzyme used was full-length His-tagged human METTL1 co-expressed with full length Flag-tagged human WDR4 in a baculovirus expression system. The enzyme complex was purified as described above. Enzymatic reactions were performed as described previously23, with the following modifications: a 25 nM final concentration of the METTL1–WDR4 complex was pre-incubated with variable compound concentrations for 10 min, followed by addition of 1 µM final concentration of the in vitro transcribed substrate pre-tRNA-Phe and 1 µM final concentration SAM. The reaction was incubated for further 90 min at 37 °C and then quenched with 40 µl 7.5% TCA with two internal product standards (D4-SAH and 13C10-SAH). After termination, the plates were sealed, centrifuged and stored at 4 °C until analysis.
METTL1–WDR4 activity was measured label-free using the RapidFire MS platform, as described previously27.
Compound IC50 values were calculated by measuring enzyme activity over a dilution series of inhibitor concentrations. The relative inhibition was normalized to control wells with full inhibition (1 µM sinefungin, Merck, S8559) and without inhibition (DMSO only). Data were evaluated using a four-parameter logistic equation in GraphPad Prism (v.10).
SPR analysis
Characterization of compound binding to METTL1 was carried out on a Biacore 8K instrument (GE HealthCare) using Series S NTA chips (GE HealthCare) at 20 °C.
For kinetic studies, approximately 2 to 3 kRU His–METTL1–WDR4 complex was immobilized to flow cell 2 of the sensor chip, preactivated with EDC/NHS according to the manufacturer’s protocol in a running buffer of 20 mmol l−1 HEPES pH 7.5, 150 mmol l−1 NaCl, 1 mmol l−1 TCEP and 0.05% Tween-20. After protein capture, surfaces were washed with 60 s pulses of ethanolamine and 10 mmol l−1 EDTA, respectively. Serial dilutions of the compound in 20 mmol l−1 HEPES pH 7.5, 150 mmol l−1 NaCl, 1 mmol l−1 TCEP, 0.05% Tween-20 and 2% DMSO were injected using single-cycle kinetics mode with 90 s association time at a flow rate of 80 μl min−1. Dissociation of the samples was monitored for 5,000 s. Data processing was performed using Biacore Insight Evaluation software v.3.0.12 (Biacore, GE HealthCare). Sensorgrams recorded on the reference flow cell 1 of the same channel without protein were subtracted from sensorgrams recorded on the METTL1–WDR4 surface. For double-referencing, injections of the running buffer without compound were run.
Solvent correction was applied to all sample sensorgrams to correct for buffer mismatches. To determine kinetic parameters, data were fitted to a simple 1:1 Langmuir interaction model.
STM9005 selectivity profiling
The selectivity profile of STM9005 was assessed by testing the level of inhibition in a panel of methyltransferases (Reaction Biology). Inhibition of five additional RNA methyltransferases was tested using RapidFire MS assay, as described previously27. Data were plotted using GraphPad Prism (v.10).
Additional selectivity profiling was performed using commercial SAFETYscan47 and KINOMEscan scanEDGE panels (Eurofins DiscoverX), according to the provider’s standard protocols.
X-ray crystallography methods
Cloning, protein expression and purification
Protein boundaries were designed to mimic those used in PDB structure 3CKK, namely residues 32–265 of METTL1, with an N-terminal His-tag followed by a TEV cleavage site. This construct was synthetically cloned into the vector pTriIJ-HV (Evotec). The protein was expressed in BL21-CodonPlus (DE3)-RIL cells, grown in Terrific broth (Teknova, T7060) at 37 °C, before induction with 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG; Melford Laboratories I56000-5.0) at an optical density at 600 nm of 1.5, and subsequent expression for 18 h at 15 °C. Cells were collected by centrifugation at 4,000g for 20 min at 4 °C, before storage at −80 °C.
Cells were lysed with a cell disruptor (Constant Systems) at 25,000 psi, 4 °C in a buffer of 20 mM HEPES, pH 7.4, 500 mM NaCl, 2 mM β-mercaptoethanol (Merck, M3148-100ML), 5% glycerol, 1 mM PMSF, 40 µl l−1 benzonase. Cell debris was removed through centrifugation at 200,000g, 4 °C for 45 min, before the remaining supernatant was applied to a 5 ml HisTrap crude FF column (Cytiva). The loaded column was then washed with 10 CV of buffer A (20 mM HEPES pH 7.4, 500 mM NaCl, 5% glycerol and 50 mM imidazole), before elution with buffer B (20 mM HEPES pH 7.4, 500 mM NaCl, 5% glycerol and 250 mM imidazole). The protein was dialysed for 18 h at 4 °C in buffer C (20 mM HEPES pH 7.4 and 250 mM NaCl) in the presence of TEV protease (Evotec). Dialysed protein was loaded onto a 5 ml HiTrap SP HP ion-exchange column washed with buffer D (20 mM PIPES pH 6.5, Merck, P6757-500G) and eluted with a linear gradient with buffer E (20 mM PIPES pH 6.5 and 500 mM NaCl). Protein was concentrated with a 10 kDa molecular weight cut-off centrifugal concentrator to 28.7 mg ml−1 and flash-frozen in liquid nitrogen until use for crystallography.
Crystallization and soaking
Protein samples were thawed to 4 °C before incubation on ice for 1 h with 10 mM compound STM00005451 (STM5451; final DMSO concentration of 2%; Supplementary Note 1c). Crystals were grown through hanging-drop vapour diffusion in a solution of 100 mM Bis-Tris pH 5.5 (Molecular Dimensions, MD2-004-PH), 200 mM LiSO4 (Molecular Dimensions, MD2-250-45) and 20% (w/v) polyethylene glycol 3350 (PEG 3350; Molecular Dimensions, MD2-250-9) (1 µl of protein + 1 µl reservoir). Seeding was used to enhance crystal reproducibility, and crystals grew overnight at 22 °C.
To replace STM5451 with STM9005, crystals were transferred into reservoir solution (2 µl) with 25 mM STM9005 (5% DMSO), before incubation for 24 h at 22 °C. Crystals were cryo-protected in soaking solution supplemented with 15% PEG before collecting into liquid nitrogen.
Data collection and structure determination
Data were collected at 100 K on ID17B at APS, with a 1.0 Å and processed with autoPROC and STARANISO38,39,40,41,42,43. An in-house generated PDB (originally derived from PDB 3CKK) was used for molecular replacement using dimple44,45. Iterative cycles of model building and refinement were performed using Coot46 and Buster47. STM9005 was fitted using restraints generated by Grade248. The final structure statistics are provided in Extended Data Table 1, with 100% Ramachandran-favoured protein residues. The structure was deposited to the PDB under accession code 9RYL. The structure was analysed using Molecular Operating Environment (Chemical Computing Group), and images were generated using PyMOL Molecular Graphics System (v.2.6.0a0, Schrödinger).
METTL1 cellular-target engagement
Cell target engagement of STM9005 was measured by thermal shift using InCELL Pulse technology (Eurofins DiscoverX). HeLa cells were transfected with 9 µg of pcDNA3.1-WDR4-P2A-METTL1-ePL and 21 µg pcDNA3.1(+) plasmid (empty vector as carrier) using FuGENE transfection reagent (Promega) according to the manufacturer’s instructions. Then, 24 h after transfection, cells were frozen and stored in liquid nitrogen until the day of the assay.
For the thermal shift assay, 100 nl compound dilutions (eleven-point threefold dilutions, top concentration in assay: 50 μM) were spotted into assay plates (Greiner, 384-well, 784201) using a Mosquito instrument (SPT Labtech), resulting in a final DMSO concentration of 0.5%. Assay-ready, frozen WDR4-P2A-METTL1-ePL HeLa cells were thawed, washed in Opti-MEM (Gibco, 11058021) and seeded into the assay plates at 1,500 cells per well in 20 µl Opti-MEM. Plates were sealed with aluminium foil and incubated for 1 h at 37 °C and 5% CO2. For the heat pulse, the plates were submerged for 5 min in a water bath at 60 °C, further incubated for 10 min at room temperature and finally centrifuged briefly to gather the liquid in the bottom of each well. Subsequently, 25 μl of detection solution (InCELL Detection Kit, DiscoverX; working solution: 16.7% EA reagent, 16.7% lysis buffer and 66.7% substrate reagent) was dispensed into each well and incubated for 30 min at room temperature in the dark. Subsequently, 42 µl of the reaction volume was transferred to the measurement plate (Corning, 384-well, 11058021), the assay plate was centrifuged for 1 min at 100g at room temperature and luminescence was measured using the EnVision multimode plate reader (Perkin Elmer). Relative luminescence values were normalized to the control wells with full stabilization (50 µM reference compound) and without stabilization (DMSO only). IC50 values were obtained by plotting data using GraphPad Prism (v.10).
Cell culture
HL-60, K562, Loucy, MOLM-13, MV-4-11, PEER, H1299, H1975, OE21 and OE33 were cultured in RPMI-1640 with l-glutamine (Gibco, 21875-034), supplemented with 10% FBS (Gibco, A5256701) and 1% penicillin–streptomycin (Sigma-Aldrich, P0781). B16F10, mouse PDAC (mPDAC1, 9091 PK –p48-Cre+/–, LSL-Kras+/–, ∆Tva+/−; mPDAC2, p48-cre+/−, LSL-Kras+/−)35, and HEK293FT cells were maintained in DMEM with l-glutamine (Gibco, 41965-039) supplemented with 10% FBS (Gibco) and 1% penicillin–streptomycin. A549, LoVo and SK-MES-1 cells were cultured in F-12K nutrient mix with l-glutamine (Gibco, 21127-022), supplemented with 10% FBS and 1% penicillin–streptomycin. SW982 was maintained using DMEM with l-glutamine, supplemented with 10% FBS, 1% sodium pyruvate (Gibco, 11360-070) and 1% penicillin–streptomycin. OCI-AML2 and OCI-AML3 cells were cultured in Alpha-MEM with l-glutamine (Gibco, 12509-069), supplemented with 20% FBS and 1% penicillin–streptomycin. HeLa cells were grown in MEM (Gibco, 11090-081), supplemented with 10% FBS, 1% l-glutamine (Gibco, 25030-149), 1% non-essential amino acids (Gibco, 11140-050).
Primary mouse MLL-AF9/Flt3ITD/+, MLL-AF4/Flt3ITD/+, Npm1cA+Flt3ITD/+ and MLL-ENL/Flt3ITD/+ cells were generated as described previously31. Npm1cA+NrasG12D/+ mice were described previously49. Primary mouse Dnmt3aR882H/+Npm1cA/+ cells were obtained by crossing Dnmt3aR992H/+ mice50 with Npm1cA/+ mice45. Finally, Npm1cA/+NrasG12D/+Idh1R132H/+ cells were obtained from mice generated by crossing Npm1cA/+NrasG12D/+ mice with Idh1R132H/+ mice51. All primary mouse cells were maintained in X-VIVO 20 (Lonza, LZBE04-448Q) supplemented with 5% FBS, 1% GlutaMax (Gibco, 35050-038), 1% penicillin–streptomycin, 50 ng ml−1 mouse stem cell factor (SCF) (Peprotech, AF-250-03), 10 ng ml−1 mouse IL-3 (Peprotech, AF-213-13) and 10 ng ml−1 mouse IL-6 (Peprotech, AF-216-16).
All cells were maintained at 37 °C, 5% CO2. MOLM-13, K562, HL-60, MV-411, PEER, Loucy, OCI-AML2 and OCI-AML3 cells were obtained from the Sanger Institute Cancer Cell Collection. OE21 and OE33 were purchased from Merck. All of the other cell lines were purchased from ATCC. All cell lines used were negative for mycoplasma.
Cloning and vectors
Cas9 was introduced into cells using the pKLV2-EF1a-Cas92ABsd-W (Addgene, 68343) lentiviral vector, as described previously52. Guide RNAs (gRNAs) were cloned into pKLV2-U6gRNA5(BbsI)-PGKpuro2AmAzamiGreen (Addgene, 67976), as described previously30. A list of all gRNA sequences used in this study is provided in Supplementary Table 15.
Overexpression vectors pLKO.1-puro-ΔU6-Arg-TCT-4-1 and control pLKO.1-puro-ΔU6 were described previously11.
The pKLV-EF1a-puro-TY1 vector was constructed as follows. First, we removed the Cas9 fragment from the pKLV-EF1a-puro2AhCas952 vector, replaced it with a 2×TY1-tag, and introduced XmaI and XbaI restriction enzyme sites. Human CDK4 and MYC genes were amplified from cDNA obtained from MOLM-13 cells; the human METTL1 gene was amplified from cDNA obtained from A549 cells. The pKLV-EF1a-puro-TY1 vector and the PCR products were digested with XmaI and XbaI and ligated (T4 DNA ligase, M0202, NEB) into the vector, resulting in pKLV-EF1a-puro2A-2×TY1-hCDK4, pKLV-EF1a-puro2A-2×TY1-hMYC and pKLV-EF1a-puro2A-2×TY1-hMETTL1, respectively. Mouse Cdk4, Myc and Mettl1 were amplified from cDNA obtained from B16F10 cells. For Cdk4 and Myc, the PCR products and pKLV-EF1a-puro-TY1 plasmid were digested with XmaI and XbaI and ligated into the vector, resulting in pKLV-EF1a-puro2A-2×TY1-mCdk4 or pKLV-EF1a-puro2A-2×TY1-mMyc, respectively. For Mettl1, the pKLV-EF1a-puro-TY1 plasmid was digested with SmaI and XbaI and the PCR product was digested using PmeI and XbaI and ligated together, resulting in pKLV-EF1a-puro2A-2×TY1-mMettl1. The METTL1-mutant vectors were created using the Q5 site-directed mutagenesis kit (E0554, NEB), by introducing the following amino acid changes into the WT overexpression vectors: human METTL1, L160A and D163A; mouse Mettl1, L154A and D157A. A list of the primer sequences used is provided in Supplementary Table 15.
Lentiviral production
Viruses for competition assays and Arg-TCT overexpression were prepared by transfecting HEK293FT cells with the appropriate gRNA lentiviral vector together with packaging plasmids pPAX2 and pMD.2G at a ratio of 1:2.5:0.05 in 12-well plates. The supernatant was collected 48 h after transfection, filtered through a 45 μm PVDF filter and stored at −80 °C until further use.
Viruses for the introduction of Cas9, CDK4, MYC and METTL1 overexpression and the relevant controls were produced by transfecting HEK293FT cells with the appropriate lentiviral vector, pPAX2 and pMD.2G at a ratio 1:1:0.2 in 15 cm dishes. The supernatant was collected at 48 and 72 h after transfection and filtered through a 0.45 μm PVDF filter. The two yields were mixed and concentrated by centrifuging overnight at 6,000g at 4 °C; after centrifugation, the viral pellets were resuspended as 10× concentrates in PBS.
Competitive co-culture assays
Suspension cells were seeded at 0.25 × 106 cells per well and transduced with appropriate viruses in a total medium volume of 2 ml and supplemented with polybrene (Merck, H9268)—8 μg ml−1 for human cells or 4 µg ml−1 for mouse cells—followed by spinfection (1 h, 900g, 32 °C). Adherent cells were seeded at 0.05–0.5 × 106 cells per well; the cells were transduced with appropriate viruses, in a total medium volume of 2 ml and passaged after reaching confluency. At day 4 after transduction, the transduction efficiency was confirmed to be 50% (±10%) using flow cytometry, and changes in the percentage of transduced cells were monitored by flow cytometry over time at days 7, 10 and 13. DAPI (Sigma-Aldrich, D9542) was used as a live/dead stain. Flow cytometry analysis of stained cells was performed using CytoFLEX (Beckman-Coulter). Further data processing and analysis were done using FlowJo (FlowJo, BD Biosciences). Gating strategies for all flow cytometry experiments are provided in Supplementary Fig. 2.
Overexpression
Cells were transduced with overexpression viruses as described above. On days 3 and 5 after transduction, cells were treated with 1 µg ml−1 puromycin (Merck, P8833) to select for transduced cells. Downstream experiments were initiated 6–10 days after transduction.
Cell proliferation assay
Suspension cells were seeded at 12,500 cells per 100 µl, and were treated for 3 days with vehicle or the indicated concentrations of STM9005 (100 μM to 781 nM). Then, 3 days after treatment, the cells were passaged and retreated with fresh compound, such that the resulting cell density matched the initial seeding density. Adherent cell lines were seeded at different densities, so they would reach 80–90% confluency on day 6 and treated with vehicle or the indicated concentrations of STM9005 (100 μM to 781 nM). All adherent cell lines were retreated on day 3. On day 6 after treatment, the CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (Promega, G5430) was used, absorbance values at 490 nm were acquired using the CLARIOStar Plus microplate reader (BMG Labtech) and these values were used to calculate the relative cell proliferation.
Cell proliferation assay (optimized)
Cells were seeded in 96-well culture plates at an initial seeding density of 400 (A549), 1,000 (OE21) or 2,500 (LoVo) cells. Cells were incubated for 18 h at 37 °C under 5% CO2. The next day, the medium was aspirated from all wells and replaced with assay medium (OE21: RPMI 1640 with GlutaMAX (Gibco, 61870-010) + 2% FBS; A549 and LoVo: F-12K nutrient mix with l-glutamine (Gibco, 21127-022) + 2.5% FBS). The cells were treated for 7 days. For quantification, the culture medium was aspirated and 50 µl PBS was added per well, before addition of 50 µl of CellTiter-Glo (Promega, G7572) working solution per well (prepared according to the manufacturer’s instructions). The plates were incubated for 30 min at room temperature and luminescence signal was read on a multimode plate reader (SpectraMax i3x, Molecular Devices or EnVision, Perkin Elmer).
Culture of primary human hepatocytes for viability testing
Human liver tissue was obtained from liver biopsies from deceased organ donors with informed consent. Liver biopsy samples were mechanically and enzymatically dissociated and plated following established protocols53. Primary human hepatocytes were cultured in modified HM medium, as described previously54. In culture, primary human hepatocytes were treated with vehicle or STM9005 for 72 h and viability was assessed using CellTiter-Glo (Promega, G7572) according to the manufacturer’s instructions. Doxorubicin (Sigma-Aldrich, D5220) was used as a positive control. Each experiment was performed in three biopsy samples from two donor livers. All liver samples were obtained with informed consent for use in research and ethical approval (Research Ethics Committee (REC) 23/NS/0106).
Methylene Blue confluency assay
Adherent cells were seeded in 12-well plates at densities that would lead to confluency in 3 days in control cells. At days 3, 6 and 9 after treatment, the cells were passaged and retreated with fresh compound, such that the resulting cell density in each well matched the initial seeding density.
At days 3, 6, 9 and 12, the plates were stained with 1% methylene blue (Merck, M9140), dissolved in 70% ethanol (Thermo Fisher Scientific, 16685992), incubated at room temperature for 20 min and washed with distilled water to remove excess staining. The plates were then dried and scanned using an Epson Perfection V850 scanner.
The acquired images were further analysed using ImageJ (v.1.54), and relative confluency for each treatment was calculated by normalizing to untreated controls.
Serial replating assays
For replating assays of mouse cells, 100,000 WT bone marrow cells or 10,000 primary mouse MLL-AF9/Flt3ITD/+ cells were resuspended in 3 ml of methylcellulose (StemCell Technologies, M3434), supplemented with 1% penicillin–streptomycin, and split between two wells in six-well SmartDish plates (StemCell Technologies, 27371). The colonies were imaged and counted 7 days later (STEMvision, StemCell Technologies), replated to match the starting density and recounted a week later. In drug treatment experiments, the semisolid medium was supplemented with 1 µM, 2.5 µM STM9005 or vehicle, and replenished with each plating.
CD34+ cord blood proliferation assay
Human methylcellulose complete media (StemCell Technologies, H4434) was selected to determine the function of STM9005 on colony-forming capacity in CD34+ hematopoietic stem/progenitor cells (HSPCs) from four healthy donors according to the manufacturer’s instructions. In brief, mononuclear cells were isolated from cord blood from three healthy donors by density centrifugation. CD34+ HSPCs were next positively selected from mononuclear cells by using the CD34+ MicroBead Kit (Miltenyi Biotec, 130-046-702). Finally, 1,000 CD34+ HSPCs, with vehicle or STM9005 treatment (2.5 μM and 5 μM), were suspended in 110 μl of IMDM (Gibco, 12440-053), supplemented with 2% FBS and seeded in 35-mm culture dishes (Thermo Fisher Scientific, 10038820) with 1.1 ml of methylcellulose complete medium. The cells were cultured for 14–16 days and the colony numbers were counted by microscopy. All human cord blood samples were obtained with informed consent under local ethical approval (REC 24/EE/0116).
Cell cycle assay
Adherent cells were seeded at different densities, so they would reach 70–80% confluency at assaying time points; cells were retreated at days 3 and 6. Cell cycle analysis was performed on days 5 and 7 after treatment, using 0.5 × 106 cells using the APC BrdU Flow Kit (BD Pharmingen, 557892), with a 2 h BrdU incubation for suspension cells and a 4 h BrdU incubation for adherent cells, according to the manufacturer’s protocol. Flow cytometry analysis of stained cells was performed using the CytoFLEX or CytoFLEX LX (Beckman-Coulter) system. Further data processing and analysis were done using FlowJo (FlowJo, BD Biosciences). Gating strategies for all flow cytometry experiments are provided in Supplementary Fig. 2.
Apoptosis assay
Suspension cells were seeded at 0.1 × 106 cells per well in 12 well plates in a total of 2 ml medium, supplemented with vehicle or 1 µM and 2 µM of STM9005; cells were retreated at day 3 and day 6. Adherent cells were seeded at different densities, so they would reach 70–80% confluency at assaying time points; cells were retreated at days 3 and day 6. Apoptosis analysis was performed on days 3, 5 and 7 after treatment, staining 0.1 × 106 cells per reaction using the APC Annexin V Apoptosis Detection Kit (BioLegend, 640930) and propidium iodide (Sigma-Aldrich, P4170), according to the manufacturer’s protocol. Flow cytometry analysis of stained cells was performed using CytoFLEX or CytoFLEX LX (Beckman-Coulter). Further data processing and analysis were done using FlowJo (FlowJo, BD Biosciences). Gating strategies for all flow cytometry experiments can be found in Supplementary Fig. 2.
Differentiation assays
For differentiation assays, cells were seeded at 0.25 × 106 cells per well and treated with vehicle or STM9005 (10 µM, 5 µM, 2.5 µM, 1 µM). On days 3 and 6, the cells were stained with 1:250 anti-mouse Ly-6G/Ly-6C (Gr-1)-APC (BioLegend, 108412, RB6-8C5) for mouse cells and 1:250 anti-human CD11b-APC (BioLegend, 982604, ICRF44, AB_2632619) for human cells. On day 3, cells were passaged and retreated with fresh compound, such that the resulting cell density matched the initial seeding density.
Total RNA extractions
Total RNA was extracted using TRIzol reagent (Invitrogen, 15596018) coupled with RNA Clean & Concentrator-100 column purifications (Zymo Research, R1019), according to the RNA Clean & Concentrator manufacturer’s instructions.
RNA quantity was measured by Qubit RNA Broad Range (BR) assay (Invitrogen, Q10211). Quality control of RNA was assessed and the RNA integrity number value was determined using either the Bioanalyzer RNA 6000 Nano kit (Agilent, 5067-1511) or the 4200 TapeStation system (Agilent).
For quantitative PCR (qPCR) analysis, RNA was extracted using GenElute Mammalian Total RNA Miniprep Kit (Merck, RTN350), according to the manufacturer’s protocol.
tRNA-Val MSD assay
Cells treated with STM9005 at the concentrations and time points indicated at figure legends were lysed using prewarmed HB42 buffer (50 mM Tris-HCl pH 7.5 (Invitrogen, 15567-027), 750 mM NaCl, 1 mM EDTA pH 8.0, 1% SDS (Merck, 05030) and 15% formamide (Merck, F7503)) and the lysates were heated for 20 min at 95 °C with agitation. The lysates were then snap-frozen and stored at −80 °C.
For the MSD assay, Streptavidin-coated plates (Mesoscale Discovery, L21SA) were blocked with 50 µl blocking buffer (5% BSA, 100 µg ml−1 salmon sperm DNA solution (Invitrogen, 15632-011) in 1× Tris-buffered saline (TBS) (Alfa Aesar, J60877.K3) containing 0.05% Tween-20 (Merck, P1379)) for 1 h at room temperature with 400 rpm agitation. Plates were then washed twice with 50 µl MSD1 buffer (1× TBS containing 0.05% Tween-20) and once with 50 µl MSD2 buffer (2 mM Tris-HCl pH 7.5, 300 mM NaCl, 1 mM EDTA pH 8.0). To each well, 20 µl of a 12.5 nM stock solution of the biotinylated capture oligo (/5Biosg/CGAACGTGATAACCACTACACTACGGAAAC) was added (final concentration of 0.25 pmol per well) and incubated for 1 h at room temperature with 450 rpm agitation. The plates were washed twice with MSD1 buffer and once with warm HB42 buffer. The previously frozen cell lysates in HB42 were first thawed to room temperature, heated to 95 °C for 10 min and then slowly cooled down to 42 °C (at a rate of 1 °C min−1). The lysates were kept at 42 °C until use. From the cell lysates, 25 µl was transferred to each well on the MSD plate and incubated at 42 °C overnight with 300 rpm agitation.
The next day, the plates were washed once with warm HB42 buffer (42 °C) supplemented with RNase inhibitors (final concentration, 0.002 U µl−1), followed by two washes with MSD1 buffer supplemented with RNase inhibitors (Invitrogen, AM2696), to a final concentration 0.002 U µl−1. The anti-m7G antibody (MBL, RN017M) was diluted to 0.2 µg ml−1 in 1% BSA, 100 µg ml−1 salmon sperm DNA, 0.05 U µl−1 RNase inhibitor in MSD1 buffer, and 30 µl per well was dispensed and incubated for 2 h at room temperature with 400 rpm agitation. The plates were washed three times with MSD1 buffer supplemented with 0.002 U µl−1 RNase inhibitors, followed by addition of 30 µl per well anti-mouse SULFO-TAG antibody (Mesoscale Discovery, R32AC-1) diluted to 1 µg ml−1 in 1% BSA, 0.05 U µl−1 RNase inhibitor in MSD1 buffer. The plates were incubated for 1 h at room temperature with 400 rpm agitation. The plates were then washed three times with MSD1 buffer supplemented with 0.002 U µl−1 RNase inhibitors, then 50 µl per well 1× MSD read buffer T (Mesoscale Discovery, R92TC-3) was added and plates were immediately read on a Sector 600 Mesoscale Discovery reader (Mesoscale Discovery). Dose–response curves were calculated by a four-parameter nonlinear regression model with variable slope using GraphPad Prism (v.10).
Alternatively, for total RNA preparations from cell lines or animal tissues and tumour samples, the tRNA-Val MSD assay was performed as described above with the following adjustments: total RNA was diluted to 20–40 ng µl−1 in HB42 buffer for 0.5–1 µg input in 25 µl per well. The diluted total RNA was heated to 95 °C for 10 min and then slowly cooled down to 42 °C (at a rate of 1 °C min−1) before being transferred to the 384-well MSD plate for the overnight incubation at 42 °C with 300 rpm agitation.
RNA nucleoside quantification by MS
Total RNA extracted from cell lines or animal tissues and tumour were subjected to SEC. The fractions were separated in isocratic 0.1 M ammonium acetate in water (pH 5.5) at 65 °C with a flow rate of 1,000 µl min−1 for 7.5 min on an Agilent SEC-5 1000 Å 4.6 × 300 mm column using the Thermo UltiMate 3000 HPLC system with on-line ultraviolet detection at 260 nm.
Nucleosides were prepared and quantified as described previously27. All data were plotted using GraphPad Prism (v.10).
Western blot analysis
For experiments in Figs. 2j,k and 4e and Extended Data Fig. 1k, whole-cell protein lysates were prepared using RIPA lysis buffer (Thermo Fisher Scientific, 89901) supplemented with protease inhibitor cocktail (Roche, 4693124001) followed by incubation on ice for 30 min. Protein lysates were cleared by centrifugation (20,000g, 15 min, 4 °C). Quantification of total protein was carried out using the Pierce BCA assay (Thermo Fisher Scientific, 23227) according to the manufacturer’s instructions.
For experiments in Extended Data Figs. 4k–m and 6b, cells were lysed using RIPA lysis buffer (Thermo Fisher Scientific, 89901) supplemented with 1× protease inhibitor cocktail (Merck, P8340), 1× phosphatase inhibitor cocktail 2 (Merck, P5044), 1× phosphatase inhibitor cocktail 3 (Merck, P5726), followed by incubation on ice for 10 min. The protein lysates were cleared by centrifugation (20,000g, 10 min, 4 °C). Total protein was quantified using the Pierce BCA assay (Thermo Fisher Scientific, 23227) according to the manufacturer’s instructions.
For experiments in Fig. 3d and Extended Data Figs. 2b, 4b,i,j, 5d and 6f,g, cells were lysed using whole-cell lysis buffer (0.2% IGEPAL CA-630 (Merck, I8896), 50 mM Tris-HCl pH 8.0, 450 mM NaCl, 1 mM EDTA, 1× protease inhibitor cocktail (Merck, P8340), 1× phosphatase inhibitor cocktail 2 (Merck, P5044), 1× phosphatase inhibitor cocktail 3 (Merck, P5726) and 1 mM DTT (Epigentek, R-1104-1)) and incubated on ice for 10 min. The lysates were then centrifuged at 20,000g and 4 °C for 10 min. The supernatant was quantified using Bradford Assay (Bio-Rad, 5000006).
After quantification, the samples were supplemented with 1× NuPAGE LDS sample buffer (Life Technologies, NP0007) and 1× NuPAGE Sample Reducing Agent (Life Technologies, NP0009). They were then incubated at 70 °C for 10 min and were snap-frozen until further use.
Western blots were performed using 10–20 µg of protein on SDS–PAGE gels and blotted onto a PVDF membrane.
Western blotting antibodies
Western blots in Extended Data Fig. 1i were performed using the following antibodies: anti-METTL1 (Proteintech, 14994-1-AP), anti-GAPDH loading control monoclonal antibody DyLight 680 (Thermo Fisher Scientific, MA5-15738-D680), donkey anti-rabbit IgG HRP (Cytiva, NA934-1ML, AB_772206) and sheep anti-mouse IgG HRP (Cytiva, NA931-1ML).
Western blots in Fig. 2j,k were performed using anti-CDC2 (CDK1, Cell Signaling Technology, 9116), anti-CDC25A (Abcam, ab92892), anti-cyclin B1 (CCNB1, Cell Signaling Technology, 12231), anti-p21WAF1/CIP1 (CDKN1A, Cell Signaling Technology, 2947), anti-human CDK4 (Cell Signaling Technology, 12790), anti-human cyclin A2 (CCNA2, Cell Signaling Technology, 67955), anti-GAPDH loading control monoclonal antibody DyLight 680 (Thermo Fisher Scientific, MA5-15738-D680), anti-alpha tubulin-Alexa Fluor 488 (Thermo Fisher Scientific, 53-4502-82), donkey anti-rabbit IgG HRP (Cytiva, NA934-1ML, AB_772206) and sheep anti-mouse IgG HRP (Cytiva, NA931-1ML).
Western blots in Extended Data Fig. 2b were performed using anti-H3 (Abcam, ab1791), anti-METTL1 (Proteintech, 14994-1-AP), anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L Alexa Fluor 680 (Abcam, ab175775), goat anti-rabbit IgG H&L Alexa Fluor 790 (Abcam, ab186697), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Fig. 3d were performed using anti-human CDK4 (Cell Signaling Technology, 12790), anti-KDM1A (Abcam, ab17721), anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 4b were performed using anti-human CDK4 (Cell Signaling Technology, 12790), anti-mouse/human CDK4 (Thermo Fisher Scientific, AHZ0202) anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 4i,j were performed using anti-mouse/human CDK4 (Thermo Fisher Scientific, AHZ0202), anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 4k,l were performed using: anti-METTL1 (Proteintech, 14994-1-AP), anti-β-Actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 4m were performed using anti-KDM1A (Abcam, ab17721), anti-mouse/human CDK4 (Thermo Fisher Scientific, AHZ0202), anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719), goat anti-mouse IgG H&L Alexa Fluor 680 (Abcam, ab175775) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 5d were performed using anti-METTL3 (Abcam, ab195352), anti-human WDR4 (Abcam, ab169526), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Extended Data Fig. 6b,f,g were performed using anti-MYC (Y69) (Abcam, ab32072), anti-β-actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Western blots in Fig. 4e were performed using anti-MYC (Y69) (Abcam, ab32072), anti-p53 (1C12, Cell Signaling Technology, 2524), anti-β-Actin (Abcam, ab8442), goat anti-mouse IgG H&L (HRP) (Abcam, ab205719) and goat anti-rabbit IgG H&L (HRP) (Abcam, ab6721).
Uncropped Western blots for all figures are provided in Supplementary Fig. 1.
Flow cytometry antibodies
For differentiated blood panel staining in bone marrow and spleen (Extended Data Fig. 8l,m, respectively) after treatment and erythrocyte lysis, cells were stained as described previously50 with the following antibodies: anti-mouse CD4-PE/Cyanine5 (PE/Cy5; BioLegend, 100514, 1:400), anti-mouse CD8a-PE/Cy5 (BioLegend, 100710, 1:400), anti-mouse/human CD45R/B220-PE/Cy5 (BioLegend, 103210, 1:400), anti-mouse Ly-6G/Ly-6C (Gr-1)-Alexa Fluor 700 (AF700; BioLegend, 108422, 1:400), anti-mouse/human CD11b-AF700 (BioLegend, 101222, 1:400) and anti-mouse/human CD45R/B220-AF700 (BioLegend, 103232, 1:400). B cells were defined as B220-PE-Cy5+B220-A700+, T cells as CD4-PE-Cy5+CD8-PE-Cy5+AF700− and myeloid cells as CD11b-AF700+Gr1-AF700+PE-Cy5−.
For analysis of the progenitor compartment (Extended Data Fig. 8n–s), freshly isolated BM cells after erythrocyte lysis were stained with the following antibodies, as described previously50: BD Pharmingen biotin mouse lineage panel (B220, TER119, CD11b, CD3e, Gr1; BD Biosciences, 559971, 1:300 each) detected with Streptavidin-Brilliant Violet (BV) 510 (BioLegend, 405233, 1:900), anti-mouse Ly-6A/E (Sca-1)-Pacific Blue (PB; BioLegend, 122520, 1:100), anti-mouse CD117 (KIT) (2B8)-APC-eFluor 780 (AF780) (Thermo Fisher Scientific, 47-1171-82, 1:100), anti-mouse CD48-BV605 (BioLegend, 103441, 1:200), anti-mouse CD150 (SLAM)-BV650 (BioLegend, 115932, 1:100), anti-mouse CD135-PE (BD Pharmingen, BD Biosciences, 553842, 1:100), anti-mouse CD34-FITC (Thermo Fisher Scientific, 11-0431-82, 1:50), anti-mouse CD16/32-PerCP-Cyanine5.5 (PerCP-Cy5.5; BioLegend, 101324, 1:100), 7AAD (BioLegend, 420404, 1:100). LK cells were defined as 7AAD−Lin−KIT-AF780+Sca1-PB−; LSK cells were defined as 7AAD−Lin−KIT-AF780+Sca1-PB+. Megakaryocyte–erythroid progenitors (MEPs), common myeloid progenitors (CMPs) and granulocyte–monocyte progenitors (GMPs) were gated on LK cells and defined as follows: MEP, CD34-FITC−CD16-32-PerCP-Cy5.5−; CMP, CD34-FITC+CD16-32-PerCP-Cy5.5−; GMP, CD34-FITC+CD16-32-PerCP-Cy5.5+. Short-term haematopoietic stem cells (ST-HSCs), long-term haematopoietic stem cells (LT-HSCs), multipotent progenitor 2 (MPP2) cells and multipotent progenitor 3 (MPP3) cells were gated on LSK cells and defined as follows: ST-HSC, CD150-BV650−CD48-BV605−; LT-HSC, CD150-BV650+CD48-BV605−; MPP2, CD150-BV650+CD48-BV605+; MPP3, CD150-BV650−CD48-BV605+. Haematopoietic stem cells (HSCs), multipotent progenitor (MPP) cells and lymphoid-primed multipotent progenitor (LMPP) cells were gated on LSK cells and defined as follows: HSC, CD135-PE−; MPP, CD135-PEmedium; LMPP, CD135-PE+.
Gating strategies are provided in Supplementary Fig. 2. Flow cytometry acquisition was performed using LSRFortessa (BD Biosciences) and BD FACSDiva Software (BD Biosciences), and the data were analysed in FlowJo (FlowJo, BD Biosciences).
qPCR with reverse transcription
cDNA was synthesized using SuperScript IV VILO Master Mix (Invitrogen, 11756500) and 500 ng of RNA in 10 µl reactions, according to the manufacturer’s protocol.
The levels of specific RNAs were measured using the PowerUp SYBR Green Master Mix (Applied Biosystems, A25742) and the QuantStudio 5 Real-Time PCR system (Applied Biosystems), according to the manufacturer’s instructions. A list of all of the primer sequences used in this study is provided in Supplementary Table 15.
mRNA-seq
Total RNA samples isolated from in vitro treatment with either vehicle or 1 µM STM9005 of OE21 and LoVo cell lines as well as from in vivo B16F10 tumours from mice treated twice daily for 5 days with either vehicle or 30 mg per kg STM9005 were subjected to eukaryotic strand-specific mRNA sequencing provided as service by Novogene. In brief, the mRNA-seq service entailed RNA sample quality control, directional mRNA library preparation (with poly(A) enrichment) and sequencing of libraries in a paired-end 150 bp run mode (PE150) on the NovaSeq X Plus instrument (Illumina). The minimum output was 11 Gb per sample (>60 million reads per sample). For each condition, three individual replicate samples were processed.
Fastq files were processed using the Nextflow nf-core RNA-seq pipeline (v.3.9 and v.3.13.2). The pipeline aligned reads to the reference genomes for human (Ensemble GRCh38) and mouse (Ensemble GRCm39) using the STAR aligner55. Transcript expression was quantified with Salmon56. Differential gene expression analysis was performed on per gene read counts by DESeq257, after the removal of genes with fewer than 10 reads. DESeq2 log2[FC] statistics of treatment versus control was used to rank genes for gene set enrichment analysis (clusterProfiler package, pAdjustMethod=“BH”, pvalue.cutoff=0.05) with the hallmark gene set from MSigDB58. All data analysis was performed using R.
Transfer RNA-seq by mim-tRNA-seq
mim-tRNA-seq was adapted from a previous study59 to enable profiling of tRNA levels as well as methylation levels at the G46 position of tRNAs. Total RNA (5 µg in 19 µl) was incubated for 45 min at 37 °C with 1.5 µl 1 M Tris-HCl (pH 9.0) and 1 µl SUPERase.In RNase Inhibitor (Invitrogen, AM2694) to deacylate tRNAs. This was then incubated for 30 min at 37 °C with 67 µl nuclease-free water, 10 µl T4 PNK buffer and 1 µl T4 PNK (NEB, M0201L) to dephosphorylate the 3′ end of tRNAs. RNA was recovered by ethanol precipitation with 10 µl 3 M sodium acetate (pH 5.2) and 2 µl GlycoBlue coprecipitant (Invitrogen, AM9516) and resuspended in 5 µl nuclease-free water. Abasic sites at m7G modified positions were generated by incubation for 40 min on ice with 3 µl 37 mM m7GTP (Jena Bioscience, NU-1122S), 5 µl 3 M Tris-HCl (pH 8.2), 2 µl nuclease-free water and 15 µl 1 M sodium borohydride15 (NaBH4, Merck, 21346). The reaction was stopped by adding quenching buffer (180 µl 1 M sodium acetate (pH 4.5), 120 µl nuclease-free water, 2 µl GlycoBlue and 990 µl ethanol). RNA was recovered by ethanol precipitation. RNA was resolved by gel electrophoresis on a 10% TBE-urea gel. tRNAs were selected by excising RNAs between 65–90 nucleotides and generated into libraries for sequencing as described previously59. The libraries were sequenced in a single-end 100 bp run mode (SE100) on the NextSeq 2000 (Illumina) system.
Fastq files were demultiplexed and trimmed using cutadapt59,60. The mimseq command (--cluster-id 0.97 --threads 15 --min-cov 0.0005 --max-mismatches 0.075) within the conda mimseq package (v.1.3.8) was then used to obtain counts for each isodecoder as well as summarized counts for each anticodon and the proportion of misincorporation at the G46 position. log2[FC] values in tRNA levels after treatment and P values, adjusted for multiple testing using the Benjamini–Hochberg method, for the significance of these changes were then calculated using the DESeq2 R package57. log2[FC] values in the proportion of misincorporation at m7G46-modified sites were calculated for each replicate, as were P values using a two-tailed t-test to check the significance of any change in mean misincorporation.
tRNA northern blotting
Up to 2 µg total RNA isolated from control or METTL1 inhibitor-treated cells or 0.5 µl of the low molecular mass marker (Affymetrix/USB, J76410) were diluted to 5 µl and mixed with equal volume of 2× Novex TBE-urea sample buffer. The samples and marker were heated to 80 °C for 10 min before loading onto 10% Novex TBE-Urea polyacrylamide gels, stained as described above and imaged on a ChemiDoc MP imaging system (Bio-Rad). Nucleic acids were transferred onto positively charged Nylon membrane (Cytiva Amersham Hybond N+, RPN203B) by semi-dry electroblotting (3.3 mA cm−2 for 1 h) using the Trans-Blot Turbo transfer system (Bio-Rad). After completion of transfer, nucleic acids were cross-linked to membranes using a Stratalinker UV crosslinker with 120,000 µJ cm−2 of total energy. Membranes were additionally baked for 10 min at 80 °C and stored at −20 °C until further use.
For hybridization with specific biotinylated oligonucleotides, membranes were cut into strips and incubated with 7.5 ml preheated 1× North2South hybridization buffer (Thermo Fisher Scientific, 37549) for 30 min at 55 °C. Biotinylated oligonucleotide probes (~30 nucleotide length) for specific tRNA molecules were diluted to 10 µM and heat denatured for 10 min at 98 °C. Per membrane in 7.5 ml buffer, 2.5 µl of the denatured oligonucleotide probe (equal to ~250 ng) was added and further incubated overnight at 55 °C. The next day, hybridization buffers were discarded, and membranes were washed three times with 7.5 ml preheated 1× North2South hybridization stringency wash buffer (Thermo Fisher Scientific, 37555) for 15–20 min each at 55 °C. After stringency washes, the membranes were processed for Streptavidin-HRP incubation and subsequent detection of biotin-streptavidin complexes using the Chemiluminescent Nucleic Acid Detection Module Kit (Thermo Fisher Scientific, 89880) according to the manufacturer’s recommendations and imaged on the ChemiDoc MP imaging system (Bio-Rad). Images were processed using ImageJ or Image Lab (Bio-Rad).
Ribosome profiling by ribosome-protected fragment sequencing
Cell culture medium was removed and cells were incubated (2 min, 37 °C) with new medium supplemented with 100 µg ml−1 cycloheximide (Merck, C4859), washed with PBS supplemented with 20 µg ml−1 cycloheximide, and lysed with 400 µl lysis buffer (20 mM Tris-HCl pH 7.4, 150 mM NaCl, 5 mM MgCl2 (Invitrogen, 89880), 1 mM dl-dithiothreitol (DTT, Merck, 43815), 1% Triton X-100 (Thermo Fisher Scientific, 10671652), 0.1% IGEPAL CA-630 (Merck, I8896), 100 µg ml−1 cycloheximide, 25 U ml−1 TURBO DNase (Invitrogen, AM2239) using a cell scraper. The lysates were triturated ten times through a 26-gauge needle and cell debris pelleted by centrifugation at 20,000g, 5 min, 4 °C. Total RNA was purified from 100 µl cleared lysate and mRNA-seq libraries were generated using the NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB, E7490S) and NEBNext Ultra II Directional RNA Library Prep Kit for Illumina (NEB, E7760S). RPFs were generated by incubating (45 min, 23 °C, 1,400 rpm) cleared lysates (120 µl) with 3 µl RNase I (Thermo Fisher Scientific, AM2294). Digestion was stopped with 4 µl SUPERase.In. Ribosomes were enriched by passing 100 µl through a pre-equilibrated Illustra MicroSpin S-400 HR column (GE Healthcare, 27-5140-01) and equilibrated by gravity flow with 3 ml equilibration buffer (20 mM Tris-HCl (pH 7.4), 150 mM NaCl, 5 mM MgCl2). RNA was extracted from the flowthrough using TRIzol reagent according to the manufacturer’s instructions.
rRNA was depleted from 6 µg of RNA using the Ribo-seq rRNA depletion kit (siTOOLs, dp-K024-000042) according to the manufacturer’s instructions and RNA was purified using the RNA Clean & Concentrator kit (Zymogen). RPFs were selected by excising RNAs between 27 and 32 nucleotides after electrophoresis on a 15% TBE-Urea gel and extracted by incubating (15 min, 70 °C, 1,400 rpm) gel pieces in RNA extraction buffer (0.3 M NaOAc (pH 5.2), 0.25% SDS, 1 mM EDTA (pH 8.0)) and precipitating the RNA with isopropanol and GlycoBlue. RNA was incubated at 37 °C for 30 min with 1 µl T4 PNK buffer, 1 µl T4 PNK and 1 µl SUPERase.In. RNA was purified using the RNA Clean & Concentrator column, and libraries were generated using the NEBNext Small RNA-seq kit for Illumina (NEB, E7330S). mRNA and RPF libraries were sequenced on a NovaSeq X Plus machine (Illumina) in a PE150 run mode to a depth of >40 million or >100 million reads, respectively.
Read 1 fastq files were adapter trimmed and filtered for reads with length between 20 and 40 nucleotides using cutadapt60. rRNA, tRNA, snoRNA and snRNA reads were removed using SortMeRNA54. Reads were then aligned to a transcriptome index generated from the MANE Select transcript set (GRCh38, Ensembl release 108) using STAR55 (--alignEndsType EndToEnd, --outFilterMultimapNmax 1, --outFilterMismatchNmax 2, --alignIntronMax 1). Bam files were then processed using riboWaltz61 to determine the codon that was at the ribosome A-site for each RPF, filtering for read lengths from 26 to 32 nucleotides and using the signal within a frame of −1 to +1 around the first nucleotide of each codon. log2[FC] values (treated/control) in Ribo-seq unit step transformation (RUST) ratios of observed/expected RPF density62 for each A-site codon transcriptome-wide, excluding the first 40 and last 20 codons of transcripts and transcripts with fewer than 80 codons, were then calculated for each replicate. This can be interpreted as the log2[FC] in ribosome occupancy when each codon is at the ribosome A-site. For the box plot, codons were assigned to m7G or non-m7G categories depending on the m7G status of the tRNA(s) that recognize them. In cases in which two tRNAs with opposing m7G statuses recognize the same codon, the m7G status of the cognate tRNA was selected unless the expression of this tRNA was undetectable in OE21 tRNA-seq data. Individual codons showing a significant (P < 0.05) increase in RUST ratio were calculated using a one-sample, one-tailed t-test. To calculate the significance of grouped m7G and non-m7G codons, a two-sample, two-tailed t-test was used.
mRNA fastq files were adapter-trimmed and gene counts were obtained using salmon56. An integrative differential expression analysis of RNA-seq and Ribo-seq gene counts was carried out to determine genes regulated at the level of transcription or translational efficiency63. This gives four categories of genes showing significant changes (adjusted P < 0.2): forwarded, significant change in the same direction and at the same rate for both translation and transcription; intensified, significant change in the same direction for both translation and transcription but with greater change in translation; exclusive, significant change only for translation; buffered, significant change in opposite directions for both translation and transcription.
Comparison analysis
Pairwise overlaps between differentially expressed gene sets (|log2[FC] | > 1, adjusted P < 0.05) were quantified using Jaccard similarity. Statistical significance of overlaps was assessed using a one-tailed hypergeometric test against a background of 20,000 protein-coding genes. The probability of observing at least k overlapping genes by chance was calculated as P(X ≥ k), where k is the observed overlap, A and B are the sizes of the two gene sets, and N represents the background gene universe. P values were adjusted for multiple testing using the Benjamini–Hochberg false discovery rate (FDR) correction. All reported overlaps remained significant after FDR correction (FDR < 0.001).
Histology
For histological analysis, samples were fixed in 10% formalin (CellStor Pot, CellPath, BAF-6000-08A), dehydrated and embedded in paraffin (Paraplast Plus, Sigma-Aldrich, P3683). Formalin-fixed paraffin-embedded (FFPE) tissue blocks were sectioned at a thickness of 5–7 μm using a microtome Leica RM2255 microtome and mounted onto microscopy slides (Superfrost Plus Adhesion Microscopy slides, Epredia, J1800AMNZ) for further analysis. Haematoxylin (Pioneer Research Chemicals, PRC/R/51/2) and eosin staining (Pioneer Research Chemicals, PRC/66/1) was performed on FFPE sections. All images were acquired using a ZeissApoTome2.0 with Zen Blue software (Zeiss). Biopsy tissue samples from both groups were reviewed by two blinded clinicians with experience in mouse histology. No gross histological differences were observed between groups.
Animal work
For all animal experiments, STM9005 was prepared using 20% (2-hydroxypropyl)-β-cyclodextrin (Merck, H105-5G) in water (Merck, W3500).
For the pharmacokinetic analysis of STM9005, male C57BL/6J mice were given i.p. injections of 10 mg per kg or 30 mg per kg STM9005 (3 mice per dose) and sampled serially up to 12 h after dosing. For oral dosing, 10 mg ml−1 STM9005 stock solution was administered at 30 mg per kg through oral gavage. Blood was collected from the tail vein at the indicated time points and collected in K2-EDTA tubes. Plasma was isolated by centrifugation, and 5 µl of sample was mixed with a precipitant solution of 200 µl acetonitrile and internal standard. The supernatant from this precipitation was transferred to a new plate and 3 μl injections were characterized by liquid chromatography coupled with MS (LC–MS) on a LC–MS/MS–AU Triple Quad 6500+ instrument attached to an LC instrument and autosampler (Waters ACQUITY UPLC I-Class PLUS System).
STM9005 was resolved on an ACQUITY UPLC HSS T3 column (Waters, 1.8 μm particles, 2.1 × 50 mm) with an increasing gradient of 15–95% B over 78 s. Mobile phases consisted of (A) 0.1% formic acid and 2 mM ammonium formate in water and of (B) 0.1% formic acid and 2 mM ammonium formate in acetonitrile/water (v/v, 95:5). The flow rate was held at 0.6 ml min−1. Retention time for STM9005 was 0.86 min, mass/transition for STM9005 was 462.1/219.1 Da. A calibration curve for STM9005 was run at 2.00–3,000 ng ml−1 and the lower limit of quantification was set at 2 ng ml−1. Data were plotted using GraphPad Prism (v.10).
For all subsequent experiments, each animal was treated twice a day i.p. with 30 mg per kg STM9005 or vehicle.
For experiments in Fig. 4a–g and Extended Data Fig. 7a,b, B16F10 cells were transplanted subcutaneously into female C57BL/6J mice with 0.5 × 106 cells per mouse. mPDAC1 (Fig. 4h–j and Extended Data Fig. 7d,e) and A549 (Fig 4k–n and Extended Data Fig. 7j) cells were transplanted into male NOD.Cg-PrkdcscidIl2rgtm1WjTg(CMV-IL3,CSF2,KITLG)1Eav/MloySzJ (NSG-SGM3) mice with 1 × 106 and 2 × 106 cells per mouse, respectively. All cells were resuspended in a 1:1 mixture of Matrigel (Corning, 354234) and PBS (Merck, P4474-500ML). Treatment for all subcutaneous models was initiated after detection of palpable tumours using digital callipers and lasted 5 days for B16F10, and 9 days for A549 and mPDAC. Tumour sizes were monitored regularly for the duration of the experiment and all animals were collected after the tumours of vehicle-treated animals reached the humane end point described in the study plan. Tumour and kidney tissue was collected for further analysis.
For experiments in Fig. 4o and Extended Data Fig. 7h, male NSG-SGM3 mice were transplanted intravenously (i.v.) with MLL-AF9/Flt3ITD/+ cells at 0.5 × 106 cells per mouse. Daily treatment for 5 days was initiated 12 days after transplantation. Disease progression was monitored by weekly bleeds. In brief, 20 μl of blood was collected from the tail vein using a capillary pipette containing anticoagulants (EDTA) (Microvette 100 EDTA K3E, Sarstedt, 20.1278). The blood samples were used to obtain a complete blood count using Vet abc Plus (scil). The mice were collected after reaching the humane end points described in the study plan. In parallel, C57BL/6J were treated for 5 days, and monitored for signs of toxicity and changes in weight (Extended Data Fig. 7g).
For experiments in Fig. 4p,q and Extended Data Fig. 7i–o, female NSG-SGM3 mice were i.v. transplanted with a patient-derived AML xenograft (AML-491 PDX)64. Daily treatments with either vehicle or STM9005 for 10 days were initiated 22 days after transplantation. Disease progression was monitored by weekly bleeds, as described above, and mice were collected when humane end points were reached. For the experiment in Fig. 4q, a subset of mice was collected on day 43 after transplantation, the bone marrow was stained using 1:250 anti-human CD45 FITC (BioLegend, 304006) and engraftment was quantified using flow cytometry.
For the experiment in Extended Data Fig. 8a, bone marrow was isolated from 4 WT CB57BL/6J mice and used for colony serial plating assays.
For the experiments in Extended Data Figs. 8d–s and 9a,b, C57BL/6J mice were treated for 12 days twice daily with either vehicle or 30 mg ml−1 STM9005. Tissues were collected at 30 days after treatment. Bone marrow was isolated and used to assess colony-forming potential. Bone marrow and spleen were used for flow cytometry interrogation of haematopoiesis. Heart, liver, kidney and spleen were used for histology.
In vivo absorption–distribution–metabolism–excretion studies were conducted at WuXi AppTec in accordance with IACUC standard animal procedures and guidelines and in compliance with the US Animal Welfare Act and Guide for the Care and Use of Laboratory Animals. All animal experiments related to Fig. 4 and Extended Data Fig. 7 and Extended Data Fig. 8 were performed under project licences PP3797858 (PDX, MLL-AF9/Flt3ITD/+), PP9816216 (A549, mPDAC1), PP3042348 (long-term haematopoiesis experiment) or PP7037913 (B16F10) issued by the United Kingdom Home Office, in accordance with the Animal Scientific Procedures Act 1986.
Statistical analysis
No statistical methods were used to predetermine sample size. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment. Statistical tests and sample sizes used are included in all relevant figure legends.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All datasets related to STM9005 used in this study can be accessed from the Gene Expression Omnibus (GEO) database under accession codes GSE302586 and GSE303386. STM9005’s structure has been deposited at the PDB (9RYL). Source data are provided with this paper.
Code availability
The code used to process and analyse the tRNA and Ribo-seq datasets is available at GitHub (https://github.com/storm-therapeutics/mettl1_stm9005_manuscript).
References
Alexandrov, A., Grayhack, E. J. & Phizicky, E. M. tRNA m7G methyltransferase Trm8p/Trm82p: evidence linking activity to a growth phenotype and implicating Trm82p in maintaining levels of active Trm8p. RNA 11, 821–830 https://doi.org/10.1261/rna.2030705 (2005).
Alexandrov, A., Martzen, M. R. & Phizicky, E. M. Two proteins that form a complex are required for 7-methylguanosine modification of yeast tRNA. RNA 8, 1253–1266 https://doi.org/10.1017/s1355838202024019 (2002).
Li, J. et al. Structural basis of regulated m7G tRNA modification by METTL1-WDR4. Nature 613, 391–397 https://doi.org/10.1038/s41586-022-05566-4 (2023).
Ruiz-Arroyo, V. M. et al. Structures and mechanisms of tRNA methylation by METTL1-WDR4. Nature 613, 383–390 https://doi.org/10.1038/s41586-022-05565-5 (2023).
Delaunay, S., Helm, M. & Frye, M. RNA modifications in physiology and disease: towards clinical applications. Nat. Rev. Genet. 25, 104–122 https://doi.org/10.1038/s41576-023-00645-2 (2024).
Suzuki, T. The expanding world of tRNA modifications and their disease relevance. Nat. Rev. Mol. Cell Biol. 22, 375–392 https://doi.org/10.1038/s41580-021-00342-0 (2021).
Chou, H. J., Donnard, E., Gustafsson, H. T., Garber, M. & Rando, O. J. Transcriptome-wide analysis of roles for tRNA modifications in translational regulation. Mol. Cell 68, 978–992 https://doi.org/10.1016/j.molcel.2017.11.002 (2017).
de Crécy-Lagard, V. et al. Matching tRNA modifications in humans to their known and predicted enzymes. Nucleic Acids Res. 47, 2143–2159 https://doi.org/10.1093/nar/gkz011 (2019).
Alexandrov, A. et al. Rapid tRNA decay can result from lack of nonessential modifications. Mol. Cell 21, 87–96 https://doi.org/10.1016/j.molcel.2005.10.036 (2006).
Jovine, L., Djordjevic, S. & Rhodes, D. The crystal structure of yeast phenylalanine tRNA at 2.0 A resolution: cleavage by Mg2+ in 15-year old crystals. J. Mol. Biol. 301, 401–414 https://doi.org/10.1006/jmbi.2000.3950 (2000).
Orellana, E. A. et al. METTL1-mediated m7G modification of Arg-TCT tRNA drives oncogenic transformation. Mol. Cell 81, 3323–3338 https://doi.org/10.1016/j.molcel.2021.06.031 (2021).
Fu, Y. et al. Perturbation of METTL1-mediated tRNA N7-methylguanosine modification induces senescence and aging. Nat. Commun. 15, 5713 https://doi.org/10.1038/s41467-024-49796-8 (2024).
Malbec, L. et al. Dynamic methylome of internal mRNA N7-methylguanosine and its regulatory role in translation. Cell Res. 29, 927–941 https://doi.org/10.1038/s41422-019-0230-z (2019).
Chu, J. M. et al. Existence of internal N7-methylguanosine modification in mRNA determined by differential enzyme treatment coupled with mass spectrometry analysis. ACS Chem. Biol. 13, 3243–3250 https://doi.org/10.1021/acschembio.7b00906 (2018).
Pandolfini, L. et al. METTL1 promotes let-7 microRNA processing via m7G methylation. Mol. Cell 74, 1278–1290 https://doi.org/10.1016/j.molcel.2019.03.040 (2019).
Zhang, L. S. et al. Transcriptome-wide mapping of internal N7-methylguanosine methylome in mammalian mRNA. Mol. Cell 74, 1304–1316 https://doi.org/10.1016/j.molcel.2019.03.036 (2019).
Barbieri, I. & Kouzarides, T. Role of RNA modifications in cancer. Nat. Rev. Cancer 20, 303–322 https://doi.org/10.1038/s41568-020-0253-2 (2020).
Delaunay, S. & Frye, M. RNA modifications regulating cell fate in cancer. Nat. Cell Biol. 21, 552–559 https://doi.org/10.1038/s41556-019-0319-0 (2019).
Roundtree, I. A., Evans, M. E., Pan, T. & He, C. Dynamic RNA modifications in gene expression regulation. Cell 169, 1187–1200 https://doi.org/10.1016/j.cell.2017.05.045 (2017).
Dai, Z. et al. N7-Methylguanosine tRNA modification enhances oncogenic mRNA translation and promotes intrahepatic cholangiocarcinoma progression. Mol. Cell 81, 3339–3355 https://doi.org/10.1016/j.molcel.2021.07.003 (2021).
Tian, Q. H. et al. METTL1 overexpression is correlated with poor prognosis and promotes hepatocellular carcinoma via PTEN. J. Mol. Med. 97, 1535–1545 https://doi.org/10.1007/s00109-019-01830-9 (2019).
Wang, C. et al. Methyltransferase-like 1 regulates lung adenocarcinoma A549 cell proliferation and autophagy via the AKT/mTORC1 signaling pathway. Oncol. Lett. 21, 330 https://doi.org/10.3892/ol.2021.12591 (2021).
Ali, R. H. et al. A methyltransferase-independent role for METTL1 in tRNA aminoacylation and oncogenic transformation. Mol. Cell https://doi.org/10.1016/j.molcel.2025.01.003 (2025).
Guirguis, A. A. et al. Inhibition of METTL3 results in a cell-intrinsic interferon response that enhances antitumor immunity. Cancer Discov. 13, 2228–2247 https://doi.org/10.1158/2159-8290.CD-23-0007 (2023).
Huang, Y. et al. Small-molecule targeting of oncogenic FTO demethylase in acute myeloid leukemia. Cancer Cell 35, 677–691 https://doi.org/10.1016/j.ccell.2019.03.006 (2019).
Moser, J. C. et al. Phase 1 dose escalation and cohort expansion study evaluating safety, PK, PD and clinical activity of STC-15, a METTL-3 inhibitor, in patients with advanced malignancies. J. Clin. Oncol. 42, 2586–2586 https://doi.org/10.1200/JCO.2024.42.16_suppl.2586 (2024).
Yankova, E. et al. Small-molecule inhibition of METTL3 as a strategy against myeloid leukaemia. Nature 593, 597–601 https://doi.org/10.1038/s41586-021-03536-w (2021).
Nai, F. et al. Small-Molecule Inhibitors of the m7G-RNA Writer METTL1. ACS Bio Med Chem Au 4, 100–110 https://doi.org/10.1021/acsbiomedchemau.3c00030 (2024).
Ren, L. et al. Targeting the METTL1/m7G axis as a therapeutic strategy in myeloid leukemia. Blood https://doi.org/10.1182/blood.2025030336 (2026).
Tzelepis, K. et al. A CRISPR dropout screen identifies genetic vulnerabilities and therapeutic targets in acute myeloid leukemia. Cell Rep. 17, 1193–1205 https://doi.org/10.1016/j.celrep.2016.09.079 (2016).
Barbieri, I. et al. Promoter-bound METTL3 maintains myeloid leukaemia by m6A-dependent translation control. Nature 552, 126–131 https://doi.org/10.1038/nature24678 (2017).
DepMap, B. DepMap 24Q4 public. Figshare https://doi.org/10.25452/figshare.plus.27993248.v1 (2024).
Pacini, C. et al. A comprehensive clinically informed map of dependencies in cancer cells and framework for target prioritization. Cancer Cell 42, 301–316 https://doi.org/10.1016/j.ccell.2023.12.016 (2024).
Zhang, X. et al. METTL1 coordinates cutaneous squamous cell carcinoma progression via the m7G modification of the ATF4 mRNA. Cell Death Discov. 11, 27 https://doi.org/10.1038/s41420-025-02304-3 (2025).
Mueller, S. et al. Evolutionary routes and KRAS dosage define pancreatic cancer phenotypes. Nature 554, 62–68 https://doi.org/10.1038/nature25459 (2018).
Rapino, F., Delaunay, S., Zhou, Z., Chariot, A. & Close, P. tRNA modification: is cancer having a wobble? Trends Cancer 3, 249–252 https://doi.org/10.1016/j.trecan.2017.02.004 (2017).
El-Hachem, N. et al. Valine aminoacyl-tRNA synthetase promotes therapy resistance in melanoma. Nat. Cell Biol. 26, 1154–1164 https://doi.org/10.1038/s41556-024-01439-2 (2024).
Vonrhein, C. et al. Data processing and analysis with the autoPROC toolbox. Acta Crystallogr. D 67, 293–302 https://doi.org/10.1107/S0907444911007773 (2011).
Kabsch, W. XDS. Acta Crystallogr. D 66, 125–132 https://doi.org/10.1107/S0907444909047337 (2010).
Evans, P. Scaling and assessment of data quality. Acta Crystallogr. D 62, 72–82 https://doi.org/10.1107/S0907444905036693 (2006).
Evans, P. R. & Murshudov, G. N. How good are my data and what is the resolution? Acta Crystallogr. D 69, 1204–1214 https://doi.org/10.1107/S0907444913000061 (2013).
Agirre, J. et al. The CCP4 suite: integrative software for macromolecular crystallography. Acta Crystallogr. D 79, 449–461 https://doi.org/10.1107/S2059798323003595 (2023).
Tickle, I. J. et al. STARANISO https://staraniso.globalphasing.org/cgi-bin/staraniso.cgi (Global Phasing, 2016).
McCoy, A. J., Grosse-Kunstleve, R. W., Storoni, L. C. & Read, R. J. Likelihood-enhanced fast translation functions. Acta Crystallogr. D 61, 458–464 https://doi.org/10.1107/S0907444905001617 (2005).
Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr. D 67, 235–242 https://doi.org/10.1107/S0907444910045749 (2011).
Emsley, P., Lohkamp, B., Scott, W. G. & Cowtan, K. Features and development of Coot. Acta Crystallogr. D 66, 486–501 https://doi.org/10.1107/S0907444910007493 (2010).
Bricogne, G. et al. BUSTER version 2.11.8 (Global Phasing, 2017).
Smart, O. S. et al. Grade2 version 1.6.0 (Global Phasing, 2021).
Dovey, O. M. et al. Molecular synergy underlies the co-occurrence patterns and phenotype of NPM1-mutant acute myeloid leukemia. Blood 130, 1911–1922 https://doi.org/10.1182/blood-2017-01-760595 (2017).
Gozdecka, M. et al. Mitochondrial metabolism sustains DNMT3A-R882-mutant clonal haematopoiesis. Nature 642, 431–441 https://doi.org/10.1038/s41586-025-08980-6 (2025).
Gupta, S. et al. Transcriptional variability accelerates preleukemia by cell diversification and perturbation of protein synthesis. Sci. Adv. 8, eabn4886 https://doi.org/10.1126/sciadv.abn4886 (2022).
Koike-Yusa, H., Li, Y., Tan, E. P., Velasco-Herrera Mdel, C. & Yusa, K. Genome-wide recessive genetic screening in mammalian cells with a lentiviral CRISPR-guide RNA library. Nat. Biotechnol. 32, 267–273 https://doi.org/10.1038/nbt.2800 (2014).
Knobeloch, D. et al. Human hepatocytes: isolation, culture, and quality procedures. Methods Mol. Biol. 806, 99–120 https://doi.org/10.1007/978-1-61779-367-7_8 (2012).
Zhang, K. et al. Expansion of human hepatocytes and their application in three-dimensional culture and genetic manipulation. Nat. Protoc. 20, 3722–3754 https://doi.org/10.1038/s41596-025-01211-2 (2025).
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 https://doi.org/10.1093/bioinformatics/bts635 (2013).
Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods 14, 417–419 https://doi.org/10.1038/nmeth.4197 (2017).
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 https://doi.org/10.1186/s13059-014-0550-8 (2014).
Liberzon, A. et al. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. 1, 417–425 https://doi.org/10.1016/j.cels.2015.12.004 (2015).
Behrens, A., Rodschinka, G. & Nedialkova, D. D. High-resolution quantitative profiling of tRNA abundance and modification status in eukaryotes by mim-tRNAseq. Mol. Cell 81, 1802–1815 https://doi.org/10.1016/j.molcel.2021.01.028 (2021).
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet. J. 17, 3 https://doi.org/10.14806/ej.17.1.200 (2011).
Lauria, F. et al. riboWaltz: optimization of ribosome P-site positioning in ribosome profiling data. PLoS Comput. Biol. 14, e1006169 https://doi.org/10.1371/journal.pcbi.1006169 (2018).
O’Connor, P. B., Andreev, D. E. & Baranov, P. V. Comparative survey of the relative impact of mRNA features on local ribosome profiling read density. Nat. Commun. 7, 12915 https://doi.org/10.1038/ncomms12915 (2016).
Chothani, S. et al. deltaTE: detection of translationally regulated genes by integrative analysis of Ribo-seq and RNA-seq data. Curr. Protoc. Mol. Biol. 129, e108 https://doi.org/10.1002/cpmb.108 (2019).
Vick, B. et al. An advanced preclinical mouse model for acute myeloid leukemia using patients’ cells of various genetic subgroups and in vivo bioluminescence imaging. PLoS ONE 10, e0120925 https://doi.org/10.1371/journal.pone.0120925 (2015).
Acknowledgements
We thank the members of the Tzelepis laboratory and Storm Therapeutics for advice and suggestions; the members of the Evotec Hamburg SPR and MS, Evotec Abingdon Chemistry and Evotec Toulouse Cell Biology teams for their support of this project; G. Yang for assistance with BioRender figure generation; I. Jeremias for providing AML-491 PDX cells; the staff at the Cambridge Blood and Stem Cell Biobank for access to cord blood samples; and the University of Cambridge Biomedical Services (UBS) staff J. Barratt, C. Townend, K. Griggs, E. Friel-Blanchard and L. Smith for their assistance with the animal work presented in this manuscript.
Funding
K.T., J.L.-L., E.S., S.E., V.T. and M.E. were supported by the Wellcome Trust (grants 203151/A/16/Z, 306752/A/23/Z and 306752/Z/23/Z) and UKRI Medical Research Council (grant 203151/Z/16/Z); K.T. and M.E. by Leukaemia UK (2022/FuF/001); J.L.-L. by a Fundación Ramón Areces Fellowship (BEVP37S21391); K.T. and J. Russell by the Cancer Research UK (SEBCATP-2022/100009); E.Y. by Rosetrees and Leukaemia UK John Goldman Fellowship (2024/JGF/009); J.T. by a Clinical Research Training Fellowship from AstraZeneca and the NIHR 442 BioResource (G119775); L.B. by and Early Detection Project Grant from Cancer Research UK (EDDCPJT\100010) awarded to G.S.V.; M.G. by a Blood Cancer United/Blood Cancer UK Specialized Centre of Research Grant (7035-24) and a Blood Cancer UK Project Grant (25029); R.W. by a Kay Kendall Leukemia Fund PhD Studentship (KKL1355). This research was supported by the NIHR Cambridge Biomedical Research Centre (137 BRC-1215-20014 and NIHR203312) and the Cancer Research UK Cambridge Centre (Cancer Research UK Major Centre Award (C9685/A25117 and CTRQQR139 2021\100012) that also part-funded G.S.V.; B.J.P.H. is part of a Specialized Center of Research (7035-24) supported by Blood Cancer United and Blood Cancer UK and was supported by the Cancer Research UK Cambridge Centre (C9685/A25117) and NIHR Cambridge Biomedical Research Centre (BRC-1215-20014). The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. J.L. and D.C.P. were supported by the William B. Harris Foundation (RG91681) and Alborada Trust (RG109433). M.P.A. acknowledges funding from the Wellcome Trust and The Royal Society (105942/Z/14/A). M.T.B. was supported by the Leverhulme Trust (RPG-2023-136) and the European Research Council (ERC) Executive Agency under HORIZON ERC Synergy Grant Programme (101167202, ClonEScape, ERC-2024-SyG, to M.P.A.); E.L. by Wellcome Trust and Royal Society (107630/Z/15/Z), Blood Cancer United (previously Leukaemia Lymphoma Society; 7035-24) and Blood Cancer UK (24008); G.M. by a Cancer Research UK Cambridge Cancer Centre PhD fellowship (CTRQQR-2021\100012); W.T.K. by CRUK Awards DCRPGF\100010, EDDAMC-2021\100006, PRCBTP-May23/100003 and ERC Advanced Grant 101199785; F.S. by a UKRI Future Leaders fellowship, by Health@InnoHK, Innovation Technology Commission, HKSAR, the Theme-based Research Scheme 2021 (T12-712/21-R) and the UEG Research Prize; T.B. by the Replacing Animal Research Innovation Grant (00593); Y.M. by a Cancer Research UK Clinical Research Training Fellowship (SEBCATP-2022/100009). This research was also funded in part by Wellcome, who supported the Cambridge Stem Cell Institute (203151/Z/16/Z and 226795/Z/22/Z). D.J.A. was supported by Cancer Research UK (C20510/A21717), Wellcome (WT206194), the Medical Research Council and the ERC (319661); and S.U. and N.J. by Open Targets (OTAR3083). For the purpose of open access, the authors have applied for a CC BY public copyright licence to any author accepted manuscript version arising from this submission.
Author information
Author notes
These authors contributed equally: Eliza Yankova, Alexandra Sapetschnig, Beth Thomas
Authors and Affiliations
Cambridge Stem Cell Institute, University of Cambridge, Cambridge, UK
Eliza Yankova, Siân Evans, Maria Eleftheriou, Malgorzata Gozdecka, Teresa Brevini, Rachael Wilson, George Giotopoulos, Ryan Asby, Laura Bond, James Russell, Judith López-Luis, Giovanna Mantica, Jake Tobin, Diego Lopez-Leon, Dean C. Pask, Vidur Tandon, Evangelia Stamou, Maria T. Bejar, Yogeshkumar Malam, Elisa Laurenti, Juan Li, Maria P. Alcolea, Walid T. Khaled, Fotios Sampaziotis, George S. Vassiliou, Brian J. P. Huntly & Konstantinos Tzelepis
Department of Haematology, University of Cambridge, Cambridge, UK
Eliza Yankova, Siân Evans, Maria Eleftheriou, Malgorzata Gozdecka, Rachael Wilson, George Giotopoulos, Ryan Asby, Laura Bond, James Russell, Judith López-Luis, Giovanna Mantica, Jake Tobin, Diego Lopez-Leon, Dean C. Pask, Vidur Tandon, Evangelia Stamou, Elisa Laurenti, Juan Li, George S. Vassiliou, Brian J. P. Huntly & Konstantinos Tzelepis
Milner Therapeutics Institute, University of Cambridge, Cambridge, UK
Eliza Yankova, Siân Evans, Maria Eleftheriou, Judith López-Luis, Evangelia Stamou & Konstantinos Tzelepis
Storm Therapeutics, Cambridge, UK
Alexandra Sapetschnig, Beth Thomas, Harry Fischl, Sarah J. Bucknell, Carmen M. Livi, Aleksandra Azevedo, Natalie Webster, Jack Rogan, Byron Andrews, Hendrik Weisser, Richard Fosbeary, Sapphire Sawyer & Oliver Rausch
Evotec, Abingdon, UK
Andrew Hutchin
Cambridge Blood and Stem Cell Biobank, Department of Haematology, University of Cambridge, Cambridge, UK
Joanna Baxter
Centre for Genomics Research, Discovery Sciences, BioPharmaceuticals R&D, AstraZeneca, Cambridge, UK
Jake Tobin
Wellcome Sanger Institute, Cambridge, UK
Sunay Usluer, Nithila Joseph, David J. Adams & Konstantinos Tzelepis
Open Targets, Wellcome Genome Campus, Cambridge, UK
Sunay Usluer & Nithila Joseph
Department of Physiology, Development and Neuroscience, University of Cambridge, Cambridge, UK
Maria T. Bejar & Maria P. Alcolea
Institute of Molecular Oncology and Functional Genomics, School of Medicine, Technische Universität München, Munich, Germany
Roland Rad
Center for Translational Cancer Research (TranslaTUM), School of Medicine, Technische Universität München, Munich, Germany
Roland Rad
German Cancer Consortium (DKTK), Heidelberg, Germany
Roland Rad
Department of Medicine II, Klinikum rechts der Isar, School of Medicine, Technische Universität München, Munich, Germany
Roland Rad
Department of Pharmacology, University of Cambridge, Cambridge, UK
Walid T. Khaled
Cambridge Liver Unit, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK
Fotios Sampaziotis
National Institute of Health Research (NIHR) Blood and Transplant Research Unit (BTRU) at the University of Cambridge in collaboration with Newcastle University and in partnership with NHS Blood and Transplant (NHSBT), Cambridge, UK
Fotios Sampaziotis
Department of Medicine, University of Cambridge, Cambridge, UK
Fotios Sampaziotis
Authors
- Eliza Yankova
- Alexandra Sapetschnig
- Beth Thomas
- Siân Evans
- Maria Eleftheriou
- Malgorzata Gozdecka
- Harry Fischl
- Teresa Brevini
- Rachael Wilson
- Sarah J. Bucknell
- Andrew Hutchin
- Carmen M. Livi
- George Giotopoulos
- Ryan Asby
- Laura Bond
- Aleksandra Azevedo
- Natalie Webster
- Jack Rogan
- James Russell
- Byron Andrews
- Hendrik Weisser
- Joanna Baxter
- Judith López-Luis
- Giovanna Mantica
- Jake Tobin
- Diego Lopez-Leon
- Dean C. Pask
- Vidur Tandon
- Richard Fosbeary
- Sapphire Sawyer
- Sunay Usluer
- Nithila Joseph
- Evangelia Stamou
- Maria T. Bejar
- Yogeshkumar Malam
- Roland Rad
- Elisa Laurenti
- Juan Li
- Maria P. Alcolea
- David J. Adams
- Walid T. Khaled
- Fotios Sampaziotis
- George S. Vassiliou
- Brian J. P. Huntly
- Oliver Rausch
- Konstantinos Tzelepis
Contributions
K.T. and O.R. conceived and supervised the study. E.Y., A.S., B.T., S.E., M.E., R.W., S.J.B., J. Rogan, A.H., J. Russell, A.A., N.W., J.B., G.M., L.B., V.T., B.A., R.F., S.S., S.U., D.L.-L., N.J., D.J.A., T.B., Y.M., F.S., J.T., M.G. and R.R. contributed to chemical, structural, cellular and molecular biology assays. K.T., E.Y., S.E., M.E., J. Russell, G.M., G.G., R.A., J.L.-L., D.C.P., M.G., E.S., M.T.B., J.L., D.C.P., E.L., M.P.A., G.S.V., W.T.K. and B.J.P.H. contributed to the animal experiments. C.M.L., H.W. and H.F. contributed to sequencing and bioinformatic analysis. E.Y., K.T., A.S., B.T. and O.R. wrote the manuscript with help from all of the authors. All of the authors discussed the results and commented on the manuscript.
Corresponding authors
Correspondence to Oliver Rausch or Konstantinos Tzelepis.
Ethics declarations
Competing interests
K.T. has received stock options, consultancy fees and research funding from Storm Therapeutics, and is a stockholder of TEP Therapeutics. A.S., B.T., H.F., S.J.B., H.W., C.M.L., A.A., N.W., J. Rogan, B.A., R.F., S.S. and O.R. were employees of Storm Therapeutics at the time the experiments were performed. G.S.V. is a consultant to STRM.BIO and holds a research grant from AstraZeneca for research unrelated to that presented here. J.T. receives research funding from AstraZeneca. E.L. holds a research grant from CSL Behring for research unrelated to that presented here. The other authors declare no competing interests.
Peer review
Peer review information
Nature thanks Pierre Close and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
Extended Data Fig. 1 STM9005 is a specific and selective small molecule catalytic inhibitor of METTL1.
a, Chemical structures of STM4529 and STM9001. b, Biochemical activity assay showing inhibition of the METTL1/WDR4 complex using a concentration-range of STM4529; mean (n = 2 independent samples). c,d, Surface plasmon resonance (SPR) assay, showing binding affinity to the METTL1/WDR4 complex is reduced in the presence of SAH, indicating SAM-competitive binding (c) and binding affinity (Kd) of STM9005 to the METTL1/WDR4 complex with single cycle binding kinetics (d). RU, relative units. e, Biochemical activity assay showing no inhibition in a panel of RNA-modifying enzymes, using a concentration-range of STM9005; mean (n = 2 independent samples). f, RNA mass spectrometry quantification of the levels of m1A and m5C on small-fraction RNA are unaffected upon treatment with vehicle or STM9005; mean ± s.d. (n = 3 independent samples); p calculated using 2-way ANOVA with multiple comparisons. g, Representative concentration-dependent reduction of tRNA-Val m7G levels from OE21 cells upon treatment with a concentration-range of STM9005 or STM9001; mean; geomean IC50 and s.d. indicated in brackets and derived from n = 15 independent experiments for STM9005; for STM9001 n = 2 technical replicates. h, RNA mass spectrometry quantification of m7G levels on 18S ribosomal RNA upon treatment of primary murine NPM1cA/+/Flt3ITD/+ AML and A549 cells with vehicle or 2 µM STM9005; mean ± s.d. (n = 3 independent samples); p calculated using an unpaired two-tailed Student’s t-test. i, Interactions between STM9005 and METTL1, as identified using Molecular Operating Environment (Chemical Computing Group ULC). j,k qPCR analysis of METTL1 levels following 4 days of treatment with vehicle, 1, or 5 µM STM9005 (j) and Western blot analysis of METTL1 levels following a 5-day treatment with vehicle, 0.3 or 1 µM STM9005 (k) in LoVo cells; mean ± s.e.m (n = 3 independent samples); p calculated using one-way ANOVA.
Extended Data Fig. 2 Response to METTL1 genetic loss in cancer cells in vitro.
a, Competitive co-culture assay of METTL1 gRNA (GFP), Mettl1 gRNA (BFP) or EMPTY gRNA-transfected cells (GFP or BFP positive, respectively) versus non-transfected cells in a panel of solid cancer cell lines; normalized to EMPTY gRNA and to %GFP+ or %BFP+ cells on day 4; mean ± s.e.m. (n = 3 independent cultures). b, Western blot validation of METTL1 gRNA knock-outs shown in a.
Extended Data Fig. 3 Pharmacological inhibition of METTL1 impacts cancer cells in vitro.
a, Normalized confluency of cell lines treated with vehicle, 0.3, 0.6, 1.25, 2.5 or 5 µM STM9005, following 9 days of treatment; mean ± s.d. (n = 4 independent cultures). p calculated using one-way ANOVA. Representative images shown under the corresponding bars. b, STM9005 and STM9001 concentration-response curves measuring cell viability (CellTiter-Glo®) in A549 cells under optimized culture conditions following 7 days of treatment. IC50 indicated in brackets (n = 2 technical replicates). c, Colony-forming efficiency of primary murine MLL-AF9/Flt3ITD/+ cells upon treatment with vehicle, 1 or 2.5 µM STM9005; mean ± s.d. (n = 3 independent cultures); d,e, Normalized confluency in cell lines resistant to METTL1-loss (OE33, left; mPDAC2, right) (d) or cell lines sensitive to METTL1-loss (B16F10, left; mPDAC1, right) (e), harbouring METTL1KO and treated with vehicle, 1.25, or 2.5 µM STM9005, following 6 days of treatment; mean ± s.d. (n = 4 independent cultures; B16F10 Mettl1 (2) and mPDAC1 Mettl1 (2), n = 3). p calculated using two-way ANOVA. f, CD11b levels in OCI-AML2 and HL-60 cells, following 6 days of treatment with vehicle, 1, 2.5 or 5 µM STM9005; mean MFI of CD11b-APC-stained cells ± s.d. (n = 3 independent cultures). p calculated using one-way ANOVA. f,g, Cell cycle analysis (f) and breakdown (g) of B16F10 melanoma (top) or mPDAC (bottom) cells following 7 days of treatment with vehicle, 1 or 2 µM STM9005; mean (f) or mean ± s.d. (g) (n = 3 independent cultures); p calculated using 2-way ANOVA with multiple comparisons. h, Annexin V analysis of apoptosis upon treatment with vehicle, 1 or 2 µM STM9005 in a panel of cell lines; mean (n = 3 independent cultures).
Extended Data Fig. 4 Transcriptional and translational effects of pharmacological inhibition of METTL1.
a, SYBR Green loading control for Northern blots presented in Fig. 3c. b, Western blot analysis of CDK4, and β-actin, following a 7-day treatment with vehicle, 1, 2.5 or 5 µM STM9005 in a panel of cancer models. c, d, qPCR analysis of CDK4 and KDM1A levels for the cell lines included in Fig. 3d (c), or CDK4 levels for the cell lines included in b (d) following 7 days of treatment with vehicle, 1, 2.5 or 5 µM STM9005; mean ± s.e.m. (n = 3 independent samples); p calculated using two-way ANOVA. e,h, Rescue of phenotype in HL-60 cells, following overexpression of Arg-TCT-tRNA (e) or in primary murine NPM1c/Flt3ITD/+ AML cells, following overexpression of Cdk4 (h) versus an EMPTY vector control. Left: Dose-response curves, following a 6-day treatment with STM9005. Right: Direct comparison of relative proliferation of cells, treated with STM9005 at indicated concentrations; bars show mean ± s.d. (n = 3 independent cultures). f,g, qPCR validation of Arg-TCT-tRNA-overexpression compared to EMPTY control in B16F10 (f) and HL-60 (g) cells; mean ± s.d. (n = 3 independent samples); for e-h p was calculated using an unpaired two-tailed Student’s t-test. i,j, Western blot validation of Cdk4-overexpression compared to EMPTY-TY1 vector in B16F10 and NPM1cA+/Flt3ITD/+ cells, at indicated time points. k,l, Validation of METTL1/Mettl1-overexpression in A549 cells (k) and mPDAC1 cells (l), shown in Fig. 3g and h, respectively. Left: Western blot validation of overexpression compared to EMPTY-TY1 vector. Right: Sanger sequencing of transcripts amplified from cells with WT or mutant METTL1/Mettl1-overexpression compared to reference sequence. m, Western blot analysis of KDM1A, CDK4 and β-actin levels in A549 cells harbouring EMPTY-TY1, METTL1WT- or METTL1MUT-overexpression 4 and 7 days post-treatment with vehicle, 1, 2.5 or 5 µM STM9005.
Extended Data Fig. 5 Impact on translation after pharmacological inhibition of METTL1.
a, Pearson correlation analysis of the change in transcription (RNA-seq) and translation (Ribo-seq) for genes in OE21 cells treated with vehicle of 1 µM STM9005. Points are coloured according to their translation efficiency change class. Related to Supplementary Table 6. b, Log2 fold change in ribosome occupancy at A-site codons in OE21 cells following 1- or 3-day treatment with vehicle or 1 µM STM9005, grouped by the codon’s association with either m7G-modified or non-m7G-modified tRNAs; mean (n = 3 independent sample), p calculated using unpaired two-tailed Welch’s t-test. Related to Supplementary Table 7. c, Detailed breakdown of the data represented in b. Purple stars represent significant increase (p < 0.05, one-tailed Student’s t-test). m7G and non-m7G tRNA labels are coloured green and blue, respectively. d, Western blot analysis of METTL3, WDR4 and β-actin levels in MV-4-11 and OCI-AML3 cells 4 and 7 days post-treatment with vehicle, 1, 2.5 or 5 µM STM9005 (3 independent experiments). e,f, Overlap between genes downregulated in Ribo-seq upon treatment of OE21 with 1 µM STM9005 and proteins downregulated in SILAC-based proteomics analysis of LNZ308 cells, harbouring shMETTL111 (e) and between genes downregulated in Ribo-seq upon treatment of OE21 with 1 µM STM9005 and proteins upregulated in SILAC-based proteomics analysis of MEF cells, harbouring METTL1/WDR4-overexpression11 (f); p calculated using one-tailed hypergeometric test. Related to Supplementary Table 8.
Extended Data Fig. 6 The impact of STM9005 on cancer proliferation pathways.
a, Gene ontology analysis of differentially expressed transcripts after in vitro treatment with either vehicle or 1 µM STM9005 for 3 days in OE21 (left) and LoVo (middle), and in vivo treatment twice daily with either vehicle or 30 mg/kg STM9005 for 5 days in B16F10 (right). Related to Supplementary Tables 9–14. b,c, Western blot (b) and qPCR (c) analysis of MYC levels in A549 and OE21 cells following 4 or 7 days of treatment with vehicle, 1, 2.5 or 5 µM STM9005; mean ± s.e.m. (n = 3 independent samples); p calculated using two-way ANOVA. d,e, Rescue in A549 (d) or mPDAC1 (e) cells, following an overexpression of MYC/Myc versus an EMPTY-TY1 control. Normalized confluency measured by quantification of methylene blue staining of cells harbouring MYC/Myc overexpression or an EMPTY-TY1 control vector, treated with vehicleorSTM9005 at indicated concentrations, following 6 days of treatment; mean ± s.d. (n = 4 independent cultures); p calculated using a 2-way ANOVA with multiple comparisons. Representative images shown under the corresponding bar. f,g, Western blot validation of MYC/Myc-overexpression compared to EMPTY-TY1 vector in A549 (f) or mPDAC1 (g) cells, at days 0 and 12 post-treatment (3 independent experiments). h, qPCR analysis of METTL1/Mettl1 levels in a panel of cell lines following 7 days of treatment with vehicle, 1, 2.5 or 5 µM STM9005; mean ± s.e.m. (n = 3 independent samples); p calculated using one-way ANOVA.
Extended Data Fig. 7 Treatment with STM9005 in vivo shows no measurable adverse signs.
a, Schematic representation of the experiment shown in Fig. 4a–d. b, Tumour volume over time in animals transplanted with B16F10 melanoma cells, treated twice daily with either vehicle (n = 6) or 30 mg/kg STM9005 (n = 9). Green line indicates ongoing treatment. c, Schematic representation of the experiment shown in Fig. 4e–g. d, Schematic representation of the experiment shown in Fig. 4h–k. e-g. Weight measurements of animals transplanted with mPDAC1 (e), A549 (f) cells or C57BL/6J mice (g) and treated twice daily with either vehicle or 30 mg/kg STM9005. Data presented,as box plots showing median, interquartile range, and extremes (whiskers extend from minimum to maximum values in all the box plots presented); significance was calculated using multiple unpaired two-tailed Student’s t-tests. ns, not significant; for e n = 5 vehicle-treated animals, and n = 9 STM9005-treated animals; for f n = 5 animals in each group; for g n = 4 vehicle-treated animals, and n = 5 STM9005-treated animals. h, Schematic representation of the experiment shown in Fig. 4l. i, Schematic representation of the experiment shown in Fig. 4m,n. j-o, Weight measurements (j), white blood counts (k), red blood counts (l), platelet counts (m), haemoglobin levels (n) and haematocrit levels (o) of AML-491 PDX animals, treated twice daily with either vehicle (n = 5; day 40/41 n = 4) or 30 mg/kg STM9005 (n = 8; day 40 n = 7); mean ± s.d.; significance was calculated using multiple unpaired two-tailed Student’s t-tests. ns, not significant. The schematics were created using BioRender; Yang, G. https://BioRender.com/qvhcwzs (a), https://BioRender.com/j8zvej0 (c), https://BioRender.com/5f5d61l (d), https://BioRender.com/75vah3j (h), https://BioRender.com/khb7w5n (i) (2026).
Extended Data Fig. 8 Treatment with STM9005 illustrates no signs of acute or lasting toxicity on normal cells.
a,b, Colony-forming efficiency of murine WT haematopoietic bone marrow cells (a) or CD34+ human cord blood cells (b) upon treatment with vehicle or STM9005 at indicated concentrations; mean ± s.d.; for a n = 4 individual animals, p calculated using 2-way ANOVA; for b, n = 4 individual donors, each marked with unique symbol, p calculated using one-way ANOVA. c, Relative viability of primary human hepatocytes upon treatment with vehicle, 1.25, 2.5 or 5 µM STM9005 or doxorubicin; mean ± s.d. (n = 6 biopsies from n = 2 individual patients, each marked with a unique symbol); p calculated using 2-way ANOVA. d, Schematic representation of experimental design to study the effects of STM9005 on long-term haematopoiesis in C57BL/6J animals, related to e-s. Created using BioRender; Yang, G. https://BioRender.com/n8zeag0 (2026). e-g, Weight measurements of spleen (e), liver (f) and kidneys (g) from h, Bone marrow cellularity. i, Representative images of colony forming assays with bone marrow cells from h. j, Colony-forming efficiency of primary murine WT haematopoietic bone marrow cells. k-n, White blood (k), red blood (l), platelet counts (m) and haemoglobin levels (n) of C57BL/6J animals, treated twice daily with vehicle or 30 mg/kg STM9005. o,p, Proportion of B, T and myeloid cells in the bone marrow (o) or spleen (p). q-v, Frequency of Lin− (q), LK (r), MEP, GMP and CMP (s), LSK (t), LMPP, MPP and HSC (u) and ST-HSC, LT-HSC, MPP2 and MPP3 (v) cells in live bone marrow. For e-s animals were treated twice daily with vehicle (n = 5) or 30 mg/kg STM9005 (n = 7); mean ± s.d.; for e-h,j-n,q,r,t p was calculated using unpaired two-tailed Student’s t-test; for o,p,s,u and v q was calculated using multiple unpaired two-tailed Student’s t-tests.
Extended Data Fig. 9 Treatment with STM9005 has no effects on normal tissue histology.
a,b, Histology sections of heart, liver, kidney and spleen from vehicle-treated C57BL/6J animals (a) or from animals treated 12 times twice-daily with 30 mg/kg STM9005 (b); n = 5 animals; scale bar = 250 µm.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Yankova, E., Sapetschnig, A., Thomas, B. et al. Pharmacological inhibition of METTL1 as a therapy against solid and blood cancers. Nature (2026). https://doi.org/10.1038/s41586-026-11178-z
Received: 23 July 2025
Accepted: 17 September 2026
Published: 28 October 2026
Version of record: 28 October 2026
DOI: https://doi.org/10.1038/s41586-026-11178-z
来源:Nature 综合主刊 · nature.com