Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Sep 1.
Published in final edited form as: Arterioscler Thromb Vasc Biol. 2018 Sep;38(9):2207–2216. doi: 10.1161/ATVBAHA.118.311339

Mechanism of increased LDL and decreased triglycerides with SGLT2 inhibition

Debapriya Basu 1, Lesley-Ann Huggins 1, Diego Scerbo 1, Joseph Obunike 1, Adam E Mullick 2, Paul L Rothenberg 3, Nicholas A Di Prospero 3, Robert H Eckel 4, Ira J Goldberg 1
PMCID: PMC6207215  NIHMSID: NIHMS1500152  PMID: 30354257

Abstract

Objective:

Sodium glucose cotransporter 2 (SGLT2) inhibition in humans leads to increased levels of LDL cholesterol and decreased levels of plasma triglyceride. Recent studies however, have shown this therapy to lower cardiovascular mortality. In this study, we aimed to determine how SGLT2 inhibition alters circulating lipoproteins.

Approach and Results:

We used a mouse model expressing human cholesteryl ester transfer protein and human apolipoprotein B100 to determine how SGLT2 inhibition alters plasma lipoprotein metabolism. The mice were fed a high fat diet and then were made partially insulin deficient using streptozotocin. SGLT2 was inhibited using a specific anti-sense oligonucleotide or canagliflozin, a clinically available oral SGLT2 inhibitor. Inhibition of SGLT2 increased circulating levels of LDL cholesterol and reduced plasma triglyceride levels. SGLT2 inhibition was associated with increased lipoprotein lipase activity in the post heparin plasma, decreased postprandial lipemia and faster clearance of radiolabeled VLDL from circulation. Additionally, SGLT2 inhibition delayed turnover of labeled LDL from circulation.

Conclusions:

Our studies in diabetic CETP-Apolipoprotein B100 transgenic mice recapitulate many of the changes in circulating lipids found with SGLT2 inhibition therapy in humans and suggest that the increased LDL cholesterol found with this therapy is due to reduced clearance of LDL from the circulation and greater lipolysis of triglyceride-rich lipoproteins. Most prominent effects of SGLT2 inhibition in the current mouse model were seen with ASO mediated knockdown of SGLT2.

Keywords: Diabetes, Lipase, Lipemia, Cardiovascular disease, Animal Models of Human Disease, Lipids and Cholesterol, Metabolism, Diabetes Type 2, Cardiovascular Disease, Genetically Altered and Transgenic Models

INTRODUCTION

Under normal physiologic conditions, plasma glucose is filtered at the renal glomerulus and then fully reabsorbed in the proximal tubule. This process of reabsorption in the tubules is mediated by two glucose transporters: SGLT1 (Sodium glucose cotransporter 1), responsible for ~10% of the reabsorption and SGLT2, responsible for ~90% of the reabsorption1, 2. When circulating levels of plasma glucose exceed the renal threshold for glucose reabsorption (referred to as RTG), glucose is excreted in the urine. In patients with type 2 diabetes mellitus (T2DM), the RTG is increased, primarily due to an increase in SGLT2 expression, which contributes to maintenance of elevated blood glucose levels2–4.

Drugs that inhibit SGLT transporters reduce plasma glucose levels. Phlorizin, which inhibits both SGLT1 and 2, produces glycosuria and normalizes circulating glucose levels in rodents5, 6. Within the past few years, a number of oral SGLT2 inhibitors (e.g., canagliflozin, dapagliflozin, and empagliflozin) have been approved and widely used as treatments for T2DM. However, there have been some concerns as these drugs have consistently been found to increase LDL cholesterol (LDL-c) levels. Although the LDL-c increase is a concern, SGLT2 inhibitors are associated with reduced cardiovascular disease (CVD), including heart failure, and also vascular disease7, 8.

Some drugs associated with increased LDL levels have been found to be either neutral or cardioprotective. Like SGLT2 inhibitors, fibric acids and fish oils reduce triglyceride (TG) levels and often raise LDL-c. This is in part due to reduced cholesteryl ester transfer protein (CETP) actions as its substrate VLDL is decreased. However, increased lipolysis of TG, as occurs with fibrates9, 10 also increases VLDL to LDL conversion. Moreover, greater lipolysis also increases HDL cholesterol (HDL-c), which is seen in some but not all SGLT2 clinical trials11, 12.

Many studies of SGLT2 inhibition in preclinical models such as mice and rats have been reported, but none of these studies explain the changes in circulating lipids that are seen in humans. Hence, an animal model with a lipoprotein profile similar to humans may be required to investigate the potential mechanism for SGLT2 inhibitor-induced lipid changes. To that end, transgenic mice overexpressing CETP and apolipoprotienB100 (ApoB100) were fed a high fat diet (HFD), made diabetic through administration of a low dose of streptozotocin (STZ), and treated with either an SGLT2 anti-sense oligonucleotide (ASO) or canagliflozin to inhibit SGLT2. We then went on to define the changes in lipoprotein physiology that resulted from inhibition of SGLT2.

MATERIALS AND METHODS

Ethical statement for animal studies:

All procedures were conducted in conformity with the National Institutes of Health’s “Public Health Service Policy on Humane Care and Use of Laboratory Animals” and the National Research Council of the National Academy of Sciences’ Guide for the Care and Use of Laboratory Animals and were approved by the New York University Langone Medical Center Institutional Animal Care and Use Committee.

Mice strains, housing, husbandry and diets:

8–10 week old male CETP-ApoB100 transgenic mice on a background of C57Bl/6 (3716-M) were purchased from Taconic Bioscience, Inc. (Hudson, NY). After 1 week of acclimatization, the mice were fed a 60% HFD (Research Diets, D12492) till the end of the study (16 weeks). These mice were used for the results shown in Figure 1–6 and SI Figure I and II. For the data presented in SI Table 1, 10–12 week old male wild type C57Bl/6J mice on laboratory chow diet were used. These mice were initially purchased from Jackson Laboratory (#000664, Bar Harbor, ME) and then bred in the mice facility at the New York University Langone Medical Center.

Figure 1. Effect of SGLT2 inhibition on glucose, body weight, insulin and plasma NEFA in diabetic CETP-ApoB100 transgenic mice.

Figure 1.

Open in a new tab

(a) Experimental outline showing timelines for diet, STZ injections and treatments in CETP-ApoB100 transgenic mice. After inducing hyperglycemia, the mice were divided in four groups: vehicle, canagliflozin, SGLT2 ASO and insulin. These treatments were carried out for another 30 days. (b) Fasting blood glucose before STZ injection (Pre-STZ), after last STZ injection but before starting any treatments (Pre-tx), after 7 days of treatment (D7-tx), 14 days of treatment (D14-tx), 21 days of treatment, (D21-tx), 28 days of treatment (D28-tx), and 30 days of treatment (D30-tx) for vehicle (closed circles), canagliflozin (open circles), SGLT2 ASO (closed triangles) or insulin (open square) treatments (n=8–10 per group). (c) Body weight of mice over time (n=8–10 per group). (d) Fasting plasma insulin levels at Pre-STZ and D30-tx (n=4–5 per group). (e) Fasting plasma NEFA levels at Pre-STZ and at D28-tx (n=8–10 per group). The groups are: Pre-STZ (checkered bar), Vehicle (black bar), canagliflozin (white bar), SGLT2 ASO (bar with horizontal stripes) and insulin (bar with diagonal stripes). Values represent mean ± SD. * denotes p<0.05 and ** denotes p<0.01 when unpaired Student’s t-test was used for analysis; # denotes p<0.05 and ### denotes p<0.001 when ANOVA was used for analysis. All comparisons are with control vehicle group. Note: Human insulin cross reacted with the mouse ELISA kit used, giving very high values.

Figure 6. Lipid levels and metabolic gene expression in the heart.

Figure 6.

Open in a new tab

Cardiac lipid levels: (a) NEFA, (b) TG, and (c) TC in the four groups of mice at the end of 30 days of treatment (n=8–10 per group). Comparison of cardiac mRNA levels using real-time PCR for (d) regulators of ANGPTL4 expression and (e) fatty acid uptake and oxidation, in the four mice groups at the end of 30 days of treatment (n=4–5 per group). Values represent mean ± SD. * denotes p<0.05 when Student’s t-test was used for analysis; # denotes p<0.05 and ### denotes p<0.001 when ANOVA was used for analysis. All comparisons are with control vehicle group.

Mice were maintained in a 12-hour light/dark cycle and given free access to food and water, except when blood samples were obtained. Mice were monitored regularly to ensure their welfare.

Induction of diabetes mellitus:

CETP-ApoB100 transgenic mice were made diabetic by intraperitoneal administration of STZ (S0130 from Sigma Aldrich, St. Louis, MO) at two concentrations: day 1 at 50 mg/kg, day 2–7 at 25 mg/kg13. Blood glucose was measured from tail bleeding using a hand held glucometer and glucose strips (OneTouch Ultra) and was monitored during the first week of STZ administration, and then once a week throughout the treatment period. Wild type C57Bl/6 mice were made diabetic by intraperitoneal injection of STZ (dose of 50 mg/kg body weight) for five consecutive days. Fasting blood glucoses were monitored during the STZ treatment period and then once a week throughout the 4-week treatment period. Animals that had fasting blood glucose concentrations higher than 250 mg/dl were considered as diabetic.

Study design and drug treatments:

A total of forty diabetic CETP-ApoB100 transgenic mice were assigned to one of four treatment groups (n=10/group). Mice in group 1 received vehicle (0.5% methocel solution) daily by oral gavage for 4 weeks. Mice in group 2 received canagliflozin daily by oral gavage at a dose of 30 mg/kg (a concentration of 6 mg/ml in 0.5% methocel solution) for 4 weeks. Canagliflozin and methocel were provided by Janssen Research & Development. Mice in group 3 were injected intraperitoneally with the SGLT2 ASO at a dose of 10mg/kg body weight, once a week for four weeks. The SGLT2 ASO (Ionis 388625) provided by Ionis Pharmaceutical, is a 12-mer phosphorothioate oligonucleotide gapmer containing 2’−0-(2-methoxyethyl)-modified ribonucleosides (2’-MOE) at the 2 flanking positions surrounding an 8 base gap containing 2’-deoxynucleosides (‘2–8–2’ design with a sequence of 5-TGTTCCAGCCCA-3.14. Mice in group 4 had insulin pellets (LinShin Canada Inc., Ontario, Canada) implanted subcutaneously to achieve a slow-release of insulin (0.1 U/day) over the treatment period15. During the 15th week of the study, one mouse in the canagliflozin group and one mouse in the insulin group were found dead, although there were no obvious changes in their appearance or behavior before that.

In another experiment (SI Table 1), a group of ten wild type C57Bl/6 male mice on chow diet was made diabetic. Another group of ten wild type C57Bl/6 male mice on chow diet was injected with vehicle only to serve as the non-diabetic control group. Next, the diabetic group and the non-diabetic group were further divided in two groups: one treated with control ASO (non-targeting) and the other treated with SGLT2 ASO (dose of 10mg/kg body weight), once a week for four weeks. Non-diabetic mice with control ASO and SGLT2 ASO had five mice in each group. Out of ten mice injected with STZ, two did not develop hyperglycemia and were excluded from the study at that point. Hence, diabetic mice with control ASO and SGLT2 ASO had four mice in each group.

Blood sampling:

All the blood samples for measuring total cholesterol (TC), TG, lipoprotein profiling, post-heparin lipoprotein lipase (LpL) activity were collected after a 4-hour fasting period. Mice were anesthetized using isoflurane and blood was collected from the retro-orbital plexus into tubes containing EDTA. Blood was centrifuged at 10,000 g for 10 minutes for collection of the plasma, which was then used immediately for analysis and/or frozen at −80°C. Plasma insulin was quantified using Mouse Ultrasensitive Insulin ELISA kit (ALPCO, Salem, NH).

Tissue collection:

At the end of the study, mice (fasted for 4 hours) were anesthetized with xylazine (10 mg/kg) and ketamine (100 mg/kg) and then perfused by heart puncture with PBS till the livers were blanched. Tissues were rapidly excised and snap frozen in liquid nitrogen unless otherwise noted.

Lipid extraction from heart:

The lipid extraction protocol was adapted from Folch et al.16. Approximately 100 mg of frozen tissue was homogenized in 1 mL of PBS using stainless steel beads for 1 minute in a bead beater homogenizer. From each sample, 50 μL were removed for protein analysis and 3 mL of 2:1 chloroform: methanol was added to the remainder of the sample and vortexed. Samples were then centrifuged for 10 minutes at 3000 rpm at 4°C. The lower, organic phase was then collected and dried under nitrogen gas. The dried lipid was then dissolved in 500 μL of 1% Triton-X 100 in chloroform, further dried and then dissolved in 100μL of double distilled water. This was then used for enzymatic detection of lipids or stored at −20C.

Lipid measurements:

TG and TC were measured using Infinity Triglyceride Reagent and Infinity Total Cholesterol Reagent (Thermo Scientific, Waltham, MA). Lipoproteins–very low density lipoprotein d < 1.006 g/ml, low density lipoprotein d 1.006–1.063 g/ml, and high density lipoprotein d 1.063–1.21 g/ml–were separated by sequential density ultracentrifugation of mouse plasma in a TLA 100 rotor (Beckmann Instruments, Palo Alto, CA). Non-esterified fatty acid (NEFA) levels were measured using a kit (Wako Diagnostics, Richmond, VA).

Western blot analysis:

Snap frozen liver samples were homogenized using a RIPA buffer with protease inhibitors. Protein concentrations were determined using Pierce BCA Protein kit (Thermo Scientific, Waltham, MA). Forty microgram samples of total protein were used for western blot analysis with mouse LDL receptor (LDLR) antibody (R&D AF2255, Carlsbad, CA) at 1:1000 dilution. HSP90 antibody (Abcam ab87133, Cambridge, MA) was used for control of protein loading at a dilution of 1:1000.

Tissue gene expression:

Total RNA was prepared using a GeneJET RNA Purification Kit (Thermo Scientific, Waltham, MA). One microgram of RNA was used for reverse transcription using the Verso cDNA synthesis kit (Thermo Scientific, Waltham, MA) or the High-Capacity cDNA Reverse Transcription Kit (Thermo Scientific, Waltham, MA). Real-time quantitative PCR was performed using an ABI 7700 (Applied Biosystems, Foster City, CA). Amplification was performed using SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA). Primers used for PCR amplification were obtained from Primer Bank (pga.mgh.harvard.edu/primerbank/). Data were normalized with hypoxanthine phosphoribosyltransferase 1(HPRT) or ribosomal protein S3 (RSP3) genes.

Hepatic TG secretion:

To measure hepatic TG production rate, mice were injected intraperitoneally with Poloxamer-407 (P-407), (Sigma-Aldrich, St. Louis, MO) at 1 g/kg (in saline) after a 4-hour fasting period. Immediately prior to injection, and at 1, 2 and 4 hours following injection, blood samples were drawn in heparinized capillary tubes, plasma was prepared, and TG concentrations were determined17.

Olive oil gavage:

After a 4-hour fast, mice were given olive oil by oral gavage at 10ml/kg (~300 μl). Blood samples were drawn at baseline, 1, 2 and 4 hours after receiving the olive oil gavage.

VLDL turnover assay:

VLDL was isolated from human plasma by density gradient ultracentrifugation.14Ctriolein (Perkin Elmer, MA) was dried down in a glass tube under a stream of N2 to obtain a dry lipid film. Isolated VLDL was added to the dried lipid film and sonicated on ice for 15–20 minutes. Approximately, 20 μCi of[14C]triolein/ml of VLDL was used for 20 mice. VLDL was then dialyzed in a buffer containing 150mM NaCl and 0.24mM EDTA at 4C and counts were measured. Radiolabeled VLDL was then diluted with PBS and injected intravenously in mice (1*106 cpm/mouse). Blood samples were collected before (0) and 30 seconds, 5 minutes, 15 minutes, 30 minutes and 60 minutes after injection. Plasma was separated and counts were measured using 10 μl plasma in a scintillation counter.

LDL turnover assay:

LDL was isolated by density gradient ultracentrifugation of plasma from LDLR knockout mice on western diet for 4 weeks. After dialysis to remove excess salt, approximately 1 ml LDL was incubated with 300μCi of[3H]cholesteryl oleate ester (dried down in a glass vial) and 1 ml of CETP-containing 1.21 bottom fraction. G418 antibiotic was added to prevent contamination. The glass vial was incubated overnight at 37°C with constant shaking by inversion. LDL was again isolated by density gradient centrifugation and dialyzed in a buffer containing150 mM NaCl and 0.24 mM EDTA at 4oC and counts were measured. Radiolabeled LDL was then diluted with PBS and injected intravenously in mice (5*106 cpm/mouse). Blood samples were collected before (0) and 30 seconds, 5 minutes, 15 minutes, 30 minutes, 60 minutes and 120 minutes after injection. Plasma was separated and counts were measured using 10 μl plasma in a scintillation counter.

VLDL and LDL were labeled separately as described above and were mixed together with PBS and injected intravenously in mice in a total volume of 100 μl.

Lipoprotein lipase (LpL) activity assay:

Post-heparin plasma LpL activity was determined as previously described18. Post-heparin plasma was obtained from 4-hour fasted mice, 5 min after intravenous injection of 10 units of heparin. To measure total lipase activity, plasma was incubated with 10% Intralipid/[3H]triolein (Perkin Elmer, MA) emulsion as substrate and heat inactivated human serum as the source of apolipoprotein CII. The contribution of hepatic lipase (HL) in the plasma was determined by including NaCl (final concentration 1M) in the assay and was subtracted from the total lipase activity to estimate specific LpL activity. Lipase activity was expressed as μmol FFA/ml/h. Aliquots of human post-heparin plasma were used for a standard curve in all the experiments.

Statistical analyses:

Data are presented as means ± SD. Statistical differences were assessed via unpaired Student’s t-test (for comparison among two groups) or a one-way/two-way ANOVA (Dunnett’s test, Fisher’s LSD or Tukey’s test for posthoc analysis where appropriate) for comparison among multiple groups, using GraphPad Prism 7. A p value of less than 0.05 was considered as significant.

RESULTS

SGLT2 inhibition lowers blood glucose in diabetic CETP-ApoB100 transgenic mice

Figure 1a shows the experimental design for this study. Forty CETP-ApoB100 transgenic mice were fed a 60% high fat diet (HFD) for 11 weeks and then diabetes was induced with a low dose of STZ. The mice developed hyperglycemia within a week (Figure 1b), and subsequently were divided into four groups: controls treated with vehicle (0.5% methocel), canagliflozin-treated, SGLT2 ASO-treated, and insulin-treated. All three interventions – canagliflozin, SGLT2 ASO and insulin – lowered blood glucose in the diabetic mice compared to vehicle; the mean glucose levels across the four groups at the end of the 4-week treatment period were 443 mg/dl in vehicle, 176 mg/dl with canagliflozin, 200 mg/dl with SGLT2 ASO and 234 mg/dl with insulin (Figure1b, p<0.05 vs vehicle). Body weight increased after HFD feeding and the mice lost weight after STZ treatments, however, there was no difference among the four treatment groups (Figure 1c). Insulin levels were low in the STZ-treated mice and canagliflozin further reduced the insulin level compared to vehicle (Figure 1d), similar to what is seen in humans treated with SGLT2 inhibitors19, 20. Both and insulin and ASO reduced circulating NEFA (Figure 1e). SGLT2 gene expression in the kidney was suppressed more than 90% by SGLT2 ASO, whereas canagliflozin and insulin treatments slightly reduced SGLT2 mRNA (SI Figure I).

SGLT2 inhibition alters plasma lipids

Next, we measured plasma lipid levels in these diabetic CETP-ApoB100 mice, which have a plasma lipoprotein distribution similar to humans. After 4 weeks of intervention, both canagliflozin and SGLT2 ASO treatments raised plasma TC, from 165 mg/dl in vehicle to 183 mg/dl with canagliflozin and 200 mg/dl with SGLT2 ASO. Insulin treatment also raised TC modestly (Figure 2a). The increase in TC seen with SGLT2 inhibition was attributed to an increase in LDL-c; the LDL-c level was 71 mg/dl in vehicle, 89 mg/dl with canagliflozin (25% increase), and 108 mg/dl with SGLT2 ASO (52% increase) (Figure 2b). HDL-c levels did not change with the SGLT2 ASO or canagliflozin, but increased with insulin treatment (Figure 2c).

Figure 2. SGLT2 inhibition and plasma lipoprotein levels.

Figure 2.

Open in a new tab

(a) Plasma TC at D28-tx. Cholesterol levels in (b) LDL fraction and (c) HDL fractions at D28-tx. (d) Plasma TG levels and (e) TG levels in VLDL fractions at D28-tx (n=8–10 per group). Values represent mean ± SD. * denotes p<0.05 and ** denotes p<0.01 when Student’s t-test was used for analysis; # denotes p<0.05 and ## denotes p<0.01 when ANOVA was used for analysis. All comparisons are with control vehicle group.

In contrast to canagliflozin, which did not change plasma TG, SGLT2 ASO and insulin reduced circulating TGs (Figure 2d) due to a reduction in VLDL TG (Figures 2e). These lipid alterations of increased LDL-c and reduced TGs, were similar to those found with SGLT2 inhibition in humans with diabetes, although in our studies the greatest changes were found with SGLT2 ASO. SGLT2 ASO treatment in diabetic or non-diabetic wild type C57Bl/6 mice lacking CETP and ApoB100 transgenes had no effect on plasma LDL-c compared to control ASO treated mice (SI Table 1). Our previous study with canagliflozin treatment also did not raise plasma LDL-c in diabetic or non-diabetic wild type mice13.

SGLT2 inhibition increases TG clearance

We next investigated the effect of SGLT2 inhibition on TG metabolism and specifically assessed whether the reduction of plasma TG in the ASO treated mice was due to reduced hepatic TG production or increased clearance from circulation. Subsets of the mice in all four groups were fasted for 4 hours and injected with P-407 to block lipolysis and to measure hepatic TG secretion. During the 4 hour study, there was no significant difference in rates of TG secretion among the groups (Figure 3a). Next, we analyzed postprandial lipemia, which is an index of blood TG clearance and often viewed as an in vivo assay of lipolytic capacity. The mice were gavaged with olive oil, and plasma TG was measured over time. SGLT2 ASO-treated but not canagliflozin-treated mice cleared the TG faster than the vehicle treated mice (Figure 3b). Insulin treatment also caused faster clearance of TG from the circulation, as previously reported21. Additionally, we studied turnover of human VLDL in a separate group of diabetic CETP-ApoB10 mice. SGLT2 ASO-treated mice cleared VLDL tracer from the circulation faster than control (Figure 3c).

Figure 3. Effect of SGLT2 inhibition on hepatic VLDL production, postprandial lipemia and VLDL turnover.

Figure 3.

Open in a new tab

(a) Hepatic VLDL production was determined in 4-hour fasted mice after injection with P-407 to inhibit lipolysis. Blood samples were collected at 0, 1, 2 and 4 hours after P-407 and processed to measure plasma TG accumulation (n=4–5 per group). (b) Postprandial TG clearance after an oral gavage of olive oil (~300 ml). TG levels were measured 0, 1, 2, and 4 hours after gavage (n=4–5 per group); (c) VLDL turnover after intravenous injection of[C14]triolein labeled human VLDL over time in CETP-ApoB100 mice treated with vehicle, canagliflozin, control ASO or SGLT2 ASO (n=4–5 per group). (d) Total TG lipase, HL and LpL activities in post heparin plasma from the four mice groups using a radiolabeled triglyceride substrate as described in Methods (n=4–5 per group). Values represent mean ± SD. * denotes p<0.05 and ** denotes p<0.01 when Student’s t-test was used for analysis; # denotes p<0.05, ## denotes p<0.01 and ### denotes p<0.001 when ANOVA was used for analysis. All comparisons are with control vehicle group.

To assess if faster TG clearance correlated with increased TG lipase activity, we measured total, HL, and LpL activities in the post-heparin plasma from these mice. SGLT2 ASO, canagliflozin, and insulin treatments increased total TG lipase activity, as well as LpL activity in the post-heparin plasma. These increases were greater with SGLT2 ASO than with canaglifozin.

SGLT2 inhibition downregulates angiopoietin-like protein 4 (ANGPTL4) expression in adipose, skeletal muscle, and heart

To determine the reason for the changes in LpL activity, we examined gene expression of LpL and its known inhibitor ANGPTL4 in white and brown adipose, skeletal muscle, and heart, the tissues with highest LpL expression22–24. LpL mRNA expression was unchanged in the three treatment groups in all these tissues (Figure 4a-d). Canagliflozin, as well as insulin, reduced ANGPTL4 expression in white adipose tissue, whereas SGLT2 ASO reduced its expression in brown adipose, skeletal muscle and heart similar to insulin (Figure 4a-d). Hepatic mRNA expression of HL and ANGPTL3, another posttranslational inhibitor of LpL, was not altered in any of the treatment groups (Figure 4e).

Figure 4. ANGPTL4 and LpL expression in adipose, skeletal muscle and heart.

Figure 4.

Open in a new tab

Comparison of mRNA of LpL and ANGPTL4 using real-time PCR in (a) white adipose, (b) brown adipose (c) skeletal muscle and (d) hearts of the four mice groups after 30 days of treatment (n=4–5 per group). (e) mRNA levels of HL and ANGPTL3 in the livers of the four mice groups after 30 days of treatment (n=4–5 per group). Values represent mean ± SD. * denotes p<0.05 and ** denotes p<0.01 when Student’s t-test was used for analysis; # denotes p<0.05 and ### denotes p<0.001 when ANOVA was used for analysis. All comparisons are with control vehicle group.

SGLT2 inhibition decreases LDL turnover

The increased LDL-c in the mice might be due to greater LpL-mediated conversion of VLDL to LDL. However, it is also possible that our treatments altered clearance of LDL from the circulation. To test this, we assessed LDL turnover. SGLT2 ASO treated mice had significantly delayed LDL turnover compared to the other groups (Figure 5a). Since hepatic LDLR is the major receptor for clearance of plasma LDL, we measured the gene expression of LDLR and its post-transcriptional modulator PCSK9 in the liver. SGLT2 ASO treatment lowered mRNA levels of both genes (Figure 5b). Total hepatic LDLR protein levels decreased modestly in SGLT2 ASO treated mice (Figure 5c, d).

Figure 5. LDL turnover and hepatic LDLR expression.

Figure 5.

Open in a new tab

(a)LDL clearance from plasma over time after intravenous injection of[H3]cholesteryl oleate labeled LDL in CETP-ApoB100 mice treated with vehicle, canagliflozin, control ASO or SGLT2 ASO (n=4–5 per group). (b) Liver LDLR and PCSK9 mRNA levels in mice after 30 days of treatment with vehicle, canagliflozin, SGLT2 ASO or insulin (n=4–5 per group). (c) Immunoblot for LDLR in the livers of CETP-ApoB100 mice treated with vehicle, canagliflozin, SGLT2 ASO or insulin (n=3–4 per group). (d) Quantification for the western blot. Values represent mean ± SD. * denotes P<0.05 using unpaired Student’s t-test and # denotes p<0.05, ### denotes p<0.001when ANOVA is used for analysis. All comparisons are with control vehicle group.

SGLT2 inhibition increases cardiac NEFA levels

Since SGLT2 inhibition treatments are also associated with reduced heart failure, we analyzed cardiac lipid levels and major genes involved in lipid metabolism. Cardiac NEFAs were increased with SGLT2 inhibition (Figure 6a). Heart TG content increased with canagliflozin only (Figure 6b), while TC levels in the heart were not altered by any of the treatments (Figure 6c). We assessed the expression of peroxisomal proliferator-activated receptors (PPARs). Canagliflozin treatment lowered PPARα gene expression; however, SGLT2 ASO did not alter mRNA expression of cardiac PPAR genes (Figure 6d). In the heart, PPARδ is the major regulator of ANGPTL4 expression22, 25, 26, however, none of the treatments altered PPARδ mRNA levels. Gene expression of FoxO1, another ANGPTL4 regulator27, was not altered with SGLT2 inhibition as well (Figure 6d). CPT1b, which is involved in fatty acid oxidation, was downregulated along with ACSL1, VLDLR and CD36 in insulin and ASO-treated mouse hearts (Figure 6e). Gene expression of BNP, a heart failure marker was lower in SGLT2 ASO treated mice hearts whereas expression of ANP, another marker of heart failure did not differ in any of the groups (SI Figure II).

DISCUSSION

SGLT2 inhibitors are rapidly becoming a preferred drug for patients with T2DM as they reduce glucose, cause weight loss, and offer protection from cardiovascular complications, especially heart failure, as recently demonstrated in the EMPA-REG and CANVAS trials7, 8. This cardiovascular benefit occurs despite an increase in LDL-c found in several clinical studies11, 12. However, like some other TG-reducing therapies, the changes in LDL-c occur in the setting of other beneficial changes in plasma lipoprotein metabolism.

For our study, we chose a mouse model of combined insulin resistance and insulin deficiency seen in patients with T2DM. Mice were first made obese on a HFD and then insulin deficient using STZ, albeit at lower doses than used in chow-fed animals, leading to hyperglycemia in concert with insulin resistance. We used mice that contained two transgenes, human ApoB100 to increase circulating VLDL and LDL, and human CETP to allow exchange of lipids between HDL and ApoB-containing lipoproteins. In contrast to the several previous studies where SGLT2 inhibition by drugs or ASO did not increase plasma LDL-c13, 28–33, SGLT2 inhibition using our novel diabetic animal model reproduced many of the lipid alterations found in humans treated with SGLT2 inhibitors, namely increased LDL-c and reduced TG. Similar changes in TG were found with insulin treatment, which is not surprising as others have shown that one effect of SGLT2 inhibition treatment is improved insulin actions20.

Unlike the weight loss seen in patients treated with SGLT2 inhibitors, we did not observe such weight reductions in our model with either ASO or canagliflozin. The general weight loss seen in all four groups of mice treated with STZ mice may have preempted any further effect of SGLT2 inhibition on body weight.

In this study, we were able to examine the reasons for changes in lipoprotein physiology that accompany SGLT2 inhibition. SGLT2 ASO and canagliflozin treatment in mice also increased total post-heparin plasma lipolytic activity and LpL activity. With greater LpL activity, TG was reduced and postprandial lipemia blunted in SGLT2 ASO-treated mice. Likely due to the less robust induction of LpL, these changes were not evident in canagliflozin-treated mice. Both postprandial TG and VLDL-TG turnover were increased in ASO-treated mice, consistent with the increased LpL activity.

The increased LpL activity was not accompanied by a corresponding change in LpL mRNA, suggesting that a posttranscriptional mechanism was operative. Specifically, we found that gene expression of ANGPTL4, which inactivates LpL activity by increasing dissociation of LpL dimers24, was markedly reduced in heart, skeletal muscle, and brown adipose tissue with SGLT2 ASO treatment. ANGPTL4 is mainly regulated by PPARs25, 26, 34, 35, which are in turn activated by fatty acids. Canagliflozin did not alter plasma NEFA levels but SGLT2 ASO lowered them. This could explain reduced ANGPTL4 expression in metabolic tissues with ASO treatment. Future studies are warranted to define the detailed mechanisms.

To our surprise, the mice treated with SGLT2 ASO had more pronounced lipid changes than mice treated with canagliflozin despite similar changes in fasting glucose. This could be due to the differences in the duration of pharmacodynamic effects of SGLT2 ASO versus canagliflozin. The estimated half-life of ASO in the kidney is expected to be at least 2–4 weeks. On the other hand, it is possible that canagliflozin’s shorter half-life, effects of the daily oral gavage (e.g., stress), or a reduced or less prolonged effect on non-renal SGLT2 transporter mitigated some of the effects compared to the ASO. Hence, we have focused on the effects of SGLT2 ASO in this report.

The increase in plasma LDL-c found in our study could be due increased VLDL to LDL conversion as well as reduced LDL clearance. Because of the rapid clearance of VLDL in the mouse, we were unable to track conversion, but did find a reduced LDL-c turnover. LDLR mRNA was reduced in the ASO-treated mice; LDLR protein was modestly reduced. The only other animal study showing elevated plasma LDL-c with SGLT2 inhibition was performed in non-diabetic HFD-fed hamsters33. The hamsters treated with empagliflozin had increased hepatic cholesterol synthesis as well as reduced LDL clearance associated with reduced LDLR levels. However, the animals were not diabetic, and changes in plasma lipids were seen only after overnight fasting, which is different than our study, where we saw lipid changes after a more physiologic 4-hour fast. Changes in circulating TG were not reported for the hamsters.

Other treatments that increase LpL activity or reduce TG, such as fibric acids and omega 3 fatty acids, also increase LDL-c levels36, 37. In some situations, these LDL-c increases accompany reduced TG levels and likely reduce the CETP-mediated transfer of LDL-c to VLDL. Reduced VLDL-TG and increased LpL activity also likely explains the HDL increase seen in humans treated with SGLT2 inhibitors. Increased LpL reduces postprandial lipemia, a well-established risk factor for CVD38, 39. Further support of the role of lipolysis in CVD risk has come from GWAS and large scale exome wide screening studies that have linked loci associating with TG levels and CVD. Changes in LpL correlate with CVD risk40.

While the reason for the marked reduction in heart failure seen in the EMPA-REG and CANVAS trials are unclear, hypotheses include an effect due to reduced blood pressure, vascular dilation, increased use of ketone bodies as a cardiac fuel source41, and reduced substrate overload due to lower circulating glucose without greater insulin-driven nutrient uptake42. One characteristic of many forms of heart failure is a deficiency in energy production43. The major substrate for heart ATP production is fatty acid and heart failure is usually associated with reduced fatty acid oxidation44, 45. While it has been argued that reduced fatty acid oxidation could be beneficial as oxidation of glucose requires less oxygen than that of fatty acids46, it is also possible that this metabolic switch exacerbates energy deficiency. If true, then the reduction of ANGPTL4 found in our studies would likely lead to increased LpL actions in the heart, and thus greater fatty acid uptake. This could explain the increase in cardiac fatty acid levels seen in our studies with SGLT2 inhibition. Others have shown that ANGPTL4 overexpression in cardiomyocytes leads to reduced LpL activity and small hearts that are prone to failure23; cardiac LpL deficiency shows a similar increase in afterload induced heart failure47. We postulate that reduced heart expression of ANGPTL4 improves heart energy production, which in turn will improve heart function.

In summary, we have defined the cause of increased LDL-c with the use of SGLT2 inhibition therapies. Our study demonstrated that LDL-c increases were associated with delayed clearance of LDL from the circulation along with increased plasma LpL activity, reduced postprandial lipemia, and lower circulating TG levels in mice treated with the SGLT2 ASO. The major molecular change was a marked reduction in ANGPLT4 mRNA in heart, adipose, and skeletal muscle allowing greater LpL activity. Genetic studies have associated reduced ANGPTL4 with reduced CVD48, 49, suggesting that the clinical benefits of SGLT2 inhibitors on CVD might be via a similar mechanism.

Supplementary Material

Graphic Abstract
SuppleMental Material

HIGHLIGHTS.

  • SGLT2 inhibition increases LDL-c in diabetic mice expressing human CETP and human apoB100

  • SGLT2 inhibition is associated with increased lipolysis and reduced LDL turnover

  • SGLT2 Inhibitor and insulin therapy reduce expression of ANGPTL4

ACKNOWLEDGEMENTS

(a) Acknowledgements: D.B. participated in concept, design of the study, carried out the experiments, and data analysis. L.A.H. performed oral gavage in mice and the lipoprotein fractionations. D.S. performed lipid extraction. J.O. assisted with LpL assay. A.E.M. supplied the SGLT2 ASO and assisted with its experimental use. R.H.E. and P.L.R. were involved in concept, experimental design and data analysis. I.J.G and N.A.D. conceived of the study, participated in its design, execution and data analysis, and drafted the manuscript.

We thank Dr. Patricia Freitas and Ms. Allison Mogul for technical assistance. We thank Ms. Stephanie Chiang for editing and proof reading the manuscript.

(b) Sources of Funding: Janssen Research & Development, LLC, Raritan, NJ to I.J.G.

National Institute of Diabetes and Digestive and Kidney Diseases (DK095684) to I.J.G

National Heart, Lung, and Blood Institute (P01 HL092969) to I.J.G.

(c) Disclosures: N.A.D. and P.L.R. are employees of Janssen R&D. A.E.M. is an employee of Ionis Pharmaceuticals.

ABBREVIATIONS

ASO

Antisense oligonucleotides

HL

Hepatic lipase

HFD

High fat diet

HPRT

Hypoxanthine-guanine phosphoribosyltransferase

LpL

Lipoprotein lipase

NEFA

Non-esterified fatty acids

P-407

Poloxamer 407

RPS3

Ribosomal protein S3

STZ

Streptozotocin

REFERENCES

  • 1.Wright EM, Loo DD, Hirayama BA. Biology of human sodium glucose transporters. Physiol Rev. 2011;91:733–794 [DOI] [PubMed] [Google Scholar]
  • 2.Rahmoune H, Thompson PW, Ward JM, Smith CD, Hong G, Brown J. Glucose transporters in human renal proximal tubular cells isolated from the urine of patients with non-insulin-dependent diabetes. Diabetes. 2005;54:3427–3434 [DOI] [PubMed] [Google Scholar]
  • 3.Abdul-Ghani MA, DeFronzo RA. Inhibition of renal glucose reabsorption: A novel strategy for achieving glucose control in type 2 diabetes mellitus. Endocr Pract. 2008;14:782–790 [DOI] [PubMed] [Google Scholar]
  • 4.Farber SJ, Berger EY, Earle DP. Effect of diabetes and insulin of the maximum capacity of the renal tubules to reabsorb glucose. J Clin Invest. 1951;30:125–129 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Ehrenkranz JR, Lewis NG, Kahn CR, Roth J. Phlorizin: A review. Diabetes Metab Res Rev. 2005;21:31–38 [DOI] [PubMed] [Google Scholar]
  • 6.Rossetti L, Smith D, Shulman GI, Papachristou D, DeFronzo RA. Correction of hyperglycemia with phlorizin normalizes tissue sensitivity to insulin in diabetic rats. J Clin Invest. 1987;79:1510–1515 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Zinman B, Wanner C, Lachin JM, Fitchett D, Bluhmki E, Hantel S, Mattheus M, Devins T, Johansen OE, Woerle HJ, Broedl UC, Inzucchi SE, Investigators E-RO. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. N Engl J Med. 2015;373:2117–2128 [DOI] [PubMed] [Google Scholar]
  • 8.Neal B, Perkovic V, Matthews DR. Canagliflozin and cardiovascular and renal events in type 2 diabetes. N Engl J Med. 2017;377:2099. [DOI] [PubMed] [Google Scholar]
  • 9.Agrawal N, Freitas Corradi P, Gumaste N, Goldberg IJ. Triglyceride treatment in the age of cholesterol reduction. Prog Cardiovasc Dis. 2016;59:107–118 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Weintraub H Update on marine omega-3 fatty acids: Management of dyslipidemia and current omega-3 treatment options. Atherosclerosis. 2013;230:381–389 [DOI] [PubMed] [Google Scholar]
  • 11.Ptaszynska A, Hardy E, Johnsson E, Parikh S, List J. Effects of dapagliflozin on cardiovascular risk factors. Postgrad Med. 2013;125:181–189 [DOI] [PubMed] [Google Scholar]
  • 12.Rodriguez-Gutierrez R, Gonzalez-Saldivar G. Canagliflozin. Cleve Clin J Med. 2014;81:87–88 [DOI] [PubMed] [Google Scholar]
  • 13.Yu T, Sungelo MJ, Goldberg IJ, Wang H, Eckel RH. Streptozotocin-treated high fat fed mice: A new type 2 diabetes model used to study canagliflozin-induced alterations in lipids and lipoproteins. Horm Metab Res. 2017 [DOI] [PubMed] [Google Scholar]
  • 14.Willecke F, Scerbo D, Nagareddy P, Obunike JC, Barrett TJ, Abdillahi ML, Trent CM, Huggins LA, Fisher EA, Drosatos K, Goldberg IJ. Lipolysis, and not hepatic lipogenesis, is the primary modulator of triglyceride levels in streptozotocin-induced diabetic mice. Arterioscler Thromb Vasc Biol. 2015;35:102–110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Renard CB, Kramer F, Johansson F, Lamharzi N, Tannock LR, von Herrath MG, Chait A, Bornfeldt KE. Diabetes and diabetes-associated lipid abnormalities have distinct effects on initiation and progression of atherosclerotic lesions. J Clin Invest. 2004;114:659–668 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Folch J, Lees M, Sloane Stanley GH. A simple method for the isolation and purification of total lipides from animal tissues. J Biol Chem. 1957;226:497–509 [PubMed] [Google Scholar]
  • 17.Millar JS, Cromley DA, McCoy MG, Rader DJ, Billheimer JT. Determining hepatic triglyceride production in mice: Comparison of poloxamer 407 with triton wr-1339. J Lipid Res. 2005;46:2023–2028 [DOI] [PubMed] [Google Scholar]
  • 18.Hocquette JF, Graulet B, Olivecrona T. Lipoprotein lipase activity and mrna levels in bovine tissues. Comp Biochem Physiol B Biochem Mol Biol. 1998;121:201–212 [DOI] [PubMed] [Google Scholar]
  • 19.Polidori D, Sha S, Mudaliar S, Ciaraldi TP, Ghosh A, Vaccaro N, Farrell K, Rothenberg P, Henry RR. Canagliflozin lowers postprandial glucose and insulin by delaying intestinal glucose absorption in addition to increasing urinary glucose excretion: Results of a randomized, placebo-controlled study. Diabetes Care. 2013;36:2154–2161 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Merovci A, Solis-Herrera C, Daniele G, Eldor R, Fiorentino TV, Tripathy D, Xiong J, Perez Z, Norton L, Abdul-Ghani MA, DeFronzo RA. Dapagliflozin improves muscle insulin sensitivity but enhances endogenous glucose production. J Clin Invest. 2014;124:509–514 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Sadur CN, Eckel RH. Insulin stimulation of adipose tissue lipoprotein lipase. Use of the euglycemic clamp technique. J Clin Invest. 1982;69:1119–1125 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Robciuc MR, Skrobuk P, Anisimov A, Olkkonen VM, Alitalo K, Eckel RH, Koistinen HA, Jauhiainen M, Ehnholm C. Angiopoietin-like 4 mediates ppar delta effect on lipoprotein lipase-dependent fatty acid uptake but not on beta-oxidation in myotubes. PLoS One. 2012;7:e46212. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Yu X, Burgess SC, Ge H, Wong KK, Nassem RH, Garry DJ, Sherry AD, Malloy CR, Berger JP, Li C. Inhibition of cardiac lipoprotein utilization by transgenic overexpression of angptl4 in the heart. Proc Natl Acad Sci U S A. 2005;102:1767–1772 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Sukonina V, Lookene A, Olivecrona T, Olivecrona G. Angiopoietin-like protein 4 converts lipoprotein lipase to inactive monomers and modulates lipase activity in adipose tissue. Proc Natl Acad Sci U S A. 2006;103:17450–17455 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Georgiadi A, Lichtenstein L, Degenhardt T, Boekschoten MV, van Bilsen M, Desvergne B, Muller M, Kersten S. Induction of cardiac angptl4 by dietary fatty acids is mediated by peroxisome proliferator-activated receptor beta/delta and protects against fatty acid-induced oxidative stress. Circ Res. 2010;106:1712–1721 [DOI] [PubMed] [Google Scholar]
  • 26.Staiger H, Haas C, Machann J, Werner R, Weisser M, Schick F, Machicao F, Stefan N, Fritsche A, Haring HU. Muscle-derived angiopoietin-like protein 4 is induced by fatty acids via peroxisome proliferator-activated receptor (ppar)-delta and is of metabolic relevance in humans. Diabetes. 2009;58:579–589 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Kuo TY, Chen TC, Yan S, Foo F, Ching C, McQueen A, Wang JC. Repression of glucocorticoid-stimulated angiopoietin-like 4 gene transcription by insulin. Journal of Lipid Research. 2014;55:919–928 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Nagareddy PR, Murphy AJ, Stirzaker RA, Hu Y, Yu S, Miller RG, Ramkhelawon B, Distel E, Westerterp M, Huang LS, Schmidt AM, Orchard TJ, Fisher EA, Tall AR, Goldberg IJ. Hyperglycemia promotes myelopoiesis and impairs the resolution of atherosclerosis. Cell Metab. 2013;17:695–708 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Honda Y, Imajo K, Kato T, Kessoku T, Ogawa Y, Tomeno W, Kato S, Mawatari H, Fujita K, Yoneda M, Saito S, Nakajima A. The selective sglt2 inhibitor ipragliflozin has a therapeutic effect on nonalcoholic steatohepatitis in mice. PLoS One. 2016;11:e0146337. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Jojima T, Tomotsune T, Iijima T, Akimoto K, Suzuki K, Aso Y. Empagliflozin (an sglt2 inhibitor), alone or in combination with linagliptin (a dpp-4 inhibitor), prevents steatohepatitis in a novel mouse model of non-alcoholic steatohepatitis and diabetes. Diabetol Metab Syndr. 2016;8:45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Terasaki M, Hiromura M, Mori Y, Kohashi K, Nagashima M, Kushima H, Watanabe T, Hirano T. Amelioration of hyperglycemia with a sodium-glucose cotransporter 2 inhibitor prevents macrophage-driven atherosclerosis through macrophage foam cell formation suppression in type 1 and type 2 diabetic mice. PLoS One. 2015;10:e0143396. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Han JH, Oh TJ, Lee G, Maeng HJ, Lee DH, Kim KM, Choi SH, Jang HC, Lee HS, Park KS, Kim YB, Lim S. The beneficial effects of empagliflozin, an sglt2 inhibitor, on atherosclerosis in apoe −/− mice fed a western diet. Diabetologia. 2017;60:364–376 [DOI] [PubMed] [Google Scholar]
  • 33.Briand F, Mayoux E, Brousseau E, Burr N, Urbain I, Costard C, Mark M, Sulpice T. Empagliflozin, via switching metabolism toward lipid utilization, moderately increases ldl cholesterol levels through reduced ldl catabolism. Diabetes. 2016;65:2032–2038 [DOI] [PubMed] [Google Scholar]
  • 34.Mandard S, Zandbergen F, Tan NS, Escher P, Patsouris D, Koenig W, Kleemann R, Bakker A, Veenman F, Wahli W, Muller M, Kersten S. The direct peroxisome proliferator-activated receptor target fasting-induced adipose factor (fiaf/pgar/angptl4) is present in blood plasma as a truncated protein that is increased by fenofibrate treatment. J Biol Chem. 2004;279:34411–34420 [DOI] [PubMed] [Google Scholar]
  • 35.Kersten S, Mandard S, Tan NS, Escher P, Metzger D, Chambon P, Gonzalez FJ, Desvergne B, Wahli W. Characterization of the fasting-induced adipose factor fiaf, a novel peroxisome proliferator-activated receptor target gene. J Biol Chem. 2000;275:28488–28493 [DOI] [PubMed] [Google Scholar]
  • 36.Vakkilainen J, Steiner G, Ansquer JC, Perttunen-Nio H, Taskinen MR. Fenofibrate lowers plasma triglycerides and increases ldl particle diameter in subjects with type 2 diabetes. Diabetes Care. 2002;25:627–628 [DOI] [PubMed] [Google Scholar]
  • 37.Harris WS. N-3 fatty acids and serum lipoproteins: Human studies. Am J Clin Nutr. 1997;65:1645S–1654S [DOI] [PubMed] [Google Scholar]
  • 38.Freiberg JJ, Tybjaerg-Hansen A, Jensen JS, Nordestgaard BG. Nonfasting triglycerides and risk of ischemic stroke in the general population. JAMA. 2008;300:2142–2152 [DOI] [PubMed] [Google Scholar]
  • 39.Nordestgaard BG, Benn M, Schnohr P, Tybjaerg-Hansen A. Nonfasting triglycerides and risk of myocardial infarction, ischemic heart disease, and death in men and women. JAMA. 2007;298:299–308 [DOI] [PubMed] [Google Scholar]
  • 40.Do R, Willer CJ, Schmidt EM, Sengupta S, Gao C, Peloso GM, Gustafsson S, Kanoni S, Ganna A, Chen J, Buchkovich ML, Mora S, Beckmann JS, Bragg-Gresham JL, Chang HY, et al. Common variants associated with plasma triglycerides and risk for coronary artery disease. Nat Genet. 2013;45:1345–1352 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Ferrannini E, Mark M, Mayoux E. Cv protection in the empa-reg outcome trial: A “thrifty substrate” hypothesis. Diabetes Care. 2016;39:1108–1114 [DOI] [PubMed] [Google Scholar]
  • 42.Abdul-Ghani M, Del Prato S, Chilton R, DeFronzo RA. Sglt2 inhibitors and cardiovascular risk: Lessons learned from the empa-reg outcome study. Diabetes Care. 2016;39:717–725 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Neubauer S The failing heart-an engine out of fuel. N Engl J Med. 2007;356:1140–1151 [DOI] [PubMed] [Google Scholar]
  • 44.Choi YS, de Mattos AB, Shao D, Li T, Nabben M, Kim M, Wang W, Tian R, Kolwicz SC Jr., Preservation of myocardial fatty acid oxidation prevents diastolic dysfunction in mice subjected to angiotensin ii infusion. J Mol Cell Cardiol. 2016;100:64–71 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45.Drosatos K, Pollak NM, Pol CJ, Ntziachristos P, Willecke F, Valenti MC, Trent CM, Hu Y, Guo S, Aifantis I, Goldberg IJ. Cardiac myocyte klf5 regulates ppara expression and cardiac function. Circ Res. 2016;118:241–253 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Lopaschuk GD, Ussher JR, Folmes CD, Jaswal JS, Stanley WC. Myocardial fatty acid metabolism in health and disease. Physiol Rev. 2010;90:207–258 [DOI] [PubMed] [Google Scholar]
  • 47.Khan RS, Lin Y, Hu Y, Son NH, Bharadwaj KG, Palacios C, Chokshi A, Ji R, Yu S, Homma S, Schulze PC, Tian R, Goldberg IJ. Rescue of heart lipoprotein lipase-knockout mice confirms a role for triglyceride in optimal heart metabolism and function. Am J Physiol Endocrinol Metab. 2013;305:E1339–1347 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Dewey FE, Gusarova V, O’Dushlaine C, Gottesman O, Trejos J, Hunt C, Van Hout CV, Habegger L, Buckler D, Lai KM, Leader JB, Murray MF, Ritchie MD, Kirchner HL, Ledbetter DH, et al. Inactivating variants in angptl4 and risk of coronary artery disease. N Engl J Med. 2016;374:1123–1133 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Myocardial Infarction G, Investigators CAEC, Stitziel NO, Stirrups KE, Masca NG, Erdmann J, Ferrario PG, Konig IR, Weeke PE, Webb TR, Auer PL, Schick UM, Lu Y, Zhang H, Dube MP, et al. Coding variation in angptl4, lpl, and svep1 and the risk of coronary disease. N Engl J Med. 2016;374:1134–1144 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Graphic Abstract
NIHMS1500152-supplement-Graphic_Abstract.pdf (188.6KB, pdf)
SuppleMental Material
NIHMS1500152-supplement-SuppleMental_Material.pdf (91KB, pdf)

RESOURCES