Fig 2: Arg2 deficiency suppresses obesity-associated mouse PDA and results in ammonia accumulation. a DNA gel confirming the genotypes of KPC cells with wild-type (WT) expression or knockout (KO) of Arg2. b Proliferation curves of cells in a (n = 6). Data are representative of three independent experiments. c Volumes of orthotopic PDA tumors from cells in b injected (250,000 per mouse) in lean or obese 11–13 week old male C57BL/6J-129 svJae mice and grown for 2 weeks (n = 6 except for Obese-Arg2+/+ n = 5). d 15N4,13C6 enrichment in plasma arginine in mice from c infused with 15N4,13C6-arginine (n = 3). e, Schematic showing that Arg2 knockdown in obese PDA causes shunting of arginine away from the urea cycle and towards creatine biosynthesis, along with transfer of nitrogen into glutamate, aspartate, glutamine, and ammonia in Arg2 -/- KPC tumors of obese mice. Red arrows indicate increases in 15N-labeling, of indicated metabolites in Arg2 -/- tumors of obese mice. Gamt, Guanidinoacetate N-Methyltransferase. f Relative abundance of total pool of the indicated metabolites from tumors in c. g Relative abundance of the indicated labeled isotopomers in tumors from c. h Relative abundance (left) and fractional labeling (right) of derivatized 15N-ammonia. Data represent the mean ± s.e.m. in b, or mean ± s.d. in c, d, f–h. * P = 0.05, **P = 0.01 and *** P = 0.001, **** P = 0.0001, by two-way ANOVA in b, and one-way ANOVA in c, f–h, followed by Tukey test

Fig 3: Obesity leads to enhanced PDA growth and transcriptional induction of nitrogen and arginine metabolism. a Schematic timeline for the generation of lean or obese orthotopic xenograft models of PDA, using 4–6 week old male Rag1 -/- mice. b Growth curves of Lean-CFP (n = 15) or Obese-CFP (n = 14) mice. Arrow indicates orthotopic injection of AsPC1-CFP cells (105 per mouse). c Volumes of orthotopic PDA tumors from mice in b. d Heatmaps (left) of genes enriched in metabolic pathways in Lean-CFP compared to Obese-CFP tumors (n = 9 per group), with corresponding GSEA plots (right). Red indicates higher expression, and blue lower expression, relative to the mean expression level within each group. NES normalized enrichment scores, FDR false discovery rate, Nom. nominal. e Immunoblots for ARG2 and ARG1 in representative orthotopic tumors from c. f Graph demonstrating that the urea levels in PDA tumors are comparable to those in the liver of lean or obese mice (n = 5); data are representative of three independent experiments. In b, data represent the mean ± s.e.m. In c, f, data represent the mean ± s.d. *P = 0.05, **P = 0.01, ****P = 0.0001, two-way ANOVA followed by Tukey test in b and t-test in c

Fig 4: Glutamine nitrogens do not integrate into the urea cycle in vitro. a Immunoblots for ARG2 and ARG1 in AsPC1-shScr and AspC1-shARG2 cells. b Proliferation curves of cells in a (n = 6). Data are representative of 4 independent experiments. c Fractional 13C-labeling of urea (M1) and ornithine (M5) in AsPC-1 cells with control (shScr) or ARG2 knockdown (shARG2, hairpins A06 and A10) grown in culture in 0.1% serum and 100 ng ml-1 IGF-1, supplemented with 1 mM 13C6-arginine for 48 h (n = 4). d Diagram of the urea cycle, with metabolites in red indicating significant 15N-labeling. e Fractional labeling of the indicated metabolites in cells described in a, grown in media supplemented with 2 mM 15N(amine)-glutamine for 1 h or 24 h (n = 4). Data are representative of two independent experiments. In b, data represent the mean ± s.e.m. In c, e, data represent the mean ± s.d. *P = 0.05, **P = 0.01, ***P = 0.001, ****P = 0.0001, by one-way ANOVA followed by Tukey test

Fig 5: AKT activation enhances in vivo PDA growth and sensitizes tumors to ARG2 knockdown. a Immunoblots for pT308-AKT, pS473-AKT and total AKT in isogenic AsPC1-CFP and AsPC1-AKT PDA cells expressing SRC myristoylation sequence-AKT fusion (myrAKT). b Volumes of orthotopic PDA tumors from cells in a (105 per mouse) grown for 6 weeks in lean (n = 14) or obese (n = 15) 12–14 week old male Rag1 -/- mice following the outline in (Fig. 1a). c, d Immunoblots for pT308-AKT, pS473-AKT, and total AKT c, ARG2 and ARG1 d in tumors from b. e ARG2 levels as quantified by ImageJ and normalized to GAPDH, from immunoblots in d, Fig. 1e, and Supplementary Fig. 10c. f Immunoblots of ARG2 and ARG1 in isogenic cells from a, harboring control Scramble (shScr) or ARG2 (shARG2) knockdown. g Proliferation curves of cells in f (n = 6). Data are representative of four independent experiments. h Volumes of orthotopic PDA xenografts from cells in f that were injected in lean 8–9 week old male Rag1 -/- mice (500,000 per mouse) and grown for 5 weeks (n = 6). i Immunoblots for ARG2, ARG1, pT308-AKT, pS473-AKT and total AKT from representative tumors in h. j, k Representative immunohistochemistry staining showing enhanced ARG2 levels in PDA tumors from patients with higher compared to lower BMI j, and overlapping pS473-AKT and ARG2 staining in sequential tumor sections from PDA patients k. Arrows indicate areas of ductal adenocarcinoma and arrowheads indicate stroma. Scale bars: 20 µm. l Multivariate regression analysis of the association of BMI and pAKT with high ARG2 levels in resected PDA immunostained in j, k. OR odds ratio, CI confidence interval, AJCC American Joint Committee on Cancer. Data represent the mean ± s.d. in b, e, h or mean ± s.e.m. in g. *P = 0.05, **P = 0.01, ***P = 0.001, one-way ANOVA followed by Tukey test