Resistance to diet-induced obesity in mice with synthetic glyoxylate shunt

Abstract

Given the success in engineering synthetic phenotypes in microbes and mammalian cells, constructing non-native pathways in mammals has become increasingly attractive for understanding and identifying potential targets for treating metabolic disorders. Here, we introduced the glyoxylate shunt into mouse liver to investigate mammalian fatty acid metabolism. Mice expressing the shunt showed resistance to diet-induced obesity on a high-fat diet despite similar food consumption. This was accompanied by a decrease in total fat mass, circulating leptin levels, plasma triglyceride concentration, and a signaling metabolite in liver, malonyl-CoA, that inhibits fatty acid degradation. Contrary to plants and bacteria, in which the glyoxylate shunt prevents the complete oxidation of fatty acids, this pathway when introduced in mice increases fatty acid oxidation such that resistance to diet-induced obesity develops. This work suggests that using non-native pathways in higher organisms to explore and modulate metabolism may be a useful approach.

Figure 1. Hepatic Glyoxyate Shunt Expression Yields Multiple Possibilities With Distinct Phenotypes

Red lines represent the metabolic cycle for complete acetyl-CoA oxidation including the glyoxylate shunt enzymes, which are shaded orange. This pathway is expected to increase fatty acid degradation by decreasing malonyl-CoA levels. The gray dashed lines represent alternative pathways, which include acetyl-CoA incorporation into oxaloacetate for gluconeogenesis and malonyl-CoA for lipid biosynthesis, as well as the part of TCA cycle that is bypassed. The dashed green line represents allosteric inhibition. Abbreviations: PDH: pyruvate dehydrogenase, CPT1a: carnitine palmitoyltransferse-1a, PEPCK: phosphoenolpyruvate carboxykinase, ME1: cytosolic malic enzyme.

Figure 2. Human HepG2 Hepatocytes Expressing the Glyoxyate Shunt Have Increased Fatty Acid Degradation

(A) Reverse transcriptase PCR products from ACE cells (lanes 1 and 4) and WT cell lines (lanes 2 and 3) were run on a 1% agarose gel. The sizes of the ACEA (1.3 kb) and ACEB (1.6kb) mRNA are shown (lanes 1 and 4 respectively). (B) Western blot analysis of protein extracts from ACE (lanes 1 and 3) and WT (lanes 2 and 4) cells indicating the 61 kDa ACEA and the 48 kDa ACEB proteins. (C) Enzyme assay shows that the glyoxylate shunt was active in ACE cells. (D) Fatty acid uptake rate over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3–6 (E) Fatty acid uptake assay using fluorescently labeled dodecanoic acid confirms that the ACE cells take up fatty acids more rapidly than WT. Image is representative of two experiments. (F) Gluconeogenesis was assessed by measuring [U-¹³C] glycerol and [U-¹³C] palmitate incorporation into glucose. The glucose pentacetate derivative containing glucose carbons 2–6 was analyzed with GC-MS. ACE cells incorporate [U-¹³C] glycerol into glucose, as evidenced by the increased ¹³C enrichment in glucose. (G) [U-¹⁴C] palmitate was used to measure palmitate oxidation to CO₂. WT was used a normalization basis. *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3 (H) Glucose uptake over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). ACE cells consume less glucose than WT cells. *p < 0.05 compared to corresponding WT control. Mean ± SEM; n = 3–6 (I) Palmitate uptake for ACE ME1 knockdown (ACE-ME1-KD) and WT ME1 knockdown (WT-ME1-KD). ME1 KD results in ACE and WT cells consuming identical amounts of palmitate. Mean ± SEM; n = 6

Figure 2. Human HepG2 Hepatocytes Expressing the Glyoxyate Shunt Have Increased Fatty Acid Degradation

(A) Reverse transcriptase PCR products from ACE cells (lanes 1 and 4) and WT cell lines (lanes 2 and 3) were run on a 1% agarose gel. The sizes of the ACEA (1.3 kb) and ACEB (1.6kb) mRNA are shown (lanes 1 and 4 respectively). (B) Western blot analysis of protein extracts from ACE (lanes 1 and 3) and WT (lanes 2 and 4) cells indicating the 61 kDa ACEA and the 48 kDa ACEB proteins. (C) Enzyme assay shows that the glyoxylate shunt was active in ACE cells. (D) Fatty acid uptake rate over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3–6 (E) Fatty acid uptake assay using fluorescently labeled dodecanoic acid confirms that the ACE cells take up fatty acids more rapidly than WT. Image is representative of two experiments. (F) Gluconeogenesis was assessed by measuring [U-¹³C] glycerol and [U-¹³C] palmitate incorporation into glucose. The glucose pentacetate derivative containing glucose carbons 2–6 was analyzed with GC-MS. ACE cells incorporate [U-¹³C] glycerol into glucose, as evidenced by the increased ¹³C enrichment in glucose. (G) [U-¹⁴C] palmitate was used to measure palmitate oxidation to CO₂. WT was used a normalization basis. *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3 (H) Glucose uptake over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). ACE cells consume less glucose than WT cells. *p < 0.05 compared to corresponding WT control. Mean ± SEM; n = 3–6 (I) Palmitate uptake for ACE ME1 knockdown (ACE-ME1-KD) and WT ME1 knockdown (WT-ME1-KD). ME1 KD results in ACE and WT cells consuming identical amounts of palmitate. Mean ± SEM; n = 6

Figure 2. Human HepG2 Hepatocytes Expressing the Glyoxyate Shunt Have Increased Fatty Acid Degradation

(A) Reverse transcriptase PCR products from ACE cells (lanes 1 and 4) and WT cell lines (lanes 2 and 3) were run on a 1% agarose gel. The sizes of the ACEA (1.3 kb) and ACEB (1.6kb) mRNA are shown (lanes 1 and 4 respectively). (B) Western blot analysis of protein extracts from ACE (lanes 1 and 3) and WT (lanes 2 and 4) cells indicating the 61 kDa ACEA and the 48 kDa ACEB proteins. (C) Enzyme assay shows that the glyoxylate shunt was active in ACE cells. (D) Fatty acid uptake rate over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3–6 (E) Fatty acid uptake assay using fluorescently labeled dodecanoic acid confirms that the ACE cells take up fatty acids more rapidly than WT. Image is representative of two experiments. (F) Gluconeogenesis was assessed by measuring [U-¹³C] glycerol and [U-¹³C] palmitate incorporation into glucose. The glucose pentacetate derivative containing glucose carbons 2–6 was analyzed with GC-MS. ACE cells incorporate [U-¹³C] glycerol into glucose, as evidenced by the increased ¹³C enrichment in glucose. (G) [U-¹⁴C] palmitate was used to measure palmitate oxidation to CO₂. WT was used a normalization basis. *p < 0.05 compared to corresponding WT control; Mean ± SEM; n = 3 (H) Glucose uptake over a range of palmitate concentrations for both ACE (closed squares ■) and WT (open squares □). ACE cells consume less glucose than WT cells. *p < 0.05 compared to corresponding WT control. Mean ± SEM; n = 3–6 (I) Palmitate uptake for ACE ME1 knockdown (ACE-ME1-KD) and WT ME1 knockdown (WT-ME1-KD). ME1 KD results in ACE and WT cells consuming identical amounts of palmitate. Mean ± SEM; n = 6

Figure 3. Post-translational and Gene Expression Changes in ACE Cells are Consistent With Increased Fatty Acid Degradation

(A) Western blot analysis of WT (left panel) and ACE (right panel) cell lines. ACE and WT cells were given DMEM supplemented with 300uM palmitate and protein was harvested at various time points over 24 hours. Blots are representative of two experiments. (B) Metabolic networks related to glucose and fatty acid metabolism. Microarray analysis showing relative (fold change) gene expression levels in the genes involved in glycolysis, fatty acid beta-oxidation, and triglyceride synthesis from ACE and WT cell lines. Positive numbers indicated that the ACE cell line had higher expression levels than WT and are indicated by red arrows. Negative numbers indicate that the ACE cell line had lower expression levels and are indicated by green arrows. Data are represented as Mean ± SEM, n = 2. Dashed green line represents allosteric regulation: malonyl-CoA inhibits CPT1a. Abbreviations are listed in Table S1. Key enzymes are shown in blue circles: malic enzyme (ME1), pyruvate dehydrogenase (PDH), pyruvate carboxylase (PC), acetyl-CoA carboxylase (ACC), and carnitine palmitoyltransferse-1a (CPT1A).

Figure 4. Mice Expressing the Glyoxyate Shunt in the Liver Resist Diet-Induced Obesity and Accumulate Less Fat Mass on a High Fat Diet

(A) aceA and aceB were present exclusively in liver tissue as determined by reverse-transcriptase PCR at the end of diet experiments. (B) The glyoxylate shunt was functional in ACE mice, as determined by enzyme assay. (C) Female ACE and lacZ injected mice after 6 weeks on HFD. Female ACE mice gain less weight than lacZ injected mice. (D) Weight gain over six weeks for LFD and HFD for female mice. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6 (E) Weight gain over six weeks for LFD and HFD for male mice. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6 (F) Body composition as assessed at the conclusion of diet experiments using NMR. Female ACE mice have decreased total fat mass after 6 weeks on both LFD and HFD. *p < 0.05 for ACE fat mass compared to lacZ injected mouse fat mass. Mean ± SEM; n = 4–6 (G) Body fat percentage for ACE and lacZ injected mice. Female ACE mice have decreased body fat percentage. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6

Figure 4. Mice Expressing the Glyoxyate Shunt in the Liver Resist Diet-Induced Obesity and Accumulate Less Fat Mass on a High Fat Diet

(A) aceA and aceB were present exclusively in liver tissue as determined by reverse-transcriptase PCR at the end of diet experiments. (B) The glyoxylate shunt was functional in ACE mice, as determined by enzyme assay. (C) Female ACE and lacZ injected mice after 6 weeks on HFD. Female ACE mice gain less weight than lacZ injected mice. (D) Weight gain over six weeks for LFD and HFD for female mice. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6 (E) Weight gain over six weeks for LFD and HFD for male mice. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6 (F) Body composition as assessed at the conclusion of diet experiments using NMR. Female ACE mice have decreased total fat mass after 6 weeks on both LFD and HFD. *p < 0.05 for ACE fat mass compared to lacZ injected mouse fat mass. Mean ± SEM; n = 4–6 (G) Body fat percentage for ACE and lacZ injected mice. Female ACE mice have decreased body fat percentage. *p < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 4–6

Figure 5. Female ACE Mice Have Smaller Visceral Fat Pads and Decreased Circulating Leptin

(A) Female ACE mice accumulated less intra-abdominal fat pads compared to lacZ injected mice. (B) Visceral fat mass expressed as percent body weight for female mice on HFD. *p < 0.05; Mean ± SEM; n = 4 (C) Female ACE mice showed lower circulating plasma leptin concentrations as determined by ELISA. *p < 0.05; Mean ± SEM; n = 3–4

Figure 6. The Glyoxylate Shunt Decreases Liver Malonyl-CoA and Triglyceride Levels

(A) Female ACE mice on HFD had lower liver malonyl-CoA levels on HFD. ACE mice may resist diet-induced obesity in part due to decreased liver malonyl-CoA and therefore decreased CPT1a inhibition. *P < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 3–4 (B) Glyoxyate shunt expression led to decreased total ATP levels in the liver, presumably due to bypassing the oxidative portion of the TCA cycle. *p < 0.05; Mean ± SEM; n = 4 (C) Female ACE mice had decreased liver TG content. *P < 0.05; Mean ± SEM; n = 4 (D) Plasma analysis of metabolites for female ACE and lacZ injected mice on HFD. Mice were fasted overnight before blood was collected. ACE mice had decreased triglyceride and total cholesterol levels. Fasting glucose levels were identical, suggesting that the glyoxyate shunt does not serve a gluconeogenic function in the liver. *p < 0.05; Mean ± SEM; n = 4–6 (E) Plasma concentrations of ketone body 3-hydroxybutyrate in female ACE and lacZ injected mice on HFD after overnight fast. *p < 0.05; Mean ± SEM; n = 3. (F) Quantitative RT-PCR analysis of female ACE and lacZ injected mice on HFD. Mice were fasted overnight and tissues were collected. ACE mice have increased mRNA levels of UCP2 and CPT1a, indicative of increased fatty acid oxidation. Additionally, ACE mice had increased PEPCK expression, suggesting that the glyoxylate shunt and PEPCK form a cycle for acetyl-CoA oxidation in vivo. *p < 0.05; Mean ± SEM; n = 3–4. Abbreviations: G6Pase: glucose-6-phosphatase, PEPCK: phosphoenolpyruvate carboxykinase, PPAR-alpha: peroxisome proliferator-activated receptor alpha, UCP2: mitochondrial uncoupling protein 2, CPT1a: carnitine palmitoyltransferse-1a, MCAD: medium-chain acyl-CoA dehydrogenase, AOX: acyl-CoA oxidase.

Figure 6. The Glyoxylate Shunt Decreases Liver Malonyl-CoA and Triglyceride Levels

(A) Female ACE mice on HFD had lower liver malonyl-CoA levels on HFD. ACE mice may resist diet-induced obesity in part due to decreased liver malonyl-CoA and therefore decreased CPT1a inhibition. *P < 0.05 compared to corresponding lacZ injected mouse on same diet; Mean ± SEM; n = 3–4 (B) Glyoxyate shunt expression led to decreased total ATP levels in the liver, presumably due to bypassing the oxidative portion of the TCA cycle. *p < 0.05; Mean ± SEM; n = 4 (C) Female ACE mice had decreased liver TG content. *P < 0.05; Mean ± SEM; n = 4 (D) Plasma analysis of metabolites for female ACE and lacZ injected mice on HFD. Mice were fasted overnight before blood was collected. ACE mice had decreased triglyceride and total cholesterol levels. Fasting glucose levels were identical, suggesting that the glyoxyate shunt does not serve a gluconeogenic function in the liver. *p < 0.05; Mean ± SEM; n = 4–6 (E) Plasma concentrations of ketone body 3-hydroxybutyrate in female ACE and lacZ injected mice on HFD after overnight fast. *p < 0.05; Mean ± SEM; n = 3. (F) Quantitative RT-PCR analysis of female ACE and lacZ injected mice on HFD. Mice were fasted overnight and tissues were collected. ACE mice have increased mRNA levels of UCP2 and CPT1a, indicative of increased fatty acid oxidation. Additionally, ACE mice had increased PEPCK expression, suggesting that the glyoxylate shunt and PEPCK form a cycle for acetyl-CoA oxidation in vivo. *p < 0.05; Mean ± SEM; n = 3–4. Abbreviations: G6Pase: glucose-6-phosphatase, PEPCK: phosphoenolpyruvate carboxykinase, PPAR-alpha: peroxisome proliferator-activated receptor alpha, UCP2: mitochondrial uncoupling protein 2, CPT1a: carnitine palmitoyltransferse-1a, MCAD: medium-chain acyl-CoA dehydrogenase, AOX: acyl-CoA oxidase.

Figure 7. ACE Mice Have Decreased RER and Increased Energy Expenditure

(A) ACE and lacZ injected mice consumed the same amount of food normalized to body weight. n=7–8 (B) ACE mice had decreased RER during light, dark and total 24-hour period, suggesting increased fat oxidation. *p < 0.05 compared to corresponding lacZ injected mouse during same period; Mean ± SEM; n = 7–8. (C) ACE mice demonstrated increased energy expenditure, which may partially explain decreased weight gain observed in ACE mice. *p < 0.05 compared to corresponding lacZ injected mouse during same period; Mean ± SEM; n = 7–8. (D) ACE and lacZ injected mice had similar oxygen consumption levels. Mean ± SEM; n = 7–8. (E) ACE and lacZ injected mice had similar levels of activity. Mean ± SEM; n = 7–8.

Figure 7. ACE Mice Have Decreased RER and Increased Energy Expenditure

(A) ACE and lacZ injected mice consumed the same amount of food normalized to body weight. n=7–8 (B) ACE mice had decreased RER during light, dark and total 24-hour period, suggesting increased fat oxidation. *p < 0.05 compared to corresponding lacZ injected mouse during same period; Mean ± SEM; n = 7–8. (C) ACE mice demonstrated increased energy expenditure, which may partially explain decreased weight gain observed in ACE mice. *p < 0.05 compared to corresponding lacZ injected mouse during same period; Mean ± SEM; n = 7–8. (D) ACE and lacZ injected mice had similar oxygen consumption levels. Mean ± SEM; n = 7–8. (E) ACE and lacZ injected mice had similar levels of activity. Mean ± SEM; n = 7–8.