Hypoxic regulation of glutamine metabolism through HIF1 and SIAH2 supports lipid synthesis that is necessary for tumor growth

Abstract

Recent reports have identified a phenomenon by which hypoxia shifts glutamine metabolism from oxidation to reductive carboxylation. We now identify the mechanism by which HIF-1 activation results in a dramatic reduction in the activity of the key mitochondrial enzyme complex α ketoglutarate dehydrogenase (αKGDH). HIF-1 activation promotes SIAH2 targeted ubiquitination and proteolysis of the 48 kDa splice variant of the E1 subunit of the αKGDH complex (OGDH2). Knockdown of SIAH2 or mutation of the ubiquitinated lysine residue on OGDH2 (336KA) reverses the hypoxic drop in αKGDH activity, stimulates glutamine oxidation, and reduces glutamine-dependent lipid synthesis. 336KA OGDH2-expressing cells require exogenous lipids or citrate for growth in hypoxia in vitro and fail to grow as model tumors in immunodeficient mice. Reversal of hypoxic mitochondrial function may provide a target for the development of next-generation anticancer agents targeting tumor metabolism.

Figure 1. Hypoxia down-regulates glutamine oxidation, see also figure S1

Panel A. Mitochondrial oxygen consumption (OCR) in control and DMOG treated SAS cells (500 μM 16h). OCR was measured in basal DMEM without serum, containing only 5 mM glucose, or only 2 mM glutamine or both as indicated. Panel B. OCR in VHL-deficient RCC4 (constitutively active HIF1) and RCC4 cells with VHL reintroduced, treated as in A. Panel C. OCR in empty vector SAS or ShHIF1α SAS treated as in A. Panel D. Pyruvate dehydrogenase J(PDH) enzyme activity in SAS, RCC4 and RCC4-VHL cells in control conditions, or after 16h of 0.5% oxygen, or 500 μM DMOG. Panel E. αKGDH enzyme activity in SAS, RCC4 and RCC4-VHL cells, treated as in D. Panel F PDH activity in ShHIF1α SAS cells, treated as in D. Panel F αKGDH activity in ShHIF1a SAS cells, treated as in D. Data are represented 3 independent replicates ± S.D.

Figure 2. SIAH2 E3 ligase is responsible for OGDH2 protein degradation after HIF1 stabilization, see also figure S2

Panel A. Western blot of SAS protein extracts from control, 0.5% O₂ conditions, and 0.5% O₂ with either 10 μM MG132, 50 μM VitK3 or 10 μM DUBI and MG132 as indicated. Panel B. αKGDH activity in SAS cells treated with either 500μM DMOG or DMOG in combination with 10μM MG132 or 50μM VitK3. Panel C. Western blot analysis of the indicted proteins in SAS, RCC4 and RCC4-VHL cells with and without ShRNA-SIAH2 as indicated cultured in normoxia or 16h 0.5% oxygen. Panel D. αKGDH activity in control and ShRNA-SIAH2 SAS, RCC4 and RCC4-VHL cells ± 16h 500μM DMOG. Panel E. Mitochondrial OCR in control and ShRNASIAH2 SAS, RCC4 and RCC4-VHL cells 16h ± 500μM DMOG measured in basal, serum-free media containing only glutamine. Data from b, d and e are two independent experiments in triplicate ± S.D.

Figure 3. Hypoxia-resistant αKGDH activity is not compatible with tumor growth, see also figure S3 and table S1

Panel A. Different splice variants of OGDH. OGDH 1/3 are 99% identical except at amino acid positions 153–169 (represented by ). OGDH2 is identical to OGDH1 until aa 403. Panel B. Western blot of lysates from SAS cells stably expressing empty vector, Flag-WT OGDH2 and Flag-336KA OGDH2 exposed to normoxia or 0.5% oxygen for 16h. Lysates were probed for proteins as indicated. Note both anti-Flag and anti-OGDH2 antibodies show OGDH2 336KA is resistant to hypoxic degradation. Panel C. αKGDH activity in SAS cell lines described in B, treated with either control, 16h 500μM DMOG, or 0.5% oxygen. Panel D. Mitochondrial OCR in SAS cell lines described in B treated with either control, 500μM DMOG, or 0.5% oxygen and measured in glutamine-only media. Panel E. Tumor volume of SAS cells expressing empty vector, OGDH-WT and OGDH-336KA grown in nude mice (n=8–10/group). Statistically significant growth differences exist between all three groups. Panel F. Western blot analysis of lysates of tumors harvested after growth as described in E. Note decrease in markers of proliferation as tumors grow more slowly. Data from c and d are mean ± S.D and tumor volumes are mean ± S.E.

Figure 4. OGDH 336KA inhibits cell growth under hypoxia by reducing glutamine derived lipid production, see also figure S4

Panel A: Relative hypoxic proliferation of SAS, RCC4 and RCC4VHL cells grown for 72 h in the indicated medias, presented as a percentage of normoxic growth. Hypoxic growth requires either glutamine (2mM), or glutamine derivatives αKG (2mM) or citrate (2mM). Panel B Hexane soluble lipids derived from a 1h pulse of 0.5 μCi ¹⁴C-glutamine in SAS cells expressing either empty vector or OGDH2-336KA, grown for 16h in normoxia or hypoxia. Panel C. Hexane soluble lipids derived from a 1h pulse of 0.5 μCi ¹⁴C-glucose in SAS cells as described in B. Panel D. Relative hypoxic proliferation of SAS cells expressing empty vector or OGDH2-336KA after 72 hours in basal glutamine-containing media and 10% charcoal-stripped serum supplemented either 2mM citrate or 2mM αKG. Note WT cells have maximal hypoxic proliferation with glutamine, but 336KA OGDH2 expressing cells require citrate. Panel E Relative hypoxic proliferation of SAS cells described in D, in basal media with 10% charcoal-stripped serum. Supplementation with absorbable lysophospholipid rescues the glutamine dependence of the WT OGDH2 cells, and the citrate dependence of the 336KA OGDH2 expressing cells. Data from panels A–E represents mean +/− S.D. Panel F. A model illustrating how hypoxic degradation of OGDH2 shifts the fate of αKG from energy production to the production of lipids.