Antisense-Mediated Lowering of Plasma Apolipoprotein C-III By

1408 Diabetes Care Volume 39, August 2016

Andres Digenio,1 Richard L. Dunbar,2 Antisense-Mediated Lowering Veronica J. Alexander,3 Marcus Hompesch,4 Linda Morrow,4 of Plasma Apolipoprotein C-III Richard G. Lee,3 Mark J. Graham,3 Steven G. Hughes,3 Rosie Yu,3 by Volanesorsen Improves Walter Singleton,3 Brenda F. Baker,3 Dyslipidemia and Insulin Sanjay Bhanot,3 and Rosanne M. Crooke3 Sensitivity in Type 2 Diabetes Diabetes Care 2016;39:1408–1415 | DOI: 10.2337/dc16-0126

OBJECTIVE To determine the effects of volanesorsen (ISIS 304801), a second-generation 2’-O-methoxyethyl chimeric antisense inhibitor of apolipoprotein (apo)C-III, on triglyceride (TG) levels and insulin resistance in patients with type 2 diabetes.

RESEARCH DESIGN AND METHODS A randomized, double-blind, placebo-controlled trial was performed in 15 adult

patients with type 2 diabetes (HbA1c >7.5% [58 mmol/mol]) and hypertriglyceridemia (TG >200 and <500 mg/dL). Patients were randomized 2:1 to receive volanesorsen 300 mg or placebo for a total of 15 subcutaneous weekly doses. Glucose handling and insulin sensitivity were measured before and after treatment using a two-step hyperinsulinemic-euglycemic clamp procedure.

RESULTS Treatment with volanesorsen signiﬁcantly reduced plasma apoC-III (288%, P =

EMERGING TECHNOLOGIES AND THERAPEUTICS 0.02) and TG (269%, P = 0.02) levels and raised HDL cholesterol (HDL-C) (42%, P = 0.03) compared with placebo. These changes were accompanied by a 57% improvement in whole-body insulin sensitivity (P < 0.001). Importantly, we found a strong relationship between enhanced insulin sensitivity and both plasma apoC-III 1Akcea Therapeutics, Cambridge, MA (r = 20.61, P =0.03)andTG(r = 20.68, P = 0.01) suppression. Improved insulin 2University of Pennsylvania, Philadelphia, PA fi fi 2 P 3Ionis Pharmaceuticals, Inc., Carlsbad, CA sensitivity was suf cient to signi cantly lower glycated albumin ( 1.7%, = 4 fi 2 m P Pro l Institute for Clinical Research, San Diego, 0.034) and fructosamine ( 38.7 mol/L, = 0.045) at the end of dosing and HbA1c CA 2 2 P ( 0.44% [ 4.9 mmol/mol], = 0.025) 3 months postdosing. Corresponding author: Andres Digenio, adigenio@ akceatx.com. CONCLUSIONS Received 20 January 2016 and accepted 9 May Volanesorsen reduced plasma apoC-III and TG while raising HDL-C levels. Impor- 2016. tantly, glucose disposal, insulin sensitivity, and integrative markers of diabetes A.D. and R.L.D. are co–primary authors. also improved in these patients after short-term treatment. Clinical trial reg. no. NCT01647308, clinicaltrials .gov. Insulin resistance states usually seen in the context of obesity and type 2 diabetes This article contains Supplementary Data online are commonly associated with a metabolic dyslipidemia that amplifies cardiovas- at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc16-0126/-/DC1. cular disease risk. Among patients with type 2 diabetes, insulin resistance impairs © 2016 by the American Diabetes Association. the ability to utilize glucose as fuel, prompting a switch toward fat storage that Readers may use this article as long as the work is promotes free fatty acid flux, hepatic triglyceride (TG) synthesis, and secretion of properly cited, the use is educational and not for large VLDL particles (1). Cholesteryl ester transfer protein-mediated exchange of profit, and the work is not altered. care.diabetesjournals.org Digenio and Associates 1409

VLDL-TG for cholesteryl esters in LDL more recent studies, it does not appear euglycemic (HE) clamp trial was conducted and HDL particles promotes atherogenic that apoC-III is directly associated with at a single center in the U.S. (ProfilIn- small-dense LDL and dysfunctional small- the incidence of nonalcoholic fatty liver stitute for Clinical Research, San Diego, dense HDL particles, which are more disease in broader patient populations CA) between 8 August 2012 and 31 Jan- readily catabolized by the kidney. The re- (14). Two recent Mendelian randomiza- uary 2014 to evaluate the effects of sulting metabolic dyslipidemia features tion studies showed that individuals with volanesorsen treatment in adult pa- a triad of high levels of VLDL-TG, low loss-of-function APOC3 variants had tients with type 2 diabetes and hyper- HDL cholesterol (HDL-C), and an in- lower TG concentrations and LDL-C, triglyceridemia. The primary objective creased prevalence of small-dense LDL higher HDL-C, and, compellingly, a 40% of this trial was to determine the effects particles despite normal or even low LDL reduction in risk of coronary heart dis- of volanesorsen versus placebo on fast- cholesterol (LDL-C) levels. The cumulative ease (15,16). While it is unclear whether ing total apoC-III. The secondary objec- result is a significant increase in the risk of the coronary heart disease risk reduction tives were to determine its effects on atherosclerosis (2). is due to suppressed TG, other TRL- other lipid parameters, glycemic con- Recommended treatments include associated factors, or apoC-III per se, trol, and insulin sensitivity. Written in- diet and exercise as well as statins that the aforementioned cardioprotective formed consent was obtained from all have the strongest evidence for im- properties suggest that apoC-III is an participants prior to participation in the provement of this condition (3–5). Be- attractive therapeutic target (17). study. The protocol was approved by cause the TG-reducing effect of statins is Potent second-generation antisense the institutional review board and con- modest at best, patient-centered guide- oligonucleotides (ASOs) have been de- ducted in compliance with the stan- lines also recommend a second drug veloped to reduce apoC-III expression dards of Good Clinical Practice and such as fibric acid derivatives, niacin, in animals and humans (18). In mice, guidelines of the 2002 Declaration of or n-3 fatty acids (3). In support of this administration of a species-specific Helsinki. approach, meta-analyses of five large tri- apoC-III ASO led to dose-dependent re- Eligible patients were 18–65 years als assessing the impact of fibrates on car- ductions in hepatic apoC-III mRNA and old; had hypertriglyceridemia (TG diovascular end points revealed benefits plasma apoC-III protein. In dyslipidemic .200 mg/dL and ,500 mg/dL) and among patients with the lipid triad (6). rodent models, this led to .50% reduc- type 2 diabetes that had been diag- While it is well established that insulin tions in both fasting and postprandial nosed for at least 6 months; had uncon- resistance drives dyslipidemia, there is plasma TG. Recapitulating preclinical trolled hyperglycemia (HbA1c .7.5% growing evidence that the elevation in models, treating normolipidemic and [58 mmol/mol]); and were on a stable plasma TGs and reduction in HDL-C ac- hypertriglyceridemic human subjects dose of metformin $1,000 mg/day. Pa- companying insulin resistance states with the human-specificapoC-IIIASO, tients with uncontrolled high blood may exacerbate the insulin-resistant volanesorsen (ISIS 304801), profoundly pressure, heart failure New York Heart phenotype, albeit by yet-to-be-defined reduced plasma apoC-III protein and Association III and IV, recent episode of mechanisms (7). In preclinical models, plasma TG levels (18,19). This included angina, prior acute myocardial infarction, HDL may improve insulin sensitivity, in- subjects with familial chylomicronemia recent coronary bypass surgery, severe di- sulin secretion through the b-cells, and syndrome (FCS), who have extreme abetes microvascular complications, kid- b-cell survival (8). Clinically, elevations hypertriglyceridemia due to loss-of- ney or liver disease, hypothyroidism, or in TG-rich lipoprotein (TRL) levels have function mutations leading to absent low platelet count were excluded from been shown to exacerbate the pathol- LPL activity (20). This latter study dem- participation in the study. Acceptable con- ogy of insulin resistance. Therefore, re- onstrates that apoC-III also inhibits an traceptive methods were required during ducing TG and increasing HDL may not LPL-independent pathway of TRL dosing and in the follow-up period. Lipid only reduce cardiovascular risk but also clearance. and blood pressure medications were al- improve insulin sensitivity and delay Given evidence that hypertriglyceri- lowed provided that they were stable b-cell failure (1,9). demia may worsen insulin sensitiv- before dosing. Oral antidiabetes medi- Apolipoprotein (apo)C-III has been ity and the data suggesting a role for cation other than metformin was pro- identified as a key modulator of plasma apoC-III in mediating insulin resistance, hibited. Men had to restrict alcohol TG concentrations in animal models and we reasoned that an intervention that intake to 2 drinks and women to 1 drink in humans. ApoC-III affects TG metabo- robustly suppressed apoC-III and plasma daily. All patients received glucose- lism by inhibiting lipoprotein lipase TG would lead to improved insulin sen- monitoring instructions and dietary ed- (LPL) activity and by interfering with sitivity. Thus, we conducted a clinical ucation based on the American Diabetes receptor-mediated uptake of TRLs by trial that allowed us to test the mecha- Association guidelines. the liver (10,11). Mice overexpressing nistic hypothesis that aggressively lower- The trial consisted of three periods apoC-III not only have high plasma TG ing TG by apoC-III ASO treatment would (Supplementary Fig. 1): screening, dos- but also display increased insulin resis- improve insulin sensitivity in patients ing (13 weeks, days 1–85), and post- tance (12). In East Indian Asians, an with diabetes with high TGs. treatment/follow-up (13 weeks, days APOC3 genotype associated with in- 86–176). Eligible patients were random- creased plasma TG is also associated with RESEARCH DESIGN AND METHODS ized 2:1 to receive volanesorsen 300 mg: insulin resistance and nonalcoholic fatty Clinical Trial Design and Participants placebo for a total of 15 subcutaneous liver disease, the hepatic manifestation of A randomized, double-blind, placebo- doses (days 1, 3, 5, 8, 15, 22, 29, 35, 42, the metabolic syndrome (13). However, in controlled, two-step hyperinsulinemic- 49, 56, 63, 70, 78, and 85). On the 1410 Volanesorsen Treatment in Patients With Diabetes Diabetes Care Volume 39, August 2016

mornings of days 1 and 92, patients under- for at least 2 h prior to starting the clamp. total apoC-III plasma measure and at went a two-step HE clamp procedure. At the initiation of the clamp, a low-dose least one postbaseline measure, and Volanesorsen is a second-generation insulin infusion of 30 mU/m2/min (step 1) did not have any significant protocol de- 2’-O-methoxyethyl chimeric ASO de- was started and maintained for 3 h. Dex- viations that would be expected to bias signed to sequence-specifically reduce trose 20% was variably infused to maintain the patients’ PD assessments. The pri- expression levels of the human apoC-III the plasma glucose target of 110 mg/dL. mary analysis was a comparison of the mRNA (18). After 3 h, the insulin infusion rate was percent changes from baseline to day 91 increased to 150 mU/m2/min (step 2) for in fasting total apoC-III levels in the vol- Pharmacodynamics an additional 3 h. Sampling for glucose anesorsen group and placebo group of The primary pharmacodynamic (PD) and insulin occurred every 10 min for the per-protocol population. The data variable was fasting plasma apoC-III the last 30 min of each step during the were analyzed using the Wilcoxon rank concentrations. Secondary PD variables steady-state plateau. The dose of insu- sum test. included whole-body insulin sensitivity lin at the first step was chosen to avoid Exploratory analyses were performed measurements including those that af- completely suppressing endogenous on the data collected from the HE clamp fect insulin sensitivity (nonesterified (principally hepatic) glucose produc- procedure (SIclamp, M, MCR, M/I). These fatty acids [NEFA]), integrative markers tion; the dose at the second step was mechanistic outcomes were analyzed of glucose control (glycosylated albumin, chosen to provide a robust stimulus to by mixed effects regression, with treat- fructosamine, glycosylated HbA1c), fasting glucose disposal, primarily in skeletal ment group, sequence (day 1, day 92), total TG, and other lipoproteins, including muscle. and their interaction as fixed effects and – VLDL apoC-III, total cholesterol, LDL-C, The primary measure of whole-body patient as a random intercept. Two pa- HDL-C, VLDL cholesterol (VLDL-C), apoB, insulin sensitivity was the insulin sensi- tients were excluded from the analysis: and non-HDL-C. ApoC-III and lipids were tivity index (SIclamp). This index is based one had a technical problem that invali- measured by MedPace Reference Labora- on the glucose infusion rate (GIR) per dated the day 92 clamp procedure, and tories (Cincinnati, OH). ApoC-III was deter- minute from the last 30 min of each one did not complete the day 92 pro- mined by rate nephelometry. HDL-C was step and the corresponding plasma con- cedure. Four patients received fewer determined after separation from apoB- centrations of insulin and blood glucose: doses than planned, so a fixed covariate containing lipoproteins (dextran sulfate was included in the clamp regressions to precipitation); VLDL-C and LDL-C were de- ¼ account for suboptimal dosing. For sen- termined after isolation by ultracentrifuga- SIclamp meanðGIRÞ 2 meanðGIRÞ sitivity analyses, we used the Wilcoxon tion. Cholesterol content was determined h step 2 i step 1 ð Þ 2 ð Þ 3 ð Þ rank sum test. To measure the strength mean I step 2 mean I step 1 mean BG steps 1 and 2 using a standard enzyme-based colorimet- of the linear relationship between vari- ric assay. Markers of glycemic control in- ables, we used the Pearson correlation cluding those collected during the clamp where I is insulin and BG is blood glucose. coefficient. procedure were measured by KineMed, Similarly, we used steady-state measures Inc. (Emeryville, CA). to calculate corroboratory parameters: 1) RESULTS the glucose disposal rate (M), the GIR Safety Assessments Patients Safety assessments included treatment- corrected for body weight; 2) the glucose Fifteen patients were randomized (5 3 emergent adverse events, vital signs metabolic clearance rate (MCR), 100 placebo and 10 volanesorsen) in this and weight, clinical laboratory tests (mean[M]/mean[BG]), and 3) the glucose study (Supplementary Fig. 2). The – – (LabCorp, San Diego, CA), physical ex- metabolism to insulin ratio (M/I): M mean (SD) age was 56.5 (7.5) years, amination, electrocardiogram, and use divided by the insulin concentration, I. and 73% were women (Table 1). All pa- of concomitant medications. Further details on these methods are tients were white, and all were over- available in the Supplementary Data. weight or obese (mean [SD] BMI 33.1 Two-Step HE Clamp Procedure [4.4] kg/m2). The mean (SD) baseline fast- A two-step HE clamp procedure was Statistical Analysis ing TG level was 249.3 (70.2) mg/dL, and performed at baseline (day 1) and end Sample size was based upon prior clini- HbA1c was 7.90 (0.62)% (63 mmol/mol). of treatment (day 92) using the Biosta- cal experience with volanesorsen (18) Eleven of the 15 patients received all tor (MTB Medizintechnik, Amstetten, and powered by 80% to detect a 45% 15 doses of treatment (7 active and 4 Germany). Insulin (Humulin R U100, Eli difference in apoC-III levels between placebo), 2 received 12 doses of active Lilly & Co., Indianapolis, IN) was infused volanesorsen and placebo-treated drug, 1 received 7 doses of active, and 1 using a precision pump at a rate calcu- groups at an a-level of 0.05, under the received 7 doses of placebo. Dosing was lated to acutely elevate serum insulin assumption of a 50% reduction in the discontinued by the sponsor for admin- concentration from basal (fasting) levels volanesorsen treatment group and 5% istrative reasons. to predefined plateaus (21,22). in the placebo group. On the evening prior to the clamp The safety population consisted of all Volanesorsen Improves Metabolic procedure, after the placement of intra- patients who were randomized and re- Dyslipidemia in Patients With Type 2 venous forearm catheters, an overnight ceived at least one dose of study drug. Diabetes variable intravenous regular human in- The per-protocol population included Administration of volanesorsen resulted sulin infusion was initiated to achieve a patients who received at least nine in a rapid and prolonged suppression of stable target glucose of 110 6 10 mg/dL doses of study drug, had a valid baseline fasting plasma apoC-III concentrations care.diabetesjournals.org Digenio and Associates 1411

(Fig. 1A). A similar response was observed in TG levels (Fig. 1B). All treated Table 1—Baseline characteristics and effect of 300 mg volanesorsen on lipid and patients achieved TG levels ,105 mg/dL lipoprotein levels and glycemic control by 4 weeks of dosing, and at the end of Placebo Volanesorsen P treatment the average TG level was 75.9 Baseline characteristics mg/dL (Table 1). In addition, HDL-C in- N 510 creased (Fig. 1C) and non-HDL-C, a mea- Sex, female:male 3:2 8:2 sure of all apoB-containing lipoproteins, Age, years 55.0 (10.0) 57.2 (6.4) BMI, kg/m2 32.5 (4.9) 33.4 (4.4) decreased (Fig. 1D). After 13 weeks, Fasting glucose, mg/dL 180.2 (31.3) 180.9 (29.3) compared with placebo, volanesorsen HbA , % 7.6 (0.3) 8.0 (0.7) fi 1c produced a statistically signi cant re- HbA1c, mmol/mol 60.0 (3.7) 64.3 (7.6) duction from baseline in apoC-III TGs 215.2 (48.6) 266.3 (75.2) (287.5%, P =0.019),TG(269.1%, P = Fasting lipids and lipoproteins 0.019), and VLDL–apoC-III (290.2%, N 49 P = 0.023) while raising HDL-C (42.5%, ApoC-III P = 0.034) (Table 1). Importantly, this Baseline, mg/dL 11.7 (2.3) 13.9 (4.4) Day 91, mg/dL 11.1 (3.4) 1.7 (0.6) marked decrease in TG was not associ- % change 27.3 (14.0) 287.5 (5.4) 0.019 ated with increases in LDL-C, while TGs non-HDL-C and apoB decreased non- Baseline, mg/dL 223.0 (52.3) 260.1 (77.0) significantly. There were no changes Day 91, mg/dL 202.8 (71.1) 75.9 (18.6) in plasma NEFA levels (P = 0.412). % change 29.9 (19.9) 269.1 (10.1) 0.019 VLDL-C Volanesorsen Improves Insulin Baseline, mg/dL 41.9 (7.3) 49.5 (20.1) Day 91, mg/dL 36.8 (18.0) 12.1 (4.8) Sensitivity in Proportion to TG % change 213.5 (40.8) 272.9 (13.0) 0.059 Suppression VLDL–apoC-III Volanesorsen improved whole-body in- Baseline, mg/dL 6.2 (2.4) 7.3 (3.1) sulin sensitivity compared with placebo Day 91, mg/dL 5.8 (2.5) 0.6 (0.4)* (Fig. 2 and Supplementary Table 1). % change 28.3 (20.4) 290.2 (7.3)* 0.023 Among those randomized to volane- HDL-C sorsen, SI improved 50%, compared Baseline, mg/dL 38.9 (6.6) 41.1 (7.7) clamp Day 91, mg/dL 36.5 (11.4) 57.8 (13.3) witha7%droponplacebo(adjusted % change 27.2 (16.5) 42.5 (32.2) 0.034 , difference 57%, P 0.001 mixed model, LDL-C P , 0.05 Wilcoxon rank sum test). Baseline, mg/dL 146.5 (22.3) 122.1 (39.4) Within the treated group, six out of Day 91, mg/dL 138.5 (23.2) 118.2 (31.3) eight patients (75%) displayed at least % change 25.5 (7.2) 0.0 (26.3) 0.706 a 20% improvement in insulin sensi- Non-HDL-C tivity. We tested whether changes in Baseline, mg/dL 188.4 (28.8) 171.6 (50.2) Day 91, mg/dL 175.3 (40.9) 130.3 (34.5) SIclamp depended on TG suppression % change 27.6 (13.8) 222.1 (18.5) 0.168 (Fig. 3 and Supplementary Table 2) and ApoB found that TG suppression strongly cor- Baseline, mg/dL 128.3 (17.3) 111.9 (31.6) relatedwithimprovementsinSIclamp (r = Day 91, mg/dL 115.8 (25.1) 86.6 (18.5) 20.68, P = 0.01). Improved whole-body % change 210.4 (8.5) 220.8 (15.9) 0.270 insulin sensitivity alsodependedonsup- NEFA Baseline, mg/dL 0.5 (0.1) 0.7 (0.3) pression of other components that Day 91, mg/dL 0.6 (0.2) 0.6 (0.2) fl re ect TLR levels, including apoC-III (r = % change 20.0 (28.3) 25.0 (36.8) 0.412 2 – 0.61, P =0.03),VLDLapoC-III (r = Glycemic control 20.61, P = 0.03), and VLDL-C (r = 20.66, N 49 P = 0.01). The relationship between im- Glycated albumin, % proved insulin sensitivity and total plasma Baseline 15.6 (0.5) 16.2 (1.7) apoC-III suppression is especially notewor- Day 91, change from baseline 0.7 (1.6) 21.7 (1.2) 0.034 2 thy as the primary effect of volanesorsen. Day 176, change from baseline 1.8 (1.8) 2.1 (2.4) 0.059 Fructosamine, mmol/L Moreover, at the high insulin rate, Baseline 244.3 (4.2) 273.8 (31.0) volanesorsen raised insulin-corrected Day 91, change from baseline 14.5 (33.2) 238.7 (22.5) 0.045 glucose disposal (M/I) 30% compared Day 176, change from baseline 47.8 (22.5) 211.6 (22.6) 0.019 with placebo (P = 0.02) (Supplementary HbA1c,% Table 1). Though glucose disposal (M) Baseline 7.8 (0.2) 7.9 (0.6) and MCR each rose about 25% on volane- Day 91, change from baseline 0.50 (0.62) 20.27 (0.50) 0.100 2 sorsen, comparison with placebo was not Day 176, change from baseline 0.78 (0.71) 0.44 (0.39) 0.025 statistically significant (P =0.08andP = Continued on p. 1412 0.06, respectively). 1412 Volanesorsen Treatment in Patients With Diabetes Diabetes Care Volume 39, August 2016

groups were mild in severity. There were Table 1—Continued no clinically relevant changes in serum Placebo Volanesorsen P chemistries, hematology, urinalysis, in-

HbA1c, mmol/mol flammatory markers, electrocardiogram, Baseline 61.5 (1.9) 62.7 (6.2) or vital signs. Day 91, change from baseline 5.5 (6.8) 22.9 (5.4) 0.100 Day 176, change from baseline 8.5 (7.8) 24.9 (4.2) 0.025 CONCLUSIONS Data are means (SD) unless otherwise indicated. Baseline data represent the safety population The administration of volanesorsen to (all randomized patients who received at least one dose of study drug); PD data represent the patients with hypertriglyceridemia and per-protocol population (patients who received at least 9 doses of study drug, had a valid baseline total apoC-III measure and at least one postbaseline measure, and did not have any poorly controlled type 2 diabetes on significant protocol deviations that would be expected to bias the patients’ PD assessments). metformin led to a profound decrease Baseline is defined as the average of all assessments prior to the first dose. Primary end point is in plasma apoC-III and TG levels, which defined as the day 91 results. For patients who terminated treatment early, the primary end point is defined as the first measurement after last dose. P value determined by the Wilcoxon was associated with a marked improve- rank sum test. *n = 8 for these values. ment in whole-body insulin sensitivity. The improvement in insulin sensitivity was strongly related to plasma apoC-III Volanesorsen Improves Clinical study (23.6%, P = 0.04 vs. placebo) and TG suppression and was reflected in Markers of Glycemic Control (Supplementary Table 3). improved moderate (glycated albumin Volanesorsen substantially improved and fructosamine) and long-term (HbA1c) integrative markers of glucose control, Tolerability and Adverse Effects indices of plasma glucose control. These lowering glycated albumin (21.73%, Therewerenodeathsinthisstudyor data suggest that volanesorsen treatment P = 0.034) and fructosamine (238.7 discontinuations of study drug due to improved both the dyslipidemia character- mmol/L, P = 0.045) 1 week after the an adverse event. There was one serious istic of type 2 diabetes and glucose control. last dose (day 91). Moreover, volane- adverse event of syncope in the volane- The effects of apoC-III inhibition on sorsen lowered HbA1c (20.44%, P = sorsen treatment group, which was the metabolic dyslipidemia of patients 0.025) at the end of the follow-up period moderate in severity and considered un- with type 2 diabetes are supported by (3 months after the last dose) (Table 1). likely to be related to the study drug. several studies. Patients with type 1 These changes correlated significantly The most common adverse events oc- diabetes and type 2 diabetes display with reductions in TG (r =0.60,P = curred at the injection site, followed by elevated apoC-III levels (23–26). Addi- 0.03) and apoC-III (r = 0.60, P =0.03) upper–respiratory tract infection (5 pa- tionally, studies have shown good cor- levels (Supplementary Table 2). Further- tients [50%]) and headache (5 patients relation between apoC-III levels and TG more, there was a moderate percentage [50%]). Overall, 15% of all injections in various populations (27–29). The po- weight loss at the end of treatment with were accompanied by local cutaneous tential mechanism by which apoC-III volanesorsen (22.1%, P = 0.05 vs. pla- reactions at the injection site. The major- inhibition may improve the metabolic cebo), which persisted to the end of the ity of adverse events in both treatment dyslipidemia of these patients is thought

Figure 1—Effect of volanesorsen 300 mg on plasma apoC-III and lipid levels over time. ApoC-III (A), TGs (B), HDL-C (C), and non-HDL-C (D). Data are shown as the mean percentage change from baseline. Error bars represent the 6 SEM. Solid blue triangles indicate dosing days. care.diabetesjournals.org Digenio and Associates 1413

resistance (9). It is thought that the in- ability of insulin resistant adipose tissue to store TG may be the initial step in the development of insulin resistance (1,31). There is molecular support for the concept that TGs provoke insulin resistance by peripheral catabolism in situ by tissues expressing LPL, causing local increases in NEFA uptake, result- ing in excess intracellular fatty acid metabolites. Intracellular fatty acid accumulation is thought to disrupt insulin receptor substrate phosphorylation, ulti- mately impairing insulin receptor actions downstream, such as glucose transport (32). Reductions in peripheral venous NEFA levels were not observed in the patients in this study. The use of circulating NEFA as a biomarker is limited by evidence suggesting that plasma NEFA sampled downstream from the site of NEFA uptake does not necessarily pre- dict intracellular NEFA accumulation. Figure 2—Volanesorsen 300 mg improves whole-body insulin sensitivity on clamp. In this box Clinical experiments have shown that ex- plot the horizontal line represents the median, the box represents the 25th and 75th percentiles, cessive plasma NEFAs are not required and the whiskers represent the 10th and 90th percentiles. The colored lines between baseline for hypertriglyceridemia to provoke in- and treated clamps demonstrate the response of individual patients. A green line indicates a sulin resistance (33), and suppressing 20% or greater improvement in insulin sensitivity, whereas a red line indicates a 20% or greater NEFA does not reverse insulin resistance worsening from baseline. when intracellular fatty acids remain elevated (34). to be enhancement of LPL activity and P = 0.01). Third, improved insulin sensitiv- The expression of apoC-III is regulated enhanced hepatic uptake of TRL (10,11,20). ity depended also on apoC-III suppression by insulin via the insulin response ele- Of note, the reduction in TG noted in as a whole by the group-wise comparison. ment encoded in the APOC3 gene (35– this study was not associated with in- Thepresenceofastrongrelationshipbe- 37). Plasma concentrations of apoC-III creasesinLDL-C,asoftenseenwith tween improved whole-body insulin sensi- were previously found to be associated TG-lowering therapies. Furthermore, tivity and apoC-III suppression across with the degree of insulin resistance as the reduction in TGs was associated individual patients strengthens the case estimated by HOMA (38). ApoC-III inhi- with increases in HDL-C, presumably that apoC-III suppression may be causally bition may also enhance the sensitivity by reducing the cholesteryl ester trans- relatedtoimprovedinsulinsensitivity. to the normal insulin-mediated suppres- fer protein-mediated exchange of HDL-C Moreover, the favorable effect on sion of APOC3 gene expression (10), for TRL TG (10). whole-body insulin sensitivity trans- which is dysregulated in patients with The most intriguing finding of the latedintoimprovementsinmedium type 2 diabetes. Finally, data also sug- current study was that profound sup- and long-term glucose handling as re- gest that circulating apoC-III itself may pression of apoC-III and TG levels by flected by significant plasma reductions act as a diabetogenic factor regulating volanesorsen improved insulin sensitiv- in percent glycated albumin, fructosamine, the preservation of pancreatic b-cells ity. Volanesorsen increased mean SIclamp and HbA1c concentrations. Recent evi- by as yet undefined mechanisms (9). .50% over baseline and raised SIclamp dence indicates that, like HbA1c,the A recent study showed that elevated at least 20% in the large majority of pa- medium-term markers also associate with apoC-III is associated with the hypertri- tients. Such robust improvements are cardiovascular risk (30,31). The reduction glyceridemia seen in both generalized not typical for current approved TG- in HbA1c, which reached statistical signifi- and partial lipodystrophy, suggesting lowering medications. The concept cance versus placebo 3 months after ces- that apoC-III may represent a therapeu- that apoC-III, independently or via hyper- sation of treatment, likely reflects the long tic target in these patients (38). Patients triglyceridemia, worsens insulin sensitiv- drug elimination half-life of ;4weeks with lipodystrophy, unable to store ex- ity is supported in several ways by the (Supplementary Table 4) and subsequent cess calories in adipocytes, develop ec- results of the current study. First, a po- prolonged duration of action. topic lipid deposits in muscle and liver, tent TG-lowering intervention enhanced Several lines of evidence have impli- whichleadtoanincreaseflux of free insulin sensitivity compared with pla- cated elevated TG and their by-products fatty acids and severe insulin resistance cebo. Second, there was a strong asso- in insulin resistance (1–3). Clinically, el- (39).Currenttherapiesusedtotreat ciation between the decrement in TG evations in TRL levels have been shown diabetes and severe insulin resistance and the increment in SIclamp (r =0.68, to exacerbate the pathology of insulin (thiazolidinediones, U-500 insulin) 1414 Volanesorsen Treatment in Patients With Diabetes Diabetes Care Volume 39, August 2016

This study has several limitations. First, the data were obtained in a very small number of subjects and need to be replicated in a larger cohort. There- fore, at this stage these data should only be considered exploratory and hypothesis generating, which is especially true of the mechanistic clamp study. Second, the current clamp design does not provide insight into the mechanism of improved insulin sensitivity: hepatic glucose production versus peripheral glucose uptake. Further understanding of this mechanism would be of interest. On the one hand there is a mechanistic basis supporting the concept that profoundly suppressing circulating TGs might improve peripheral glucose uptake. On the other hand, an association between apoC-III and hepatic glucose production has been reported (12). In conclusion, inhibition of apoC-III with a second-generation ASO improved the metabolic dyslipidemia of patients with type 2 diabetes. ApoC-III inhibition also improved whole-body insulin sensitivity and clinical markers of glucose handling. Improved insulin sensitivity was tightly correlated with TG suppression, consistent with the concept that hypertriglyceridemia exacerbates insulin resistance. These findings support further investigation to clarify the mechanism and better define the population that may benefit. Clinically, robust TG suppression by inhibition of apoC-III could complement diabetes management.

Figure 3—Improved insulin sensitivity depends on suppression of TRLs and apoC-III. Absolute change from baseline at end of treatment in TGs (A), apoC-III (B), VLDL–apoC-III (C), and VLDL-C Acknowledgments. The authors thank the (D) vs. posttreatment change in insulin sensitivity index. Shaded area represents the 95% CI. patients who participated in this study; Dr. Sotirios Tsimikas from Ionis Pharmaceuticals and University of California, San Diego, and Dr. and/or high TG (fibrates, niacin, fish oils) (NCT02211209), will help answer some Joseph L. Witztum from University of California, are not very efficacious in this patient pop- of the questions posed above. San Diego, for critical review of the manuscript; ulation (40). By reducing TG levels and Whether pharmacological inhibition Wei Cheng and QingQing Yang for statistical support; Nguyen Pham for technical support; improving insulin sensitivity, apoC-III in- of apoC-III could result in an improve- and Tracy Reigle for graphics support (all from hibition may improve the metabolic pro- ment in cardiovascular outcomes is un- Ionis Pharmaceuticals, Inc.). file of patients with lipodystrophy, as known but merits further consideration Duality of Interest. A.D. is an employee of well as those with FCS, reducing their if we take into account that individuals Akcea Therapeutics, a subsidiary of Ionis Phar- maceuticals, Inc. R.L.D. is an investigator for an risk of acute pancreatitis and other com- with a loss-of-function mutation in ongoing volanesorsen clinical trial (clinical trial plications associated with diabetes and APOC3 have 40% lower TG levels and a reg. no. NCT02211209). V.J.A., R.G.L., M.J.G., potentially of cardiovascular events. A 40% reduction in coronary artery dis- S.G.H., R.Y., W.S., B.F.B., S.B., and R.M.C. are employees of the study sponsor, Ionis Pharma- recently started trial evaluating the ease risk (15,16). This consideration is ceuticals, Inc. M.H. and L.M. are employees of effectsofvolanesorseninpatients further supported by the TG-dependent the study center, Profil Institute for Clinical with partial lipodystrophy, referred to association found recently between Research, Inc. No other potential conflicts of as BROADEN (NCT02527343), and the apoC-III and coronary artery calcification, interest relevant to this article were reported. Author Contributions. A.D. and R.L.D. are placebo-controlled trial in patients a measure of atherosclerosis, in patients co–primary authors based on contributions to with FCS, referred to as APPROACH with type 2 diabetes (27). manuscript content. V.J.A., M.H., L.M., S.G.H., R.Y., care.diabetesjournals.org Digenio and Associates 1415

W.S., and S.B. contributed to the study design. 11. Ooi EMM, Barrett PHR, Chan DC, Watts GF. families in NIDDM. Diabetes 1992;41(Suppl. 2): R.G.L., M.J.G., and R.M.C. provided the original Apolipoprotein C-III: understanding an emerg- 18–25 concept. M.H. and L.M. acquired data. A.D., R.L.D., ing cardiovascular risk factor. Clin Sci (Lond) 26. Blackett P, Sarale DC, Fesmire J, Harmon J, V.J.A., M.H., L.M., R.Y., W.S., B.F.B., and S.B. 2008;114:611–624 Weech P, Alaupovic P. Plasma apolipoprotein analyzed and/or interpreted data. All authors 12. Lee HY, Birkenfeld AL, Jornayvaz FR, et al. C-III levels in children with type I diabetes. South were involved in drafting and/or critical revision Apolipoprotein CIII overexpressing mice are Med J 1988;81:469–473 of the manuscript. A.D. is the guarantor of this predisposed to diet-induced hepatic steatosis 27. Qamar A, Khetarpal SA, Khera AV, Qasim A, work and, as such, had full access to all the data and hepatic insulin resistance. Hepatology Rader DJ, Reilly MP. Plasma apolipoprotein C-III in the study and takes responsibility for the in- 2011;54:1650–1660 levels, triglycerides, and coronary artery calcifi- tegrity of the data and the accuracy of the data 13. Petersen KF, Dufour S, Hariri A, et al. Apo- cation in type 2 diabetics. Arterioscler Thromb analysis. lipoprotein C3 gene variants in nonalcoholic Vasc Biol 2015;35:1880–1888 Prior Presentation. Results from this study fatty liver disease. N Engl J Med 2010;362: 28. Batal R, Tremblay M, Barrett PH, et al. fi – were presented at the 73rd Scienti c Sessions 1082 1089 Plasma kinetics of apoC-III and apoE in normo- of the American Diabetes Association, Chicago, 14. Verrijken A, Beckers S, Francque S, et al. A lipidemic and hypertriglyceridemic subjects. IL, 21–25 June 2013, and the Scientific Sessions gene variant of PNPLA3, but not of APOC3, is J Lipid Res 2000;41:706–718 of the National Lipid Association, New Orleans, associated with histological parameters of 29. Schonfeld G, George PK, Miller J, Reilly P, LA, 19–22 May 2016. NAFLD in an obese population. Obesity (Silver Witztum J. Apolipoprotein C-II and C-III levels in Spring) 2013;21:2138–2145 hyperlipoproteinemia. Metabolism 1979;28: 15. Crosby J, Peloso GM, Auer PL, et al.; TG and References 1001–1010 HDL Working Group of the Exome Sequencing 1. von Eckardstein A, Sibler RA. Possible contri- 30. Selvin E, Rawlings AM, Lutsey PL, et al. Fruc- Project, National Heart, Lung, and Blood Insti- butions of lipoproteins and cholesterol to the tosamine and glycated albumin and the risk of pathogenesis of diabetes mellitus type 2. Curr tute. Loss-of-function mutations in APOC3, triglycerides, and coronary disease. N Engl J Med cardiovascular outcomes and death. Circulation Opin Lipidol 2011;22:26–32 – 2014;371:22–31 2015;132:269 277 2. Wu L, Parhofer KG. Diabetic dyslipidemia. 31. Ginsberg HN. Insulin resistance and cardio- Metabolism 2014;63:1469–1479 16. Jørgensen AB, Frikke-Schmidt R, Nordestgaard vascular disease. J Clin Invest 2000;106:453– 3. Jacobson TA, Ito MK, Maki KC, et al. Na- BG, Tybjærg-Hansen A. Loss-of-function mutations 458 tional Lipid Association recommendations in APOC3 and risk of ischemic vascular disease. N 32. Shulman GI. Cellular mechanisms of insulin for patient-centered management of dyslipide- Engl J Med 2014;371:32–41 resistance. J Clin Invest 2000;106:171–176 mia: part 1 - executive summary. J Clin Lipidol 17. Cohen JC, Stender S, Hobbs HH. APOC3, 33. Pedrini MT, Niederwanger A, Kranebitter 2014;8:473–488 coronary disease, and complexities of Mende- 4. Stone NJ, Robinson JG, Lichtenstein AH, lian randomization. Cell Metab 2014;20:387– M, et al. Postprandial lipaemia induces an acute et al.; American College of Cardiology/American 389 decrease of insulin sensitivity in healthy men Heart Association Task Force on Practice Guide- 18. Graham MJ, Lee RG, Bell TA 3rd, et al. An- independently of plasma NEFA levels. Diabeto- lines. 2013 ACC/AHA guideline on the treatment tisense oligonucleotide inhibition of apolipo- logia 2006;49:1612–1618 of blood cholesterol to reduce atherosclerotic protein C-III reduces plasma triglycerides in 34. Johnson NA, Stannard SR, Rowlands DS, cardiovascular risk in adults: a report of the rodents, nonhuman primates, and humans. Circ et al. Short-term suppression of plasma free American College of Cardiology/American Heart Res 2013;112:1479–1490 fatty acids fails to improve insulin sensitivity Association Task Force on Practice Guidelines. 19. Gaudet D, Alexander VJ, Baker BF, et al. An- when intramyocellular lipid is elevated. Diabet Circulation 2014;129(Suppl. 2):S1–S45 tisense inhibition of apolipoprotein C-III in pa- Med 2006;23:1061–1068 5. Catapano AL, Reiner Z, De Backer G, et al.; tients with hypertriglycereidemia. N Engl J Med 35. Caron S, Verrijken A, Mertens I, et al. Tran- European Society of Cardiology (ESC); European 2015;373:438–447 scriptional activation of apolipoprotein CIII ex- Atherosclerosis Society (EAS). ESC/EAS Guide- 20. Gaudet D, Brisson D, Tremblay K, et al. Tar- pression by glucose may contribute to diabetic lines for the management of dyslipidaemias geting APOC3 in the familial chylomicronemia dyslipidemia. Arterioscler Thromb Vasc Biol The Task Force for the management of dyslipi- syndrome. N Engl J Med 2014;371:2200–2206 2011;31:513–519 daemias of the European Society of Cardiology 21. Muniyappa R, Lee S, Chen H, Quon MJ. Cur- 36. Altomonte J, Cong L, Harbaran S, et al. (ESC) and the European Atherosclerosis Society rent approaches for assessing insulin sensitivity Foxo1 mediates insulin action on apoC-III and (EAS). Atherosclerosis 2011;217:3–46 and resistance in vivo: advantages, limitations, triglyceride metabolism. J Clin Invest 2004; 6. Bruckert E, Labreuche J, Deplanque D, and appropriate usage. Am J Physiol Endocrinol 114:1493–1503 – Touboul PJ, Amarenco P. Fibrates effect on car- Metab 2008;294:E15 E26 37. Chan DC, Watts GF, Barrett PHR, Mamo JCL, diovascular risk is greater in patients with high 22. Krentz AJ, Heinemann L, Hompesch M. Redgrave TG. Markers of triglyceride-rich lipo- Methods for quantifying insulin sensitivity and triglyceride levels or atherogenic dyslipidemia protein remnant metabolism in visceral obesity. fi determining insulin time-action profiles. In pro le: a systematic review and meta-analysis. Clin Chem 2002;48:278–283 – Translational Research Methods for Diabetes, J Cardiovasc Pharmacol 2011;57:267 272 38. Kassai A, Muniyappa R, Levenson AE, et al. 7. Ginsberg HN, Huang LS. The insulin resis- Obesity, and Cardiometabolic Drug Develop- Effect of leptin administration on circulating tance syndrome: impact on lipoprotein metab- ment: Focus on Early Phase Studies. Krentz AJ, apolipoprotein CIII levels in patients with lipo- olism and atherothrombosis. J Cardiovasc Risk HeinemannL,HompeschM,Eds.NewYork, dystrophy. J Clin Endocrinol Metab 2016;101: 2000;7:325–331 Springer, 2015, p. 3–44 – 8. Petremand´ J, Puyal J, Chatton JY, et al. HDLs 23. Hokanson JE, Kinney GL, Cheng S, Erlich HA, 1790 1797 protect pancreatic b-cells against ER stress by Kretowski A, Rewers M. Susceptibility to type 1 39. Heilbronn L, Smith SR, Ravussin E. Failure restoring protein folding and trafficking. Diabe- diabetes is associated with ApoCIII gene haplo- of fat cell proliferation, mitochondrial func- tes 2012;61:1100–1111 types. Diabetes 2006;55:834–838 tion and fat oxidation results in ectopic fat 9. Holmberg R, Refai E, Hoög¨ A, et al. Lowering 24. Florez H, Mendez A, Casanova-Romero P, storage, insulin resistance and type II diabetes apolipoprotein CIII delays onset of type 1 diabe- et al. Increased apolipoprotein C-III levels asso- mellitus. Int J Obes Relat Metab Disord 2004; tes. Proc Natl Acad Sci U S A 2011;108:10685– ciated with insulin resistance contribute to 28(Suppl. 4):S12–S21 10689 dyslipidemia in normoglycemic and diabetic 40. Handelsman Y, Oral EA, Bloomgarden ZT, 10. Chan DC, Chen MM, Ooi EMM, Watts GF. subjects from a triethnic population. Athero- et al.; American Association of Clinical Endocri- An ABC of apolipoprotein C-III: a clinically useful sclerosis 2006;188:134–141 nologists. The clinical approach to the detection new cardiovascular risk factor? Int J Clin Pract 25. Alaupovic P, Bard JM, Tavella M, Shafer D. of lipodystrophy - an AACE consensus state- 2008;62:799–809 Identification of apoB-containing lipoprotein ment. Endocr Pract 2013;19:107–116