The Concentration of Phosphatidylethanolamine in Mitochondria Can Modulate ATP Production and Glucose Metabolism in Mice

2620 Diabetes Volume 63, August 2014

Jelske N. van der Veen,1,2 Susanne Lingrell,1,2 Robin P. da Silva,1,3 René L. Jacobs,1,3 and Dennis E. Vance1,2

Diabetes 2014;63:2620–2630 | DOI: 10.2337/db13-0993

Phosphatidylethanolamine (PE) N-methyltransferase (PEMT) of type 2 diabetes is increasing, and therefore under- catalyzes the synthesis of phosphatidylcholine (PC) standing the underlying physiology is of major impor- in the liver. Mice lacking PEMT are protected against tance. Up to 75% of patients with obesity and diabetes diet-induced obesity and insulin resistance. We in- also suffer from nonalcoholic fatty liver disease (NAFLD) vestigated the role of PEMT in hepatic carbohydrate me- (1,2), indicating that the liver plays an important role in tabolism in chow-fed mice. A pyruvate tolerance test the etiology of obesity-associated diabetes. Steatosis in ﬁ revealed that PEMT de ciency greatly attenuated gluco- the liver is often associated with hepatic IR; the exact neogenesis. The reduction in glucose production was mechanisms by which these conditions are related remain speciﬁc for pyruvate; glucose production from glycerol METABOLISM unclear. IR may cause accumulation of triacylglycerol in was unaffected. Mitochondrial PC levels were lower and the liver (3,4). In an IR state, elevated plasma concentra- PE levels were higher in livers from Pemt2/2 compared +/+ tions of both glucose and fatty acids can lead to increased with Pemt mice, resulting in a 33% reduction of the 2/2 hepatic de novo lipogenesis and steatosis (3,4). Intracel- PC-to-PE ratio. Mitochondria from Pemt mice were also smaller and more elongated. Activities of cytochrome c lular accumulation of lipids, such as diacylgycerol or oxidase and succinate reductase were increased in mito- ceramide, can activate protein kinase C and subsequently chondria of Pemt2/2 mice. Accordingly, ATP levels in hep- impair insulin signaling (5,6). Hence, metabolism of fat atocytes from Pemt2/2 mice were double that in Pemt+/+ and glucose in the liver is connected, particularly in the hepatocytes. We observed a strong correlation between mitochondria, where fatty acid oxidation supplies acetyl- mitochondrial PC-to-PE ratio and cellular ATP levels in hep- CoA to the tricarboxylic acid (TCA) cycle for ATP produc- atoma cells that expressed various amounts of PEMT. tion. The TCA cycle also provides a source of carbons for Moreover, mitochondrial respiration was increased in cells gluconeogenesis. Therefore, abnormal lipid and glucose lacking PEMT. In the absence of PEMT, changes in mito- metabolism associated with IR and type 2 diabetes may chondrial phospholipids caused a shift of pyruvate toward be related to abnormalities in mitochondria. decarboxylation and energy production away from the car- Hepatic phosphatidylcholine (PC) synthesis also has boxylation pathway that leads to glucose production. a role in the development of IR (7). PC is synthesized via the cytidine diphosphate–choline pathway (8). Alterna- tively, hepatocytes can methylate phosphatidylethanol- Obesity is often accompanied by hepatic steatosis and amine (PE) via PE N-methyltransferase (PEMT) (9). The insulin resistance (IR) or type 2 diabetes. The prevalence PEMT pathway is responsible for ;30% of hepatic PC

1Group on Molecular and Cell Biology of Lipids, the Alberta Diabetes Institute, the Received 25 June 2013 and accepted 18 March 20014. Mazankowski Alberta Heart Institute, Edmonton, Alberta, Canada © 2014 by the American Diabetes Association. Readers may use this article as 2 Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada long as the work is properly cited, the use is educational and not for proﬁt, and 3 Department of Agricultural, Food and Nutritional Science, University of Alberta, the work is not altered. Edmonton, Alberta, Canada Corresponding author: Dennis E. Vance, [email protected]. diabetes.diabetesjournals.org van der Veen and Associates 2621

synthesis; the remaining 70% is generated via the cytidine 5 mmol/L KCl, and 7.5 mmol/L MgCl2 (pH 7.4). The diphosphate–choline pathway (10). Unexpectedly, we re- homogenate was centrifuged at 1,000g for 10 min at ported that mice lacking PEMT are protected against 4°C. The pellet was resuspended in 10 mmol/L Tris-HCl, high-fat diet (HFD)-induced obesity and IR (7). This pro- 10 mmol/L KCl, and 0.15 mmol/L MgCL2 (pH 6.7) and tection was completely blunted when the diet was supple- frozen at 280°C. Thawed cells were disrupted using mented with choline, leading to the conclusion that PEMT a Dounce homogenizer, mixed with one-seventh the vol- provides an endogenous source of choline that supports ume of 2 mol/L sucrose and centrifuged at 1,000g for 10 2 2 normal weight gain (7). Although Pemt / mice were com- min. The supernatant was centrifuged at 12,000g for 15 pletely protected against IR, hepatomegaly and steatosis min at 4°C. The pellet was resuspended in 10 mmol/L Tris- developed when the mice were fed an HFD. This protection HCl, 0.15 mmol/L MgCl2, and 0.25 mol/L sucrose (pH 6.7). 2 2 against IR in the presence of hepatic steatosis led us to Mitochondria from livers of Pemt+/+ and Pemt / mice investigate the role of PEMT in hepatic carbohydrate me- were isolated as previously described (11). 3 tabolism, independent of body weight or fatty liver. In Vivo 2-[ H]deoxy-D-Glucose Uptake 2/2 We report that in livers of Pemt mice, the concen- Nonfasted animals were injected with 0.1 units/kg insulin tration of PE in mitochondria is increased, leading to in- 3 i.v., 2 g/kg glucose i.v., and 300 mCi/kg 2-[ H]deoxy-D- fl creased ux of pyruvate into the TCA cycle for subsequent glucose i.v. (12). Fifteen minutes after injection, blood energy production. Consequently, hepatic glucose produc- was collected by cardiac puncture into EDTA-containing tion from pyruvate was strongly diminished, suggesting tubes, plasma was prepared by centrifugation, and organs that these changes may contribute to the protection 2 2 were harvested and snap-frozen in liquid nitrogen. Plasma against the development of IR in Pemt / mice. proteins were precipitated with 0.3 N Ba(OH)2 and 0.3 N RESEARCH DESIGN AND METHODS ZnSO4. Radioactivity in the supernatant was divided by the plasma glucose concentration to calculate the specific Animals 2 2 radioactivity of glucose. White adipose, skeletal muscle, Female C57Bl/6 Pemt+/+ and Pemt / mice (backcrossed and liver were homogenized in 0.5% perchloric acid, and .7 generations) (7) were fed standard rodent chow (cat. the homogenate was centrifuged for 20 min at 2,000g. no. 5001; LabDiet) ad libitum with free access to water. Half of the supernatant was treated with Ba(OH) and Experimental procedures were approved by the University 2 ZnSO4. Radioactivity was measured in both aliquots, of Alberta’s Institutional Animal Care Committee in ac- 3 and radioactivity in 2-[ H]deoxy-D-glucose was calculated cordance with the Canadian Council on Animal Care. as the difference between the supernatants divided by In Vivo Tolerance Tests sample weight. This value was divided by the specific ra- Mice (4–6 months old) were nonfasted before the gluca- dioactivity of glucose to calculate glucose uptake per gram gon tolerance tests, fasted for 4 h before insulin tolerance of tissue (micromoles per gram of tissue). tests and for 12 h before glucose, pyruvate, and glycerol Phospholipid Labeling in Primary Hepatocytes tolerance tests. Subsequently, the mice were injected in- 2 2 Hepatocytes were isolated from Pemt+/+ and Pemt / mice traperitoneally with bovine insulin (1 units/kg body wt after perfusion of the liver with collagenase (13) and cul- i.p.), glucagon (140 mg/kg body wt i.p.), sodium pyruvate tured in 60-mm collagen-coated dishes in Dulbecco’smod- (2 g/kg body wt i.p.), or glycerol (2 g/kg body wt i.p.) or ified Eagle’s medium (DMEM) containing 10% FBS for 18 h. alternatively, glucose (2 g/kg body wt i.p.) was adminis- tered by oral gavage. Blood glucose levels were measured Cells were washed twice with serum-free DMEM and pulsed m 3 using a glucometer (Accu-Chek). with 10 Ci [ H]serine/dish for 1 h, followed by a chase of 0, 1, 2, or 4 h in the presence of 1 mmol/L hydroxylamine Real-Time Quantitative PCR and Analysis of to inhibit phosphatidylserine decarboxylase (14). Mitochon- Mitochondrial DNA Content dria were isolated and lipids extracted with chloroform/ RNA isolation, cDNA synthesis, and real-time quantitative methanol (2:1 v/v). Phospholipids were separated by thin- PCR were performed as described (7). mRNA levels were layer chromatography (15), followed by methanolysis with normalized to cyclophilin mRNA. For determination of mi- 3 CH3OH/sulphuric acid at 80°C for 2 h. Since [ H]serine is tochondrial DNA copy number, total DNA was extracted also incorporated into fatty acids (16), radioactivity in only from livers using a DNEasy kit (Qiagen). The mitochon- the phospholipid head groups was measured. drial DNA content was calculated using real-time quan- Analytical Procedures titative PCR by measuring a mitochondrial gene (Nd1, Mitochondrial PC, PE, and phosphatidylserine were forward: ACA CTT ATT ACA ACC CAA GAA CAC AT, re- quantified by phosphorous assay (15) after separation verse: TCA TAT TAT GGC TAT GGG TCA GG) versus by thin-layer chromatography. ATP in McArdle-RH7777 a nuclear gene (Lpl, forward: GAA AGG TGT GGG GAG cells and primary hepatocytes was measured using a lucif- ACA AG, reverse: TCT GTC AAA GGC ACT GAA CG). erase assay kit (Sigma-Aldrich) according to the manufac- Isolation of Mitochondria turer’s instructions. Hepatic glycogen content was measured McArdle-RH7777 hepatoma cells were collected and as previously described (17). Plasma insulin concentrations homogenized in 1 mmol/L Tris-HCl, 0.13 mmol/L NaCl, were measured using a multispot assay system (Meso Scale 2622 Mitochondrial PE Increases ATP Production Diabetes Volume 63, August 2014

Discovery). Plasma glucagon levels were determined using 19.6 6 0.45 g, respectively, P , 0.05), and the livers of 2 2 Millipore RIA GL-32K. Samples were prepared for elec- Pemt / mice were not steatotic (205.3 6 146.7 vs. tron microscopy (7). Mitochondrial size and shape were 117 6 64.5 nmol/mg protein for Pemt+/+). Fasting blood 2 2 quantified using ImageJ software. Pyruvate dehydroge- glucose levels were slightly reduced in Pemt / mice 2 2 nase (PDH) activity was measured using a microplate as- (Pemt+/+ mice, 4.86 6 0.13 mmol/L; Pemt / mice, say kit (Abcam) in the presence of phosphatase inhibitors. 4.31 6 0.10 mmol/L, P , 0.05); fasting plasma insulin For measurement of PEMT activity, 50-mg cell or liver levels were unaltered by PEMT deficiency (Pemt+/+ mice, 2 2 homogenate was incubated with phosphatidylmonome- 420.5 6 93.1 pg/mL; Pemt / mice, 424.2 6 78.1 pg/mL). thylethanolamine and S-adenosyl[methyl-3H]methionine, Next, we performed tolerance tests for glucose, insulin, 2 2 and the incorporation of radiolabel was measured as de- pyruvate, and glycerol (Fig. 1). Pemt+/+ and Pemt / mice scribed previously (18). Proteins in total liver homogenates that were fed a chow diet were similarly sensitive to in- or isolated mitochondria were resolved by SDS-PAGE. Im- sulin (Fig. 1A). There was a slightly improved glucose 2 2 munoreactive proteins were detected using the enhanced tolerance in Pemt / mice compared with Pemt+/+ mice chemiluminescence system (GE Healthcare) according to (Fig. 1B). This improvement appeared to be due to a lower the manufacturer’s instructions, and protein levels were peak in blood glucose reached after the oral gavage of quantified using ImageJ software. glucose rather than enhanced glucose clearance. We in- Activities of Oxidative Phosphorylation Complexes vestigated whether the improvement in glucose tolerance could be explained by a difference in gluconeogenesis. Enzyme activities of oxidative phosphorylation complexes +/+ Sodium pyruvate was injected intraperitoneally, and the were measured in mitochondria isolated from Pemt and 2/2 rise in blood glucose levels was assessed over 2 h. Glucose Pemt livers (11). Mitochondria were homogenized and production from pyruvate was profoundly attenuated in sonicated before the assay. Succinate and NADH cyto- 2 2 Pemt / mice (Fig. 1C). This reduction was specific for chrome c reductase were measured (19) in assay mixtures pyruvate because the response to an intraperitoneal in- containing 0.5 mmol/L potassium cyanide (KCN), 0.1 jection of glycerol, another precursor for glucose, was not mmol/L cytochrome c, and 300 mmol/L succinate or 10 +/+ 2/2 mmol/L NADH as substrates. Substrates were added at different between Pemt and Pemt mice (Fig. 1D). Since this large difference in glucose production trans- 37°C, and the reduction of cytochrome c was measured spec- lated into only a minor decrease in fasting blood glucose, trophotometrically at 550 nm. Cytochrome c oxidase activity we measured hepatic glycogen levels. Under fed condi- was measured similarly, using reduced ferrocytochrome c as tions, hepatic glycogen levels were not different between a substrate in an assay mixture that lacked KCN and cyto- 2 2 Pemt+/+ and Pemt / mice (29.0 6 6.66 and 31.5 6 3.71 chrome c but contained 0.015% Triton X-100 (20). mg/g liver, respectively). After an overnight fast, glycogen 2 2 Respirometry levels were 50% lower in Pemt / mice than in Pemt+/+ 2 2 Oxygen consumption was measured at 30°C with a Clark- mice (Fig. 2A). Apparently, Pemt / mice use hepatic gly- type oxygen electrode (Qubit Systems, Kingston, ON, cogen to maintain normal blood glucose as compensation 6 Canada). Cells were diluted to 4 3 10 cells/mL in for reduced gluconeogenesis. Uptake of glucose into the 3 DMEM. Substrate-driven respiration was measured in cells liver, measured in vivo by injection of 2-[ H]deoxy-D- 2 2 permeabilized for 5 min with digitonin (20 mg/mL/23 glucose, was 45% lower in Pemt / mice (Fig. 2B); uptake 6 10 cells) on ice and then resuspended in buffer (4 3 of glucose into muscle (Fig. 2C) and white adipose tissue 6 10 cells/mL) containing 125 mmol/L KCl, 20 mmol/L (Fig. 2D) was unaffected. Thus, hepatic gluconeogenesis HEPES, 2 mmol/L MgCl, 2.5 mmol/L KH2PO4,0.1%BSA, from pyruvate was reduced in mice lacking PEMT without and 2 mmol/L ADP. Respiratory substrates were 1.25 hypoglycemia because of an increase in glycogenolysis and mmol/L pyruvate, 1.25 mmol/L malate (complex I), and a decrease in hepatic glucose uptake. 5 mmol/L succinate (complex II) plus 2 mmol/L rotenone (complex I inhibitor). Hormonal and Transcriptional Regulation of Hepatic fi Statistical Analysis Gluconeogenesis Is Not Altered by PEMT De ciency fi Data were analyzed with GraphPad Prism software. All To determine why PEMT de ciency reduced gluconeogen- values are means 6 SEM. Statistical analyses of the tol- esis in mice, we measured plasma glucagon that promotes erance tests were performed using ANOVA, and correla- gluconeogenesis and glycogenolysis. Plasma glucagon con- +/+ tions were determined by the Pearson correlation. For all centrations were not different in fasted mice (Pemt mice, 6 2/2 6 other comparisons, a Student t test was performed. Level 51.5 4.08 pg/mL; Pemt mice, 51.2 3.30 pg/mL). of significance of differences was P , 0.05. Furthermore, blood glucose increased similarly in both mouse models after an intraperitoneal injection of gluca- RESULTS gon (Fig. 3A), indicating equal sensitivity to glucagon. Glucose Production From Pyruvate Is Reduced in Hepatic mRNAs encoding pyruvate carboxylase (Pcx), Chow-Fed Mice Lacking PEMT phosphoenolpyruvate carboxykinase (Pepck), glucose-6- Under chow feeding, there was only a small difference in body phosphatase (G6Pase), enolase 1 (Eno1), and fructose- 2 2 weight between Pemt+/+ and Pemt / mice (21.1 6 0.54 and 1,6-bisphosphatase 1 (Fbp1), involved in hepatic glucose diabetes.diabetesjournals.org van der Veen and Associates 2623

Figure 1—Reduced glucose production from pyruvate in Pemt2/2 mice. Insulin (A), glucose (B), pyruvate (C), and glycerol (D) tolerance tests were performed in Pemt+/+ mice and Pemt2/2 mice. Values are means 6 SEM (n =4–5 per group for insulin and glucose tolerance tests, and n =9–10 per group for pyruvate and glycerol tolerance tests). *P < 0.05.

2 2 production, were not affected by PEMT deficiency (Fig. molar ratio PC-to-PE in mitochondria of Pemt / mice 3B). Consistently, protein levels of Pepck were not differ- was 33% lower than in Pemt+/+ mice (Fig. 4C). The num- 2 2 ent between Pemt+/+ and Pemt / mice (Fig. 3C and D). ber of mitochondria per cell, as assessed by mitochondrial The protein levels of peroxisome proliferator–activated DNA copy number, was not affected by PEMT deficiency receptor g coactivator (Pgc)1a, a transcription factor (Fig. 4D). Electron microscopy revealed that mitochondria 2 2 that influences hepatic glucose production (21), were in livers from Pemt / mice were smaller than in Pemt+/+ also unaffected by PEMT deficiency (Fig. 3C and D). These mice (Fig. 4E and F). Moreover, mitochondria in PEMT- data suggest that the reduction in glucose production deficient livers were more elongated than in livers of from pyruvate in PEMT-deficient mice is not due to tran- Pemt+/+ mice (Fig. 4E and G). Hepatic protein levels of scriptional regulation of these genes. mitofusin (Mfn)1 and 2 and optic athrophy protein 2 2 (Opa)1 were not different between Pemt+/+ and Pemt / PEMT Deficiency Alters Mitochondrial Phospholipid mice (Fig. 4H and I), indicating that fusion and fission Composition and Increases Oxidative Phosphorylation of mitochondria were not affected by the lack of PEMT. and ATP Levels We determined whether these changes in phospholipid Since mRNAs and proteins involved in hepatic glucose composition and morphology affected mitochondrial production were not altered by PEMT deficiency (Fig. 3B function. Initially, we examined the hepatic expression and C), we investigated hepatic pyruvate metabolism. In of genes involved in the TCA cycle and oxidative addition to glucose production, pyruvate can be converted phosphorylation (Fig. 5A). The levels of mRNAs encoding into acetyl-CoA for energy production. We, therefore, de- cytochrome synthase (CS-1), isocitrate dehydrogenase termined whether any changes occurred in hepatic mito- (Idh3a), ATP synthase (Atp5b), malate dehydrogenase chondria in response to PEMT deficiency. Since PEMT (Mdh2), cytochrome c oxidase (Cox4i1), succinate dehy- activity is present in mitochondrial associated membranes drogenase (Sdha), and PDH E1a1(Pdha1) were not al- (MAM) (22), we analyzed the phospholipid composition tered, whereas the expression of cytochrome c (Cytc) 2 2 of hepatic mitochondria. PC levels were lower in mito- was slightly increased in Pemt / mice (Fig. 5A). Pdha1 2 2 chondria of Pemt / compared with Pemt+/+ mice (Fig. is part of the PDH complex that converts pyruvate into 4A), whereas the amount of mitochondrial PE was in- acetyl-CoA that is used by the TCA cycle or for fatty acid 2 2 creased by PEMT deficiency (Fig. 4B). Consequently, the synthesis. PDH activity was 25% lower in Pemt / mice 2624 Mitochondrial PE Increases ATP Production Diabetes Volume 63, August 2014

Figure 2—Reduced hepatic glycogen levels and glucose uptake in Pemt2/2 mice. A: Hepatic glycogen levels in Pemt+/+ mice (black bars) and Pemt2/2 mice (white bars) after an overnight fast. B–D: Glucose uptake by liver (B), muscle (C), and white adipose tissue (D)ofPemt+/+ mice 2/2 3 and Pemt mice, measured 15 min after injection of 2-[ H]deoxy-D-glucose i.v. Values are means 6 SEM (n = 6 per group). *P < 0.05.

than in Pemt+/+ mice (Fig. 5F). The activities of cyto- phosphorylation complexes than cells that lack PEMT. chrome c oxidase (complex IV) and succinate:cytochrome Importantly, O2 consumption from endogenous sub- c reductase (complex II) were higher in mitochondria from strates (basal respiration in intact cells) was 29% lower PEMT-deficient livers, whereas the activity of NADH cyto- in cells that expressed PEMT (Fig. 6B). Furthermore, the chrome c reductase (complex I) was not changed (Fig. rates of substrate-driven respiration through complexes I 5C–E). The apparent increase in oxidative phosphoryla- and II (from pyruvate/malate or succinate, respectively), tion was underscored by increased amounts of proteins were 31% and 32% lower in p38 cells (Fig. 6B). These of complexes I–V in PEMT-deficient mitochondria (Fig. results clearly show that PEMT expression directly reduces 5G and H). The amount of ATP was twofold higher in mitochondrial respiration in hepatocytes. 2/2 +/+ hepatocytes from Pemt compared with Pemt mice Deficiency of PEMT Leads to PE Accumulation in (Fig. 5B). The cells were cultured in DMEM (25 mmol/L Mitochondria glucose) without addition of fatty acids; thus, higher To investigate the mechanism by which PEMT deficiency ATP levels are likely due to increased glucose/pyruvate affected mitochondrial phospholipid composition, we 2 2 oxidation rather than fatty acid oxidation. Thus, mito- used primary hepatocytes from Pemt+/+ and Pemt / chondrial morphology and phospholipid composition mice. Mitochondrial PE is made on mitochondrial inner fi fi fi are signi cantly modi ed by PEMT de ciency with in- membranes by phosphatidylserine decarboxylase (PSD) creased activity of the mitochondrial electron transport (23). We pulse-labeled hepatocytes with [3H]serine to la- chain. bel the mitochondrial PE pool. During the 4-h chase pe- PEMT Deficiency Increases Mitochondrial Respiration riod, PSD activity was inactivated by hydroxylamine to fi For assessment of the effect of PEMT de ciency on O2 prevent formation of additional mitochondrial PE. During consumption, basal and substrate-driven respiration was the chase period, radioactivity in mitochondrial PE 2 2 measured in McArdle-RH7777 hepatoma cells with or decayed in Pemt+/+ hepatocytes—not in Pemt / without PEMT expression. Cells with PEMT activity com- hepatocytes (Fig. 7A). In Pemt+/+ hepatocytes, radioactiv- parable with liver (p38 cells) (Fig. 6A) were compared with ity in PC increased, while no additional radiolabel was 2 2 control cells that did not express PEMT (pCI). Figure 6C incorporated into PC in Pemt / hepatocytes (Fig. 7B). and D show that, similar to what we observed in livers, These data indicate that in Pemt+/+ hepatocytes, mito- cells that express PEMT have lower expression of oxidative chondrial PE is exported to the endoplasmic reticulum diabetes.diabetesjournals.org van der Veen and Associates 2625

Figure 3—Hormonal or transcriptional activation of gluconeogenesis is not altered by PEMT deficiency. A: Glucagon tolerance test. Blood glucose levels were monitored after injection of glucagon (140 mg/kg body wt i.p.) into nonfasted Pemt+/+ mice and Pemt2/2 mice. B: Gluconeogenic gene expression in livers of Pemt+/+ and Pemt2/2 mice was determined using quantitative PCR. Data are relative to cyclophilin mRNA. C: Protein expression of PEPCK and PGC1a in livers of Pemt+/+ mice and Pemt2/2 mice. Protein disulfide isomerase (PDI) was used as a loading control. D: Quantification of protein levels from C. Values are means 6 SEM (n = 6 per group).

(ER)/MAM for methylation by PEMT. In contrast, in membrane phospholipid composition, which results in PEMT-deficient hepatocytes, PE is not methylated to increased respiratory capacity. An increased utilization of PC, resulting in reduced PE export and thus accumulation pyruvate for ATP production leads to lower substrate of PE in mitochondria. availability for hepatic glucose production. The data fi PE Concentration Correlates With ATP indicate that elevated hepatic PE may have bene cial To test whether mitochondrial phospholipid composition glucose-lowering effects in IR or diabetic patients. and ATP levels are directly correlated, we used McArdle- PEMT-mediated PC synthesis contributes to the de- RH7777 hepatoma cells that stably express different velopment of IR (7). When mice lacking PEMT were fed amounts of PEMT (13). We analyzed phospholipid com- an HFD, the mice were protected against obesity and IR; position of mitochondria as well as ATP levels. As however, they developed hepatic steatosis (7). Since stea- expected, increased PEMT activity was accompanied by tosis is often associated with IR, we hypothesized that fi an increased PC-to-PE ratio in mitochondria (r2 = 0.9717). lack of PEMT may have direct bene cial effects on hepatic A striking correlation was observed between ATP levels glucose metabolism, which may compensate for the det- and PEMT activity (r2 = 0.671) (Fig. 8A) and between ATP rimental effects of steatosis. Indeed, the data suggest that and the PC-to-PE ratio in mitochondria (r2 = 0.632) (Fig. the profound reduction in hepatic glucose production fi 8B). Moreover, cellular ATP correlated with mitochondrial induced by PEMT de ciency can contribute to the pro- PE concentrations (r2 = 0.568) (Fig. 8C) but not with tection against diet-induced IR. Elevated gluconeogen- mitochondrial PC (Fig. 8D). Thus, PEMT-mediated PC syn- esis is often a consequence of IR rather than a cause. thesis can regulate the phospholipid composition of mito- Under the conditions used in this study, both animal chondria, and PE and/or the PC-to-PE ratio can affect ATP models are normally sensitive to insulin. Therefore, it fi levels in hepatocytes. appears that PEMT de ciency has a direct effect on lowering hepatic glucose production, which could be benefi- DISCUSSION cial for ameliorating hyperglycemia in IR or diabetic Elevated gluconeogenesis contributes to the pathogenesis subjects. of IR and diabetes. Thus, inhibition of gluconeogenesis may Thus, the question remained how hepatic phospholip- be a pharmaceutical target for treatment of these disor- ids might directly influence carbohydrate metabolism in ders. We report that PEMT deficiency alters mitochondrial the liver. Hepatic lipid and carbohydrate metabolism are 2626 Mitochondrial PE Increases ATP Production Diabetes Volume 63, August 2014

Figure 4—PEMT deficiency alters hepatic mitochondrial phospholipid composition and morphology. A–C: Levels of PC, PE, and the PC-to- PE ratio were determined in mitochondria from livers of Pemt+/+ mice and Pemt2/2 mice (n = 6/group). D: Mitochondrial copy number in livers of Pemt+/+ mice and Pemt2/2 mice were determined by the ratio of mitochondrial DNA to nuclear DNA (n = 6/group). E: Transmission electron microscope images of mitochondria in liver. White arrows indicate mitochondria. LD, lipid droplet; N, nucleus. The size (F) and circularity (G) of individual mitochondria (n = 150–350/group) were quantified using ImageJ software. H: Protein expression of mitofusin Mfn1 and -2 and Opa1 in mitochondria from livers of Pemt+/+ mice and Pemt2/2 mice. Voltage-dependent anion channel (VDAC) was used as a loading control. I: Quantification of protein levels from H. Values are means 6 SEM. *P < 0.05.

highly interconnected, particularly in mitochondria, respiratory capacity, ATP production, and electron trans- where both fatty acids and carbohydrates can be used port chain complex activities and profoundly altered mi- for energy production. PEMT activity is present in MAM tochondrial morphology (20). Our data show that an (22). One function proposed for MAM is a membrane increase in mitochondrial PE positively correlates with bridge between the ER and mitochondria, mediating electron transport chain complex activities, ATP levels, lipid transfer between these two organelles (24). Mito- and respiration in hepatocytes. chondrial PE is preferentially synthesized in mitochon- How can increased mitochondrial energy production dria via PSD rather than being imported from the ER. ultimately reduce gluconeogenesis? The TCA cycle and Moreover, PSD-derived PE is exported from mitochon- electron transport chain directly link the metabolism of dria to the MAM/ER for methylation to PC via PEMT. lipids and glucose. Thus, defects in hepatic glucose Therefore, it is not surprising that the lack of PEMT production can be secondary to changes in fatty acid alters the phospholipid composition in hepatic mito- oxidation or ﬂux through the TCA cycle. For example, 2 2 2 2 chondria. The importance of phospholipids for mito- both Ppara / and Pgc-1a / mice have defects in mitochondrial function has been recognized. Cardiolipin, chondrial metabolism and fatty acid oxidation (26,27) as amitochondrial-speciﬁc phospholipid, is important for well as defects in gluconeogenesis, despite normal expres- mitochondrial function such as cellular respiration and sion of gluconeogenic enzymes (26–28). Consistent with 2 2 2 2 ATP production (25). Moreover, modest depletion of mi- our observations in Pemt / mice, both Ppara / and Pgc- 2 2 tochondrialPEinChinesehamsterovarycellsdecreased 1a / mice are resistant to HFD-induced IR (28–31), diabetes.diabetesjournals.org van der Veen and Associates 2627

Figure 5—Increased oxidative phosphorylation and ATP production in Pemt2/2 mice. A: mRNA expression of genes involved in the TCA cycle in livers of Pemt+/+ mice and Pemt2/2 mice was quantiﬁed using quantitative PCR. Data are relative to cyclophilin mRNA. Values are means 6 SEM (n = 6/group). B: ATP levels in primary hepatocytes isolated from Pemt+/+ mice and Pemt2/2 mice. Values are means 6 SEM of four independent experiments. C–E: Activities of complexes I, II, and IV in mitochondria isolated from livers of Pemt+/+ mice and Pemt2/2. Values are means 6 SEM (n = 6/group). F: Hepatic PDH activity in Pemt+/+ mice and Pemt2/2 mice after an overnight fast. Values are means 6 SEM (n = 12/group). G: Protein expression of oxidative phosphorylation complexes I–IV in mitochondria from Pemt+/+ mice and Pemt2/2 livers. Expression was quantiﬁed using ImageJ software (H). Values are means 6 SEM (n = 4/group). *P < 0.05.

which can partly be explained by impairment of hepatic reduced substrate availability. The elevated flux through the 2 2 glucose production. PGC-1a deficiency directly decreased TCA cycle in Pemt / mice shifts the utilization of pyruvate hepatic expression of enzymes of the TCA cycle and the toward energy rather than glucose production. Surprisingly, electron transport chain but had no impact on the expres- PDH activity was slightly reduced by PEMT deficiency. Al- sion of genes involved in fatty acid oxidation or gluconeo- though seemingly counterintuitive, it could be a response to genesis (27). Nevertheless, gluconeogenesis was diminished high levels of ATP, which can activate pyruvate dehydroge- 2 2 in Pgc-1a / mice, indicating that impaired hepatic energy nase PDH kinase, which phosphorylates and inactivates PDH. production, as a result of the defects in the TCA cycle, Mitochondrial dysfunction, as exemplified by impaired inhibited hepatic glucose production. Cytosolic glucose pro- energy generation, is associated with IR in skeletal muscle duction from glycerol was unaffected by PGC-1a deficiency (32–34). Muscle mitochondrial activity is reduced in di- 2 2 (27). In contrast, instead of a decrease, Pemt / mice abetic patients (32,35) as well as in IR children of type 2 exhibited an increase in hepatic levels of proteins involved diabetic patients (34). Recent reports show that also in in the electron transport chain and an increase in ATP levels. livers of diabetic patients, energy homeostasis is abnormal Thus,decreasedglucoseproductioninPEMT-deficient mice is as indicated by lower rates of hepatic ATP synthesis (36). notcausedbyimpairedhepaticenergyproduction.Therefore, Importantly, reduced hepatic ATP concentrations related attenuated glucose production from pyruvate is caused by to IR independent of hepatic lipid content (37). Thus, 2628 Mitochondrial PE Increases ATP Production Diabetes Volume 63, August 2014

Figure 6—Increased respiration in hepatoma cells lacking PEMT. A: PEMT activity in McArdle-RH7777 cells without (pCI) or with (p38) PEMT expression compared with normal mouse liver homogenate. B: Basal rate of O2 consumption in intact cells from endogenous substrates (intact) and substrate-driven O2 consumption in digitonin-permeabilized cells supplied with substrates: pyr/mal, pyruvate plus malate (complex I); Suc, succinate (complex II) plus rotenone (inhibitor of complex I). Values are means 6 SEM of four independent experiments, each performed in duplicate. C: Protein expression of oxidative phosphorylation complexes I–IV in mitochondria from pCI and p38 McArdle-RH7777 cells. Expression was quantiﬁed using ImageJ software (D). *P < 0.05.

increasing hepatic energy production by PEMT inhibition tissue in the regulation of glucose homeostasis was could improve hepatic insulin sensitivity. reported (40). Transplantation of brown adipose tissue Brown adipose tissue is a highly energetic, mitochondria- into mice improved metabolic parameters such as glucose dense organ responsible for diet- and cold-induced thermo- tolerance and insulin sensitivity. Consequently, defects in genesis (38,39). Recently, a direct role for brown adipose mitochondrial function in brown adipose tissue could lead

Figure 7—Metabolism of mitochondrial PC and PE in hepatocytes. A: Decay of radiolabeled PE in mitochondria in primary hepatocytes from Pemt+/+ and Pemt2/2 mice. Cells were pulse labeled with [3H]serine for 1 h, followed by a 4-h chase period. B: Incorporation of [3H] serine into the choline head group of PC in Pemt+/+ and Pemt2/2 hepatocytes. Values are means 6 SEM of three independent experiments. diabetes.diabetesjournals.org van der Veen and Associates 2629

Figure 8—ATP levels correlate with mitochondrial PE content of hepatoma cells. Correlation of cellular ATP concentration with PEMT activity (A), mitochondrial PC-to-PE ratio (B), mitochondrial PE (C), and mitochondrial PC (D) in McArdle-RH7777 hepatoma cells that express different amounts of PEMT. Correlations were determined by Pearson correlation. Values are means 6 SEM of four independent experiments, each performed in triplicate.

to metabolic diseases including obesity and type 2 diabe- Funding. This research was supported by grants from the Canadian Institutes tes. Even in white adipose tissue, where mitochondria are of Health Research (to R.L.J.) and the Canadian Diabetes Association (to D.E.V.). less abundant, impaired mitochondrial activity appears to Part of this research was accomplished during the tenure of a Postdoctoral predispose mice to obesity and type 2 diabetes (41). Thus, Fellowship to J.N.v.d.V. from Alberta Innovates-Health Solutions. R.L.J. is a elevation of PE or reduction of PC-to-PE ratio in mitochon- New Investigator of the Canadian Institutes of Health. D.E.V. was a scientist of Alberta Innovates-Health Solutions. dria of other tissues could also have beneficial metabolic Duality of Interest. No potential conflicts of interest relevant to this article effects. were reported. In conclusion, mitochondrial PE levels and/or PC-to-PE Author Contributions. J.N.v.d.V. designed the experiments, performed ratio dictate energy production in hepatocytes. We the experiments, and wrote the first draft of the manuscript. S.L. and R.P.d.S. observed a strong correlation between cellular ATP levels performed the experiments and read and revised the manuscript. R.L.J. and D.E.V. and mitochondrial PE or PC-to-PE ratio. Moreover, designed the experiments and read and revised the manuscript. J.N.v.d.V. and respiratory capacity is increased when PEMT is lacking D.E.V. are the guarantors of this work and, as such, had full access to all the data and the PC-to-PE ratio is low. This could cause an in the study and take responsibility for the integrity of the data and the accuracy increased flux of pyruvate through the TCA cycle, re- of the data analysis. ducing the availability of pyruvate for hepatic glucose References production. Since elevated gluconeogenesis contributes to the pathogenesis of IR and type 1 and type 2 diabetes, 1. McCullough AJ. Update on nonalcoholic fatty liver disease. J Clin Gastro- enterol 2002;34:255–262 understanding the role of mitochondrial phospholipids in 2. Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002;346: respiratory capacity and hepatic glucose production may 1221–1231 offer the prospect of more effective therapeutic 3. Marchesini G, Brizi M, Morselli-Labate AM, et al. Association of nonalcoholic approaches for treatment or prevention of these diseases. fatty liver disease with insulin resistance. Am J Med 1999;107:450–455 4. Sanyal AJ, Campbell-Sargent C, Mirshahi F, et al. Nonalcoholic steatohep- atitis: association of insulin resistance and mitochondrial abnormalities. Gastro- Acknowledgments. The authors are grateful to Dr. Ariel D. Quiroga, Dr. enterology 2001;120:1183–1192 Martin Hermansson, and Dr. Jean Vance of the University of Alberta for helpful 5. Jornayvaz FR, Shulman GI. Diacylglycerol activation of protein kinase C´ discussions. and hepatic insulin resistance. Cell Metab 2012;15:574–584 2630 Mitochondrial PE Increases ATP Production Diabetes Volume 63, August 2014

6. Samuel VT, Petersen KF, Shulman GI. Lipid-induced insulin resistance: 24. Vance JE. Phospholipid synthesis in a membrane fraction associated with unravelling the mechanism. Lancet 2010;375:2267–2277 mitochondria. J Biol Chem 1990;265:7248–7256 7. Jacobs RL, Zhao Y, Koonen DP, et al. Impaired de novo choline synthesis 25. Jiang F, Ryan MT, Schlame M, et al. Absence of cardiolipin in the crd1 null explains why phosphatidylethanolamine N-methyltransferase-deficient mice are mutant results in decreased mitochondrial membrane potential and reduced protected from diet-induced obesity. J Biol Chem 2010;285:22403–22413 mitochondrial function. J Biol Chem 2000;275:22387–22394 8. Kennedy EP, Weiss SB. The function of cytidine coenzymes in the 26. Bandsma RH, Van Dijk TH, Harmsel At At, et al. Hepatic de novo synthesis of biosynthesis of phospholipides. J Biol Chem 1956;222:193–214 glucose 6-phosphate is not affected in peroxisome proliferator-activated receptor 9. Ridgway ND, Vance DE. Purification of phosphatidylethanolamine N- alpha-deficient mice but is preferentially directed toward hepatic glycogen stores methyltransferase from rat liver. J Biol Chem 1987;262:17231–17239 after a short term fast. J Biol Chem 2004;279:8930–8937 10. DeLong CJ, Shen YJ, Thomas MJ, Cui Z. Molecular distinction of phos- 27. Burgess SC, Leone TC, Wende AR, et al. Diminished hepatic gluconeo- phatidylcholine synthesis between the CDP-choline pathway and phosphatidyl- genesis via defects in tricarboxylic acid cycle flux in peroxisome proliferator- ethanolamine methylation pathway. J Biol Chem 1999;274:29683–29688 activated receptor gamma coactivator-1alpha (PGC-1alpha)-deficient mice. J Biol 11. Croze EM, Morré DJ. Isolation of plasma membrane, golgi apparatus, and Chem 2006;281:19000–19008 endoplasmic reticulum fractions from single homogenates of mouse liver. J Cell 28. Lin J, Wu PH, Tarr PT, et al. Defects in adaptive energy metabolism with Physiol 1984;119:46–57 CNS-linked hyperactivity in PGC-1alpha null mice. Cell 2004;119:121–135 12. Raines SM, Richards OC, Schneider LR, et al. Loss of PDGF-B activity 29. Finck BN, Bernal-Mizrachi C, Han DH, et al. A potential link between muscle increases hepatic vascular permeability and enhances insulin sensitivity. Am peroxisome proliferator- activated receptor-alpha signaling and obesity-related J Physiol Endocrinol Metab 2011;301:E517–E526 diabetes. Cell Metab 2005;1:133–144 13. Noga AA, Zhao Y, Vance DE. An unexpected requirement for phosphati- 30. Guerre-Millo M, Rouault C, Poulain P, et al. PPAR-alpha-null mice are pro- dylethanolamine N-methyltransferase in the secretion of very low density lipo- tected from high-fat diet-induced insulin resistance. Diabetes 2001;50:2809–2814 proteins. J Biol Chem 2002;277:42358–42365 31. Leone TC, Lehman JJ, Finck BN, et al. PGC-1alpha deficiency causes multi- 14. Shiao YJ, Lupo G, Vance JE. Evidence that phosphatidylserine is imported system energy metabolic derangements: muscle dysfunction, abnormal weight into mitochondria via a mitochondria-associated membrane and that the majority control and hepatic steatosis. PLoS Biol 2005;3:e101 of mitochondrial phosphatidylethanolamine is derived from decarboxylation of 32. Kelley DE, He J, Menshikova EV, Ritov VB. Dysfunction of mitochondria in phosphatidylserine. J Biol Chem 1995;270:11190–11198 human skeletal muscle in type 2 diabetes. Diabetes 2002;51:2944–2950 15. Zhou X, Arthur G. Improved procedures for the determination of lipid 33. Koves TR, Ussher JR, Noland RC, et al. Mitochondrial overload and in- phosphorus by malachite green. J Lipid Res 1992;33:1233–1236 complete fatty acid oxidation contribute to skeletal muscle insulin resistance. Cell 16. Vance JE, Vance DE. Specific pools of phospholipids are used for lipoprotein Metab 2008;7:45–56 secretion by cultured rat hepatocytes. J Biol Chem 1986;261:4486–4491 34. Petersen KF, Dufour S, Befroy D, Garcia R, Shulman GI. Impaired mito- 17. Oosterveer MH, van Dijk TH, Grefhorst A, et al. Lxralpha deficiency hampers the chondrial activity in the insulin-resistant offspring of patients with type 2 di- hepatic adaptive response to fasting in mice. J Biol Chem 2008;283:25437–25445 abetes. N Engl J Med 2004;350:664–671 18. Ridgway ND, Vance DE. Phosphatidylethanolamine N-methyltransferase 35. Szendroedi J, Schmid AI, Chmelik M, et al. Muscle mitochondrial ATP from rat liver. Methods Enzymol 1992;209:366–374 synthesis and glucose transport/phosphorylation in type 2 diabetes. PLoS Med 19. Sottocasa GL, Kuylenstierna B, Ernster L, Bergstrand A. An electron- 2007;4:e154 transport system associated with the outer membrane of liver mitochondria. A 36. Schmid AI, Szendroedi J, Chmelik M, Krssák M, Moser E, Roden M. Liver biochemical and morphological study. J Cell Biol 1967;32:415–438 ATP synthesis is lower and relates to insulin sensitivity in patients with type 2 20. Tasseva G, Bai HD, Davidescu M, Haromy A, Michelakis E, Vance JE. diabetes. Diabetes Care 2011;34:448–453 Phosphatidylethanolamine deficiency in Mammalian mitochondria impairs oxi- 37. Szendroedi J, Chmelik M, Schmid AI, et al. Abnormal hepatic energy ho- dative phosphorylation and alters mitochondrial morphology. J Biol Chem 2013; meostasis in type 2 diabetes. Hepatology 2009;50:1079–1086 288:4158–4173 38. Lowell BB, Spiegelman BM. Towards a molecular understanding of adaptive 21. Yoon JC, Puigserver P, Chen G, et al. Control of hepatic gluconeogenesis thermogenesis. Nature 2000;404:652–660 through the transcriptional coactivator PGC-1. Nature 2001;413:131–138 39. Rothwell NJ, Stock MJ. Effects of age on diet-induced thermogenesis and 22. Cui Z, Vance JE, Chen MH, Voelker DR, Vance DE. Cloning and expression of brown adipose tissue metabolism in the rat. Int J Obes 1983;7:583–589 a novel phosphatidylethanolamine N-methyltransferase. A specific biochemical 40. Stanford KI, Middelbeek RJ, Townsend KL, et al. Brown adipose tissue and cytological marker for a unique membrane fraction in rat liver. J Biol Chem regulates glucose homeostasis and insulin sensitivity. J Clin Invest 2013;123: 1993;268:16655–16663 215–223 23. Zborowski J, Dygas A, Wojtczak L. Phosphatidylserine decarboxylase is 41. De Pauw A, Tejerina S, Raes M, Keijer J, Arnould T. Mitochondrial (dys) located on the external side of the inner mitochondrial membrane. FEBS Lett function in adipocyte (de)differentiation and systemic metabolic alterations. Am J 1983;157:179–182 Pathol 2009;175:927–939