ARTICLE https://doi.org/10.1038/s41467-021-26258-z OPEN Inhibition of CBP synergizes with the RNA- dependent mechanisms of Azacitidine by limiting protein synthesis Jeannine Diesch1,2, Marguerite-Marie Le Pannérer 1, René Winkler 1, Raquel Casquero1, Matthias Muhar3, Mark van der Garde4,5, Michael Maher1, Carolina Martínez Herráez 6,7, Joan J. Bech-Serra1, Michaela Fellner3, Philipp Rathert 3,11, Nigel Brooks8, Lurdes Zamora9, Antonio Gentilella 6,7, Carolina de la Torre 1, Johannes Zuber 3,10, Katharina S. Götze4 & Marcus Buschbeck 1,2 ✉ 1234567890():,; The nucleotide analogue azacitidine (AZA) is currently the best treatment option for patients with high-risk myelodysplastic syndromes (MDS). However, only half of treated patients respond and of these almost all eventually relapse. New treatment options are urgently needed to improve the clinical management of these patients. Here, we perform a loss-of- function shRNA screen and identify the histone acetyl transferase and transcriptional co- activator, CREB binding protein (CBP), as a major regulator of AZA sensitivity. Compounds inhibiting the activity of CBP and the closely related p300 synergistically reduce viability of MDS-derived AML cell lines when combined with AZA. Importantly, this effect is specific for the RNA-dependent functions of AZA and not observed with the related compound deci- tabine that is only incorporated into DNA. The identification of immediate target genes leads us to the unexpected finding that the effect of CBP/p300 inhibition is mediated by globally down regulating protein synthesis. 1 Cancer and Leukaemia Epigenetics and Biology Program, Josep Carreras Leukaemia Research Institute (IJC), Campus ICO-GTP-UAB, Badalona, Spain. 2 Program for Predictive and Personalized Medicine of Cancer, Germans Trias i Pujol Research Institute (PMPPC-IGTP), Badalona, Spain. 3 Research Institute of Molecular Pathology (IMP), Vienna BioCenter (VBC), Vienna, Austria. 4 Department of Medicine III, Klinikum rechts der Isar, Technische Universität München, Munich, Germany. 5 German Cancer Consortium (DKTK), Partner Site Munich, Heidelberg, Germany. 6 Laboratory of Cancer Metabolism, ONCOBELL Program, Institut d’Investigació Biomèdica de Bellvitge (IDIBELL), L’Hospitalet de Llobregat, Spain. 7 Faculty of Pharmacy, Department of Biochemistry and Physiology, Universitat de Barcelona, Barcelona, Spain. 8 CellCentric, Ltd, Chesterford Research Park, Little Chesterford, Cambridge CB10 1XL, UK. 9 Hematology Laboratory Service, ICO Badalona‐Hospital Germans Trias I Pujol, Josep Carreras Leukemia Research Institute (IJC), Badalona, Spain. 10 Medical University of Vienna, Vienna BioCenter (VBC), Vienna, Austria. 11Present address: Biochemistry Department, University Stuttgart, Stuttgart, Germany. ✉email: mbuschbeck@carrerasresearch.org NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 1 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z M yelodysplastic syndromes (MDSs) are a heterogeneous regulating properties can be attributed to three main functions. group of clonal hematopoietic stem cell diseases char- First, they bridge sequence-specific DNA-binding proteins to the acterized by dysplasia and ineffective haematopoiesis basal transcription machinery. Second, they act as transcriptional affecting one or more myeloid cell lineages1. MDSs are part of a co-activators of a large number of transcription factors. Finally, spectrum of myeloid diseases that evolve from asymptomatic they are enzymes with lysine acetyltransferase activity that acet- clonal haematopoiesis with indeterminate potential and trans- ylate histone and non-histone proteins. In haematopoiesis, CBP form into secondary acute myeloid leukemia (sAML) in 30–40% and p300 contribute to cell differentiation and stem cell main- of cases2. Environmental or therapy-related factors accelerate the tenance through their interaction with key transcription factors development of MDS3,4. MDS are among the most frequent such as GATA130, MYB31 and AML-132. In AML and chronic haematological disorders in the elderly5 and sAML has particu- myelomonocytic leukemia (MLL) translocations of MLL to CBP larly poor prognosis6. The genetic alterations in MDSs are well are common, leading to gain-of-function of MLL33. Similarly, described and include mutations in genes involved in epigenetic MOZ-CBP and MOZ-p300 translocations can drive AML regulation and RNA processing7,8. leukemogenesis34. In addition, many oncogenes interact with Currently, the azanucleosides azacitidine (AZA) and decitabine CBP and p300, making it important factors in malignant have multiple clinical applications and are, for instance, the transformation27. The deregulation of CBP and p300 in blood alternative standard treatment options for high-risk MDS patients cancers has prompted the development of inhibitors35 and one of that are not eligible for an allogeneic transplant9. However, the these, CCS1477, has entered clinical development in an ongoing response is only partial: 40–50% of treated patients show hae- first in human trial in haematological cancers (NCT04068597). matological improvements and a complete response is limited to Here we report the results of a genetic loss-of-function screen as few as 10–15%9,10. Furthermore, AZA is also used for the of chromatin regulators that leads to the identification of CBP as supportive treatment of a subset of AML cases11. Azanucleosides a combinatorial drug target for AZA-based treatments in vitro. are frequently referred to as hypomethylating agents and epige- We dissect the synergistic interaction between AZA and phar- netic drugs, as they interfere with DNA methylation12. In additon macologic CBP/p300 inhibitors, and are able to link it to the to reducing global DNA methylation, azanucleosides are general RNA-dependent functions of AZA. Moreover, we demonstrate a nucleotide analogues that affect multiple cellular functions. After previously unrecognized but prominent interference with protein cellular uptake, azanucleosides are metabolically converted into synthesis by CBP/p300 inhibitors. their active forms and incorporated into newly synthesized DNA and RNA13. AZA and decitabine differ in that decitabine is exclusively incorporated into DNA, whereas 80–90% of AZA Results becomes part of RNA14. Incorporation into DNA leads to the An improved genetic loss-of-function screen identifies chro- formation of covalent adducts between DNA and DNA methyl- matin regulators that affect the cell-intrinsic response to AZA. transferases (DNMTs), particularly DNMT115, for which there The need for improvement of the response rate and duration of are two consequences. First, these adducts elicit the DNA damage AZA-based therapy has motivated us to search for potential response in an ataxia-telangiectasia mutated (ATM)/ataxia tel- combinatorial drug targets. As AZA is a chromatin-modifying angiectasia and Rad3-related protein (ATR)-dependent manner agent11, we set out to determine the effect of inhibiting chromatin and, if not properly repaired, cause cytotoxicity16. Second, regulators on treatment response in vitro. We have opted for an the inhibition of the cross-linked DNMTs leads to global RNA interference-mediated knockdown approach that is more DNA hypomethylation that induces the expression of a subset similar than a knockout approach to the level of inhibition that of genes promoting cell differentiation17. In addition, the can be expected to be reached with drugs in the clinic. As a hypomethylation-induced transcription of endogenous retro- cellular model, we have chosen the MDS-derived sAML cell line viruses mimics a viral infection and activates the innate immune SKK-1 that has been well characterized in our laboratory36. SKK- response18. 1 cells have been isolated from a patient with MDS-derived sAML The incorporation of AZA into RNA has additional effects. and harbour mutations in splicing factor U2AF1 and the chro- Analogous to the inhibition of DNMTs, AZA inhibits a subset of matin regulator BCOR37. To test chromatin regulators in a most RNA methyltransferases that methylate position 5 of cytosine in comprehensive manner, we have generated a lentiviral library transfer RNAs (tRNAs) and other RNAs19,20. AZA sensitivity is targeting 912 human chromatin regulatory genes with 8 inde- best illustrated for DNMT221. As a consequence, AZA but not pendent short hairpin RNAs (shRNAs) per gene (Supplementary decitabine inhibits protein synthesis22 and this is at least in part Data 1 and Fig. 1a). shRNAs were embedded in an optimized mediated by tRNAs23. AZA also impacts on messenger RNA microRNA backbone, allowing efficient knockdown of target (mRNA) stability. In this context, AZA treatment leads to the genes after single-copy integration38,39. To ensure efficient downregulation of ribonucleotide reductase by destabilizing the shRNA expression and prevent transgene repression, we further mRNA encoding an essential component of the enzymatic optimized the previously described lentiviral pRRL-SFFV-GFP- complex24. Thereby, AZA reduces the conversion of ribonu- miRE-PGK-Puro (SGEP) backbone38 by inserting a ubiquitous cleotides to deoxyribonucleotides required for DNA synthesis, chromatin opening element (UCOE, Fig. 1b). The UCOE origi- which includes its own conversion to decitabine24. Furthermore, nates from the CBX3 promoter and has anti-silencing AZA disrupts the RNA-dependent interaction of the complex properties40. We have tested the functionality of the modified containing hnRNPK, DNMT2 and NSUN3 with Polymerase 2 on vector, termed cSGEP, by confirming the cytotoxic effects of actively transcribed genes25. Taken together, these additional shRNAs targeting essential genes (Supplementary Fig. S1a). RNA-dependent functions of AZA provide part of an explanation Specifically, we have targeted the transcriptional coactivator for why AZA and decitabine cause largely non-overlapping BRD4, the hematopoietic transcription factor MYB and the core changes in mRNA levels22,26. sub-unit of RNA polymerase 2 RPA3, which are known to be The CREB-binding protein (CBP) and its paralogue p300 are essential in AML cell lines. highly conserved transcriptional co-activators that are involved in For the loss-of-function screening, SKK-1 cells were infected many physiological processes, such as proliferation, differentia- with the cSGEP shRNA library at low viral titre to favour single- tion and apoptosis27,28. CBP and p300 have a high degree of copy integration (Fig. 1c), puromycin selected and half of the cells homology and their functions are largely redundant29. Their were treated for 2 weeks with partial lethal doses of AZA that we 2 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE a hEPI9 library: 912 chromatin regulators, x8 shRNAs b Nucleid Acid Binding HNRPA2B1 CBX3 Transcription factor alternative Transferase Exon1 Exon1 Exon1 Hydrolase Ligase Enzyme modulator Signaling molecule 674bp Other 0 10 20 30 40 50 cSGEP UCOE SFFV GFP PGK Puro Composition [%] c Pooled library e Genes required for survival Genes inhibit survival 14 days in presence of AZA in presence of AZA Transduction Puro 4 CREBBP NAA15 selection SKK-1 Untreated gDNA sequencing log2FC≥1and 3 -log10(PValue) p-val<0.01 AZA HMGA2 d SKK-1 log2FC≤-1 and CBX3 p-val<0.01 100 2 80 % live cells 60 Replicate1 1 40 Replicate2 Replicate3 EP300 20 0 0 0 2 4 6 8 10 12 14 -2 -1 0 1 2 AZA treatment [days] log2FC f g h SKK-1 SKK-1 100 *** 100 *** 100 *** 80 80 80 % live cells % live cells % GFP 60 60 60 40 cSGEP 40 40 cSGEP shNAA15-1 shCBP-1 20 shNAA15-2 20 20 shCBP-2 0 0 0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 AZA[ M] AZA[ M] AZA[ M] Fig. 1 Chromatin regulators affecting AZA sensitivity are identified by an optimized loss-of-function shRNA screen. a Composition of the hEPI9 library. b Illustration of the human HNRPA2B1-CBX3 locus including the ubiquitous chromatin opening element (UCOE) and schematic of the cSGEP backbone vector containing the UCOE, spleen focus-forming virus (SFFV) promoter, green fluorescent protein (GFP) marker gene, shRNA integration site, phosphoglycerate kinase (PGK) promoter and puromycin (Puro) resistance gene. c Schematic workflow of the shRNA screen-targeting genes encoding chromatin regulators. d Percentage of live SKK-1 cells determined at different time points throughout the shRNA screen after initiation of AZA treatment (0.25 µM every 2 days for 2 weeks) in three biological replicates. e Volcano plot of genes corresponding to enriched (“Genes inhibit survival in presence of AZA”) or depleted (“Genes required for survival in presence of AZA”) shRNAs. Selected genes are named. FC, fold change. CREBBP, P = 0.0001; NAA15, P = 0.0001; HMGA2, P = 0.0032; CBX3, P = 0.0061; EP300, P = 0.7405. f Percentage of live SKK-1 cells stably expressing cSGEP, shNAA15-1 or shNAA15-2 determined after 4 days of treatment with indicated concentrations of AZA. g, h SKK-1 cells stably expressing cSGEP, shCBP-1 or shCBP-2 were treated for 4 days with indicated concentrations of AZA and the percentage of live cells (g) or the percentage of GFP-positive cells was assessed (h). f–h Data represent the mean ± SEM of four independent experiments. Statistical analysis was performed using two-way ANOVA. ***p-value < 0.0001. Source data are provided as a Source Data file and in Supplementary Data 1 and 3. previously determined in dilution series (Supplementary Fig. S1b). gene encoding for the N-α-acetyltransferases, NAA15, which is a Treatment of cells led to a rapid reduction in the percentage of component of the NatA complex binding to ribosomes and co- live cells after 4 days followed by subsequent recovery, suggesting translationally acetylating proteins41. the expansion of cell clones with a growth and survival advantage Depleted hairpins were of particular interest to us, as they (Fig. 1d). To identify genes affecting the response to AZA, we indicate potential combinatorial drug targets. Thus, we selected have compared the abundance of all shRNA clones in control and the top depleted hit encoding for CBP for further analysis and treated cells by massive parallel sequencing of the transgenes in included the top enriched gene NAA15 as control. To validate the genomic DNA purified from cell populations at the end of the results of the shRNA screen, we generated polyclonal SKK-1 cell treatment. To identify relevant genes, we only considered those lines with stable single knockdowns of CBP or NAA15 selecting genes for which at least five of the eight shRNAs showed the same two independent shRNAs per gene (Supplementary Fig. S1c, d). trend and calculated their average fold change (Supplementary We assessed viability of these cells after 4 days of treatment with Data 3). Using this approach, we were able to identify 451 genes, various concentrations of AZA ranging around the previously whose shRNAs were significantly depleted, and 252 genes, whose calculated half maximal inhibitory concentration (IC50) (Supple- shRNAs were enriched (Fig. 1e). A decrease of shRNAs in the mentary Fig. S2a). In line with the results of the shRNA screen, treated cells indicated that the knockdown of the corresponding we observed a pronounced increase in viability in the NAA15- genes was not tolerated, and that the gene products were required deficient cells reflecting decreased sensitivity to the drug (Fig. 1f). for survival in the presence of AZA treatment. Among the top CBP-deficient cell lines showed the opposite behaviour with hits were CREBBP encoding for the transcriptional coactivator decreased cell viability in the presence of AZA (Fig. 1g). This CBP. Conversely, genes associated with increased sensitivity to increased sensitivity of cells with the CBP knockdown was also AZA were presented by enrichment of their shRNAs, as their visible when monitoring the percentage of infected and non- knockdown reduced the sensitivity and thus increased the infected cells among viable cells grown together. In such a survival of the cells in the presence of AZA. This included the competitive growth condition, CBP knockdown cells depleted NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 3 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z a SKK-1 b MOLM-13 100 *** 100 *** 80 80 ± ± % live cells % live cells DMSO 60 DMSO 60 ++ ++ 5 M C646 ++ 5 M C646 40 ++ 7.5 M C646 40 ++ 7.5 M C646 20 ++ Synergism 20 ++ Synergism ± Nearly additive ± Nearly additive 0 0 0 2 4 6 8 0 2 4 6 8 AZA[ M] AZA[ M] c MOLM-13 0.5 μM A-485 20 μM iCBP-112 0.75 μM CCS1477 0.5 μM JQ-1 10 nM ABT-199 100 *** 100 *** 100 *** 100 * 100 *** ++ ++ 80 ++ 80 ++ 80 ++ 80 ± 80 ++ ++ ++ % live cells ++ ++ 60 ++ 60 ++ 60 ++ ++ 60 ± 60 ++ ++ 40 ++ 40 DMSO 40 40 40 Combination 20 20 20 20 20 ++ Synergism ± Nearly additive 0 0 0 0 0 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 8 0 2 4 6 AZA[ M] AZA[ M] AZA[ M] AZA[ M] AZA[ M] d e Leukemic blast cells F-36P MDS-L Patient 1 Patient 2 Patient 3 100 *** 100 *** 100 100 100 80 ± 80 ++ % apoptotic cells 80 80 80 ± % live cells ++ ++ 60 60 ++ 60 60 60 DMSO 40 40 ± 7.5 M C646 40 40 40 20 20 ++ Synergism 20 20 20 ± Nearly additive 0 0 0 0 0 0 2 4 6 8 0 5 10 15 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 AZA[μM] AZA[μM] AZA [ M] AZA [ M] AZA [ M] DMSO CCS1477 A485 Fig. 2 Inhibition of CBP/p300 increases sensitivity towards AZA in AML cells. a, b Percentage of live SKK-1 (a) or MOLM-13 cells (b) determined by DAPI/MitoTracker staining after 4 days of treatment with indicated concentrations of AZA in combination with C646 or 0.075% DMSO as vehicle control. c Percentage of live MOLM-13 cells after 4 days of treatment with different concentrations of AZA in combination with indicated concentrations of the CBP/p300 inhibitors A-485, iCBP-112 and CCS1477, the BET bromodomain inhibitor JQ-1, the BCL-2 inhibitor ABT-199 or 0.03% DMSO as vehicle control. d Percentage of live F-36P or MDS-L cells after 4 days of treatment with indicated concentrations of AZA in combination with C646 or DMSO (0.075%) as vehicle control. e Percentage of apoptotic leukemic blast cells from three AML patients determined by Annexin V staining after 3 days of treatment with different concentrations of AZA and 2 days with 500 nM CCS1477 or 8 µM A-485, or 0.08% DMSO as vehicle control. a–d Combination index (CI) was calculated according to the Chou–Talalay method43. ++ synergism (CI: 0.3–0.85), ± additive effect (CI: 0.9–1.1). Data represent the mean ± SEM of four independent experiments. Statistical analysis was performed using two-way ANOVA. *p-value < 0.05; ***p-value < 0.0001. Source data are provided as a Source Data file. more rapidly in the presence of AZA than control cells (Fig. 1h). indicated a synergistic interaction between both compounds Accordingly, we observed an increase in apoptosis in CBP (Fig. 2a). This was also reflected in a substantial increase in the knockdown cells treated with AZA (Supplementary Fig. 1e). number of apoptotic cells when treating with a combination of Taken together, employing a genetic loss-of-function screen of both drugs (Supplementary Fig. S3a). C646 treatment also further chromatin regulators allowed the identification of several genes increased the response of CBP-depleted cells, suggesting that the affecting the sensitivity to AZA in SKK-1 cells. Specifically, we inhibition of p300 or residual CBP contributed to the observed identified the histone acetylase CBP as a potential drug target for effect (Supplementary Fig. S3b). combinatorial treatments. We next assessed the effects of CBP/p300 inhibition on AZA response in an additional cell line. MOLM-13 is another MDS- Inhibition of CBP/p300 increases sensitivity towards AZA. To derived sAML cell line but carries different driver mutations than further validate the impact of CBP on AZA sensitivity, we took SKK-136. These include some genetic alterations that are more advantage of the availability of small compound inhibitors of prevalent in other AML subtypes such as FLT3 internal tandem CBP/p300. The molecule C646 was the first inhibitor of CBP and repeats, an MLL-AF9 fusion and mutations in CBL, KMT2A and p300 to be developed, and functions by competing with the co- NF1 genes. As MOLM-13 cells are more commonly used in in vitro substrate acetyl-CoA for binding to the enzymatic KAT experiments and their characteristics are more comparable to the domain42. We first determined the IC50 for C646 in SKK-1 cells ones seen in MDS and sAML patients, we decided to further focus on (Supplementary Fig. S2a). To assess the combinatorial effect of this cell line. First, we determined the IC50 for AZA and C646 in cell C646 with AZA, cells were treated with a range of AZA con- viability assays (Supplementary Fig. S2b). Similar to the results in centrations alone or in combination with C646 at a concentration SKK-1 cells, we observed a significant and synergistic reduction in close to its IC50 or slightly below. Similar to the effect of CBP cell viability when combining C646 and AZA (Fig. 2b). Again, this knockdown, we could observe a significant reduction in cell was also reflected by increased apoptosis in response to the viability in response to the combination of C646 with AZA in combinatorial treatment compared to the individual treatments comparison to AZA treatment alone (Fig. 2a). Calculating the (Supplementary Fig. S3c). Next, we tested additional CBP/p300 combination index (CI) using the Chou–Talalay method43 inhibitors in combination with AZA in MOLM-13 cells. For this, we 4 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE chose the potent and highly specific KAT domain inhibitor A-48544 genes (Fig. 3d) (Supplementary Data 5). The inhibition of genes and the CBP/p300-specific bromodomain inhibitors iCBP-112 and involved in protein translation and the upregulation of chaperone CCS147745,46. Importantly, CCS1477 is the first inhibitor of CBP/ genes could be confirmed by real-time quantitative PCR (RT- p300 to be tested first in human clinical trials in AML, multiple qPCR) after 4 h of C646 treatment (Fig. 3e, f). Importantly, inhi- myeloma and non-Hodgkin lymphoma (NCT04068597), and in bition of CBP/p300 using another more specific inhibitor, A-485, castration-resistant prostate cancer (NCT03568656)47. We further confirmed the downregulation of genes involved in protein trans- included the more general Bromodomain and extraterminal domain lation (Fig. 3g). The inhibition of a subset of genes in response to (BET) inhibitor JQ-1, which inhibits among others the bromodo- C646 could be further validated in SKK-1 cells (Fig. 3h). mains of downstream effectors BRD2 and BRD448. As a benchmark Taken together, our results suggest that in MDS-derived sAML and reference, we used the BCL-2 inhibitor ABT-199 (venetoclax)49, cells, CBP/p300 promotes the expression of genes involved in which is US Food and Drug Administration (FDA) approved, in protein synthesis and cellular homoeostasis. combination with AZA or decitabine for the treatment of AML in adults of 75 years or older (https://www.fda.gov/drugs/resources- information-approved-drugs/fda-grants-regular-approval-venetoclax- Inhibition of CBP/p300 limits protein synthesis. Through its combination-untreated-acute-myeloid-leukemia). We determined the incorporation in RNA, AZA has previously been shown to inhibit IC50 values for these compounds in MOLM-13 cells (Supplementary protein synthesis13. This led us to wonder whether the synergistic Fig. S2c). Combination of AZA with concentrations of A-485, iCBP- effect of CBP/p300 inhibition with AZA in cell viability could be due 112 and CCS1477 close to their IC50 demonstrated a significant to potentiating the RNA-dependent function of AZA and by limiting decrease in cell viability and synergy with AZA (Fig. 2c). Notably, the protein synthesis to a level that is not tolerable for leukemia cells. In combination of AZA with CBP inhibitors was more effective than the order to test this intriguing hypothesis, we analysed the combination combination with the BET inhibitor JQ-1 and in a similar range as of the CBP/p300 inhibitor C646 with the AZA analogue decitabine. In contrast to AZA, its analogue decitabine is not incorporated into the combination with ABT-199 (Fig. 2c). We further confirmed the RNA and hence does not affect protein synthesis22. Indeed, we found increased response to AZA in the two additional sAML cell lines, that the combination of C646 and decitabine on cell viability of F36-P and MDS-L, co-treated with C646 (Fig. 2d). Next, we used MOLM-13 and SKK-1 cells was neither additive nor synergistic and, primary AML blasts isolated from three independent patients in the case of SKK-1 cells, even strongly antagonistic (Fig. 4a). To (patient characteristics in Supplementary Data 2) and treated them make sure that SKK-1 and MOLM-13 cells lines are responsive to with A-485 and the clinically more relevant inhibitor CCS1477. In all decitabine treatment, we tested its inhibitory effect on LINE-1 DNA three cases, the combination was more effective than the AZA mono- methylation (Supplementary Fig. S4). treatment (Fig. 2e). Next, we wished to better understand to which extent CBP/ In summary, these results indicate that inhibitors of CBP and p300 inhibition affects protein synthesis. Similar to SLAM-seq for p300 increase the response of AML cells to AZA. Remarkably, the detecting mRNA, we used metabolic labelling to identify and combination reaches a similar synergy in cell lines as the FDA- quantify newly synthesized proteins by mass spectrometry (MS, approved combination of AZA and ABT-199. Fig. 4b). Specifically, we have chosen a short treatment with C646 of 4 h to avoid secondary effects such as those from apoptotic SLAM-seq identified direct target genes of CBP/p300 relevant signalling. As shown in Fig. 4c, we observed a prominent for protein synthesis. Although AZA has been widely used and reduction of the total number of detected nascent proteins from studied during the last decades, CBP/p300 inhibitors have only 2807 to 638 (average of replicates) by CBP/p300 inhibition recently gained clinical interest and the biological consequences of (Supplementary Data 6). Only 160 proteins were detected in the inhibiting CBP/p300 are not fully known. To further explore the background control without labelling and were excluded from effects of CBP/p300 inhibition and to understand the mechanism further analysis, with the exception of a few proteins that were causing the synergistic effect of AZA and C646, we set out to detected at 10-fold or higher level in untreated samples. identify the direct target genes of CBP/p300. For this, we quantified Furthermore, we decided to focus our analysis on the 2423 newly synthesized mRNA after short-term inhibition of CBP/p300 proteins that were detected in all DMSO control samples. The and metabolic labelling of mRNA with a method termed number of peptide spectrum matches (PSMs) per protein was thiol(SH)–linked alkylation for the metabolic sequencing of RNA reduced after CBP/p300 inhibition, indicating a reduction in (SLAM-seq)50. Specifically, we analysed MOLM-13 cells treated or abundance of detected proteins (Fig. 4d). Using a reduction in not treated with C646 for 2 h coinciding with metabolic labelling PSMs of 75% or more as cutoff indicated that 2125 proteins, during the last hour (Fig. 3a). Calling differentially expressed genes corresponding to 68% of all proteins detected in DMSO samples, from reads of newly synthesized transcripts defined by T > C con- were affected by CBP/p300 inhibition (Fig. 4e). When comparing versions, we were able to detect a prominent and broad repression the SLAM-seq (Fig. 3b) and proteomics results, we found partial of nascent transcripts in treated cells (Fig. 3b, right panel) (Sup- overlap of 29% and 28% of the proteomics and SLAM-seq hits, plementary Data 4). This was not detectable when analysing all respectively, between newly synthesized mRNA transcripts and reads and plotting global mRNA levels (Fig. 3b, left panel) (Sup- corresponding newly synthesized proteins following CBP/p300 plementary Data 4). After applying stringent cutoffs (false disovery treatment (Fig. 4f). This suggests that the impact of CBP/p300 rate (FDR) ≤ 0.01, log2FC ≤ −1/log2FC ≥ 1) to the newly synthe- inhibition on protein synthesis cannot be solely attributed to a sized transcripts, we identified 220 upregulated genes and 2253 reduction of mRNAs but rather to a global downregulation of the downregulated genes in C646-treated cells compared to the dime- protein translation machinery. Closer inspection of the list of thyl sulfoxide (DMSO) control. The strong over-representation of affected proteins indicated that the reduction of mRNAs downregulated genes was in line with the role of CBP/p300 as encoding components of the translation machinery was also transcriptional coactivator. By performing functional Gene Ontol- reflected on the protein level, which could be confirmed by Gene ogy analysis on the downregulated genes, we found an enrichment Ontology analysis of affected proteins, demonstrating an enrich- of genes that are important for RNA biogenesis and protein ment of proteins involved in mRNA processing and splicing synthesis (Fig. 3c). Gene ontologies related with protein folding, (Fig. 4g and Supplementary Data 7). heat shock proteins and chaperone binding and protein stabiliza- Taken together, our results show that CBP/p300 inhibition has tion were enriched among the tenfold smaller set of upregulated a pronounced impact on protein synthesis that is at least in part a NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 5 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z a C646/ 4sU collect and DMSO label sequence RNA MOLM-13 0 1 2 hours b All reads SLAM-seq reads 10 220 up 10 2253 down log2FC (C646/DMSO) 5 5 0 0 -5 -5 0 2 4 Density 0 2 4 Density log10RPM log10RPM c Down-regulated genes d Up-regulated genes Structural constituent of ribosome Response to unfolded protein Mitochondrial translational elongation Unfolded protein binding Mitochondrial translational termination Protein processing in ER Structural constituent of ribosome Heat shock protein binding Translation Protein refolding Ribosome 0 2 4 6 8 10 12 0 2 4 6 8 10 12 14 16 18 -log10(pvalue) -log10(pvalue) e f * 1.00 40 Relative expression Relative expression 0.75 * 30 * * * 0.50 * * * * * 20 * DMSO DMSO 0.25 C646 10 C646 0.00 0 LU 1 P9 M P1 P O 1C M 26 PO U1 3H PO 1A PO 1 B 1E H PS C R R N LR LR LR LR S LR R R AL PI R PO R SE Translation genes Chaperones g h SKK-1 1.00 1.00 Relative expression Relative expression * * * * 0.75 * * * 0.75 * * * 0.50 0.50 DMSO A-485 DMSO 0.25 0.25 C646 0.00 0.00 P9 M P1 P O 1C M 26 PO U1 3H P9 P1 PO 1A PO 1 B 1C 1E 26 1A PS PS R R R R LR LR LR LR LR S LR LR R R R R AL R R PO PO PO M Translation genes Translation genes Fig. 3 Immediate target genes of CBP/p300 identified by SLAM-seq. a Schematic workflow of the SLAM-seq sample preparation protocol for the metabolic labelling of newly synthesized RNAs. MOLM-13 cells were treated with 10 µM C646 or DMSO, labelled with 4sU after 1 h, the RNA was collected after 2 h and subsequently sequenced. b Plots show the fold change (FC) in the abundance of total mRNA or newly synthesized mRNA in relation with their normalized baseline expression in reads per million (RPM). Genes significantly changed (FDR < 0.01) are highlighted in blue and red, respectively. c Top enrichment clusters of downregulated or d upregulated transcripts as determined by functional annotations clustering using DAVID68,69. e, f Relative expression of selected CBP/p300 target genes by RT-qPCR in MOLM-13 cells treated for 4 h with 10 µM C646 or 0.1% DMSO. g Relative expression of selected CBP/p300 target genes by RT-qPCR in MOLM-13 cells treated for 4 h with 1 µM A-485 or 0.01% DMSO. h Relative expression of selected CBP/ p300 target genes by RT-qPCR in SKK-1 cells treated for 4 h with 10 µM C646 or 0.1% DMSO. c, d Statistical analysis was performed by Fisher’s exact test. e–h Data represent the mean ± SEM of at least three independent experiments. Statistical analysis was performed by two-sided Student’s T-test compared to untreated samples, *p-value < 0.05. Source data are provided as a Source Data file and in Supplementary Data 4 and 5. 6 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE a MOLM-13 SKK-1 100 100 80 80 – % live cells % live cells 60 60 – – – 40 40 DMSO 20 20 5 M C646 – Antagonism 0 0 0 1 2 3 4 0 1 2 3 4 Decitabine [ M] Decitabine [ M] b c MOLM-13 Newly synthesised proteins C646/ met-free 3000 DMSO media AHA collect cells 2000 0 1.5 2 4 1000 Enrichment of newly synthesized proteins 0 1 2 3 1 2 l tro on Mass spectometry DMSO C646 C d Abundance e Newly synthesized proteins not affected Mean PSM [log2] 100 n=298 10 n=2125 1 affected ≥75% decrease (PSM) DMSO C646 f g Overlap proteomics-SLAMseq mRNA processing mRNA splicing Spliceosome Nucleotide-binding n=623 ATP-binding Cell cycle Cell division 2125 2253 Mitosis affected down-regulated 0 10 20 30 40 50 proteins mRNA -log(pvalue) Fig. 4 CBP/p300 inhibition reduces protein synthesis but does not synergize with decitabine. a Percentage of live MOLM-13 or SKK-1 cells after 4 days of treatment with indicated concentrations of AZA in combination with 5 µM decitabine or DMSO (0.05%) as vehicle control. Data represent the mean ± SEM of four independent experiments. Statistical analysis was performed using two-way ANOVA. Calculating the combination index (CI) indicated antagonism (CI > 1). b Schematic workflow of the detection of nascent proteins. MOLM-13 cells were treated with 10 µM C646 or 0.1% DMSO, the media was changed to methionine (met)-free media after 1.5 h, metabolic AHA label was added for 2 h and cells were collected after a total of 4 h of inhibitor treatment. Newly synthesized proteins were affinity purified following the ClickIT protocol. c The number of detected newly synthesized proteins in untreated and treated samples is plotted. An unlabelled sample served as negative control. d Violin plot presenting the mean peptide spectrum matches (PSM) of all detected proteins. e Pie chart indicating that the majority of detected proteins was affected as defined by a reduction in PSM of 75% or more. f Overlap between significantly decreased newly synthesized proteins (≥75% reduction in PSM) determined by ClickIT enrichment and significantly downregulated nascent mRNA (log2FC ≤ −1, FDR ≤ 0.1) determined by SLAM-seq. g Ontology analysis of significantly reduced proteins (≤75% reduction compared to untreated samples). Statistical analysis was performed by Fisher’s exact test. Source data are provided as a Source Data file and in Supplementary Data 6 and 7. NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 7 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z a Polysome profile in MOLM-13 b Global protein synthesis 0.25 80S CD34+ cells from healthy donors Polysome abundance 60 * median intensity 1.0 Absorbance (260nm) 0.20 40 ** ** ** ** 0.15 60S 20 0.5 0.10 40S Polysomes 0 DMSO 0.0 SO 6 64 0.05 C646 0 1 5 10 20 M C D C646 [ M] 0.00 c d Global protein synthesis Global protein synthesis MOLM-13 SKK1 Leukemic blast cells * * * * * * * * * * * * median intentsity 1.0 1.0 median intensity 1.0 0.5 0.5 0.5 0.0 0.0 0.0 - + - + - + - + AZA - + - + - + - + AZA - + - + - + - + AZA DM X DM X DM X CH CH CH SO 6 5 7 SO 6 CC 85 7 4 48 47 SO 6 CC 85 4 47 7 C6 C6 4 4 47 A- S1 A- C6 4 S1 A- S1 CC e CBP/p300 inhibition AZA DNA RNA Ribosomal genes defective tRNAs and mRNAs Protein synthesis Cell death Fig. 5 CBP/p300 inhibition impairs active protein translation, which can be further reduced by azacitidine. a Representative polysome profile in MOLM- 13 treated with 10 µM C646 for 4 h. Polysomal lysates were fractionated by ultracentrifugation on sucrose gradients and the ribosomal RNA content was monitored across the gradient by UV absorbance (260 nm). The polysome abundance was normalized to the baseline absorbance. Data represent the mean ± SEM of three independent experiments. Statistical analysis was performed by two-sided Student’s T-test compared to untreated samples, *p- value = 0.0267. b Median fluorescence intensity of CD34+ cells from healthy donors after 8 h of treatment with increasing concentrations of C646 followed by Protein Synthesis Assay. Data represent the mean ± SEM of five independent experiments. Statistical analysis was performed by unpaired two- sided T-test compared to untreated samples, **p-value < 0.001. c Median fluorescence intensity of MOLM-13 or SKK-1 cells treated for 6 h with DMSO (0.1%), C646 (10 µM), A-485 (1 µM or 2.5 µM) or CCS1477 (1 µM or 2.5 µM) in the absence or presence of AZA (2 or 1 µM) followed by Click-iT™ HPG Alexa Fluor™ 594 assay. Cells treated 1.5 h with cycloheximide (CHX) were used as positive control. d Representative images of median fluorescence intensity of leukemic blast cells from AML patients treated for 3 h with the indicated concentrations of DMSO (0.075%), C646 (7.5 µM), A-485 (1 µM) or CCS1477 (0.25 µM) in the absence or presence of 0.25 µM AZA followed by Click-iT™ HPG Alexa Fluor™ 594 assay. Cells treated 1.5 h with cycloheximide (CHX) were used as positive control. e Mechanistic model suggesting that the synergy between CBP/p300 inhibition and AZA is caused by the reduction of protein synthesis. c, d Data represent the mean ± SEM of at least three independent experiments. Statistical analysis was performed by two-sided Student’s T-test, *p-value < 0.05. Source data are provided as a Source Data file and in Supplementary Data 8. consequence of transcriptional downregulation of target genes cellular polysomal content as readout for the activity of the cel- encoding components of the translation machinery. lular transcriptome, which represents the protein synthetic capacity of the cell51. As shown in Fig. 5a (Supplementary CBP/p300 inhibition reduces active protein translation and Data 8), treatment of MOLM-13 cells with the CBP/p300 inhi- enhances the RNA-dependent functions of AZA. In order to bitor C646 strongly decreased the amount of poly-ribosomes further explore the effects of CBP/p300 inhibition on protein compared to untreated cells, and indicates a severe impact of synthesis, we assessed by sucrose gradient ultracentrifugation the C646 on global protein translation. To investigate this further, we 8 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE performed flow cytometric assays to globally monitor protein between CBP/p300 inhibition and decitabine, our results suggest synthesis. We were able to confirm the reduction of protein that CBP/p300 inhibition synergized with the RNA-dependent synthesis in response to CBP/p300 inhibition in primary hema- effects of AZA. Both AZA and decitabine are approved for the topoietic stem and progenitor cells isolated from the bone mar- treatment of haematological diseases including high-risk MDS. row of five healthy donors (Donor details in Supplementary Currently, there is no biomarker or patient stratification for Data 2). As shown in Fig. 5b, the treatment of these primary cells choosing one treatment over the other. Although there is no with C646 lead to a dose-dependent partial reduction in protein randomized clinical trial that has directly compared both drugs in synthesis. In MOLM-13 and SKK-1 cells, the impact on protein the same cohort of patients, several studies suggested that the synthesis was much more pronounced after treatment with any of overall response of patients to decitabine and AZA is highly the three CBP/p300 inhibitors (Fig. 5c). A less pronounced similar9,54–57. The partially different mechanisms of action might, reduction was visible in the cells stably depleted of CBP (Sup- however, become relevant when considering response predicting plementary Fig. S5). Furthermore, we were able to confirm the parameters, combinations with other drugs or mechanisms of inhibitory effect of AZA on translation (Fig. 5c). The combination resistance. Our study points at a beneficial use of CBP inhibitors of AZA with the CBP/p300 inhibitors C646, A-485 or CCS1477 primarily in combination with AZA and not decitabine. Others resulted in a significant further decrease in protein synthesis, in have recently shown that resistance to AZA was associated with some cases approaching the levels of a total shutdown achieved changes in RNA-dependent chromatin structures and an with cycloheximide (CHX, Fig. 5c). Similar to sAML cell lines, increased interaction of RNA Polymerase 2 with NSUN1 and Brd4, and loss of interactions with NSUN3 and DNMT225. leukemic blast cells isolated from the bone marrow of AML Importantly, these changes have been attributed to the RNA- patients displayed a strong sensitivity to CBP/p300 inhibitors on dependent functions of AZA and create a vulnerability of AZA- the level of protein synthesis, which was further increased in the resistant cells for Brd4 inhibition that would not be expected for presence of AZA (Fig. 5d). decitabine-resistant cells. Revisiting our screening results, the Taken together, CBP/p300 inhibition has a pronounced knockdowns of NSUN3 and Brd4 had also opposing effects albeit inhibitory influence on protein synthesis. Based on our data, we much less pronounced than those observed by knockdown of propose that this mechanism of action of CBP/p300 inhibitors CBP or NAA15 (see Supplementary Data 3). In support of the explains their synergy with the RNA-dependent effects of AZA study by Cheng at al.25, we found that knockdown of NSUN3 (Fig. 5e). increased AZA sensitivity. An unexpected and important finding of our study is that CBP/ Discussion p300 inhibition affects protein synthesis. As a possible explana- In our study, we examined the influence of chromatin regulators tion, we found that CBP/p300 positively regulates several genes on the cell-intrinsic response to AZA by performing a genetic encoding core components of the translational machinery, and loss-of-function screen in an MDS-derived sAML cell line. Using that CBP/p300 inhibition drastically reduced global protein this unbiased approach, we identified the N-α-acetyltransferase synthesis. Genes encoding chaperones dealing with unfolded NAA15 and the transcriptional coactivator CBP as a major reg- proteins were upregulated in response to CBP/p300 inhibition, ulator of AZA resistance and sensitivity, respectively. NAA15 is most likely as a stress response to problems in protein translation. an N-terminal α-amino-acetylase that, as part of the NatA However, based on the fast and pronounced impact of CBP/p300 complex, promotes sensitivity to apoptotic stimuli by acetylation inhibition on protein synthesis, we cannot exclude the possibility of key apoptotic mediators52. This could explain the resistance to of an additional direct mechanism. Taken together, our results AZA treatment observed in the NAA15 knockdown cells. In suggest that CBP/p300 inhibition potentiates the RNA-dependent contrast, the effect of CBP on AZA sensitivity was unexpected but functions of AZA. The convergence of AZA and CBP/p300 confirmed by pharmacologic inhibition of CBP and its closely inhibition on reducing protein synthesis is a likely explanation for related protein p300 in several AML cell lines. Our approach was the observed synergy. Here we have primarily used the tool limited to the cell-intrinsic response and has focused on cell compound C646, which was the first selective inhibitor of CBP/ survival as a primary parameter to assess drug response. It is p300 to be developed42. However, it lacks potency and its phar- important to keep in mind that our approach has modelled only macokinetic properties are unfavourable for use in vivo44. To one aspect of the mechanism underlying the function of AZA in overcome these problems, new inhibitors of CBP/p300 with the clinic. Future studies will need to evaluate the validity of CBP/ improved characteristics have been generated. A-485 is such an p300 as a combinatorial drug target in mouse models and co- inhibitor and, similar to C646, inhibits the catalytic domain44. culture experiments that better represent the complexity of MDS Another promising CBP/p300 inhibitor is CCS1477 from Cell- and sAML. This is particularly important, as part of the clinical Centric, Ltd46,58, which, in contrast to C646 and A-485, targets efficacy of AZA has been attributed to activation of an interferon the conserved bromodomain of p300 and CBP. CCS1477 is response through viral mimicry, making disease cells visible to the currently being evaluated first in human clinical trials in immune system18. If further validated in pre-clinical studies, the AML, multiple myeloma and non-Hodgkin lymphoma availability of biosafe inhibitors of CBP/p300 will warrant their (NCT04068597), and in castration-resistant prostate cancer combination with AZA to be tested in a clinical trial assessing (NCT03568656). Our results demonstrate that irrespective of the improvements of the current therapy of high-risk MDS. mode of inhibition, all tested CBP/p300 inhibitors synergized Importantly, here we show that the synergistic effect of CBP/ with AZA and induced a strong reduction of protein synthesis. p300 inhibition is specific for AZA and not observed in combi- The insight that CBP/p300 inhibition affects translation opens nation with the closely related drug decitabine. The main dif- an avenue for the rational design of combinatorial treatments. In ference between both drugs is that 80%–90% of AZA is future studies, we will test potential synthetic lethal combinations incorporated into RNA and only the remaining 10%–20% into of CBP/p300 inhibitors with drugs such as the eIF2 translation DNA, whereas decitabine is exclusively incorporated into DNA14. factor inhibitor salubrinal. Furthermore, our data suggest that A number of previous studies have reported differences in the CBP or p300-deficient and haplo-sufficient cells might be parti- cellular response to decitabine and AZA22,26,53, and this included cularly sensitive to AZA or treatments directly limiting protein the finding that AZA, but not decitabine, reduces protein synthesis. This is relevant, as loss of p300 is associated with synthesis in AML cell lines22. As we did not observe a synergy disease progression59. NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 9 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z In conclusion, our study provides important insight into drugs Loss-of-function screen. During the whole experiment, we aimed to have a 1000× relevant for haematological diseases. CBP/p300 inhibition representation of the hEpi9 library containing 7296 different plasmids. Taking into account the infection efficiency of 10% determined prior to the experiment to synergistically enhances the cell-intrinsic response to AZA, which achieve single-copy integration, 7.3 × 107 cells were transduced in triplicates as is currently used to treat high-risk MDS patients. The synergy is described above. After puromycin treatment, each replicate was split in two and specific for AZA and its RNA-dependent functions, and is not one half treated with 0.25 µM AZA every 2 days. During the treatment, cells were observed with the alternative FDA-approved drug decitabine. kept at a density of ~0.5 × 106 cells/ml, ensuring that the cell numbers never CBP/p300 inhibitors have entered clinical trials and the here- dropped under 7.3 × 106 to preserve library representation. Fourteen days after initiation of the treatment, the cells were collected and prepared for sequencing as recognized link to protein synthesis is informative for the previously described39. The analysis of the sequencing results was performed using development of potent drug combinations and response the R package EdgeR (R Core team 2019) as described63. Gene-level analysis was predictors. performed with the function Roast64 and the results ranked on a gene-by-gene level (Supplementary Data 3). Methods Flow cytometric analysis of viability and apoptosis. Following 4 days incu- Plasmids including hEPI9 library. The cSGEP vector was generated using stan- bation with the indicated inhibitors, cell viability was assessed by flow cyto- dard cloning techniques to insert the described UCO element40 into the previously metry of cells stained with 1 μg/mL DAPI (4′,6-Diamidino-2-phenylindole described mir-E shRNA expression vector SGEP (Addgene #111170)38. The dihydrochloride) (Thermo Scientific), and 100 μM MitoTracker® Red CMXRos shRNA library hEPI9 consisting of 7296 shRNA targeting 912 different chromatin (Thermo Scientific), using the LSR Fortessa cytometer and the BD FACSDi- genes (8 shRNAs per target) (Supplementary Data 1) was generated by cloning a vaTM software (BD Biosciences, Franklin Lakes, New Jersey) (see gating PCR-amplified oligo nucleotide pool (Agilent Technologies, Palo Alto, CA) into strategy in Supplementary Fig. S6). Statistical analysis (analysis of variance SGEP and cSGEP following previously described procedures60. Individual shRNAs test) and IC50 values were calculated using GraphPad Prism software (version including positive and negative controls (Supplementary Data 2) were inserted into 6). Chou–Talalay’s CI method using the CompuSyn software (version 1) was cSGEP following the described cloning procedure39. For the production of lenti- used to calculate the additive effect (CI = 1), synergy (CI < 1) or antagonism viral particles, we used packaging plasmids psPax2 (Addgene #12260) and pCMV- (CI > 1) in the drug combinations65. The percentage of apoptotic cells was VSV-G (Addgene #8454). determined by flow cytometry of cells stained with Annexin V-APC (BD Biosciences, 1 : 200 dilution) and 1 μg/mL DAPI. Cell culture and drug treatments. MOLM-13 (DMSZ # ACC554), F36-P (DMSZ # ACC543) and SKK-1 have been obtained from DSMZ as a collaboration with Identification of direct target genes using SLAM-seq analysis. Next, 5 × 106 Hans Drexler and have been characterized in detail36. SKK-1 cells are available MOLM-13 cells were either treated with 10 µM C646 or DMSO (three replicates from the original authors37. MDS-L cells were kindly provided by Tohyama and each) for 1 h followed by labelling with 100 µM 4sU for another hour. RNA was colleagues61. extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany) and SLAM-seq Cells were maintained in RPMI 1640 (Gibco, Thermo Scientific, Waltham, MA) was performed using the SLAM-seq Anabolic Kinetics Kit (Lexogen Gmbh, Aus- supplemented with 10% of fetal bovine serum (FBS), 1% penicillin–streptomycin tria) followed by library preparation using the QuantSeq 3′mRNA-Seq for Illumina and 1% L-Gutamine (Gibco) at 37 °C in 5% CO2. HEK293T (ATCC #CRL-11268) (FWD) and PCR Add-on Kit for Illumina (Lexogen Gmbh, Austria) according to and HS-5 (ATCC #CRL-11882) were obtained from American Type Culture the manufacturer’s instructions. The subsequent data analysis has essentially been Collection and cultured in Dulbecco’s modified Eagle’s media (Gibco) performed as described in Muhar et al.50 using the software SLAM-Dunk v0.4.166 supplemented with 10% FBS, 1% penicillin–streptomycin and 1% L-Glutamine (http://t-neumann.github.io/slamdunk/index.html) and the R package DEseq267. (Gibco) at 37 °C in 5% CO2. Cells were authenticated and passaged for <6 months. In brief, the reads were aligned to the human genome (GRCh38) and annotated CD34+ cells were isolated from mononuclear cells of healthy donors (see to the 3′-untranslated region (UTR) reference using the slamdunk all function with Supplementary Data 2 for donor characteristics) and cultured in suspension as default settings50. For the differential expression analysis, reads mapping to described62. For combinatorial treatments, cells were treated with AZA (Sigma- different UTR annotations of the same genes were collapsed and differential Aldrich, St. Louis, MO), decitabine (Sigma-Aldrich), C646 (Sigma-Aldrich), A-485 expression called on SLAM-seq reads with ≥2 T > C conversions using the R (Tocris, Bristol, UK), iCBP-112 (Tocris), CCS1477 (CellCentric, Ltd, Cambridge, package DESeq2 version 1.26.067 (Supplementary Data 4). For comparison, UK), JQ-1 (Sigma-Aldrich), DMSO (Sigma-Aldrich) or CHX (Sigma-Aldrich) at differential expression analysis was also performed on all reads (Supplementary the indicated amounts. Data 4). Gene Ontology analysis was performed on the most significant deregulated genes (FDR ≤ 0.01, log2FC ≤ −1 or ≥1) using DAVID 6.868,69 (Supplementary Data 5). Viral transduction and generation of stable cell lines. SKK-1 or MOLM-13 cells were transduced using standard lentiviral transduction procedures. In brief, viral supernatants generated in transfected HEK293T cells were mixed with 8 µg/ml Analysis of de novo protein synthesis. De novo protein synthesis was assessed polybrene (Sigma-Aldrich) and used to infect 2 × 106 cells/well in six-well plates by using the ClickIT Enrichment Kit (Thermo Scientific) and ClickIT-AHA (L-Azi- centrifuging at 500 × g for 30 min at 37 °C. This procedure was repeated once dohomoalanine) (Thermo Scientific) according to the ClickIT® Protein Enrich- before starting selection with 1 µg/ml puromycin 48 h after the first infection. ment Kit instructions. In brief, 10 × 106 MOLM-13 cells were treated with 10 µM C646 for 2 h or 0.1% DMSO, AHA labelled (in methionine-free media) for another 2 h and subsequently processed using the indicated kits. Liquid chromatography- Antibodies and protein analysis. SKK-1 cells were lysed in RIPA lysis buffer MS/MS was performed employing an Advion TriVersa NanoMate (Advion, Ithaca, containing protease and phosphatase inhibitors. Protein quantification was done NY) fitted on an Orbitrap Fusion Lumos™ Tribrid mass spectrometer (Thermo using the BCA Protein Assay Kit (Thermo Scientific) and after boiling in Laemli Scientific). The data were acquired with Xcalibur software vs. 4.0.27.10 (Thermo buffer, equal protein concentrations were loaded and separated on SDS- Scientific) and analysed using the MaxQuant software (1.6.7.0). The data were polyacrylamide gel electrophoresis prior to transfer overnight to nitrocellulose aligned to the Swissprot/Uniprot human database downloaded from the membranes. After blocking with non-fat milk, membranes were incubated with www.uniprot.org website (October 2019). A target and decoy database were used to primary antibodies and then with appropriate secondary fluorophore-conjugated assess the FDR, which was set to 1% at both peptide and protein level. Trypsin was secondary antibodies (LI-COR Biosciences, IRDye, 1 : 20,000). The dried mem- chosen as an enzyme and a maximum of two missed cleavages were allowed. branes were scanned with an Odyssey CLx Imager. Protein levels were quantified Carbamidomethylation (c) was set as a fixed modification, whereas oxidation (M) using the Odyssey and the band density normalized using the density of the histone and acetylation (N-terminal) were used as variable modifications. Searches were H3 signal. The following antibodies were used: anti-CREBBP (Abcam, Cambridge, performed using a peptide tolerance of 7 p.p.m. and a product ion tolerance of UK, ab2832, 1 : 1000 dilution) and anti-histone H3 (Abcam, ab1791, 1 : 10,000 0.5 Da. dilution). Proteins were filtered by removing ‘Potential Contaminant’, ‘Reverse’ and ‘Identified Only by Site’, background proteins appearing in the negative control (cells non-labelled with AHA), as well as proteins with PSM = 0 in one or more Real-time quantitative PCR. For RT-qPCR, RNA was extracted using the Max- DMSO control samples. Finally, proteins disappearing due to the treatment and the well® RSC simplyRNA Cells Kit (Promega, Madison, Wisconsin) and reverse proteins with a decrease of >75% of the number of PSMs in the treated cells transcribed with the First strand cDNA synthesis kit (Thermo Scientific). The compared to DMSO control were included in the final list of candidates cDNA was PCR amplified in triplicate using the Fast SYBR green dye on the (Supplementary Data 6). Gene Ontology analysis of genes corresponding to the Applied BiosystemsTM QuantStudioTM 7 Flex Real-Time PCR System. Relative significantly affected proteins (≥75% reduction compared to DMSO control) was expression was determined using parental cells or cells expressing the control performed using DAVID 6.868,69 (Supplementary Data 7). cSGEP vector as reference samples and GAPDH as the internal control. Sequences of primers are provided in Supplementary Data 2. Statistical analysis was per- formed by two-sided Student’s T-test using the GraphPad Prism software Bone marrow samples. All experiments and methods were performed in accor- (version 6). dance with relevant guidelines and regulations. Specifically, experiments using 10 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE human leukemic blast cells from AML patients were approved by the Ethics References committee of the Hospital Germans Trias i Pujol (Badalona, Spain; CEIm 1. Menssen, A. J. & Walter, M. J. Genetics of progression from MDS to IRB00002131) for the protocol RESPONSE (PIE16/00011). Bone marrow samples secondary leukemia. Blood 136, 50–60 (2020). from healthy donors were obtained from femoral heads of patients undergoing hip 2. Steensma, D. P. et al. Multicenter study of decitabine administered daily for replacement surgery and the experiments approved by the Ethics commission of 5 days every 4 weeks to adults with myelodysplastic syndromes: the the Faculty of Medicine (Technical University of Munich, Germany). Written Alternative Dosing for Outpatient Treatment (ADOPT) Trial. J. Clin. Oncol. informed consent in accordance with the Declaration of Helsinki was obtained 27, 3842–3848 (2009). from all patients. 3. Greenberg, P. L., Young, N. S. & Gattermann, N. Myelodysplastic syndromes. Bone marrow samples from AML patients (see Supplementary Data 2 for patient Hematol. Am. Soc. Hematol. Educ. Prog. 2002, 136–161, https://doi.org/ details) have been obtained from the IJC-Campus ICO-GTP Biological Sample 10.1182/asheducation-2002.1.136 (2002). Collection. Only samples from patients with high blast count (>75) were used. 4. Vardiman, J. W. et al. The 2008 revision of the World Health Organization Samples were thawed, seeded on top of HS-5 stroma cells and a few hours later (WHO) classification of myeloid neoplasms and acute leukemia: rationale and treated with different concentrations of AZA. One day later, cells were treated with important changes. Blood 114, 937–951 (2009). different concentrations of A-485, CCS1477 or DMSO for two additional days. The 5. Bejar, R. et al. TET2 mutations predict response to hypomethylating agents in number of apoptotic cells was determined using Annexin V-APC as described myelodysplastic syndrome patients. Blood 124, 2705–2712 (2014). above. To exclude the signal from co-cultured HS-5 cells, samples were co-stained 6. Lindsley, R. C. et al. Acute myeloid leukemia ontogeny is defined by distinct with Pharmingen™ PE Mouse Anti-Human CD90 Clone: 5E10 (BD Biosciences, somatic mutations. Blood 125, 1367–1376 (2015). Catalogue number: 561970, 1 : 400 dilution) and only the CD90-negative fraction 7. Papaemmanuil, E. et al. Clinical and biological implications of driver used for further analysis (see gating strategy in Supplementary Fig. S6). mutations in myelodysplastic syndromes. Blood 122, 3616–3627 (2013). quiz CD34+ cells were isolated from healthy donors (see Supplementary Data 2 for patient details) with immunomagnetic beads (Miltenyi Biotec, Bergisch Gladbach, 3699. Germany) and cultured for 48 h. 8. Haferlach, T. et al. Landscape of genetic lesions in 944 patients with myelodysplastic syndromes. Leukemia 28, 241–247 (2014). 9. Fenaux, P. et al. Efficacy of azacitidine compared with that of conventional Flow cytometric analysis of protein synthesis. CD34+ cells were washed and care regimens in the treatment of higher-risk myelodysplastic syndromes: a resuspended in media with different concentrations of C646 (Sigma-Aldrich) and randomised, open-label, phase III study. Lancet Oncol. 10, 223–232 (2009). incubated for another 8 h after which protein synthesis was analysed with the 10. Garcia-Manero, G. & Fenaux, P. Hypomethylating agents and other Protein Synthesis Assay Kit Red (Abcam) according to the manufacturer’s novel strategies in myelodysplastic syndromes. J. Clin. Oncol. 29, 516–523 instructions. The samples were analysed on a Beckman Coulter Cyan ADP (2011). (Beckman Coulter) and with FlowJo v10 (FlowJo LLC). The median fluorescence 11. Diesch, J. et al. A clinical-molecular update on azanucleoside-based therapy intensity of each condition was normalized to the DMSO control. Statistical ana- for the treatment of hematologic cancers. Clin. Epigenet. 8, 71 (2016). lysis (unpaired t-test) were calculated using GraphPad Prism software (version 6). 12. Berdasco, M. & Esteller, M. Clinical epigenetics: seizing opportunities for Leukemic blast cells or SKK-1 and MOLM-13 cells were treated with different translation. Nat. Rev. Genet. 20, 109–127 (2019). concentrations of AZA, C646, A-485, CCS1477 or DMSO and incubated for 3 or 13. Stresemann, C. & Lyko, F. Modes of action of the DNA methyltransferase 6 h, respectively. Cells treated 1.5 h with CHX were used as a positive control. The inhibitors azacytidine and decitabine. Int. J. Cancer 123, 8–13 (2008). samples were then processed using the Click-iT™ HPG Alexa Fluor™ 594 kit (Thermo Scientific) according to the manufacturer’s instructions and the protein 14. Li, L. H., Olin, E. J., Buskirk, H. H. & Reineke, L. M. Cytotoxicity and mode of synthesis analysed using the LSR Fortessa cytometer (BD Biosciences) and action of 5-azacytidine on L1210 leukemia. Cancer Res. 30, 2760 (1970). FlowLogic v8 (Inivai Technologies, Victoria, Australia). 15. Liu, K., Wang, Y. F., Cantemir, C. & Muller, M. T. Endogenous assays of DNA methyltransferases: evidence for differential activities of DNMT1, DNMT2, and DNMT3 in mammalian cells in vivo. Mol. Cell. Biol. 23, 2709–2719 Polysome profiling. Analysis of distribution of RNA across sucrose gradients (2003). was performed as described earlier70, except for minor modifications. Briefly, 16. Kiziltepe, T. et al. 5-Azacytidine, a DNA methyltransferase inhibitor, induces 5 × 106 MOLM-13 cells were treated with 10 µM C646 or DMSO and incubated ATR-mediated DNA double-strand break responses, apoptosis, and for 4 h. Equal amounts of polysomal lysate (170–500 μg) were loaded on synergistic cytotoxicity with doxorubicin and bortezomib against multiple 10–50% sucrose linear gradients generated with a BIOCOMP gradient master myeloma cells. Mol. Cancer Ther. 6, 1718–1727 (2007). and containing 80 mM NaCl, 5 mM MgCl2, 20 mM Tris HCl pH 7.4, 1 mM 17. Taylor, S. M. 5-Aza-2′-deoxycytidine: cell differentiation and DNA dithiothreitol, 10 U/ml Rnase inhibitor. Gradients were centrifuged on a methylation. Leukemia 7(Suppl 1), 3–8 (1993). SW40Ti rotor for 2 h 30 min at 218,000 × g and analysed on a BIOCOMP 18. Jones, P. A., Ohtani, H., Chakravarthy, A. & De Carvalho, D. D. Epigenetic gradient station. Quantification of polysomal absorbance at 260 nm was nor- therapy in immune-oncology. Nat. Rev. Cancer 19, 151–161 (2019). malized to the baseline absorbance. 19. Lu, L. J. & Randerath, K. Mechanism of 5-azacytidine-induced transfer RNA cytosine-5-methyltransferase deficiency. Cancer Res. 40, 2701–2705 (1980). LINE-1 methylation assay. Response to decitabine was validated using the 20. Lu, L. W., Chiang, G. H., Medina, D. & Randerath, K. Drug effects on nucleic Global DNA Methylation LINE-1 kit (Active Motif, Carlsbad, CA). For this, acid modification. I. A specific effect of 5-azacytidine on mammalian transfer MOLM-13 cells were treated with 1 µM decitabine or DMSO and 24 h later the RNA methylation invivo. Biochem. Biophys. Res. Commun. 68, 1094–1101 genomic DNA collected and processed according to the manufacturer’s (1976). instructions. 21. Schaefer, M., Hagemann, S., Hanna, K. & Lyko, F. Azacytidine inhibits RNA methylation at DNMT2 target sites in human cancer cell lines. Cancer Res. 69, Statistical analysis. Statistical analyses were performed using the GraphPad Prism 8127–8132 (2009). software (version 6) and suitable tests as indicated in legends. 22. Hollenbach, P. W. et al. A comparison of azacitidine and decitabine activities in acute myeloid leukemia cell lines. PLoS ONE 5, e9001 (2010). 23. Momparler, R. L., Siegel, S., Avila, F., Lee, T. & Karon, M. Effect of tRNA from Reporting summary. Further information on research design is available in the Nature 5-azacytidine-treated hamster fibrosarcoma cells on protein synthesis in vitro Research Reporting Summary linked to this article. in a cell-free system. Biochem. Pharmacol. 25, 389–392 (1976). 24. Aimiuwu, J. et al. RNA-dependent inhibition of ribonucleotide reductase is a Data availability major pathway for 5-azacytidine activity in acute myeloid leukemia. Blood The data that support this study are available from the corresponding author upon 119, 5229–5238 (2012). reasonable request. The proteomics data generated in this study have been deposited in 25. Cheng, J. X. et al. RNA cytosine methylation and methyltransferases mediate the PRIDE database under accession code PXD021738. The shRNA screen data chromatin organization and 5-azacytidine response and resistance in generated in this study have been deposited in the GEO database under accession code leukaemia. Nat. Commun. 9, 1163 (2018). GSE158738. The SLAM-Seq data generated in this study have been deposited in the GEO 26. Flotho, C. et al. The DNA methyltransferase inhibitors azacitidine, decitabine database under accession code GSE158813. The protein reference used in this study is and zebularine exert differential effects on cancer gene expression in acute available in the Swissprot/Uniprot human database (www.uniprot.org website, October myeloid leukemia cells. Leukemia 23, 1019–1028 (2009). 2019). The human genome reference GRCh38 used in this study is available at https:// 27. Goodman, R. H. & Smolik, S. CBP/p300 in cell growth, transformation, and www.ncbi.nlm.nih.gov/. Source data are provided with this paper. development. Genes Dev. 14, 1553–1577 (2000). 28. Wang, F., Marshall, C. B. & Ikura, M. Transcriptional/epigenetic regulator CBP/p300 in tumorigenesis: structural and functional versatility in target Received: 21 November 2020; Accepted: 22 September 2021; recognition. Cell. Mol. Life Sci. 70, 3989–4008 (2013). 29. Kalkhoven, E. CBP and p300: HATs for different occasions. Biochem. Pharmacol. 68, 1145–1155 (2004). NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 11 ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z 30. Boyes, J., Byfield, P., Nakatani, Y. & Ogryzko, V. Regulation of activity of the 59. Cheng, G. et al. Loss of p300 accelerates MDS-associated leukemogenesis. transcription factor GATA-1 by acetylation. Nature 396, 594–598 (1998). Leukemia 31, 1382–1390 (2017). 31. Tomita, A. et al. c-Myb acetylation at the carboxyl-terminal conserved domain 60. Rathert, P. et al. Transcriptional plasticity promotes primary and acquired by transcriptional co-activator p300. Oncogene 19, 444–451 (2000). resistance to BET inhibition. Nature 525, 543–547 (2015). 32. Kitabayashi, I., Yokoyama, A., Shimizu, K. & Ohki, M. Interaction and 61. Matsuoka, A. et al. Lenalidomide induces cell death in an MDS-derived cell functional cooperation of the leukemia-associated factors AML1 and p300 in line with deletion of chromosome 5q by inhibition of cytokinesis. Leukemia myeloid cell differentiation. EMBO J. 17, 2994–3004 (1998). 24, 748–755 (2010). 33. Satake, N. et al. Novel MLL-CBP fusion transcript in therapy-related chronic 62. Wenk, C. et al. Direct modulation of the bone marrow mesenchymal stromal myelomonocytic leukemia with a t(11;16)(q23;p13) chromosome cell compartment by azacitidine enhances healthy hematopoiesis. Blood Adv. translocation. Genes Chromosomes Cancer 20, 60–63 (1997). 2, 3447–3461 (2018). 34. Borrow, J. et al. The translocation t(8;16)(p11;p13) of acute myeloid leukaemia 63. Dai, Z. et al. edgeR: a versatile tool for the analysis of shRNA-seq and fuses a putative acetyltransferase to the CREB–binding protein. Nat. Genet. 14, CRISPR-Cas9 genetic screens. F1000Research 3, 95 (2014). 33–41 (1996). 64. Wu, D. et al. ROAST: rotation gene set tests for complex microarray 35. Furdas, S. D., Carlino, L., Sippl, W. & Jung, M. Inhibition of bromodomain- experiments. Bioinformatics 26, 2176–2182 (2010). mediated protein-protein interactions as a novel therapeutic strategy. 65. Chou, T. C. & Talalay, P. Quantitative analysis of dose-effect relationships: the Medchemcomm 3, 123–134 (2012). combined effects of multiple drugs or enzyme inhibitors. Adv. Enzym. Regul. 36. Palau, A. et al. Immunophenotypic, cytogenetic, and mutational 22, 27–55 (1984). characterization of cell lines derived from myelodysplastic syndrome patients 66. Neumann, T. et al. Quantification of experimentally induced nucleotide after progression to acute myeloid leukemia. Genes Chromosomes Cancer 56, conversions in high-throughput sequencing datasets. BMC Bioinformatics 20, 243–252 (2017). 258 (2019). 37. Matsuoka, H. et al. Establishment of a human myeloid cell line with trisomy 8 67. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and derived from overt leukemia following myelodysplastic syndrome. dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014). Haematologica 90, 981–982 (2005). 68. Huang da, W., Sherman, B. T. & Lempicki, R. A. Systematic and integrative 38. Fellmann, C. et al. An optimized microRNA backbone for effective single- analysis of large gene lists using DAVID bioinformatics resources. Nat. Protoc. copy RNAi. Cell Rep. 5, 1704–1713 (2013). 4, 44–57 (2009). 39. Fellmann, C. et al. Functional identification of optimized RNAi triggers using 69. Huang da, W., Sherman, B. T. & Lempicki, R. A. Bioinformatics enrichment a massively parallel sensor assay. Mol. Cell 41, 733–746 (2011). tools: paths toward the comprehensive functional analysis of large gene lists. 40. Müller-Kuller, U. et al. A minimal ubiquitous chromatin opening element Nucleic Acids Res. 37, 1–13 (2009). (UCOE) effectively prevents silencing of juxtaposed heterologous promoters 70. Fumagalli, S., Ivanenkov, V. V., Teng, T. & Thomas, G. Suprainduction of p53 by epigenetic remodeling in multipotent and pluripotent stem cells. Nucleic by disruption of 40S and 60S ribosome biogenesis leads to the activation of a Acids Res. 43, 1577–1592 (2015). novel G2/M checkpoint. Genes Dev. 26, 1028–1040 (2012). 41. Polevoda, B. & Sherman, F. N-terminal acetyltransferases and sequence requirements for N-terminal acetylation of eukaryotic proteins. J. Mol. Biol. 325, 595–622 (2003). 42. Bowers, E. M. et al. Virtual ligand screening of the p300/CBP histone Acknowledgements acetyltransferase: identification of a selective small molecule inhibitor. Chem. This project was supported by the FEDER/Ministerio de Ciencia e Innovación - Agencia Biol. 17, 471–482 (2010). Estatal de Investigación through the grant PIE16/00011 RESPONSE (to M.B.), the 43. Chou, T. C. Drug combination studies and their synergy quantification using European Research Council ERC-StG-336860 (to J.Z.), the grant Juan de la Cierva— the Chou-Talalay method. Cancer Res. 70, 440–446 (2010). Formación FJCI-2014-22983 (to J.D.), the grant Sara Borrell CD17/00084 (to J.D.), the 44. Lasko, L. M. et al. Discovery of a selective catalytic p300/CBP inhibitor that Marie Skłodowska Curie Training network “ChroMe” H2020-MSCA-ITN-2015-675610 targets lineage-specific tumours. Nature 550, 128–132 (2017). (to M.M.), the FPI predoctoral fellowship BES-2016–077251 (to M.M.L.P.), the SFB 1243 45. Picaud, S. et al. Generation of a selective small molecule inhibitor of the CBP/ DFG (to K.S.G.), the Austrian Science Fund SFB-F4710 (to J.Z.) and the Deutsche José p300 bromodomain for leukemia therapy. Cancer Res. 75, 5106–5119 (2015). Carreras Leukämie Stiftung DJCLS 14R/2018 (to K.G. and M.B.). Research in the 46. Brooks, N. et al. CCS1477: a novel small molecule inhibitor of p300/CBP Buschbeck lab is further supported by the following grants: MINECO grant RTI2018- bromodomain for the treatment of acute myeloid leukaemia and multiple 094005-B-I00 (to M.B.), the Marie Skłodowska Curie Training network ‘INTERCEPT- myeloma. Blood 134, 2560–2560 (2019). MDS’ H2020-MSCA-ITN-2015-953407 (to M.B. and K.S.G.); AGAUR 2017-SGR-305 (to 47. Welti, J. et al. Targeting the p300/CBP axis in lethal prostate cancer. Cancer M.B.) and Fundació La Marató de TV3 257/C/2019 (to M.B.). Research at IMP is Discov. 11, 1118 (2021). supported by Boehringer Ingelheim, the Austrian Research Promotion Agency (Head- 48. Filippakopoulos, P. et al. Selective inhibition of BET bromodomains. Nature quarter grant FFG-852936) and the Austrian Academy of Sciences. Research at the IJC is 468, 1067–1073 (2010). generously supported by the ‘La Caixa’ Foundation, the Fundació Internacional Josep 49. Pan, R. et al. Selective BCL-2 inhibition by ABT-199 causes on-target cell Carreras, Celgene Spain and the CERCA Programme/Generalitat de Catalunya. A.G. death in acute myeloid leukemia. Cancer Discov. 4, 362–375 (2014). received funds by “Agencia Estatal de Investigación” (AEI) through the Plan Nacional 50. Muhar, M. et al. SLAM-seq defines direct gene-regulatory functions of the “Excelencia” grant number SAF2017-84301-P, by the “Associación Española Contra el BRD4-MYC axis. Science 360, 800–805 (2018). Cancer” (AECC) grant number LABAE20040GENT and by the Agency for Management 51. Gentilella, A. et al. Autogenous control of 5′TOP mRNA stability by 40S of University and Research Grants (AGAUR) of the Catalan Government grant ribosomes. Mol. Cell 67, 55–70.e54 (2017). 2017SGR01743. Proteomic analyses were performed in the IJC Proteomic Unit, which 52. Yi, C. H. et al. Metabolic regulation of protein N-alpha-acetylation by Bcl-xL are part of Proteored PRB3 and are supported by grant PT17/0019 from the PE I + D + i promotes cell survival. Cell 146, 607–620 (2011). 2013-2016, funded by ISCIII and ERDF. We thank Kaoru Tohyama for providing MDS- 53. Nguyen, A. N. et al. Azacitidine and decitabine have different mechanisms of L cells, members of the Buschbeck lab, the RESPONSE network (PIE16/00011), Blanca action in non-small cell lung cancer cell lines. Lung Cancer 1, 119–140 (2010). Xicoy and Francesc Solé for valuable discussions. We thank the IJC Biobanking unit for 54. Kantarjian, H. et al. Decitabine improves patient outcomes in myelodysplastic sample preparation and storage, Bernat Cucurull from the IJC Proteomics unit for the syndromes: results of a phase III randomized study. Cancer 106, 1794–1803 ClickIT sample preparation, Marco Fernandez for advice and training in flow cytometry (2006). and all other staff of IJC and IGTP core facilities for excellent support. 55. Silverman, L. R. et al. Randomized controlled trial of azacitidine in patients with the myelodysplastic syndrome: a study of the cancer and leukemia group B. J. Clin. Oncol. 20, 2429–2440 (2002). Author contributions 56. Lübbert, M. et al. Low-dose decitabine versus best supportive care in elderly Conceived and designed the experiments: J.D. and M.B. Designed and generated essential patients with intermediate- or high-risk myelodysplastic syndrome (MDS) tools: M.F., P.R., J.Z. and N.B. Performed the experiments: J.D., R.W., M.M., M.M.L.P., ineligible for intensive chemotherapy: final results of the randomized phase III R.C., M.V.D.G., C.D.L.T., C.M.H. and A.G. Analysed the data: J.D., M.M.L.P., C.D.L.T., study of the European Organisation for Research and Treatment of Cancer J.B., C.M.H. and A.G. Contributed reagents/materials/analysis tools: M.a.M., K.S.G., L.Z., Leukemia Group and the German MDS Study Group. J. Clin. Oncol. 29, J.Z. and N.B. Wrote the paper: J.D. and M.B. 1987–1996 (2011). 57. Zeidan, A. M. et al. Long-term survival of older patients with MDS treated with HMA therapy without subsequent stem cell transplantation. Blood 131, 818–821 (2018). Competing interests 58. Pegg, N. et al. Characterisation of CCS1477: a novel small molecule inhibitor K.S.G. has received support for independent research from Celgene Corp. N.B. is an of p300/CBP for the treatment of castration resistant prostate cancer. J. Clin. employee and stockholder in CellCentric, Ltd. The other authors declare no competing Oncol. 35, 11590–11590 (2017). interests. 12 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-26258-z ARTICLE Additional information Open Access This article is licensed under a Creative Commons Supplementary information The online version contains supplementary material Attribution 4.0 International License, which permits use, sharing, available at https://doi.org/10.1038/s41467-021-26258-z. 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 Correspondence and requests for materials should be addressed to Marcus Buschbeck. Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless Peer review information Nature Communications thanks Olatoyosi Odenike, and the indicated otherwise in a credit line to the material. If material is not included in the other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer article’s Creative Commons license and your intended use is not permitted by statutory reviewer reports are available. regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/ Reprints and permission information is available at http://www.nature.com/reprints licenses/by/4.0/. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. © The Author(s) 2021 NATURE COMMUNICATIONS | (2021)12:6060 | https://doi.org/10.1038/s41467-021-26258-z | www.nature.com/naturecommunications 13