Deacetylation of ACLY Mediates RNA M6A-Modification of NOXA and Promotes Chemoresistance of Colorectal Cancer

Abstract

Chemoresistance is a major challenge for colorectal cancer (CRC) therapy and is a leading cause of cancer mortality, yet the underlying molecular mechanism remains unclear. ATP citrate lyase (ACLY), a rate-limiting enzyme of de novo lipid synthesis, plays an important role in tumor progression and chemotherapy. Here, It is demonstrated that deacetylation of ACLY is critical for chemoresistance in CRC. Through proteomic screening acetylated proteins in chemoresistant patient-derived cells, It is identified that ACLY is deacetylated at K978 site, which induces the relocation of ACLY to the nucleus and promotes its binding to RNA-binding protein 15 (RBM15). This facilitates N⁶-methyladenosine (m⁶A) methylation of NOXA (also known as PMAIP1, phorbol-12-myristate-13-acetate-induced protein 1) and decreases the stability of NOXA mRNA, resulting in chemoresistance. With the selective inhibitor Santacruzamate A, targeting the deacetylase histone deacetylase 2 (HDAC2) to inhibit the acetylation may enhance the sensitivity of chemoresistance. These findings provide new insights into the mechanism of ACLY deacetylation promoting chemoresistance and suggest a potential therapeutic strategy to mitigate the chemoresistant effects.

Keywords: ACLY; acetylation; chemoresistance; colorectal cancer; m6A.

Figure 1

Lysine acetylation profile in cancer chemoresistance and the important role of lysine deacetylation modification of ACLY. a) Schematic representation of quantitative proteomics and acetylation modification analysis of eight postoperative tumor specimens. b) Numbers of differentially expressed acetylated proteins and sites identified by acetylation analysis in sensitive and non‐sensitive groups. c, d) Gene enrichment of biological processes (c) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis (d) of the proteomics revealed that the Tricarboxylic acid cycle pathway ranked the top significant pathway using a two‐sided hypergeometric test, with P values adjusted for multiple comparisons via the Benjamini–Hochberg False Discovery Rate (FDR) method. e) Protein‐protein interaction network of differentially expressed acetylated proteins visualized using Cytoscape software (v3.10.2). f) LC‐MS/MS spectrum of the acetylated ACLY peptide (K978). g) Western blot analysis of K978 acetylation in HCT8 cells transfected with ACLY‐WT/K978R after 5‐FU (1 µM) treatment for 24 h. h, i) Western blot analysis of K978 acetylation in the parental and 5‐FU resistant cells treated with different concentrations of 5‐FU for 24 h. j) Representative immunohistochemical staining for ACLY‐K978ac in tissue microarrays from 65 CRC patients, showing decreased ACLY‐K978ac expression in tumor compared to adjacent normal tissue. k) Comparison of immunohistochemical staining scores of ACLY‐K978ac between tumor and adjacent tissues, ^** p < 0.01. l) Kaplan‐Meier overall survival analysis using the expression level of ACLY‐K978ac in 65 CRC patients, log‐rank test. Figure 1a was created using Figdraw (www.figdraw.com) and is accompanied by proper copyright authorization.

Figure 2

ACLY‐K978 deacetylation promotes chemoresistance. a) Apoptosis‐associated markers were evaluated by western blot in HCT8 and HCT15 cells stably expressing shNC, shACLY, or shACLY‐WT/K978R. b) Cell growth of HCT8 and HCT15 cells stably expressing shNC or shACLY‐WT/K978R (n = 6). c) Representative images of colony formation assays with DMSO or 5‐FU (1 µm) treatment in two CRC cell types. d) Quantification of the colonies (c) (n = 3). e) Apoptotic cells induced by 5‐FU treatment were analyzed by flow Jo (n = 3). f) Immunofluorescence staining of γH2AX (red) and DAPI (blue) in HCT8 shNC/shACLY cells expressing WT/KR (scale bars, 50 µm). g) Representative images of immunohistochemical staining of tumors (scale bars, 50 µm). h) HCT8 tumors were dissected from the BALB/c‐nu mice treated with PBS, or 5‐FU (n = 5/group, scale bars, 5 mm). i) Tumor growth of HCT8 cells in male BALB/c‐nude mice, with PBS or 5‐FU (50 mg kg⁻¹) intraperitoneally every three days, beginning on day 7. j) Tumors were dissected and the weight was measured on day 19. Error bars indicate mean ± SD. ^* p < 0.05, ^** p < 0.01, and ^*** p < 0.001.

Figure 3

K978‐deacetylated ACLY preferentially binds to RBM15 protein. a) Western blot analysis of ACLY protein localization in WT or KR cells. b) Western blot to detect ACLY protein localization in HCT8 and HCT8R cells. c, d) Immunofluorescence staining showing the subcellular localization of ACLY in CRC cells (scale bars, 20 µm). e) Venn diagram showing the overlap of proteins binding ACLY in different groups (HEK293T cells transfected with Flag‐ACLY‐WT/KR; HCT8 and HCT8R cells transfected with Flag‐ACLY). f) Western blot of input and anti‐ACLY immunoprecipitate (IP) analysis of endogenous ACLY interaction with RBM15 in HCT8 and HCT8R cells. g) Western blot assay to detect interaction between ACLY and RBM15 in HCT8/shACLY cells stably expressing ACLY‐WT/KR. h, i) in situ proximity ligation assay (PLA) demonstrated interaction between ACLY and RBM15 in HCT8 and HCT8R cells (h) or cells stably expressing ACLY‐WT/K978R (i). Positive PLA signals showed ACLY/RBM15 complex, shown as red clusters, with cell nuclei stained blue (scale bars, 20 µm).

Figure 4

ACLY‐K978 deacetylation promotes chemoresistance depending on RBM15. a) Comparison of RBM15 mRNA expression between normal tissue (n = 349) and CRC (n = 275) from GEPIA2 dataset. b) RBM15 mRNA level in different tumor grades of CRC from UALCAN dataset. c) Cell proliferation assays conducted using HCT8 cells transfected with siNC or siRBM15 #1/2 (n = 6). d, e) Western blot showing the efficiency of RBM15 knockdown and apoptosis‐related protein expression levels after 5‐FU treatment for 24 h. f, g) Apoptotic cells induced by RBM15 knockdown or 5‐FU treatment were analyzed by flow cytometry and flow Jo (n = 3). h) Western blot analysis of apoptosis‐associated proteins in ACLY‐WT/K978R‐expressing HCT8/HCT15 cells with RBM15 knockdown. i) Representative images of colony formation assays in two CRC cell lines transfected with siNC or siRBM15 following 5‐FU treatment. j) Quantification of the colonies in (i) (n = 3). Error bars represent the mean ± SD. For comparisons among multiple groups, one‐way ANOVA was employed. ^* p < 0.05, ^** p < 0.01, and ^*** p < 0.001.

Figure 5

ACLY participates in the m⁶A methylation of NOXA. a) Venn diagram showing the overlap of meRIP‐identified mRNAs and corresponding mRNA levels with significant change in four groups. b) Pathway enrichment analysis of the overlap 39 genes screened from (a). c) Expression levels of mRNAs measured in WT, KR, and relative RBM15 knockdown cells (n = 3). d) MeRIP qRT‐PCR assays measuring m⁶A modification levels in WT, and relative RBM15 knockdown cells, and KR cells (n = 3). e) NOXA mRNA stability analysis in WT/KR or NC/siRBM15 HCT8 cells using actinomycin D treatment (0/2/4/6/8 h) (n = 3). f) Prediction m⁶A modification sites in region‐3 by SRAMP (www.cuilab.cn/sramp) and mutation luciferase reporter plasmids of very high‐confidence prediction sites. g) The luciferase activity in WT/KR cells co‐transfected with wide type/mutated luciferase reporter vectors (WT1/2/3, MUT1/2/3) and si‐NC/siRBM15‐1 (n = 3). h) The luciferase activity in HCT8 cells co‐transfected with wide type/mutated luciferase reporter vectors (WT1/2, MUT1/2) and si‐NC/siRBM15#1 (n = 3). i) Western blot to detect the NOXA protein expression in WT/KR mutation cells with or without RBM15 knockdown. j) RIP assay and subsequent qRT‐PCR analysis of NOXA mRNA were conducted in indicated cells using antibodies against RBM15 or ACLY. Error bars represent the mean ± SD. For comparisons among multiple groups, one‐way ANOVA was employed. ^* p < 0.05, ^** p < 0.01, and ^*** p < 0.001; ns, no significance.

Figure 6

ACLY‐K978 deacetylase HDAC2 enhances the sensitivity of 5‐FU. a) Western blot to detect the ACLY‐K978 acetylation level after treatment with DMSO, NAM, or TSA. b‐d) HCT8 cells were treated with siRNA targeting HDAC2 (b) or Santacruzamate A (SCA) (c) or transfected with HA‐HDAC2 (d), with the acetylation level analyzed by western blot. e, f) IP and western blot to analyze the interaction of ACLY and HDAC2 proteins. g) IP analysis of the interaction between HDCA2 and ACLY proteins in HCT8 cells, with or without 5‐FU treatment (1 µm for 24 h). h) Schematic of GST‐ACLY (full length) and series of deletion mutants was shown. i) GST‐pulldown assay was conducted using recombinant GST‐ACLY deletion mutants (D1‐ D4) and total cell lysates from HCT8. j) Representative images of colony formation assay in HCT8 cells with DMSO, 5‐FU (1 µM), SCA (60 µM), or both treatment (n = 3). k) Quantification of the colonies in (j). l) Calculation and visualization of synergy scores for drug combinations of 5‐FU and SCA by SynergyFinder 2.0 (https://synergyfinder.fimm.fi). m) HCT8 tumors were dissected from the nude mice treated with PBS, 5‐FU (50 mg kg⁻¹, every three days), SCA (10 mg kg⁻¹, every three days), or both (n = 5/group) intraperitoneally, beginning on day 7. n) Tumors were dissected and the weight was measured on day 19. o) Tumor size was measured every three days during drug treatment. p) The proposed working model was created using Adobe Illustrator 2019. Error bars indicate mean ± SD. For comparisons among multiple groups, one‐way ANOVA was employed. ^* p < 0.05, ^** p < 0.01, and ^*** p < 0.001; ns, no significance.