Histone lactylation-driven feedback loop modulates cholesterol-linked immunosuppression in pancreatic cancer

Abstract

Background: Pancreatic cancer exhibits limited clinical responses to immunotherapy, highlighting the need for new strategies to counteract its immunosuppressive microenvironment. Although metabolic reprogramming and epigenetic changes contribute to malignancy, the impact of lactate-driven histone lactylation on the tumour microenvironment (TME) has not been fully explored.

Objective: This study aims to investigate the role of histone lactylation in pancreatic cancer, focusing on its effects on cholesterol metabolism and antitumour immunity.

Designs: Global lactylome profiling was conducted to identify novel epigenetic mechanisms driven by lactate-induced histone lactylation. Mechanisms were investigated via RNA sequencing, CUT&Tag, immunoprecipitation-mass spectrometry and GST-pull down. Mass cytometry by time-of-flight, in vitro co-culture system, orthotopic pancreatic cancer models and flow cytometry were used to explore Acetyl-CoA acetyltransferase (ACAT2) functions. A proteolysis-targeting chimaera (PROTAC) was developed to degrade ACAT2.

Results: Global lactylome profiling revealed that lactate-driven histone lactylation, particularly H3K18la, promotes the transcriptional activation of ACAT2. ACAT2 acetylates mitochondrial carrier homolog 2 (MTCH2), stabilising it and disrupting oxidative phosphorylation, which increases lactate production and fuels a positive feedback loop in pancreatic cancer. This loop facilitates the delivery of cholesterol via small extracellular vesicles (sEVs), polarising tumour-associated macrophages toward an immunosuppressive M2 phenotype. Additionally, the PROTAC targeting ACAT2 enhanced the efficacy of immune checkpoint blockade therapy in vivo.

Conclusions: Our findings highlight the critical role of the H3K18la/ACAT2/sEV-cholesterol axis in TME reprogramming. Targeting this pathway may improve anti-PD-1 therapy response in pancreatic cancer, providing a novel therapeutic strategy by linking histone lactylation, cholesterol metabolic reprogramming and immune modulation.

Keywords: GENE TARGETING; IMMUNE RESPONSE; LIPID METABOLISM; MACROPHAGES; PANCREATIC CANCER.

Figure 1. H3K18la is the most prevalent histone modification with an unfavourable prognosis in pancreatic cancer. (A) Schematic overview of the lysine lactylome and proteome analysis in pancreatic cancer (n=5) and paired normal tissues (n=5) using label-free quantification. (B) Overview of Kla identifications in pancreatic cancer (blue, n=5) and non-tumour (red, n=5) samples. Paired samples are connected by grey straight lines, and the shaded areas beneath the lasso curves represent the 95% CIs. (C) Quantitative analysis of the overall abundance of Kla in pancreatic samples (n=5 per group). Two-tailed paired t-test. (D) Number of Kla sites per protein in pancreatic cancer and paired normal tissues (n=5 per group). (E) Quantification of Kla sites on different histones in pancreatic samples (n=10). (F) Lactylation levels on specific histones are shown. Two-tailed unpaired t-test. (G) Violin plot showing the identified histone Kla sites between pancreatic cancer and adjacent normal tissues (n=5 per group). Two-tailed paired t test. (H, I) Heatmap (H) and volcano plot (I) analysis showing Kla sites altered on histones in pancreatic cancer (n=5 per group). (J, K) Representative images (J) and quantification (K) of H3K18la levels visualised by immunofluorescence staining in tumour and normal tissues (n=174). Scale bar: left panel, 100 µm; right panel, 20 µm. Two-tailed paired t-test. (L) Kaplan-Meier analysis of the overall survival of patients with pancreatic cancer from FUSCC based on H3K18la levels (n=174). Log rank test. (M) H3K18la levels in pancreatic cancer patients at AJCC stages T1 to T3. Two-tailed unpaired t-test. (N–P) Western blot analysis of H3K18la levels measured from PANC-1 and CFPAC-1 cells cultured in different concentrations (0 mM, 1 mM, 5 mM, 10 mM, 20 mM) of lactate (N), 2-DG (O) and oxamate (P) for 24 hours. Data are represented as the mean±SD. *p<0.05; **p<0.01; ***p<0.001; ns, not significant. 2-DG, 2-deoxy-D-glucose; AJCC, the american joint committee on cancer.

Figure 2. Genome-wide analysis of the transcriptional consequences of H3K18la in pancreatic cancer. (A) The binding density of H3K18la visualised using deepTools, showing the CUT&Tag tag counts at different H3K18la binding peaks in PDAC cell lines treated with or without 20 mM oxamate for 24 hours. (B) Stacked bar plots showing genomic annotation of H3K18la binding regions in different groups. (C) Bar plot showing the KEGG analysis of the candidate target genes with H3K18la binding peaks in both PANC-1 and CFPAC-1 cells. (D) Heatmap indicating the expression of DEGs involved in different groups via RNA-seq. (E) Venn Diagram of CUT&Tag and RNA-seq between PANC-1 and CFPAC-1 cell lines. (F) IGV snapshots showing H3K18la-binding signals in the TSS region of ACAT2 in different groups. (G) ChIP-qPCR analysis of the indicated promoters was performed using antibodies against H3K18la in PANC-1 and CFPAC-1 cells treated with or without 20 mM oxamate for 24 hours. Two-tailed unpaired t test. (H) Immunoblotting of ChIP assays illustrating the binding sites at the ACAT2 promoters against H3K18la. (I, J) Representative images (I) and quantification (J) of ACAT2 levels visualised by immunofluorescence staining in tumour and normal tissues (n=174). Scale bar: left panel, 100 µm; right panel, 20 µm. Two-tailed paired t-test. (K) Correlation between H3K18la levels and ACAT2 expression (n=174). Spearman correlation analysis. (L, M) Western blot analysis of the indicated proteins from PANC-1 cells with different treatments. Lactate 20 mM, C646 100 mM or TSA 1 mM for 24 hours. (N) Western blot analysis of EP300, H3K18la and ACAT2 in PANC-1 cell lines. (O) ChIP-qPCR analysis of the indicated promoters was performed using antibodies against H3K18la in PANC-1 and CFPAC-1 cells with different treatments. Two-tailed unpaired t-test. (P) Immunoblotting of ChIP assay results illustrating the binding sites at the ACAT2 promoters against H3K18la. Data are represented as the mean±SD. *p<0.05; **p<0.01. PDAC, pancreatic ductal adenocarcinoma; TSA, trichostatin a.

Figure 3. ACAT2 induces MTCH2 acetylation and enhances its protein stability. (A, B) Sliver staining and mass spectrum analysis of ACAT2 interacting proteins. (C) Co-precipitation of endogenous MTCH2 with ACAT2 from PANC-1 and CFPAC-1 cells. (D) Co-precipitation of exogenous MTCH2 with ACAT2 from HEK 293 T cells. (E) HEK 293 T cell lysates were incubated with purified GST-MTCH2 (31–170 amino acids), and bound ACAT2 was detected by western blot using anti-ACAT2. (F) Co-precipitation of GST-tagged MTCH2 truncats with Flag-ACAT2 from HEK-293T cells. (G, H) Acetylation of MTCH2 in PDAC cells detected by western blot after co-precipitation. Flag-tagged ACAT2 (H) was co-expressed in ACAT2 knockout cell lines. IP with MTCH2 antibody and IB with Ace-lys antibody. (I) In vitro MTCH2 acetylation assay. Purified GST-MTCH2 was incubated with Flag-ACAT2 immunoprecipitated from HEK 293 T cells, in the presence or absence of acetyl-CoA. (J) Alignment of MTCH2 amino acid sequence from various species. (K, L) Acetylation of MTCH2 mutants expressed in HEK 293 T cells. HA-tagged MTCH2 mutants cells were transfected with or without Flag-ACAT2. (M) MTCH2 protein expression in ACAT2-knockout or control PANC-1 cells. Bottom: The dashed line intersects the degradation curves at the half-life. (N, O) Co-precipitation of ubiquitylated proteins with HA-MTCH2. HA-MTCH2 was immunoprecipitated by anti-HA from HEK 293 T cells transiently expressing HA-tagged WT MTCH2 or each of the MTCH2 mutants, and co-precipitated ubiquitylated proteins were detected using anti-Myc. (P) HEK 293 T cells were coexpressed with different vectors. IP and IB with indicated antibodies. IB, immunoblotting; IP, immunoprecipitation; PDAC, pancreatic ductal adenocarcinoma.

Figure 4. Lactate/H3K18la/ACAT2/MTCH2 forms a positive feedback loop in pancreatic cancer. (A, B) OCR in MTCH2-knockdown or control PDAC cells. Two-tailed unpaired t test. (C) Intracellular lactate levels in PDAC cells with or without MTCH2 knockdown. Two-tailed unpaired t-test. (D, E) OCR in ACAT2 knockdown, ACAT2 knockdown co-transfected with MTCH2 overexpression or control PDAC cells. Two-tailed unpaired t-test. (F) Intracellular lactate levels in ACAT2 knockdown, ACAT2 knockdown co-transfected with MTCH2 overexpression or control PDAC cells. Two-tailed unpaired t-test. (G) qChIP analysis of the indicated promoters was performed using antibodies against H3K18la in ACAT2 knockdown, ACAT2 knockdown co-transfected with MTCH2 overexpression or control PDAC cells. (H) Western blot analysis shows the levels of indicated proteins in ACAT2 knockdown, MTCH2 overexpression, ACAT2 knockdown co-transfected with MTCH2 overexpression or control PDAC cells. (I) Schematic illustration of a positive feedback loop between H3K18la and ACAT2 in pancreatic cancer. Data are represented as the mean±SD. *p<0.05; **p<0.01; ***p<0.001. PDAC, pancreatic ductal adenocarcinoma.

Figure 5. Acat2 deficiency induces an immunosuppressive tumour-associated macrophage compartment and CD8⁺ T-cell dysfunction in pancreatic cancer tumour microenvironment. (A) Diagram outlining the approach to testing inhibition of ACAT2 in immune-competent and immune-deficient tumour-bearing mice. (B–D) The effects of Acat2 knockout on Panc02 tumour growth were evaluated in NCG mice (n=5 mice per group). Representative images of tumours (B), growth curves (C) and tumour weight (D) for indicated groups were presented. Two-tailed unpaired t-test. (E–G) The effects of Acat2 knockout on Panc02 tumour growth were evaluated in C57BL/6 mice (n=5 mice per group). Representative images of tumours (E), growth curves (F) and tumour weight (G) for indicated groups were presented. Two-tailed unpaired t test. (H) Schematic illustration of CyTOF data acquisition (n=5 per group). (I) Heatmap of the expression of the 42 markers for immune cell lineage in the 12 cell populations obtained by merging of the 29 meta-clusters. (J, K) t-SNE plot of Panc02-OVA tumour-infiltrating leukocytes subpopulations overlaid with colour-coded clusters. (L) Frequency of tumour-infiltrating immune cell clusters in the Acat2^WT and Acat2^KO group (n=5 per group). (M) Representative mIHC images of ACAT2, CD8 and CD206 in human PDAC tissues. Scale bar: upper, 100 µm; bottom, 20 µm. (N, O) 5×10⁵Acat2^WT or Acat2^KO Panc02-OVA cells were pancreatic orthotopic injected into C57BL/6 mice (n=5). Representative flow cytometry plots and quantification of tumour-infiltrating CD8⁺ T cells, granzyme B⁺CD8⁺ T cells, IFN-γ⁺CD8⁺ T cells (N) and M2-like macrophages (CD206⁺) (O) for the indicated groups were presented. Two-tailed unpaired t-test. (P) Experimental setup and flow cytometry validation (bottom) of orthotopic pancreatic tumour-bearing C57BL/6 mice with depletion of macrophages. (Q, R) 5×10⁵Acat2^WT or Acat2^KO Panc02-OVA cells were pancreatic orthotopic injected into C57BL/6 mice with or without macrophages depletion (n=5). Representative images of tumours (Q) and flow cytometry of tumour-infiltrating CD8⁺ T cells for the indicated groups (R) were presented. Two-tailed unpaired t test. (S) 5×10⁵Acat2^WT or Acat2^KO Panc02-OVA cells were pancreatic orthotopic injected into NCG mice. Quantification of tumour-infiltrating M2-like macrophages was presented. Two-tailed unpaired t test. (T) The schematic of the OT-1 CD8⁺ T cells treatment schedule. (U, V) 5×10⁵Acat2^WT or Acat2^KO Panc02-OVA cells were pancreatic orthotopic injected into NCG mice with or without OT-1 CD8⁺ T cells adoptively transferred. Tumour-infiltrating activated CD8⁺ T cells were assessed by flow cytometry (U) and tumour weight (V) was shown. Two-tailed unpaired t test. Data are represented as the mean±SD. *p<0.05; **p<0.01; ***p<0.001. CyTOF, cytometry by time-of-flight; mIHC, multiplex immunohistochemistry; ns, not significant; OVA, ovalbumin; PDAC, pancreatic ductal adenocarcinoma; s.c., subcutaneous;.

Figure 6. ACAT2-induced increase in cholesterol delivery to TAMs causes an immunosuppressive TME. (A) Schematic of in vitro macrophage system. (B) BMDMs isolated from C57BL/6 mice were cultured in DMEM medium with 20 ng/mL M-CSF. On day 6, cells were treated with TCM from Panc02-OVA (A) for 48 hours. The MFI of M2-like macrophages in each group of macrophages was detected by flow cytometry (n=3). Two-tailed unpaired t-test. (C) Heatmap showing the expression levels of marker genes in the identified macrophage clusters from indicated groups. (D) The IL-10 and TGF-β levels of macrophages cultured with FBS-free medium for 24 hours, which were prior treated with TCM from Panc02-OVA and fresh medium at a ratio of 1:1 for 24 hours using ELISA. Two-tailed unpaired t-test. (E) Schematic of in vitro co-culture system. (F) Flow cytometry of activated CD8⁺ T cells co-cultured with macrophages from (E). (G) Filipin III staining of Acat2^WT, Acat2^KO and Acat2^OE Panc02-OVA cells were detected by flow cytometry (n=3). (H) BMDMs isolated from C57BL/6 mice were cultured in DMEM medium with 20 ng/mL M-CSF. On day 6, cells were treated with TCM (A) for 48 hours. The Filipin III staining in each group of macrophages was detected by flow cytometry (n=3). Two-tailed unpaired t-test. (I) Cholesterol content of Panc02-OVA cells-derived sEVs. Two-tailed unpaired t test. (J) Schematic of in vitro co-culture system. (K, L) BMDMs isolated from C57BL/6 mice were cultured in DMEM medium with 20 ng/mL M-CSF. On day 6, cells were treated with sEVs for 48 hours. The Filipin III staining (K) and the quantification of M2-like MFI (L) in each group of macrophages were detected by flow cytometry (n=3). Two-tailed unpaired t test. (M) Schematic of in vitro co-culture system. (N) Flow cytometry of activated CD8⁺ T cells co-cultured with macrophages from (M). Two-tailed unpaired t test. (O, P) Pancreatic orthotopic tumour-bearing mice were intravenously injected with 100 µg sEVs on days 7 and 14 after tumour cell inoculation. Representative images of tumours (O), representative flow cytometry plots and quantification of tumour-infiltrating CD8⁺ T cells, granzyme B⁺CD8⁺ T cells and IFN-γ⁺CD8⁺ T cells (P) for the indicated groups were presented. Two-tailed unpaired t test. (Q) BMDMs cultured in medium with 20 ng/mL M-CSF for 6 days were treated with 2 µg PKH67-labelled sEVs±amiloride (10 mM), chlorpromazine (10 mM) or EIPA (5 mM) for 12 hours. The MFI of PKH67 was assessed by flow cytometry. Two-tailed unpaired t-test. (R) BMDMs cultured in medium with 20 ng/mL M-CSF for 6 days were treated with 2 µg sEVs isolated from TCM derived from Panc02-OVA. The cells were cultured with LPDS±10 µM GW4869 for 2 days. The MFI of CD206 was assessed by flow cytometry. Two-tailed unpaired t test. (S) BMDMs were cultured in medium with 20 ng/mL M-CSF for 6 days, followed by a medium containing LPDS±chlorpromazine (10 mM) and 2 µg of sEVs from Acat2^WT, Acat2^KO and Acat2^OE Panc02-OVA cells for 48 hours. Indicated proteins of macrophages were detected by western blot. Data are represented as the mean±SD. *p<0.05; **p<0.01; ***p<0.001. BMDM, bone marrow-derived macrophages; DMEM, dulbecco's modified eagle's medium; LPDS, lipoprotein- and sEV-depleted fetal bovine serum; M-CSF, macrophage colony-stimulating factor; MFI, mean fluorescence intensity; sEVs, small extracellular vesicles; TAMs, tumour-associated macrophages; TCM, tumour cell-derived conditioned medium; TME, tumour microenvironment.

Figure 7. Targeted degradation of Acat2 or Acat2 deficiency sensitises ICB therapy in pancreatic cancer. (A) The schematic of the αPD-1 treatment schedule. (B–D) 5×10⁵Acat2^WT or Acat2^KO Panc02-OVA cells were pancreatic orthotopic injected into C57BL/6 mice with or without αPD-1 treatment (n=5). Representative images of tumours (B), tumour weight (C) and bioluminescence images (D) for the indicated groups were presented. One-way ANOVA. (E) Percentage of tumour-infiltrating CD8⁺ T cells, granzyme B⁺CD8⁺ T cells, IFN-γ⁺CD8⁺ T cells by flow cytometry. One-way ANOVA. (F) Representative flow cytometry plots and quantification of M2-like macrophages (CD206⁺). One-way ANOVA. (G) IHC staining of CD8⁺ and CD206⁺ cells in tumour. Number of CD8-positive cells and CD206-positive cells per HPF was counted in tumour sections from each group. Five random HPFs were selected for analysis on each slide. Scale bar, 50 µm. One-way ANOVA. (H) Survival curve. Log rank test. (I) Schematic illustration of the design strategy of ACAT2 PROTAC. (J) Predicted binding model of compound pumecitinib with ACAT2. (K) Chemical structures of the PROTAC AP1. (L) DC50 of Panc02 cells treated with AP1 with various concentrations for 24 hours. (M) The schematic of the AP1 and αPD-1 treatment schedule. (N–P) 5×10⁵Acat2^WT Panc02-OVA cells were pancreatic orthotopic injected into C57BL/6 mice with or without AP1/αPD-1 treatment (n=5). Tumour weight (N), percentage of CD8⁺ T cells, GZMB⁺CD8⁺ T cells, IFN-γ⁺CD8⁺ T cells (O) and quantification of CD206 (P) for the indicated groups were presented. One-way ANOVA. Data are represented as the mean±SD. *p<0.05; **p<0.01; ***p<0.001. ANOVA, analysis of variance; HPF, High Power Field; ICB, immune checkpoint blockade; IHC, immunohistochemistry; PROTAC, proteolysis-targeting chimaera.

Figure 8. Schematic cartoon illustrating the roles of histone lactylation-induced ACAT2/MTCH2/lactate positive feedback loop in orchestrating cholesterol-mediated immunosuppressive reprogramming in pancreatic cancer. PROTAC, proteolysis-targeting chimaera; sEVs, small extracellular vesicles; TAM, tumour-associated macrophages.