A nuclear branched-chain amino acid catabolism pathway controls histone propionylation in pancreatic cancer

Abstract

Branched-chain amino acid (BCAA) catabolism contributes prominently to the TCA cycle in the healthy pancreas but is suppressed in pancreatic ductal adenocarcinoma (PDA). The impact of this metabolic remodeling on cancer phenotypes remains poorly understood. Here, we find that the BCAA isoleucine is a primary source of propionyl-CoA in PDA cells. Reduction of propionyl-CoA availability by either genetic perturbation or isoleucine and valine starvation decreases histone propionylation (Kpr) without impacting histone acetylation on specific lysine sites, correlating with reduced transcription of certain lipid- and immune-related genes. Mechanistically, we find that multiple enzymes of isoleucine catabolism unexpectedly localize to and carry out multi-step isoleucine oxidation within the nuclei of PDA cells. Importantly, nuclear localization of the rate-limiting branched-chain alpha ketoacid dehydrogenase (BCKDH) complex is essential for isoleucine-dependent Kpr and gene regulation. Moreover, we demonstrate that isoleucine-sensitive Kpr and its associated gene expression are driven by the MYST family of lysine acyltransferases (KATs), and that the BCKDHA subunit of the BCKDH complex interacts with KAT7 within the nuclear compartment. BCAA catabolism enzymes are apparent in the nuclei of PanIN lesions in mice and PDA tumors in patients, contrasting that in healthy pancreatic acinar and ductal cells. Collectively, these findings unveil a nuclear isoleucine catabolism pathway and highlight its role in controlling histone Kpr and tumorigenic transcriptional programs in PDA.

Fig. 1.. BCAA catabolism regulates histone propionylation and tumor growth in PDA.

(A) Bar graph showing percent labeling of succinyl-CoA M+3, acetyl-CoA M+2 and propionyl-CoA M+3 from 6 h [U-¹³C]-isoleucine tracing in a panel of cells. The schematic represents carbon labeling contribution from Ile to the indicated CoA metabolites (B) Pr-CoA absolute abundance measured by LC-MS after 24 h culture of KPC2838c3 cells in complete media (Fed), starvation (−Ile-Val), or 2 mM propionate rescue (−Ile-Val + Propionate). (C to G) Western blots of acid extracted histones after 24 h culture +/− Ile and Val, +/− 2 mM propionate in indicated cell lines. Dialyzed serum was used in all experiments to reduce compensation from other metabolites, such as propionate, that are present in the normal calf serum (H) Schematic diagram of the BCAA Ile catabolic pathway. (I) [U-¹³C]-BCAAs (valine, isoleucine, leucine) were traced into pr-CoA M+3 in DBT-KO (clones 7 and 11) or NC control KPC2838c3 cells. (J) Pr-CoA relative abundance measured by LC-MS in whole-cell DBT-KO11 or NC KPC2838c3. (K to L) Western blots of histones extracted from NC or DBT-KO7/11 cells and treated with propionate. The propionate concentration used for cell treatment was 2 mM unless otherwise specified. (M) Western blot in whole-cell extracts or acid extracted histones from NC, DBT-KO7/11 and DBT-KO11 KPC2838c3 cells reconstituted with the human DBT-WT-Flag-Myc DBT protein. (N) Proliferation of NC, DBT-KO7, and DBT-KO11 KPC2838c3 cell lines over 3 days of culture. (O) Representative images of colony formation assay over 10 days of culture of the indicated KPC2838c3 cell lines. (P) Tumor volume of NC and DBT-KO KPC2838c3 cells subcutaneously implanted in C57BL/6J mice. (Q) Tumor weight of excised NC and DBT-KO allograft tumors at end point. All experiments were repeated at least twice. Data are presented as mean ± s.d. from independent replicates; Statistical significance was assessed using two-tailed unpaired Student’s t-test; *p<0.05, **p<0.01, ***p<0.001.

Fig. 2.. BCAA metabolic enzymes localize to and function in the nuclei of PDA cells.

(A) Western blot of proteins isolated from the non-nuclear, nuclear, and chromatin compartments of NC and DBT-KO KPC2838c3 cells through biochemical fractionation. Whole cell extract (WCE) was used as a control. DBT-KO11 clone was used in the functional studies unless otherwise indicated. (B) Representative immunofluorescence images of the indicated proteins (pink) in NC and DBT-KO KPC2838c3 cells. DNA was stained with DAPI (blue). Scale bars, 30 μm. (C) Nuclei were isolated from KPC2838c3 cells and incubated with 13C6, 15N1-isoleucine (Ile) for 18 h, with or without boiling. ¹³C-labeled metabolites of the Ile catabolic pathway were measured and presented as the area under the curve (AUC) in control or boiled nuclear fractions. (D) AUC of 13C6, 15N1-Ile, α-KMV M+6, and 2-methylbutyryl-CoA M+5 measured by LC-MS in isolated control or boiled nuclei from NC and DBT-KO KPC2838c3 cells incubated with 13C6, 15N1-Ile for 18 h. (E) Immunostaining of Flag tag (green) in DBT-KO, DBT-WT-Flag-Myc addback, and nuclear excluded DBT-NES-Flag-Myc KPC2838c3 cells. Scale bars, 20 μm. (F) Tracing of ¹³C carbon labeling to propionyl-CoA M+3, succinyl-CoA M+3 and acetyl-CoA M+2 in DBT-WT-Flag-Myc and DBT-NES-Flag-Myc cells incubated with [U-¹³C]-isoleucine for 6h. (G) Histones were extracted from the indicated KPC2838c3 engineered cell lines and analyzed by western blot for the proteins listed on the left of each pane. Data are mean ± s.d. from independent biological replicates. Significance was determined by two-tailed unpaired Student’s t-test; *p<0.05, **p<0.01, ***p<0.001.

Fig. 3.. BCAA catabolic enzymes are enriched in the nuclei of PanIN lesions and PDA tumors.

(A) Immunofluorescence staining of BCKDHA (red), ductal marker CK19 (green) and Lamin A/C (yellow) in the pancreas tissue from WT, 8 weeks old Ptf1α-Cre;Kras^LSL-G12D, and 25 weeks Ptf1α-Cre;Kras^LSL-G12D old C57BL/6J mice. Scale bars, 30 μm. (B to C) Subcellular fractionation and western blot analysis of (B) PCRC38 patient-derived cells and (C) PCRC46 patient-derived organoids using antibodies against the indicated proteins. Whole cell extract (WCE) was used as a control. (D to E) Representative immunofluorescence images of the indicated proteins in (D) PCRC38 patient-derived cells and (E) PCRC28 patient-derived xenograft tumor. (F to G) Representative immunostaining images of (F) BCKDHA (red) and (G) HSP60 (red) in normal and PDAC clinical samples (n=4 biologically independent patients). The CK19 ductal cell marker is shown in green and Lamin A/C control is indicated in yellow color. Nuclei were stained with DAPI (blue). At least 6 areas were imaged in each individual tissue. Scale bars, 30 μm.

Fig. 4.. Nuclear BCAA catabolism promotes lipid metabolism and immune related gene signatures in PDA cells.

(A to B) Venn diagrams showing the overlap between downregulated genes in DBT-KO7 and DBT-KO11 with (A) BCAA starvation or (B) Propionate rescue. (C) Pathway analysis of all differentially expressed genes that are downregulated in BCAA starvation and DBT-KO (Kpr-correlated genes). (D) qRT-PCR analysis of representative Kpr-correlated genes in NC, DBT-KO7 +/− propionate, and DBT-KO11 +/− propionate cells. (E) qRT-PCR analysis of representative Kpr-correlated genes in NC, DBT-KO7, DBT-KO11, DBT-KO11 + DBT-WT-Flag-Myc, and DBT-KO11 + DBT-NES-Flag-Myc KPC2838c3 cells. (F) ChIP-qPCR analysis on Ptges, Sh3rf2, Ripor2, and CS genomic loci using the H3K23pr antibody. Primers for each locus were designed near the H3K4me3 ChIP-seq peak using the UCSC genome browser. Relative enrichment was normalized to input and the Ct values of rabbit IgG. Data are presented as mean ± s.d. from three independent experiments. Statistical analysis was performed using two-tailed unpaired Student’s t-test; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=no significance.

Fig. 5.. MYST family KATs regulate BCAA-dependent Kpr and associated transcription.

(A) Western blots of histone proteins extracted from KPC2838c3 cells that were Ile/Val starved for 24 h and then refed for 6 h during which inhibitors targeting the indicated KAT families or DMSO control were added. (B) Western blot of H3K23pr in KPC2838c3 cells that were Ile/Val starved for 24 h and then refed for 16 h in the presence of p300/CBPi [CPI-1612] (0.5 μM), KAT6i [WM-1119] (4 μM), and KAT6/7i [WM-3835] (10 μM) inhibitors. (C) Western blot of extracted histones from Fed, −Ile-Val (24h), and Refed (16h) KPC2838c3 cells that were treated with KAT6/7i inhibitor WM-3835 (10 μM) or DMSO control. (D) Meta-analysis of KAT7, KAT6A, KAT6B, and KAT2A expression levels in PDA patient tissues and matched normal samples from publicly available RNA-seq data extracted from the Gepia data portal (http://gepia2.cancer-pku.cn/#index). (E) Western blot of histone proteins extracted from BCAA-refed KPC2838c3 cells treated with the BRPF1/2/3 inhibitor, OF-1, using antibodies against the indicated histone marks. (F) Immunoprecipitation of BCKDHA protein or IgG control in isolated nuclei from the KPC2838c3 cells followed by western blot analysis probed with the indicated antibodies. (G) qRT-PCR analysis of Ripor2 and Lgi2 genes in −Ile-Val and Refed cells treated with the indicated inhibitors or DMSO control. (H) Model showing the regulation of Kpr and transcription by the nuclear isoleucine catabolic pathway in PDA. Generated using Biorender. Data are from at least three experiments. Statistical significance was determined by Student’s t-test; error is reported as SD (*p<0.05, **p<0.01, ***p<0.001, ****p<0.0001, ns=no significance).