Advertisement

Metabolic Rewiring by Loss of Sirt5 Promotes Kras-Induced Pancreatic Cancer Progression

Open AccessPublished:July 07, 2021DOI:https://doi.org/10.1053/j.gastro.2021.06.045

      Background & Aims

      SIRT5 plays pleiotropic roles via post-translational modifications, serving as a tumor suppressor, or an oncogene, in different tumors. However, the role SIRT5 plays in the initiation and progression of pancreatic ductal adenocarcinoma (PDAC) remains unknown.

      Methods

      Published datasets and tissue arrays with SIRT5 staining were used to investigate the clinical relevance of SIRT5 in PDAC. Furthermore, to define the role of SIRT5 in the carcinogenesis of PDAC, we generated autochthonous mouse models with conditional Sirt5 knockout. Moreover, to examine the mechanistic role of SIRT5 in PDAC carcinogenesis, SIRT5 was knocked down in PDAC cell lines and organoids, followed by metabolomics and proteomics studies. A novel SIRT5 activator was used for therapeutic studies in organoids and patient-derived xenografts.

      Results

      SIRT5 expression negatively regulated tumor cell proliferation and correlated with a favorable prognosis in patients with PDAC. Genetic ablation of Sirt5 in PDAC mouse models promoted acinar-to-ductal metaplasia, precursor lesions, and pancreatic tumorigenesis, resulting in poor survival. Mechanistically, SIRT5 loss enhanced glutamine and glutathione metabolism via acetylation-mediated activation of GOT1. A selective SIRT5 activator, MC3138, phenocopied the effects of SIRT5 overexpression and exhibited antitumor effects on human PDAC cells. MC3138 also diminished nucleotide pools, sensitizing human PDAC cell lines, organoids, and patient-derived xenografts to gemcitabine.

      Conclusions

      Collectively, we identify SIRT5 as a key tumor suppressor in PDAC, whose loss promotes tumorigenesis through increased noncanonic use of glutamine via GOT1, and that SIRT5 activation is a novel therapeutic strategy to target PDAC.

      Graphical abstract

      Keywords

      Abbreviations used in this paper:

      AOA (aminooxyacetate), Asp (aspartate), CS (Pdx1-Cre, Sirt5fl/fl), GLUD1 (glutamate dehydrogenase 1), GOT (glutamate-oxaloacetate transaminase), GSH (glutathione), IC50 (half maximal inhibitory concentration), KCS (Pdx1-Cre, LSL-KrasG12D, Sirt5fl/fl), KPC (KrasG12D/+, p53R172H/+, Pdx1-Cretg/+), KPCS (Pdx1-Cre, LSL-KrasG12D, LSL-Tp53R172H, Sirt5fl/fl), ME1 (malic enzyme 1), mRNA (messenger RNA), NAD (nicotinamide adenine dinucleotide), NADPH (reduced nicotinamide adenine dinucleotide phosphate), PanIN (pancreatic intraepithelial neoplasia), PCR (polymerase chain reaction), PDAC (pancreatic ductal adenocarcinoma), PDX (patient-derived xenografts), SIRT (sirtuin), UNMC (University of Nebraska Medical Center)
      See Covering the Cover synopsis on page 1352; See editorial on page 1376.

       Background and Context

      Pancreatic ductal adenocarcinoma (PDAC) cells with activating KRAS mutations use glutamine in a noncanonic way via the GOT2-GOT1-ME1 pathway. SIRT5 has pleiotropic roles, serving as an oncogene and tumor suppressor in different cancers.

       New Findings

      SIRT5 is a novel tumor suppressor for pancreatic cancer and genetically targeting SIRT5 makes PDAC tumors more aggressive. SIRT5 negatively regulates noncanonic glutamine metabolism via direct post-translational modifications of GOT1. Activation of SIRT5 with a novel small molecule is a potential new therapy for PDAC with low-SIRT5 levels.

       Limitations

      This study was conducted in PDAC cell lines, genetically engineered mice models, and human tissue samples (tissue microarrays, PDAC organoids, and patient-derived xenografts). More work needs to be done in human patients.

       Impact

      SIRT5 activation (small-molecule SIRT5-activator MC3138) combined with gemcitabine may lead to a novel safe and effective therapy for patients with PDAC with low–tumoral SIRT5 expression.
      A remarkable feature of pancreatic ductal adenocarcinoma (PDAC) is the hypoxic, harsh, and nutrient-deficient tumor microenvironment.
      • Derle A.
      • De Santis M.C.
      • Gozzelino L.
      • et al.
      The role of metabolic adaptation to nutrient stress in pancreatic cancer.
      Accordingly, PDAC cells tend to reprogram their metabolic pathways to survive and grow under these harsh conditions.
      • Derle A.
      • De Santis M.C.
      • Gozzelino L.
      • et al.
      The role of metabolic adaptation to nutrient stress in pancreatic cancer.
      • Vernucci E.
      • Abrego J.
      • Gunda V.
      • et al.
      Metabolic alterations in pancreatic cancer progression.
      • Chaika N.V.
      • Yu F.
      • Purohit V.
      • et al.
      Differential expression of metabolic genes in tumor and stromal components of primary and metastatic loci in pancreatic adenocarcinoma.
      Therefore, a deeper understanding of the biological features and molecular regulation of metabolic dependencies of PDAC is strongly expected to provide novel targets for therapeutic interventions.
      Unlike many other cells using glutamine to feed the tricarboxylic acid cycle via glutamate dehydrogenase 1 (GLUD1), PDAC cells metabolize glutamine in a noncanonic way.
      • Gebregiworgis T.
      • Purohit V.
      • Shukla S.K.
      • et al.
      Glucose limitation alters glutamine metabolism in MUC1-overexpressing pancreatic cancer cells.
      ,
      • Son J.
      • Lyssiotis C.A.
      • Ying H.
      • et al.
      Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
      Mitochondrial aspartate aminotransferase (glutamate-oxaloacetate transaminase 2 [GOT2]) converts glutamine to aspartate (Asp). Glutamine-derived Asp is transported to the cytoplasm. Next, cytosolic Asp aminotransferase (GOT1) converts glutamine-derived Asp to oxaloacetate, which is converted to malate by cytosolic malate dehydrogenase. Then, the cytoplasmic malic enzyme 1 (ME1) catalyzes the oxidative decarboxylation of malate to pyruvate, with a concomitant production of reduced nicotinamide adenine dinucleotide phosphate (NADPH), sustaining the redox homeostasis in PDAC cells. This noncanonic glutamine metabolic pathway is dispensable in normal cells, but it has the potential to serve as a promising therapeutic target in PDAC.
      Sirtuins (SIRT1–SIRT7) are a class of evolutionarily conserved nicotinamide adenine dinucleotide (NAD+)–dependent enzymes that possess deacetylase, desuccinylase, deglutarylase, and demalonylase activity.
      • Li M.
      • Chiang Y.L.
      • Lyssiotis C.A.
      • et al.
      Non-oncogene addiction to SIRT3 plays a critical role in lymphomagenesis.
      ,
      • Winnik S.
      • Auwerx J.
      • Sinclair D.A.
      • et al.
      Protective effects of sirtuins in cardiovascular diseases: from bench to bedside.
      Sirtuins have been implicated in cancer progression due to their roles in regulating cancer cell metabolism, tumor microenvironment, and genome stability.
      • Chalkiadaki A.
      • Guarente L.
      The multifaceted functions of sirtuins in cancer.
      Among the 7 mammalian sirtuins, SIRT5 has been shown to be located in both the mitochondria and the cytosol.
      • Matsushita N.
      • Yonashiro R.
      • Ogata Y.
      • et al.
      Distinct regulation of mitochondrial localization and stability of two human Sirt5 isoforms.
      • Park J.
      • Chen Y.
      • Tishkoff D.X.
      • et al.
      SIRT5-mediated lysine desuccinylation impacts diverse metabolic pathways.
      • Nishida Y.
      • Rardin M.J.
      • Carrico C.
      • et al.
      SIRT5 regulates both cytosolic and mitochondrial protein malonylation with glycolysis as a major target.
      Given the pleiotropic role of SIRT5 in regulating cancer cell metabolism, it potentially can act as a tumor suppressor, or in contrast as an oncogene, in different cancer types and microenvironments by mediating the post-translational modification of its target substrates.
      • Yang X.
      • Wang Z.
      • Li X.
      • et al.
      SHMT2 desuccinylation by SIRT5 drives cancer cell proliferation.
      • Wang Y.Q.
      • Wang H.L.
      • Xu J.
      • et al.
      Sirtuin5 contributes to colorectal carcinogenesis by enhancing glutaminolysis in a deglutarylation-dependent manner.
      • Li F.
      • He X.
      • Ye D.
      • et al.
      NADP(+)-IDH mutations promote hypersuccinylation that impairs mitochondria respiration and induces apoptosis resistance.
      • Chen X.F.
      • Tian M.X.
      • Sun R.Q.
      • et al.
      SIRT5 inhibits peroxisomal ACOX1 to prevent oxidative damage and is downregulated in liver cancer.
      • Greene K.S.
      • Lukey M.J.
      • Wang X.
      • et al.
      SIRT5 stabilizes mitochondrial glutaminase and supports breast cancer tumorigenesis.
      • Yan D.
      • Franzini A.
      • Pomicter A.D.
      • et al.
      Sirt5 is a druggable metabolic vulnerability in acute myeloid leukemia.
      • Giblin W.
      • Bringman-Rodenbarger L.
      • Guo A.H.
      • et al.
      The deacylase SIRT5 supports melanoma viability by influencing chromatin dynamics.
      • Abril Y.L.N.
      • Fernandez I.R.
      • Hong J.Y.
      • et al.
      Pharmacological and genetic perturbation establish SIRT5 as a promising target in breast cancer.
      Yet, the function of SIRT5 in the initiation and progression of PDAC remains obscure.
      The present study revealed SIRT5 to be down-regulated in both human PDAC tissues and murine pancreatic tumors. Furthermore, we found SIRT5 deletion to accelerate tumor growth and to correlate with poor survival in both patients with PDAC and genetically engineered mouse models. By integrating mass spectrometry–based metabolomics and proteomics with biochemical assays, GOT1 was identified as the critical target substrate in SIRT5-regulated PDAC progression. Mechanistically, SIRT5 deletion enhanced the enzyme activity of GOT1 by facilitating GOT1 acetylation. Importantly, SIRT5 activation with a novel small-molecule activator MC3138 inhibited proliferation in SIRT5-low PDAC cell lines and organoids. Furthermore, SIRT5 activation decreased nucleotide levels and facilitated synergy with gemcitabine in human PDAC organoids and patient-derived xenografts (PDX). Thus, the present study demonstrates SIRT5 activation, in combination with gemcitabine, as a potential therapeutic strategy in PDAC.

      Materials and Methods

       Cell Culture

      Human PDAC cell lines Capan1, Capan2, AsPC1, PANC-1, CFPAC-1, SW1990, and MIAPaCa-2, and HEK 293T were obtained from the American type culture collection (ATCC, Rockville, MD). The T3M4, S2-007, HPAF-II, Patu8902, Colo 357 (FG), and S2-013 cell lines were provided by Dr Michael A. Hollingsworth (Eppley Institute, University of Nebraska Medical Center [UNMC], Omaha, NE). The murine PDAC cell lines KPC1245 derived from KrasG12D/+, p53R172H/+, Pdx1-Cretg/+ (KPC) mice were provided by Dr David Tuveson (Cold Spring Harbor Laboratory; CSHL). Mouse PDAC cell lines KPC7460, KPC7472, and KPCS8508, KPCS8009 were derived from KPC and KPCS mice, respectively. Cells were cultured in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin and incubated at 37°C in a 5% CO2-humidified incubator. The cells were authenticated using Short Tandem Repeat (STR) profiling and tested for mycoplasma contamination using a polymerase chain reaction (PCR)-based Mycoplasma Detection Kit.

       Human Studies

      The UNMC Tissue Bank provided the pancreatic cancer tissues used for PanIN and tumor progression studies, and PDX models. The UNMC Institutional Review Board Committee approved the studies related to the use of human samples for IHC in pancreatic intraepithelial neoplasia (PanIN) lesions. The UNMC Institutional Review Board Committee also approved the informed consent waiver. The pancreatic cancer tissue microarray used for survival analysis was obtained from Shanghai Outdo Biotech (Shanghai, China).

       Mouse Strain Generation

      The conditional LSL-KrasG12D, Pdx1-Cre,
      • Hingorani S.R.
      • Petricoin E.F.
      • Maitra A.
      • et al.
      Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.
      and LSL-Tp53R172H
      • Olive K.P.
      • Tuveson D.A.
      • Ruhe Z.C.
      • et al.
      Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome.
      stains were purchased from the National Cancer Institute. Floxed Sirt5 (Sirt5fl/fl) mice were provided by Dr Johan Auwerx.
      • Du J.
      • Zhou Y.
      • Su X.
      • et al.
      Sirt5 is a NAD-dependent protein lysine demalonylase and desuccinylase.
      These mice strains were interbred to produce the following experimental cohorts: Pdx1-Cre, Sirt5fl/fl (CS); Pdx1-Cre, LSL-KrasG12D (KC); Pdx1-Cre, LSL-KrasG12D, Sirt5fl/fl (KCS); Pdx1-Cre, LSL-KrasG12D, LSL-Tp53R172H (KPC); and Pdx1-Cre, LSL-KrasG12D, LSL-Tp53R172H, Sirt5fl/fl (KPCS). For Sirt5 loss-of-function studies, wild-type, Sirt5fl/fl, or Pdx1-Cre littermates were used as controls. These experiment cohort strains were genotyped using PCR amplification methods. PCR primers used in genotyping are listed in Supplementary Table 4. For caerulein-induced pancreatitis, 7-week-old mice of indicated genotypes were intraperitoneally injected with caerulein (50 μg/kg) for 2 consecutive days every hour for 6 hours each day.
      Additional methods are provided in the Supplementary Material.

      Results

       SIRT5 Down-Regulation Correlates With Disease Progression, Adverse Patient Outcomes, and Increased Tumor Cell Growth

      The sirtuin family of proteins represents an important node in the regulation of cellular transcriptional response when presented with systemic metabolic cues or metabolic/nutritional changes.
      • Dasgupta A.
      • Shukla S.K.
      • Vernucci E.
      • et al.
      SIRT1-NOX4 signaling axis regulates cancer cachexia.
      To investigate whether sirtuins are dysregulated in PDAC, the present study evaluated the messenger RNA (mRNA) expression of all the 7 sirtuins in matched human normal and PDAC tissues from the GEO database (GDS4103). Results showed that expression levels of SIRT3, SIRT5, and SIRT6 were significantly decreased in PDAC tissues, whereas expression levels of SIRT1 and SIRT7 were remarkably increased in PDAC (Figure 1A, Supplementary Figure 1A–F). Among the dysregulated sirtuins, SIRT5 remains the only sirtuin that has not been investigated for its role in PDAC. To corroborate the expression data from GEO, the public database Oncomine was queried and that further showed SIRT5 expression to be down-regulated in PDAC tissues (Supplementary Figure 1G–H).
      Figure thumbnail gr1al
      Figure 1SIRT5 down-regulation in PDAC correlates with disease progression, poor survival outcomes, and enhanced tumor cell growth. (A) SIRT5 mRNA levels in pancreatic cancer tissues and the paired adjacent normal tissues (39 pairs) from GEO database. (B–D) Sirt5 mRNA (B) and protein (C–D) levels in 25-week-old Cre control and KPC pancreatic cancer tissues. (E–F) Immunohistochemistry (IHC) staining and quantification for SIRT5 expression in adjacent normal tissues, low-grade and high-grade PanINs, and pancreatic tumors. Scale bar 100 μm. (G) Representative IHC staining for SIRT5 expression in pancreatic cancer tissue microarrays. Scale bar 500 μm. (H–I) Survival analysis of patients with pancreatic cancer categorized by low- and high-SIRT5 expression. (J–K) Representative images and cell viability for control and SIRT5-knockdown organoids cultured for 7 days. Scale bar 1000 μm. (L–N) Representative image, tumor volume (mean ± SE) and tumor weight of control and SIRT5-knockdown tumors. (O–P) Ki67 staining and quantitation in control and SIRT5-knockdown tumors. Scale bar 100 μm. For all in vitro studies, n ≥ 3. The data are represented as mean ± standard deviation (SD). Paired Student t test (A, B, D), 1-way analysis of variance (ANOVA) with Bonferroni test (K, N, P) or Tukey test (F), 2-way ANOVA with Bonferroni test (M), log-rank test (H–I), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      Figure thumbnail gr1mp
      Figure 1SIRT5 down-regulation in PDAC correlates with disease progression, poor survival outcomes, and enhanced tumor cell growth. (A) SIRT5 mRNA levels in pancreatic cancer tissues and the paired adjacent normal tissues (39 pairs) from GEO database. (B–D) Sirt5 mRNA (B) and protein (C–D) levels in 25-week-old Cre control and KPC pancreatic cancer tissues. (E–F) Immunohistochemistry (IHC) staining and quantification for SIRT5 expression in adjacent normal tissues, low-grade and high-grade PanINs, and pancreatic tumors. Scale bar 100 μm. (G) Representative IHC staining for SIRT5 expression in pancreatic cancer tissue microarrays. Scale bar 500 μm. (H–I) Survival analysis of patients with pancreatic cancer categorized by low- and high-SIRT5 expression. (J–K) Representative images and cell viability for control and SIRT5-knockdown organoids cultured for 7 days. Scale bar 1000 μm. (L–N) Representative image, tumor volume (mean ± SE) and tumor weight of control and SIRT5-knockdown tumors. (O–P) Ki67 staining and quantitation in control and SIRT5-knockdown tumors. Scale bar 100 μm. For all in vitro studies, n ≥ 3. The data are represented as mean ± standard deviation (SD). Paired Student t test (A, B, D), 1-way analysis of variance (ANOVA) with Bonferroni test (K, N, P) or Tukey test (F), 2-way ANOVA with Bonferroni test (M), log-rank test (H–I), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      To establish whether SIRT5 plays a causative role in the pathogenesis of PDAC, we used the autochthonous LSL-KrasG12D; LSL-Trp53R172H; Pdx1-Cre (KPC) mouse model and observed that Sirt5 mRNA and protein levels were significantly decreased in KPC pancreatic cancer tissues compared with the age-matched normal pancreas tissues (Figure 1B–D). A decrease in SIRT5 expression was noted during the progression from adjacent normal tissues to PanIN and PDAC in KPC mouse models (Supplementary Figure 1I–J) and human samples (Figure 1E–F). Furthermore, low-SIRT5 expression correlated with tumor recurrence in human patients (Supplementary Table 1). Low expression of SIRT5 was associated with decreased survival of patients with PDAC (Figure 1G–I). Moreover, low-SIRT5 expression independently correlated with high mortality risk in patients with PDAC (Supplementary Figure 1E and Supplementary Tables 1–3).
      To examine the effect of SIRT5 on PDAC aggressiveness, SIRT5 expression was evaluated in 13 PDAC cell lines (Supplementary Figure 1L). Cells with relatively high expression of SIRT5 (ie, T3M4, Capan2, PANC1, and S2-007) were used to generate stable SIRT5 knockdowns, whereas those with low SIRT5 (ie, S2-013 and Capan1) were chosen to generate SIRT5-overexpressing cell lines. The efficiency of SIRT5 knockdown and overexpression was verified using quantitative PCR and immunoblotting (Supplementary Figure 1M–Q). SIRT5 knockdown enhanced PDAC cell proliferation, colony formation, and sphere formation (Supplementary Figure 2A–L). Conversely, SIRT5 overexpression inhibited PDAC cell growth, whereas the catalytically inactive mutant SIRT5-H158Y
      • Nishida Y.
      • Rardin M.J.
      • Carrico C.
      • et al.
      SIRT5 regulates both cytosolic and mitochondrial protein malonylation with glycolysis as a major target.
      did not exhibit significant inhibitory effects on PDAC cell growth (Supplementary Figure 2M–P). The expression levels of SIRT5 were also investigated in a panel of 6 human PDAC tumor-derived organoid lines, and SIRT5 expression was knocked down in the PA901 and PA717 organoids with high endogenous SIRT5 levels (Supplementary Figure 1R–S). Knockdown of SIRT5 also increased human PDAC organoid growth (Figure 1J–K). Moreover, SIRT5 knockdown accelerated T3M4 tumor growth upon orthotopic implantation (Figure 1L–P).

       Genetic Ablation of Sirt5 Promotes Pancreatic Tumorigenesis in Cooperation With Oncogenic Mutations

      To determine the role of SIRT5 in PDAC pathogenesis, genetically engineered mice were generated that carried pancreas-specific ablation of Sirt5, along with oncogenic mutations in Kras alone or Kras and Trp53. Pdx1-Cre; Sirt5fl/fl (CS) were generated by crossing Pdx1-Cre mice with Sirt5fl/fl mice (Supplementary Figure 3A–B). As controls, histologic analysis of pancreata isolated from CS and Pdx1-Cre mice did not show any apparent abnormalities (Supplementary Figure 3C–D). Furthermore, there was no significant difference in the body weight changes between these 2 groups (Supplementary Figure 3E). KrasG12D; Sirt5fl/fl mice were crossed with Pdx1-Cre; Sirt5fl/+ mice to generate mice with mutant Kras and heterozygous/homozygous Sirt5 loss, ie, Pdx1-Cre; KrasG12D; Sirt5fl/+; (KCShet) and Pdx1-Cre; KrasG12D; Sirt5fl/fl (KCS), respectively (Supplementary Figure 3A–B and Figure 2A). The cholecystokinin analog caerulein is a widely used drug to induce pancreatitis.
      • Strobel O.
      • Dor Y.
      • Alsina J.
      • et al.
      In vivo lineage tracing defines the role of acinar-to-ductal transdifferentiation in inflammatory ductal metaplasia.
      Wild type, Pdx1-Cre, and CS mice exhibited expected overall nontransformed pancreatic histology 21 days post–caerulein injection (Supplementary Figure 4A–B). Conversely, KCS mice had a larger neoplastic area, increased acinar cell loss (decreased amylase positivity and increased CK19 positivity), and more high-grade PanIN lesions than the Pdx1-Cre; KrasG12D (KC) and KCShet mice (Figure 2B–E). SIRT5 IHC staining of pancreatic tissue from KCS mice confirmed the efficient deletion of Sirt5 (Supplementary Figure 4C). Lesions from the KCS mice displayed a significantly enhanced cell proliferation index, as indicated by increased Ki67-staining (Supplementary Figure 4C–D). Thus, Sirt5 deletion accelerates KrasG12D-triggered pancreatic tumorigenesis in the context of a caerulein-induced pancreatitis model.
      Figure thumbnail gr2ae
      Figure 2Sirt5 deficiency accelerates acinar-to-ductal metaplasia, PanIN formation, and pancreatic tumorigenesis. Cre: Pdx1-Cre; KC: KrasG12D, Pdx1-Cre; KCShet: KrasG12D, Pdx1-Cre, Sirt5fl/+; KCS: KrasG12D, Pdx1-Cre, Sirt5fl/fl. (A) Genetic strategy for investigating the function of Sirt5 in KrasG12D-driven pancreatic tumorigenesis. (B) Intraperitoneally caerulein-injected KC (n = 9), KCShet (n = 9), and KCS (n = 10) mice were euthanized at day 21. (C) H&E, CK19/Amylase, and Alcian blue/eosin staining of pancreatic tissue from caerulein-injected KC, KCShet, and KCS mice. Scale bars 2000 μm (H&E), 100 μm (immunofluorescence), or 500 μm (Alcian blue/eosin staining). (D) The percentage of total PanIN area over the whole pancreatic tissue section from caerulein-treated mice. (E) Histopathologic analysis of pancreatic tissue sections from caerulein-treated mice. (F and I) H&E, CK19/Amylase, and Alcian blue/eosin staining of pancreatic tissue sections from 4-month-old (F) and 8-month-old (I) KC, KCShet, and KCS mice. (G) Percentage of PanIN lesions over the whole pancreatic tissue section from 4-month-old KC (n = 5), KCShet (n = 5), and KCS mice (n = 6). (H and K) Histopathologic analysis of 4-month-old (H) and 8-month-old (K) KC, KCShet, and KCS mice tissue slides. (J) Percentage of PanIN lesion area over the whole pancreatic tissue section from 8-month-old KC (n = 6), KCShet (n = 6), and KCS mice (n = 6). The data are represented as mean ± SD. One-way ANOVA with Tukey test was used for all panels, ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      Figure thumbnail gr2fk
      Figure 2Sirt5 deficiency accelerates acinar-to-ductal metaplasia, PanIN formation, and pancreatic tumorigenesis. Cre: Pdx1-Cre; KC: KrasG12D, Pdx1-Cre; KCShet: KrasG12D, Pdx1-Cre, Sirt5fl/+; KCS: KrasG12D, Pdx1-Cre, Sirt5fl/fl. (A) Genetic strategy for investigating the function of Sirt5 in KrasG12D-driven pancreatic tumorigenesis. (B) Intraperitoneally caerulein-injected KC (n = 9), KCShet (n = 9), and KCS (n = 10) mice were euthanized at day 21. (C) H&E, CK19/Amylase, and Alcian blue/eosin staining of pancreatic tissue from caerulein-injected KC, KCShet, and KCS mice. Scale bars 2000 μm (H&E), 100 μm (immunofluorescence), or 500 μm (Alcian blue/eosin staining). (D) The percentage of total PanIN area over the whole pancreatic tissue section from caerulein-treated mice. (E) Histopathologic analysis of pancreatic tissue sections from caerulein-treated mice. (F and I) H&E, CK19/Amylase, and Alcian blue/eosin staining of pancreatic tissue sections from 4-month-old (F) and 8-month-old (I) KC, KCShet, and KCS mice. (G) Percentage of PanIN lesions over the whole pancreatic tissue section from 4-month-old KC (n = 5), KCShet (n = 5), and KCS mice (n = 6). (H and K) Histopathologic analysis of 4-month-old (H) and 8-month-old (K) KC, KCShet, and KCS mice tissue slides. (J) Percentage of PanIN lesion area over the whole pancreatic tissue section from 8-month-old KC (n = 6), KCShet (n = 6), and KCS mice (n = 6). The data are represented as mean ± SD. One-way ANOVA with Tukey test was used for all panels, ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      Next, the role of Sirt5 in KrasG12D-triggered pancreatic tumorigenesis was examined in the absence of caerulein. The pancreata from 4-month-old KCS mice exhibited significantly larger neoplastic areas as well as higher-grade PanIN lesions compared with age-matched KC and KCShet mice (Figure 2F–H). Ki67 staining results showed that the PanIN ductal cells from KCS mice exhibited an enhanced proliferative index (Supplementary Figure 4E–F). To identify the role of Sirt5 at a later stage of KrasG12D-triggered pancreatic tumorigenesis, pancreatic tissues were collected from 8-month-old KC, KCShet, and KCS mice. All KCS mice obtained numerous PanIN lesions at the age of 8 months, and 2 of 6 KCS mice developed PDAC (Figure 2I–K). The area, grade, and proliferative potential of PanIN lesions in KC as well as KCShet mice were significantly lower than the age-matched KCS mice (Figure 2I–K, Supplementary Figure 4G–H). These data indicate Sirt5 loss facilitates the oncogenic KrasG12D-driven initiation of pancreatic tumorigenesis in the absence of caerulein.
      To further examine the effect of Sirt5 deficiency on pancreatic cancer progression and survival, KPC (Pdx1-Cre; KrasG12D; Trp53R172H) mice with heterozygous/homozygous Sirt5 loss (Pdx1-Cre; KrasG12D; Trp53R172H; Sirt5fl/+, KPCShet and Pdx1-Cre; KrasG12D; Trp53R172H; Sirt5fl/fl, KPCS) were generated (Supplementary Figure 3A–B and Figure 3A). KPCS mice at 15 weeks of age developed large intraperitoneal pancreatic tumor masses, whereas the corresponding age-matched KPC mice had only small pancreatic tumor nodules (Figure 3B). Age-matched KPC and KPCS mice were assessed at different time points to monitor the pancreatic tumorigenesis (Figure 3C). At 6 weeks, KPCS mice had a significantly increased percentage of neoplastic area, and more than half of the KPCS mice exhibited high-grade PanIN, which was significantly faster than the age-matched KPC and KPCShet mice (Figure 3C–E). Complete scans of pancreatic tissue sections indicated that pancreatic tumors developed in KPCS mice as early as 10 weeks (40%). Moreover, 5 of 6 15-week-old KPCS mice developed pancreatic cancer (Figure 3F). Furthermore, Ki67 staining results demonstrated Sirt5 deletion significantly promoted pancreatic cancer cell proliferation (Figure 3G–H). Because Sirt5 deletion in mice with the KPC background accelerated pancreatic cancer progression, KPC, KPCShet, and KPCS mice were then used to investigate the impact of Sirt5 loss on survival. The KPCS mice had a dramatically shortened median survival time at 116 days, whereas the median survival times of KPC and KPCShet mice were 173 days and 148.5 days, respectively (Figure 3I). Therefore, Sirt5 deletion, in cooperation with Kras and Tp53 mutations, accelerates pancreatic cancer progression and shortens the survival time of the spontaneous PDAC progression mouse model.
      Figure thumbnail gr3
      Figure 3Genomic ablation of Sirt5 promotes PDAC progression. (A) Genetic strategy for investigating the function of the Sirt5 in KrasG12D and Trp53R172H-driven pancreatic tumorigenesis. (B) Representative images of pancreatic tumors in 15-week-old KPC and KPCS mice. (C) Representative H&E-stained images of the pancreatic tissue from age-matched KPC (n = 5), KPCShet (n = 4), and KPCS (n = 5) mice at 6, 10, and 15 weeks. Scale bars 2000 μm. (D) Percentage of PanIN and PDAC lesion area over the whole pancreatic tissue sections from KPC, KPCShet, and KPCS mice at the indicated age. (E) Histopathologic analysis of pancreatic tissues from KPC, KPCShet, and KPCS mice at indicated age. (F) Tumor incidence in KPC, KPCShet, and KPCS cohorts at the indicated age. (G–H) Ki67 staining and quantification in pancreatic tumor tissues from KPC, KPCShet, and KPCS mice. Scale bars 100 μm. (I) Kaplan-Meier survival analysis of the KPC, KPCShet, and KPCS mice (log-rank test). The data are represented as mean ± SD. One-way ANOVA with Tukey test for panels D and H, ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

       SIRT5 Modulates Pancreatic Cancer Cell Growth by Suppressing Glutamine and Glutathione Metabolism

      Previous studies have demonstrated that SIRT5 regulated tumorigenesis of multiple cancer types by mediating their metabolic reprogramming.
      • Yang X.
      • Wang Z.
      • Li X.
      • et al.
      SHMT2 desuccinylation by SIRT5 drives cancer cell proliferation.
      • Wang Y.Q.
      • Wang H.L.
      • Xu J.
      • et al.
      Sirtuin5 contributes to colorectal carcinogenesis by enhancing glutaminolysis in a deglutarylation-dependent manner.
      • Li F.
      • He X.
      • Ye D.
      • et al.
      NADP(+)-IDH mutations promote hypersuccinylation that impairs mitochondria respiration and induces apoptosis resistance.
      • Chen X.F.
      • Tian M.X.
      • Sun R.Q.
      • et al.
      SIRT5 inhibits peroxisomal ACOX1 to prevent oxidative damage and is downregulated in liver cancer.
      • Greene K.S.
      • Lukey M.J.
      • Wang X.
      • et al.
      SIRT5 stabilizes mitochondrial glutaminase and supports breast cancer tumorigenesis.
      As such, the present study next investigated whether SIRT5 could regulate PDAC metabolism. liquid chromatography-coupled tandem-mass spectrometry (LC-MS/MS)-based metabolite profiles revealed distinct metabolic alterations between control and SIRT5-knockdown PDAC cells (Figure 4A–B, Supplementary Figure 5). Moreover, metabolite set enrichment analysis indicated that SIRT5 knockdown led to significant changes in metabolites involved in glutathione metabolism, glutamine metabolism, pyrimidine metabolism, and the tricarboxylic acid cycle (Figure 4C–D). SIRT5-knockdown significantly increased the relative abundance of glutamine metabolism intermediates (ie, malate, fumarate, and citrate), glutathione metabolism pathway metabolites (ie, reduced glutathione (GSH) and NADPH), and the pyrimidine metabolism intermediates (ie, deoxycytidine monophosphate [dCMP], cytidine triphosphate [CTP], deoxythymidine triphosphate [dTTP], deoxyuridine monophosphate [dUMP], uridine triphosphate [UTP]) (Figure 4E–H, Supplementary Figure 5A–B). SIRT5-knockdown also led to increased reduction potential as noted by increased reduced glutathione/oxidized glutathione ratio (GSH/GSSG) and NADPH/nicotinamide adenine dinucleotide phosphate (NADP) ratios (Figure 4E–F). Furthermore, the glutamine uptake was measured in control and SIRT5-knockdown PDAC cells. These assays revealed the uptake of glutamine remarkably increased in SIRT5-knockdown cells (Figure 4I). Interestingly, the differences in cell growth rates between the control and SIRT5-knockdown PDAC cells were much greater under the low-glucose (1.25 mmol/L; 5%) conditions than those in normal culture conditions (Figure 4J–M). These data suggest SIRT5 loss increases glutamine use to maintain redox balance and support PDAC cell growth even under glucose-limited conditions.
      Figure thumbnail gr4
      Figure 4SIRT5 suppresses glutamine and glutathione metabolism, and regulates cellular redox homeostasis. (A–B). Principal component analysis of metabolic profiles from control and SIRT5-knockdown cells (N = 5). (C–D) Metabolic pathway analysis of significantly dysregulated metabolites in control and SIRT5-knockdown cells. (E–F) Significantly different metabolites in glutamine and glutathione metabolism from control and SIRT5-knockdown cells. (G–H) Significantly different metabolites in pyrimidine metabolism from control and SIRT5-knockdown cells. (I) Relative glutamine uptake in control and SIRT5-knockdown cells. (J–K) Growth curves of control and SIRT5-knockdown cells cultured under low-glucose conditions (1.25 mmol/L). (L–M) Colony formation assays for control and SIRT5-knockdown cells under low-glucose conditions (1.25 mmol/L). For all in vitro studies, N ≥ 3. The data are represented as mean ± SD. One-way ANOVA with Bonferroni test (E–I, M), 2-way ANOVA with Bonferroni test (J–K), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

       SIRT5 Induces Dysregulated Glutamine and Glutathione Metabolism via GOT1

      Next, we investigated the mechanistic basis of SIRT5-mediated dysregulation of glutamine and glutathione metabolism in PDAC. KRAS-mutated PDAC cells rely on noncanonic glutamine metabolism (Figure 5A).
      • Son J.
      • Lyssiotis C.A.
      • Ying H.
      • et al.
      Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
      There were no significant differences in mRNA or protein levels of noncanonic glutamine metabolic pathway enzymes upon SIRT5 knockdown (Supplementary Figure 6A–C). As such, the present study measured the enzymatic activities of enzymes including glutaminase (GLS), GLUD1, and GOT from control and SIRT5-knockdown PDAC cell extracts. Interestingly, suppression of SIRT5 in PDAC cells significantly promoted the enzyme activity of GOT, but not GLS or GLUD1 (Figure 5B and Supplementary Figure 6D–E). Next, the control and SIRT5-knockdown PDAC cells were treated with inhibitors of glutamine pathway metabolic enzymes to examine their selective responsiveness to these inhibitors (Supplementary Figure 6F). Results showed SIRT5-knockdown PDAC cells were more sensitive to GOT-inhibitor aminooxyacetate (AOA) than control cells (Figure 5C–D), but there was no significant difference between control and SIRT5-knockdown cells when treated with GLS-inhibitor BPTES and GLUD1-inhibitor EGCG (Supplementary Figure 6G–J). Moreover, orthotopic implantation studies demonstrated SIRT5-knockdown T3M4 cells were more vulnerable to AOA treatment (Figure 5E–G, Supplementary Figure 6K–L). These data indicate SIRT5 inhibits glutamine/glutathione metabolism primarily by suppressing the enzyme activity of GOT.
      Figure thumbnail gr5
      Figure 5SIRT5 inhibits glutamine and glutathione metabolism by decreasing GOT1 enzyme activity. (A) The schematic illustration of the Kras-regulated glutamine and glutathione metabolism in mutant Kras-driven PDAC. (B) Relative GOT enzyme activity in control and SIRT5-knockdown cells. (C–D) Relative survival of control and SIRT5-knockdown cells treated with GOT inhibitor AOA. Data are normalized to the respective untreated group. (E–G) Effect of AOA treatment on T3M4 shScramble and shSIRT5 cells in vivo. Tumor growth rates at indicated time points (E). Tumor volume (F) and tumor weight in each group upon necropsy (G). (H–I) Cell growth of control and SIRT5-knockdown cells transfected with control or GOT1 short hairpin RNA (shRNA) under low-glucose conditions (1.25 mmol/L). Experiments for scrambled control and SIRT5-knockdown cells transfected with GOT1 or GOT2 shRNA were set up together with common controls. (J–K) Relative reduced nicotinamide adenine dinucleotide phosphate/oxidized nicotinamide adenine dinucleotide phosphate (NADPH/NADP) ratio (J) or GSH/GSSG ratio (K) in control, GOT1-knockdown, SIRT5-knockdown, and SIRT5/GOT1-double-knockdown cells. GSH, reduced glutathione; GSSH, oxidized glutathione. (L) Relative intracellular reactive oxygen species (ROS) levels in the indicated cells. ROS-insensitive carboxy-DCFDA (CDCFDA) dye was used as a negative control. (M) Kinetic flux analysis of 13C-labeled glutamine carbon incorporation into downstream metabolites in control, GOT1-knockdown, SIRT5-knockdown, and SIRT5/GOT1-double-knockdown T3M4 cells. Asp, aspartate; Glu, glutamate. For all in vitro studies, N ≥ 3. The data are represented as mean ± SD. One-way ANOVA with Bonferroni test (B) or Tukey test (F–L), 2-way ANOVA with Tukey test (E) or Bonferroni test (M), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      To further identify which GOT isoform (GOT1 or GOT2) is the primary target of SIRT5, GOT1 and GOT2 were knocked down in both control and SIRT5-knockdown PDAC cells (Supplementary Figure 6M–P). GOT1 silencing abolished SIRT5-knockdown-induced PDAC cell growth (Figure 5H–I), whereas knockdown of GOT2 did not significantly affect cell growth in all of the SIRT5-knockdown cells under low-glucose (1.25 mmol/L) conditions (Supplementary Figure 6Q–R). These data preliminarily indicate that GOT1 plays a major role in SIRT5-loss–mediated cell growth. To confirm the regulation of GOT1 by SIRT5, metabolites from control, GOT1-knockdown, SIRT5-knockdown, and SIRT5/GOT1-double-knockdown cells were subjected to LC-MS/MS-based metabolomic analysis. SIRT5-knockdown cells demonstrated significantly increased ratios of reduced-glutathione to oxidized-glutathione (GSH/GSSG) and cellular reduced nicotinamide adenine dinucleotide phosphate/oxidized nicotinamide adenine dinucleotide phosphate (NADPH/NADP), which was abolished by GOT1 knockdown (Figure 5J–K). GOT1 catalyzes the conversion reactions of 2-ketoglutarate and aspartate into glutamate and oxaloacetate. The metabolomics data indicated SIRT5 knockdown increased the abundance of the GOT1 downstream intermediates (ie, malate, fumarate). Contrastingly, SIRT5-loss–induced increase in malate and fumarate levels was abolished by the GOT1 knockdown, with a concomitant increase in the levels of the precursor metabolite 2-ketoglutarate (Supplementary Figure 6S-T). As the reducing equivalent NADPH and GSH are important for cellular redox homeostasis, SIRT5 knockdown was posited to increase the generation of NADPH and GSH to maintain redox balance and promote cell proliferation in PDAC cells. As expected, GOT1 silencing abrogated a SIRT5-knockdown–mediated decrease of the intracellular reactive oxygen species (ROS) levels (Figure 5L). Furthermore, [U-13C5] glutamine-based kinetic flux analysis demonstrated SIRT5 knockdown significantly increased levels of 13C-labeled malate and fumarate; these increased levels were abolished by GOT1 knockdown (Figure 5M). Taken together, these findings indicate that SIRT5 inhibits glutamine/glutathione metabolism and regulates cellular redox homeostasis by attenuating the enzymatic activity of GOT1.

       SIRT5 Inhibits GOT1 Enzymatic Activity by Catalyzing its Lysine Deacetylation

      We then investigated the mechanistic basis of SIRT5-mediated modulation of GOT1 activity. Considering SIRT5 has been found to regulate metabolic enzymatic activity by altering the post-translational modifications in enzymes,
      • Nishida Y.
      • Rardin M.J.
      • Carrico C.
      • et al.
      SIRT5 regulates both cytosolic and mitochondrial protein malonylation with glycolysis as a major target.
      ,
      • Nakagawa T.
      • Lomb D.J.
      • Haigis M.C.
      • et al.
      SIRT5 Deacetylates carbamoyl phosphate synthetase 1 and regulates the urea cycle.
      • Rardin M.J.
      • He W.
      • Nishida Y.
      • et al.
      SIRT5 regulates the mitochondrial lysine succinylome and metabolic networks.
      • Sadhukhan S.
      • Liu X.
      • Ryu D.
      • et al.
      Metabolomics-assisted proteomics identifies succinylation and SIRT5 as important regulators of cardiac function.
      • Zhou L.
      • Wang F.
      • Sun R.
      • et al.
      SIRT5 promotes IDH2 desuccinylation and G6PD deglutarylation to enhance cellular antioxidant defense.
      the present study examined whether SIRT5 potentially catalyzes lysine post-translational modifications in GOT1. Interestingly, the lysine acetylation level of GOT1 decreased upon SIRT5 overexpression in PDAC cells, whereas the lysine glutarylation, succinylation, or malonylation levels of GOT1 remained unaltered or undetectable (Figure 6A). In contrast, SIRT5-knockdown significantly increased the acetylation level of both exogenous GOT1 (Figure 6B) and endogenous GOT1 (Supplementary Figure 7A). Moreover, the catalytically inactive mutant SIRT5-H158Y
      • Nishida Y.
      • Rardin M.J.
      • Carrico C.
      • et al.
      SIRT5 regulates both cytosolic and mitochondrial protein malonylation with glycolysis as a major target.
      failed to induce the deacetylation of GOT1 protein compared with the wild-type SIRT5 (Figure 6C). To further identify the SIRT5-dependent deacetylation sites on GOT1, GOT1 protein was purified from the control and SIRT5-knockdown T3M4 cells, and post-translational modifications were assessed using mass spectrometry. As shown in Figure 6D and Supplementary Figure 7B–G, 3 acetylation sites (lysine 276, lysine 290, and lysine 369) were found in both control and SIRT5-knockdown cells. In addition, a comparison of the lysine acetylation intensities of GOT1 between control and SIRT5-knockdown cells indicated lysine 369 as the major lysine target residue that SIRT5 deacetylates (Figure 6D).
      Figure thumbnail gr6
      Figure 6SIRT5 inhibits GOT1 enzymatic activity by catalyzing its lysine deacetylation. (A) The lysine acetylation, glutarylation, succinylation, and malonylation levels of GOT1 protein immunoprecipitated from control and SIRT5-overexpressing PDAC cells. Inputs are shown below. (B) The lysine acetylation level of GOT1 protein immunoprecipitated from control and SIRT5-knockdown T3M4/Capan2 cells. Inputs are shown below. (C) The lysine acetylation level of GOT1 protein immunoprecipitated from PDAC cells transfected with vector, SIRT5, and mutant SIRT5-H158Y plasmid. Inputs are shown below. (D) Exogenous GOT1 protein immunoprecipitated from control and SIRT5-knockdown T3M4 cells was subjected to proteomic analysis. Top, schematic representation of 3 acetylation sites identified in GOT1; bottom, the mass spectrometric signal intensity of indicated acetylation sites in GOT1. (E–F) Wild-type GOT1, GOT1-K276R, GOT1-K290R, and GOT1-K369R mutant plasmids were transfected into control and SIRT5-knockdown T3M4 cells. The lysine acetylation levels of GOT1 protein immunoprecipitated from above cells; inputs are shown below (E). The quantification of relative lysine acetylation level of indicated acetylation sites (F). (G–H) The immunoblotting of SIRT5 and GOT1 levels in scrambled control cells, SIRT5-knockdown cells, and SIRT5-knockdown/GOT1-knockout cells transfected with vector, GOT1, or GOT1-K369R plasmid. (I) The GOT enzyme activity in scrambled control cells, SIRT5-knockdown cells, and SIRT5-knockdown/GOT1-knockout cells transfected with vector, GOT1, or GOT1-K369R plasmids. (J–K) Cell growth analysis of scrambled control cells, SIRT5-knockdown cells, and SIRT5-knockdown/GOT1-knockout cells transfected with vector, GOT1, or GOT1-K369R plasmids, cultured under low-glucose conditions (1.25 mmol/L). (L) Kinetic flux analysis of 13C-labeled glutamine carbon incorporation into downstream metabolites in scrambled control cells, SIRT5-knockdown cells, or SIRT5-knockdown/GOT1-knockout cells transfected with vector, wild-type GOT1, or GOT1-K369R plasmids. For all in vitro studies, N ≥ 3. The data are represented as mean ± SD. Student t-test (D, F), 1-way ANOVA with Tukey test (I–K), 2-way ANOVA with Tukey test (L), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.
      To identify if SIRT5-catalyzed post-translational modifications of GOT1 regulated GOT1 activity, we performed site-directed mutagenesis. The lysine to arginine (ie, K-to-R) mutation preserves the positive charge but cannot be acetylated, and is frequently used as a deacylated lysine mimetic.
      • Schwer B.
      • Bunkenborg J.
      • Verdin R.O.
      • et al.
      Reversible lysine acetylation controls the activity of the mitochondrial enzyme acetyl-CoA synthetase 2.
      Hence, 3 mutant HA-tagged GOT1 plasmids were generated, in which the lysine 276, lysine 290, or lysine 369 residues were substituted with arginine (R). The wild-type GOT1 plasmid, GOT1-K276R, GOT1-K290R, and GOT1-K369R mutant plasmids were transfected into control and SIRT5-knockdown T3M4 cells. Immunoblots of the purified GOT1 protein from cells expressing wild-type GOT1, GOT1-K276R, GOT1-K290R, and GOT1-K369R mutants were used to detect the acetylation levels of GOT1. The K369R mutation resulted in a significant decrease in the acetylation level of GOT1 (Figure 6E–F). Furthermore, SIRT5-knockdown significantly increased the lysine acetylation levels of wild-type GOT1, GOT1-K276R, and GOT1-K290R mutants, but not the GOT1-K369R mutant (Figure 6E–F). These data indicate lysine 369 in GOT1 protein is the primary deacetylation target of SIRT5. To further evaluate the function of GOT1-K369 acetylation under SIRT5-knockdown conditions, first the CRISPR/Cas9 technique was used to generate GOT1-knockout/SIRT5-knockdown cells. The above GOT1-knockout/SIRT5-knockdown cells were then engineered to re-express vector, wild-type GOT1, or GOT1-K369R mutant plasmids (Figure 6G–H). GOT1 knockout abolished SIRT5-knockdown–mediated increase of GOT enzymatic activity, while re-expressing wild-type GOT1, but not GOT1-K369R, restored this activity (Figure 6I). Thus, K369 deacetylation by SIRT5 is primarily responsible for SIRT5-mediated abrogation of GOT activity in PDAC cells.
      GOT1 knockout inhibited the effects of SIRT5-knockdown on cell growth, whereas re-expression of wild-type GOT1 rescuedcell growth. However, re-expression of GOT1-K369R failed to restore cell growth in GOT1-knockout/SIRT5-knockdown PDAC cells under low-glucose (1.25 mmol/L) conditions (Figure 6J–K). Consistently, the [U-13C5] glutamine-based kinetic flux analysis indicated GOT1 knockout abolished the increased 13C-labeled malate and fumarate levels in SIRT5-knockdown cells. Moreover, re-expressing wild-type GOT1 could restore the 13C-labeled malate and fumarate levels in GOT1-knockout/SIRT5-knockdown cells, whereas expressing GOT1-K369R was unable to rescue this phenomenon (Figure 6L). Taken together, these results strongly indicate that SIRT5 suppresses GOT1 enzymatic activity by deacetylating GOT1 at lysine 369, which leads to the inhibition of tumor cell growth.

       Targeting PDAC With Small-Molecule SIRT5-Activator MC3138 Improves Therapeutic Efficiency

      Given the important role of SIRT5 in the development of PDAC, it has the potential to serve as a promising drug target. A small-molecule SIRT5 activator, a 1,4-dihydropyridine, termed MC3138, was developed (Figure 7A) and tested for potential effects on human PDAC cell lines, organoids, and PDX tumors. The biochemical enzymatic assay results indicated that MC3138 exhibited a selective activation of SIRT5 compared with the selective SIRT1 activator SRT2104 and SIRT3 activator Honokiol (Supplementary Figure 8A–C). Moreover, treatment with MC3138 mimicked the deacetylation effect mediated by SIRT5 overexpression (Supplementary Figure 8D). MC3138 treatment decreased lysine acetylation of the GOT1 protein and inhibited its enzymatic activity (Figure 7B–C). The cell viability assay showed MC3138 reduced PDAC cell viability, with half maximal inhibitory concentration (IC50) values ranging from 25.4 μmol/L–236.9 μmol/L (Figure 7D and Supplementary Figure 8E–F). To validate whether the inhibitory effect of MC3138 was associated with activation of SIRT5, a correlation analysis between the IC50 of MC3138 and SIRT5 protein levels was performed. The IC50 value of MC3138 positively correlated with SIRT5 protein levels (Figure 7E). Moreover, MC3138 was tested in mouse KPC and KPCS cells. Because KPCS cells do not express SIRT5, these cells were, as expected, resistant to MC3138 treatment (Figure 7F). Conversely, the SIRT1 activator SRT2104 and SIRT3 activator Honokiol did not exhibit selective growth inhibition between KPC and KPCS cells (Supplementary Figure 8G–H). Because SIRT5 knockdown significantly increased the metabolite levels in glutamine, glutathione, and pyrimidine metabolism pathways (Figure 4E–H), these metabolites levels were also investigated in PDAC cells treated using MC3138. MC3138 treatment decreased the metabolite levels in these pathways (Supplementary Figure 8I–L). Our previous studies identified that the increase of endogenous pyrimidine pools can diminish the therapeutic efficiency of pyrimidine analog gemcitabine by molecular competition.
      • Shukla S.K.
      • Purohit V.
      • Mehla K.
      • et al.
      MUC1 and HIF-1alpha signaling crosstalk induces anabolic glucose metabolism to impart gemcitabine resistance to pancreatic cancer.
      Next, the combinational effect of MC3138 and gemcitabine, a first-line Food and Drug Administration–approved chemotherapeutic to treat patients with PDAC, was investigated. A combination of gemcitabine and MC3138 at different dosages exhibited synergistic effects in CFPAC-1 and Colo357/FG cells (Supplementary Figure 8M–P). Likewise, gemcitabine combined with MC3138 was synergistic at different concentrations in human PDAC organoids with low-SIRT5 expression (Figure 7G–L). The MC3138 pharmacokinetics showed that the maximum concentration and half-life of MC3138 in plasma was around 230 μmol/L and 5.059 hours, respectively (Supplementary Figure 8Q). To further investigate the effect of gemcitabine combined with MC3138 in vivo, PDX tumors (PA137) were orthotopically implanted in NOD-SCID mice. The concentration of MC3138 in the MC3138-treated tumors ranged from around 100–200 μmol/L, with the average as 143.57 μmol/L, indicating that MC3138 was efficiently delivered to the tumor tissues (Supplementary Figure 8R). The combination of gemcitabine and MC3138 significantly decreased tumor size, tumor weight, and tumor proliferation cell index in mice (Figure 7M–O, Supplementary Figure 8S–T). In addition, the GOT activity was significantly decreased in the MC3138-treated tumors (Supplementary Figure 8U). More importantly, the administered drug combination was well tolerated in mice because no significant alterations were observed in body weight and blood biochemistry indices (Supplementary Figure 8V–W). Collectively, these data indicate that gemcitabine combined with SIRT5 activator MC3138 can be a potential therapeutic option for PDAC with low-SIRT5 expression.
      Figure thumbnail gr7
      Figure 7SIRT5 activator MC3138 exhibits antitumor effects and synergism with gemcitabine in human PDAC cells, organoids, and PDX models. (A) The chemical structure of SIRT5 activator MC3138 and an enlarged view of SIRT5 docked with MC3138. Protein is shown in purple flat ribbon and compound is depicted via a stick model. (B) The lysine acetylation level of GOT1 protein immunoprecipitated from PDAC cells treated with dimethyl sulfoxide (DMSO) or 10 μmol/L MC3138 for 24 h. Inputs are shown below. (C) The GOT enzyme activity in PDAC cells treated with DMSO or 10 μmol/L MC3138 for 24 h. (D) IC50 data of 10 wild-type PDAC cell lines treated using MC3138. (E) Pearson correlation analysis between SIRT5 protein level (from L) and IC50 of MC3138 in PDAC cell lines. (F) Relative cell survival of KPC cell and SIRT5-knockout KPC cell (KPCS) treated with MC3138. (G–H, J–K) Representative images and cell viability of PDAC organoids PA417 (G–H) and PA137 (J–K) treated with indicated concentration of gemcitabine and MC3138 for 72 h. (I, L) Combination index (CI) of gemcitabine and MC3138 at indicated concentrations. “Effect” in the Table refers to the relative cell survival on the combination therapy treatment. (M–O) Effect of MC3138 in combination with gemcitabine on PDX (PA137) model. Representative tumor images on necropsy (M). Tumor volumes are represented as mean ± SEM (N). Tumor weight on necropsy (O). Gem, gemcitabine; MC, MC3138. For all in vitro studies, N ≥ 3. The data are represented as mean ± SD. Student t test (C), 1-way ANOVA with Tukey test (H, K, O), 2-way ANOVA with Tukey test (N), ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

      Discussion

      In this work, we established a novel tumor suppressor function of SIRT5 in PDAC cell lines, organoids, orthotopic and spontaneous models, and patients. SIRT5 knockdown increased PDAC cell growth in normal and nutrient-limited conditions, whereas SIRT5 overexpression diminished PDAC cell growth. Genetic ablation of Sirt5 in the mouse spontaneous tumor progression model accelerated acinar-to-ductal metaplasia, PanIN formation, and pancreatic tumorigenesis, which results in decreased survival.
      Previous studies demonstrate that Kras-mutated PDAC can reprogram its metabolism to adapt to the harsh microenvironment.
      • Son J.
      • Lyssiotis C.A.
      • Ying H.
      • et al.
      Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
      However, the molecular regulators downstream of oncogenic Kras that mediate metabolic adaptations remain poorly characterized. Our present study shows Sirt5 mRNA and protein levels are decreased in KPC tumor tissues. In the genetically engineered mouse model with the oncogenic KrasG12D mutation, Sirt5 ablation accelerated PanIN formation and PDAC progression. Hence, Sirt5 loss potentially functions downstream of Kras, and facilitates the oncogenic Kras mutation-driven PDAC initiation and development.
      Previous studies reveal oncogenic Kras mutation-driven PDAC cells use glutamine in a noncanonic way via the GOT2-GOT1-ME1 pathway.
      • Son J.
      • Lyssiotis C.A.
      • Ying H.
      • et al.
      Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
      ,
      • Abrego J.
      • Gunda V.
      • Vernucci E.
      • et al.
      GOT1-mediated anaplerotic glutamine metabolism regulates chronic acidosis stress in pancreatic cancer cells.
      However, the mechanistic basis of Kras-mediated induction of the pathway is not fully understood. Accordingly, our present study shows that SIRT5 can suppress PDAC progression by inhibiting the GOT1-mediated glutamine/glutathione metabolic pathway via post-translational modification of GOT1. Notably, SIRT5 knockdown sensitizes PDAC cells to the GOT-inhibitor AOA both in vitro and in vivo. Although the GOT-inhibitor AOA is a pan-aminotransferase inhibitor,
      • Korangath P.
      • Teo W.W.
      • Sadik H.
      • et al.
      Targeting glutamine metabolism in breast cancer with aminooxyacetate.
      ,
      • Hao Y.
      • Samuels Y.
      • Li Q.
      • et al.
      Oncogenic PIK3CA mutations reprogram glutamine metabolism in colorectal cancer.
      SIRT5-knockdown PDAC cells with significantly higher GOT1 activity were much more sensitive to AOA treatment both in vitro and in vivo (Figure 5B–G). Combined with increased glutamine uptake of SIRT5 knockdown cells, these results indicate SIRT5-mediated deacetylation of GOT1 as a metabolic switch that must be inactivated to facilitate Kras-induced glutamine addiction. Notably, very few studies have reported direct regulation of glutamine metabolism in cancer cells via post-translational modifications that provide nutritional flexibility. Yang et al
      • Yang H.
      • Zhou L.
      • Shi Q.
      • et al.
      SIRT3-dependent GOT2 acetylation status affects the malate-aspartate NADH shuttle activity and pancreatic tumor growth.
      showed that SIRT3 regulated the malate-aspartate shuttle activity by deacetylating GOT2 in pancreatic cancer. However, unlike GOT1, GOT2 knockdown could not abolish SIRT5 knockdown-induced increased cell growth under low-glucose conditions. Two other SIRT5-mediated post-translational modifications have been reported to target glutamine metabolism via SIRT5-mediated desuccinylation of GLS in breast cancer and deglutarylation of GLUD1 in colorectal cancer that increased glutamine uptake and metabolism to support proliferation in other tumor models.
      • Wang Y.Q.
      • Wang H.L.
      • Xu J.
      • et al.
      Sirtuin5 contributes to colorectal carcinogenesis by enhancing glutaminolysis in a deglutarylation-dependent manner.
      ,
      • Greene K.S.
      • Lukey M.J.
      • Wang X.
      • et al.
      SIRT5 stabilizes mitochondrial glutaminase and supports breast cancer tumorigenesis.
      Conversely, we report increased glutamine uptake and enhanced cell growth in PDAC upon SIRT5 knockdown. Of note, in a previous study, GLUD1 knockdown failed to diminish Kras-induced PDAC cell growth; however, GOT1 knockdown significantly inhibited tumor cell growth.
      • Son J.
      • Lyssiotis C.A.
      • Ying H.
      • et al.
      Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
      These studies are further supported by our findings and lead to the notion that the mechanistic regulation of glutamine metabolism is distinct in Kras-driven PDAC tumors as compared with tumors of other origin and noncancer cells that are dependent on the canonic glutamine metabolism. Thus, our study demonstrates the mechanistic regulation of noncanonic glutamine metabolism via post-translational modifications in PDAC cells with activating KRAS mutations.
      Our current study provides the first evidence in support of SIRT5 activation as a therapeutic strategy. We developed a novel first-in-class small-molecule SIRT5 activator, MC3138, which shows selectivity over SIRT1 and SIRT3. Furthermore, MC3138 synergized with gemcitabine in human PDAC cell lines, organoids, and PDX tumors with low-SIRT5 expression. A SIRT5 loss-mediated increase in pyrimidine nucleotide pools provides a potential mechanistic basis for the synergistic effect of the combination with gemcitabine. These results are in line with previous studies demonstrating that molecular competition with increased endogenous pyrimidine pools can outcompete pyrimidine analog gemcitabine and impart resistance.
      • Shukla S.K.
      • Purohit V.
      • Mehla K.
      • et al.
      MUC1 and HIF-1alpha signaling crosstalk induces anabolic glucose metabolism to impart gemcitabine resistance to pancreatic cancer.
      Of note, the toxicity assays indicated MC3138 combined with gemcitabine was well tolerated in mice. These results demonstrate that MC3138 combined with gemcitabine can be a safe and effective therapeutic option for PDAC with low-SIRT5 expression.
      In conclusion, our study identifies a vital role SIRT5 plays in regulating glutamine/glutathione metabolism during the initiation and progression of PDAC. Moreover, our data indicate a synergistic effect of the small-molecule SIRT5-activator MC3138 combined with gemcitabine in human PDAC organoids and PDX models. As such, these findings shed new light on the potential use of SIRT5 activation as a novel therapeutic strategy for patients with PDAC with decreased SIRT5 expression.

      CRediT Authorship Contributions

      Tuo Hu, PhD (Conceptualization: Lead; Investigation: Lead; Writing – original draft: Equal).
      Surendra Shukla, PhD (Investigation: Supporting; Resources: Supporting; Writing – review & editing: Supporting).
      Enza Vernucci, PhD (Conceptualization: Supporting; Formal analysis: Supporting; Investigation: Supporting; Writing – review & editing: Supporting).
      Chunbo He, BS (Investigation: Supporting).
      Dezhen Wang, PhD (Investigation: Supporting).
      Ryan King, PhD (Investigation: Supporting).
      Kanupriya Jha, MS (Investigation: Supporting; Visualization: Supporting).
      Kasturi Siddhanta, MS (Formal analysis: Supporting; Investigation: Supporting).
      Nicholas Mullen, BS (Investigation: Supporting; Writing – review & editing: Supporting).
      Kuldeep Attri, PhD (Investigation: Supporting).
      Divya Murthy, PhD (Investigation: Supporting).
      Nina Chaika, MS (Investigation: Supporting).
      Ravi Thakur, PhD (Investigation: Supporting).
      Scott Mulder, PhD (Investigation: Supporting).
      Camila Pacheco, BS (Methodology: Supporting).
      Xiao Fu, PhD (Investigation: Supporting).
      Robin Fu, MA (Formal analysis: Supporting).
      Fang Yu, PhD (Formal analysis: Lead).
      Audrey Lazenby, MD (Formal analysis: Supporting; Methodology: Supporting).
      Clemens Steegborn, PhD (Investigation: Supporting).
      Ping Lan, PhD (Resources: Supporting).
      Kamiya Mehla, PhD (Investigation: Supporting; Methodology: Supporting).
      Dante Rotili, PhD (Investigation: Supporting).
      Sergio Valente, PhD (Resources: Supporting).
      Sarika Chaudhary, PhD (Data curation: Supporting; Supervision: Supporting).
      Marco Tafani, PhD (Resources: Supporting).
      Antonello Mai, PhD (Resources: Supporting; Supervision: Supporting).
      Johan Auwerx, PhD (Resources: Supporting).
      Eric Verdin, PhD (Resources: Supporting).
      David Tuveson, PhD (Resources: Supporting).
      Pankaj Singh, PhD (Conceptualization: Lead; Funding acquisition: Lead; Supervision: Lead; Writing – original draft: Supporting; Writing – review & editing: Lead).

      Supplementary Material

      References

        • Derle A.
        • De Santis M.C.
        • Gozzelino L.
        • et al.
        The role of metabolic adaptation to nutrient stress in pancreatic cancer.
        Cell Stress. 2018; 2: 332-339
        • Vernucci E.
        • Abrego J.
        • Gunda V.
        • et al.
        Metabolic alterations in pancreatic cancer progression.
        Cancers (Basel). 2019; 12
        • Chaika N.V.
        • Yu F.
        • Purohit V.
        • et al.
        Differential expression of metabolic genes in tumor and stromal components of primary and metastatic loci in pancreatic adenocarcinoma.
        PLoS One. 2012; 7e32996
        • Gebregiworgis T.
        • Purohit V.
        • Shukla S.K.
        • et al.
        Glucose limitation alters glutamine metabolism in MUC1-overexpressing pancreatic cancer cells.
        J Proteome Res. 2017; 16: 3536-3546
        • Son J.
        • Lyssiotis C.A.
        • Ying H.
        • et al.
        Glutamine supports pancreatic cancer growth through a KRAS-regulated metabolic pathway.
        Nature. 2013; 496: 101-105
        • Li M.
        • Chiang Y.L.
        • Lyssiotis C.A.
        • et al.
        Non-oncogene addiction to SIRT3 plays a critical role in lymphomagenesis.
        Cancer Cell. 2019; 35: 916-931.e9
        • Winnik S.
        • Auwerx J.
        • Sinclair D.A.
        • et al.
        Protective effects of sirtuins in cardiovascular diseases: from bench to bedside.
        Eur Heart J. 2015; 36: 3404-3412
        • Chalkiadaki A.
        • Guarente L.
        The multifaceted functions of sirtuins in cancer.
        Nat Rev Cancer. 2015; 15: 608-624
        • Matsushita N.
        • Yonashiro R.
        • Ogata Y.
        • et al.
        Distinct regulation of mitochondrial localization and stability of two human Sirt5 isoforms.
        Genes Cells. 2011; 16: 190-202
        • Park J.
        • Chen Y.
        • Tishkoff D.X.
        • et al.
        SIRT5-mediated lysine desuccinylation impacts diverse metabolic pathways.
        Mol Cell. 2013; 50: 919-930
        • Nishida Y.
        • Rardin M.J.
        • Carrico C.
        • et al.
        SIRT5 regulates both cytosolic and mitochondrial protein malonylation with glycolysis as a major target.
        Mol Cell. 2015; 59: 321-332
        • Yang X.
        • Wang Z.
        • Li X.
        • et al.
        SHMT2 desuccinylation by SIRT5 drives cancer cell proliferation.
        Cancer Res. 2018; 78: 372-386
        • Wang Y.Q.
        • Wang H.L.
        • Xu J.
        • et al.
        Sirtuin5 contributes to colorectal carcinogenesis by enhancing glutaminolysis in a deglutarylation-dependent manner.
        Nat Commun. 2018; 9: 545
        • Li F.
        • He X.
        • Ye D.
        • et al.
        NADP(+)-IDH mutations promote hypersuccinylation that impairs mitochondria respiration and induces apoptosis resistance.
        Mol Cell. 2015; 60: 661-675
        • Chen X.F.
        • Tian M.X.
        • Sun R.Q.
        • et al.
        SIRT5 inhibits peroxisomal ACOX1 to prevent oxidative damage and is downregulated in liver cancer.
        EMBO Rep. 2018; 19: e45124
        • Greene K.S.
        • Lukey M.J.
        • Wang X.
        • et al.
        SIRT5 stabilizes mitochondrial glutaminase and supports breast cancer tumorigenesis.
        Proc Natl Acad Sci U S A. 2019; 116: 26625-26632
        • Yan D.
        • Franzini A.
        • Pomicter A.D.
        • et al.
        Sirt5 is a druggable metabolic vulnerability in acute myeloid leukemia.
        Blood Cancer Discov. 2021; 2: 266-287
        • Giblin W.
        • Bringman-Rodenbarger L.
        • Guo A.H.
        • et al.
        The deacylase SIRT5 supports melanoma viability by influencing chromatin dynamics.
        J Clin Invest. 2021; 131: e138926
        • Abril Y.L.N.
        • Fernandez I.R.
        • Hong J.Y.
        • et al.
        Pharmacological and genetic perturbation establish SIRT5 as a promising target in breast cancer.
        Oncogene. 2021; 40: 1644-1658
        • Hingorani S.R.
        • Petricoin E.F.
        • Maitra A.
        • et al.
        Preinvasive and invasive ductal pancreatic cancer and its early detection in the mouse.
        Cancer Cell. 2003; 4: 437-450
        • Olive K.P.
        • Tuveson D.A.
        • Ruhe Z.C.
        • et al.
        Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome.
        Cell. 2004; 119: 847-860
        • Du J.
        • Zhou Y.
        • Su X.
        • et al.
        Sirt5 is a NAD-dependent protein lysine demalonylase and desuccinylase.
        Science. 2011; 334: 806-809
        • Dasgupta A.
        • Shukla S.K.
        • Vernucci E.
        • et al.
        SIRT1-NOX4 signaling axis regulates cancer cachexia.
        J Exp Med. 2020; 217: e20190745
        • Strobel O.
        • Dor Y.
        • Alsina J.
        • et al.
        In vivo lineage tracing defines the role of acinar-to-ductal transdifferentiation in inflammatory ductal metaplasia.
        Gastroenterology. 2007; 133: 1999-2009
        • Nakagawa T.
        • Lomb D.J.
        • Haigis M.C.
        • et al.
        SIRT5 Deacetylates carbamoyl phosphate synthetase 1 and regulates the urea cycle.
        Cell. 2009; 137: 560-570
        • Rardin M.J.
        • He W.
        • Nishida Y.
        • et al.
        SIRT5 regulates the mitochondrial lysine succinylome and metabolic networks.
        Cell Metab. 2013; 18: 920-933
        • Sadhukhan S.
        • Liu X.
        • Ryu D.
        • et al.
        Metabolomics-assisted proteomics identifies succinylation and SIRT5 as important regulators of cardiac function.
        Proc Natl Acad Sci U S A. 2016; 113: 4320-4325
        • Zhou L.
        • Wang F.
        • Sun R.
        • et al.
        SIRT5 promotes IDH2 desuccinylation and G6PD deglutarylation to enhance cellular antioxidant defense.
        EMBO Rep. 2016; 17: 811-822
        • Schwer B.
        • Bunkenborg J.
        • Verdin R.O.
        • et al.
        Reversible lysine acetylation controls the activity of the mitochondrial enzyme acetyl-CoA synthetase 2.
        Proc Natl Acad Sci U S A. 2006; 103: 10224-10229
        • Shukla S.K.
        • Purohit V.
        • Mehla K.
        • et al.
        MUC1 and HIF-1alpha signaling crosstalk induces anabolic glucose metabolism to impart gemcitabine resistance to pancreatic cancer.
        Cancer Cell. 2017; 32: 71-87.e7
        • Abrego J.
        • Gunda V.
        • Vernucci E.
        • et al.
        GOT1-mediated anaplerotic glutamine metabolism regulates chronic acidosis stress in pancreatic cancer cells.
        Cancer Lett. 2017; 400: 37-46
        • Korangath P.
        • Teo W.W.
        • Sadik H.
        • et al.
        Targeting glutamine metabolism in breast cancer with aminooxyacetate.
        Clin Cancer Res. 2015; 21: 3263-3273
        • Hao Y.
        • Samuels Y.
        • Li Q.
        • et al.
        Oncogenic PIK3CA mutations reprogram glutamine metabolism in colorectal cancer.
        Nat Commun. 2016; 7: 11971
        • Yang H.
        • Zhou L.
        • Shi Q.
        • et al.
        SIRT3-dependent GOT2 acetylation status affects the malate-aspartate NADH shuttle activity and pancreatic tumor growth.
        EMBO J. 2015; 34: 1110-1125