Diabetes & Metabolism 43 (2017) 2S3-2S12
`
`After the LEADER trial and SUSTAIN-6, how do we explain
`the cardiovascular benefits of some GLP-1 receptor agonists?
`
`B. Vergès1,2,*, B. Charbonnel3
`
`1Department of Endocrinology-Diabetology, University Hospital, 2 bd Maréchal de Lattre de Tassigny, F-21000 Dijon, France
`2INSERM CRI 866, Dijon, France
`3Department of Endocrinology, University of Nantes, Nantes, France
`
`Abstract
`
`Recent cardiovascular outcome trials –  the LEADER with liragutide and SUSTAIN-6 with semaglutide  – have shown significant
`reductions of major cardiovascular (CV) events with these glucagon-like peptide (GLP)-1 receptor agonists. Progressive separation of
`the treatment and placebo curves, starting clearly between 12 and 18 months of the trial period, and significant reductions in the risk
`of myocardial infarction and stroke, indicate that the beneficial CV effects observed with GLP-1 receptor agonists could be due to an
`antiatherogenic effect. So far, the reasons for such an effect of GLP-1 receptor agonists have not been entirely clear, although several
`hypotheses may be proposed. As the reductions in glycated haemoglobin and systolic blood pressure (SBP) in these trials were modest,
`and both trials lasted only a short period of time, reductions in hyperglycaemia and SBP are unlikely to be involved in the beneficial
`CV effects of GLP-1 receptor agonists. On the other hand, their effect on lipids and, in particular, the dramatic decrease in postprandial
`hypertriglyceridaemia may explain their beneficial CV actions. Reduction of body weight, including a significant decrease in visceral
`fat in patients using GLP-1 receptor agonists, may also have beneficial CV effects by reducing chronic proatherogenic inflammation. In
`addition, there are in-vitro data showing a direct anti-inflammatory effect with these agents that could also be involved in their beneficial
`CV effects. Moreover, studies in humans have shown significant beneficial effects on ischaemic myocardium after a very short treatment
`period, suggesting a direct effect of GLP-1 receptor agonists on myocardium, although the precise mechanism remains unclear. Finally, as
`a reduction in insulin resistance has been associated with a decrease in CV risk, it cannot be ruled out that the lowering of insulin resistance
`induced by GLP-1 receptor agonists might also be involved in their beneficial CV actions.
`© 2017. Elsevier Masson SAS. All rights reserved
`
`Key words: Cardiovascular; Myocardial infarction; GLP-1; Liraglutide; Semaglutide
`
`1. Introduction
`
`Glucagon-like peptide (GLP)-1 receptor agonists are
`effective hypoglycaemic agents that are widely used. In recent
`years, considerable data have suggested that GLP-1 receptor
`agonists may have effects beyond their glucose-lowering
`actions, including a possible cardioprotective effect [1,2]. Some
`animal studies showed that GLP-1 receptor agonists could
`reduce the size of myocardial infarction (MI) [3,4] while, in
`humans, limited studies have reported reduced MI size after
`administration of these drugs, suggesting beneficial effects
`on the ischaemic heart [5–7]. Furthermore, the Liraglutide
`Effect and Action in Diabetes: Evaluation of Cardiovascular
`Outcome Results (LEADER) trial and the Trial to Evaluate
`Cardiovascular and Other Long-term Outcomes with
`Semaglutide in Subjects with Type 2 Diabetes (SUSTAIN-6)
`have recently provided clear evidence of cardiovascular (CV)
`
`benefit with these GLP-1 receptor agonists. Both studies were
`conducted in patients with type 2 diabetes mellitus (T2DM)
`and a history of previous CV events (82–83%) or high CV risk
`(17–18%) [8,9]. In the LEADER trial, 3.5 years of treatment
`with liraglutide 1.8 mg/day was associated with a significant
`13% reduction in the primary outcome (time to first major CV
`event: CV death, non-fatal MI, non-fatal stroke; p = 0.01),
`and a significant 14% reduction in MI (fatal and non-fatal;
`p = 0.046), 22% reduction in CV-related death (p = 0.007)
`and 15% reduction in total mortality (p = 0.02; Table 1)
`[8]. In SUSTAIN-6, 2 years of treatment with semaglutide,
`a long-acting GLP-1 receptor agonist administered once a
`week, resulted in a significant 26% reduction in the primary
`outcome (time to first major CV event: CV death, non-fatal
`MI, non-fatal stroke; p = 0.02), 39% reduction in non-fatal
`stroke (p = 0.04) and 35% reduction in revascularization
`procedures (p = 0.003; Table 1) [9].
`
`*Corresponding author.
`E-mail address: bruno.verges@chu-dijon.fr (B. Vergès).
`
`© 2017 Elsevier Masson SAS. All rights reserved.
`
`278911ANN_DM_NOVO_CS6_PC.indb 3
`
`14/03/2017 13:08:44
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00001
`
`

`

`2S4
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`Table 1
`Effects of glucagon-like peptide (GLP)-1 receptor agonists on primary and secondary cardiovascular (CV) outcomes in the LEADER trial and
`SUSTAIN-6
`
`Study duration (years)
`
`GLP-1 receptor agonist: molecule
`
`GLP-1 receptor agonist: dose
`
`Patients (n)
`
`Major CV eventsa
`
`Myocardial infarction
`
`Non-fatal stroke
`
`LEADER
`
`3.5
`
`Liraglutide
`
`1.8 mg/day
`
`9340
`
`SUSTAIN-6
`
`2.0
`
`Semaglutide
`
`0.5 or 1.0 mg/week
`
`3297
`
`↓ 13% (p = 0.01)
`
`↓ 26% (p = 0.02)
`
`↓ 14% (p = 0.046)
`
`↓ 15% (NS; p = 0.38)
`
`↓ 11% (p = 0.30; NS)
`
`↓ 39% (p = 0.04)
`
`Coronary revascularization
`
`↓ 9% (p = 0.18; NS)
`
`Not available
`
`Coronary + peripheral revascularization
`
`Not available
`
`↓ 35% (p = 0.003)
`
`Hospitalization for heart failure
`
`↓ 13% (p = 0.14; NS)
`
`→ (p = 0.57; NS)
`
`CV death
`
`Total mortality
`
`↓ 22% (p = 0.007)
`
`→ (p = 0.92; NS)
`
`↓ 15% (p = 0.02)
`
`→ (p = 0.79; NS)
`
`aCV death, non-fatal myocardial infarction, non-fatal stroke.
`LEADER: Liraglutide Effect and Action in Diabetes: Evaluation of Cardiovascular Outcome Results; SUSTAIN-6: Trial to Evaluate Cardiovascular and Other Long-term
`Outcomes with Semaglutide in Subjects with Type 2 Diabetes.
`↓: decrease; →: no effect; NS: not significant
`
`There are some similarities, but also some differences,
`between the results of these two trials (Table 1). One major
`similarity was the significant reduction in major CV outcomes
`with liraglutide. However, the significant reduction in CV
`death and total mortality in LEADER was not observed in
`SUSTAIN-6. This difference could be due to both the shorter
`duration of SUSTAIN-6 and its smaller number of included
`patients, leading to considerably fewer deaths compared
`with the LEADER trial (122 vs 497, respectively). A notable
`difference between the two trials was the significant reduction
`in non-fatal stroke in SUSTAIN-6, but not in LEADER.
`Although the reasons for this discrepancy are still unknown,
`it may be supposed that the greater reduction in systolic blood
`pressure in SUSTAIN-6 compared with LEADER (-2.6 mmHg
`vs -1.2 mmHg, respectively) is perhaps part of the explanation.
`Nevertheless, further studies are needed to clarify the
`dramatic effect of semaglutide on stroke. It is important to
`note that, in LEADER, the curves for liraglutide and placebo
`diverged at between 12 and 18 months, which is similar to what
`is observed in clinical prospective trials of statins, whereas in
`SUSTAIN-6, the curves for semaglutide and placebo diverged
`progressively throughout the study. While this suggests that the
`decrease in major CV events observed with GLP-1 receptor
`agonists could be due to an antiatherogenic effect, so far, the
`reasons behind this beneficial effect have not been entirely
`elucidated, although several hypotheses may be considered.
`Thus, the present review discusses the potential mechanisms
`that might explain the CV benefits of GLP-1 receptor agonists
`summarized in Fig. 1.
`
`2. Effects on CV risk factors
`
`2.1. Lipids
`
`2.1.1. Effects of liraglutide on fasting lipids
`
`Significant variations in lipid parameters are observed in
`type 2 diabetes mellitus (T2DM) patients treated with GLP-1
`receptor agonists. In a 3.5-year open-label study, exenatide
`b.i.d. reduced low-density lipoprotein (LDL) cholesterol
`by 6% and triglycerides (TGs) by 12%, while increasing
`high-density lipoprotein (HDL) cholesterol by 24% [10].
`Five-year data from the DURATION study showed a significant
`reduction in LDL cholesterol (-9.8%) and TGs (-12%), and a
`significant increase in HDL cholesterol (+4.3%), with 2-mg
`exenatide once a week [11], while a meta-analysis of six trials
`of liraglutide reported reductions in total cholesterol (-0.13
`mmol; p < 0.01), LDL cholesterol (-0.20 mmol; p < 0.0001),
`free fatty acids (-0.09 mmol; p < 0.0001) and TGs (-0.20 mmol;
`p < 0.01) compared with baseline in the intention-to-treat
`population [12].
`
`2.1.2. Effects of liraglutide on postprandial lipids
`
`The most striking effect of GLP-1 receptor agonists on
`lipids is the significant reduction in postprandial hypertriglyc-
`eridaemia. In healthy volunteers, GLP-1 infusion abolished
`postprandial lipidaemia [13]. In subjects with impaired glucose
`tolerance and recent-onset T2DM, a single subcutaneous
`
`278911ANN_DM_NOVO_CS6_PC.indb 4
`
`14/03/2017 13:08:44
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00002
`
`

`

`
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`2S5
`
`GLP-1 receptor agonists
`
` Blood pressure
`
` Lipids
`
` Hyperglycemia
`
` Insulin resistance
`
`direct?
`
`direct?
`
` Body weight
`
` in(cid:31)ammation
`
`• likely
`• mechanism unclear
`
`Endothelium
`
`direct effect unlikely effect
`via  PP lipids?
`Fig. 1. Summary of the mechanisms that may explain the cardiovascular benefits of GLP-1 receptor agonists. Solid lines indicate the likely mechanisms;
`dotted lines indicate those unlikely to play major roles.
`
`GLP-1
`agonists
`
`injection of the GLP-1 agonist exenatide (10 µg) was shown to
`markedly reduce postprandial increases in TGs, apolipoprotein
`(Apo) B48 and ApoC-III compared with a placebo [14],
`and as this effect was observed after just a single exenatide
`injection, it indicates that it was independent of its effect on
`body weight. On the other hand, 3 weeks of treatment with
`liraglutide (1.8 mg/day) compared with a placebo in patients
`with T2DM significantly reduced postprandial excursions
`of TGs and ApoB48 after a fat-rich meal, independently of
`gastric-emptying [15]. In fact, in hamsters and mice, exenatide
`decreased plasma TG-rich lipoprotein (TRL)-containing
`ApoB48, and reduced the secretion of ApoB48 in hamster
`enterocyte cultures [16]. Conversely, blockade of GLP-1
`receptor signaling by the antagonist exendin-(9-39) or by
`genetic elimination of GLP-1 signaling in GLP-1 receptor
`knock-out (KO) mice enhanced ApoB48 TRL secretion [16].
`In healthy humans, 4–6 weeks of treatment with exenatide
`significantly suppressed plasma concentrations and produc-
`tion rates of ApoB48 TRL [17]. In patients with T2DM, it
`has recently been reported that 6 months of treatment with
`liraglutide significantly reduced ApoB48 production and
`increased ApoB48 catabolism, leading to significant decreases
`in plasma ApoB48 [18].
`Although GLP-1 receptor agonists have only a relatively
`modest effect on LDL cholesterol, they can induce a major
`reduction in postprandial hyperlipidaemia, an important
`feature of diabetic dyslipidaemia [19], as it is known to be
`
`atherogenic [20]. Thus, the effect of GLP-1 receptor agonists
`on postprandial hyperlipidaemia could be one factor involved
`in their beneficial CV effects.
`
`2.2. Blood pressure
`
`In human trials, treatment with GLP-1-receptor agonists
`is associated with reductions in blood pressure (BP). Over
`82 weeks of exenatide b.i.d. treatment, systolic/diastolic BP
`fell significantly vs baseline in 314 overweight patients with
`T2DM. Average decreases were -1.3/-2.7 mmHg (95% CI: -3.1
`to +0.5/-3.8 to -1.7 mmHg), with even greater changes observed
`in the quartile of patients who lost the most weight (on average:
`-3.9/-4.4 mmHg) [21]. A large meta-analysis of patients
`using exenatide reported significant reductions in systolic BP
`compared with both placebo (-5.24 mmHg, p < 0.00001) and
`insulin glargine (-3.46 mmHg, p < 0.00001), and in diastolic
`BP compared with placebo (-5.91 mmHg, p < 0.00001) and
`sitagliptin (-0.99 mmHg, p < 0.00001) [22]. In another large
`meta-analysis, liraglutide at 1.2 mg/day lowered systolic BP
`compared with placebo and glimepiride treatment, with mean
`differences of -5.60 mmHg (p < 0.00001) and -2.38 mmHg
`(p = 0.05), respectively. In addition, liraglutide at 1.8 mg/day
`also reduced systolic BP vs placebo and glimepiride, with mean
`differences of -4.49 mmHg (p < 0.00001) and -2.62 mmHg
`(p < 0.00001), respectively [22]. In the LEADER trial, a
`mean systolic BP reduction of 1.3 mmHg vs placebo was
`
`278911ANN_DM_NOVO_CS6_PC.indb 5
`
`14/03/2017 13:08:45
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00003
`
`

`

`2S6
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`observed while, in SUSTAIN-6, systolic BP was -2.5 mmHg
`and -0.9 mmHg lower in patients using semaglutide 1 mg or
`0.5 mg, respectively, vs placebo [8,9].
`Weight loss may have contributed to the BP decrease
`observed with GLP-1 receptor agonists. However, systolic BP
`changes were seen early in these trials and preceded weight
`loss, suggesting a direct effect of GLP-1 receptor agonists
`on BP. It has been suggested that the BP-lowering effect of
`GLP-1 receptor agonists could be due to direct stimulation by
`GLP-1 of atrial natriuretic peptide (ANP) secretion, leading
`to increased natriuresis [23].
`Nevertheless, the reduction in systolic BP observed with
`GLP-1 receptor agonists appears to be too modest to be a major
`factor behind the significant decrease in major CV events
`noted with liraglutide. However, the BP-lowering observed
`with semaglutide may have contributed to its overall benefit,
`and especially the risk of stroke, as seen in SUSTAIN-6.
`Yet, it has recently been shown that the BP-lowering effect
`of sodium – glucose cotransporter 2 (SGLT2) inhibitors (which
`is more pronounced than with GLP-1 receptor agonists) may
`only partially explain the cardioprotective effects observed with
`empagliflozin in the Empagliflozin Cardiovascular Outcome
`Event Trial in Type 2 Diabetes Mellitus Patients (EMPA-
`REG OUTCOME) [24,25]. This reinforces the idea that the
`BP-lowering effect of GLP-1 receptor agonists is likely to be
`only a minor factor in any explanation of their CV benefits.
`
`3. Reduction of hyperglycaemia
`
`In LEADER, the mean HbA1c level with liraglutide was
`0.4% lower than with placebo whereas, in SUSTAIN-6, the
`mean HbA1c level was 0.7% lower with semaglutide 0.5 mg
`and 1% lower with semaglutide 1 mg than with placebo [8,9].
`Previous prospective studies have shown that the reduction in
`hyperglycaemia needs time to induce a significant decrease
`in CV events [26]. For instance, in the United Kingdom
`Prospective Diabetes Study (UKPDS), a significant reduc-
`tion in MI was observed only in the long-term report after a
`median follow-up of 17 years [27] and, in the Veterans Affairs
`Diabetes Trial (VADT), a significant 17% decrease in the
`primary CV outcome (heart attack, stroke, new or worsening
`congestive heart failure, amputation for ischaemic gangrene
`or CV death) was reported only after a median follow-up of
`9.8 years [28], whereas the first VADT report, after a median
`6.25 years of follow-up, showed no significant effects on this
`primary outcome [29].
`Nevertheless, the effect of a decrease in hyperglycaemia
`on the reduction of CV events cannot be totally ruled out in
`either LEADER or SUSTAIN-6, despite its being an unlikely
`major contributing factor, given the short durations of those
`studies: 3.5 years and 2 years, respectively.
`
`agonists [30–34]. Exendin-4 directly reduced lipopolysac-
`charide (LPS)-induced secretion of cytokines [tumour necrosis
`factor (TNF)-α, interleukin (IL)-1β and IL-10] in human
`monocytes from non-diabetic individuals, effects that were
`blocked by coadministration of the GLP-1 receptor antagonist
`exendin-(9-39), suggesting that GLP-1 had a direct effect on the
`immune system [30]. In-vitro liraglutide reduced the expression
`of vascular cell adhesion molecule (VCAM)-1 in human aortic
`endothelial cells after stimulation by LPS or TNF-α through
`a calcium- and adenosine monophosphate-activated protein
`kinase (AMPK)-dependent mechanism [31], and decreased
`monocyte adhesion to TNF-α-activated endothelial cells
`[32]. Also, GLP-1 and GLP-1 receptor agonists both reduced
`vascular monocyte adhesion and foam-cell formation in mice
`[32–34], while the anti-inflammatory action of GLP-1 was
`abolished by coadministration of the GLP-1 receptor antagonist
`exendin-(9-39), suggesting a direct effect of GLP-1 [34].
`Liraglutide administered for 7 days to C57BL/6 mice fed a
`high-fat diet reduced heart inflammation and lipid accumulation
`with no significant weight loss [32], whereas treatment with
`taspoglutide, another GLP-1 agonist, did not significantly
`change plaque area and macrophage accumulation in ApoE
`KO mice [35].
`Some data also suggest anti-inflammatory actions with both
`GLP-1 and GLP-1 receptor agonists in humans [36]. Infusions
`of native GLP-1 in patients with type 1 diabetes mellitus
`(T1DM) reduced the plasma increases of IL-6, intercellular
`adhesion molecule (ICAM)-1 and markers of oxidative stress
`[nitrotyrosine, 8-iso-prostaglandin F2α (PGF2α)] induced by
`both hypoglycaemia and hyperglycaemia [37]. In non-obese
`patients with T2DM, GLP-1 reduced plasma levels of IL-6,
`ICAM-1, PGF2α and nitrotyrosine [38] whereas, in obese
`T2DM patients, exenatide reduced circulating levels of IL-2,
`monocyte chemotactic protein (MCP)-1, serum amyloid A
`and matrix metallopeptidase (MMP)-9, with no significant
`weight loss [39]. Eight weeks of treatment with liraglutide
`reduced soluble cluster of differentiation 163 (sCD163) and
`the production of proinflammatory cytokines (IL-1β, IL-6,
`TNF-α) in peripheral blood cells in obese patients with T2DM
`and psoriasis, and increased levels of the anti-inflammatory
`adipokine adiponectin, independently of reductions in body
`weight, fructosamine and HbA1c [40].
`Although there are data indicating that the anti-inflamma-
`tory effects of GLP-1 and GLP-1 receptor agonists may be
`direct, it should be borne in mind that many of their observed
`anti-inflammatory effects may have been confounded by
`parallel decreases in body weight, blood glucose and free fatty
`acids, as reported in several studies. Thus, GLP-1 receptor
`agonists reduce chronic CV inflammation through both direct
`and indirect effects, and their anti-inflammatory effects could
`be part of their beneficial CV actions.
`
`4. Anti-inflammatory actions of GLP-1
`
`5. Effect on weight loss
`
`Several in-vitro and animal studies have shown anti-
`inflammatory effects with both GLP-1 and GLP-1 receptor
`
`In addition to their effects on blood glucose control, GLP-1
`receptor agonists have demonstrated positive effects on body
`
`278911ANN_DM_NOVO_CS6_PC.indb 6
`
`14/03/2017 13:08:45
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00004
`
`

`

`
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`2S7
`
`weight. A meta-analysis of 27 trials showed significant mean
`weight loss with GLP-1 receptor agonists vs placebo: exenatide
`2 mg/week: -1.62 kg; exenatide 20 μg: -1.37 kg; liraglutide 1.2
`mg: -1.01 kg; and liraglutide 1.8 mg: -1.51 kg [41]. Another
`meta-analysis of 18 trials involving T2DM patients reported
`a mean body weight decrease of -2.8 kg with GLP-1 receptor
`agonists vs control groups [42], while Robinson et al. [43],
`in a larger meta-analysis of 32 trials of either exenatide or
`liaglutide, reported a mean body weight decrease of -3.31 kg
`vs an active control and -1.22 kg vs placebo.
`In the Liraglutide Effect and Action in Diabetes (LEAD-2)
`trial, the weight loss associated with liraglutide treatment was
`primarily the result of decreases in both visceral and subcutane-
`ous fat tissue [44]. Six months of treatment with exenatide also
`significantly reduced both visceral and subcutaneous fat in
`drug-naïve T2DM patients [45]. It is well known that adipose
`tissue, particularly visceral adipose tissue, is associated with
`increased chronic inflammation and that a modest elevation
`of inflammation-related molecules in the circulation can
`contribute to a substantially increased risk of CV disease
`[46]. Indeed, it has also been shown that weight loss improves
`the inflammatory profile of obese subjects by decreasing
`proinflammatory factors and increasing anti-inflammatory
`molecules [47]. In an interventional study of obese women,
`body-weight reduction was associated with a significant fall
`in serum concentrations of IL-6, IL-18 and C-reactive protein
`(CRP), whereas adiponectin levels were significantly increased
`[48]. After 1 year of a multidisciplinary programme of weight
`reduction in obese subjects who achieved a loss of ≥ 10%
`of their original weight, a significant reduction in plasma
`cytokines (TNF-α, IL-6) and vascular adhesion molecules
`was observed, along with an improved vascular response to
`L-arginine [49].
`Thus, the possibility that the significant reduction in body
`weight associated with liraglutide treatment may have a
`beneficial CV effect by reducing chronic proatherogenic
`inflammation cannot be excluded, although the effect is likely
`to be minor. Indeed, in the Look AHEAD trial, an intensified
`lifestyle intervention reduced body weight (-2.6 kg vs control
`group), but with no significant reduction in CV outcomes [50].
`
`6. Direct effects on myocardium
`
`6.1. Effects of native GLP-1
`
`Several studies have shown a beneficial effect of native
`GLP-1 on the heart [51]. GLP-1 in vitro increased intracellular
`cyclic AMP in rat cardiomyocytes [52] while, in murine car-
`diomyocytes, GLP-1 protected cells against apoptosis induced
`by staurosporine, palmitate or ceramide, a cytoprotective effect
`mainly mediated by phosphatidylinositol 3-kinase (PI3K)
`and partially dependent on extracellular signal-regulated
`kinase (ERK) 1/2 [53]. In wild-type mouse hearts subjected
`to ischaemia – reperfusion, GLP-1 significantly increased
`functional recovery and cardiomyocyte viability [54]. Such
`
`effects were also observed in GLP-1-Receptor KO mice, sug-
`gesting that some cardioprotective effects of native GLP-1 may
`be mediated through a mechanism independent of the known
`GLP-1 receptors [54]. Indeed, in that study the GLP-1 metabo-
`lite, GLP-1-(9-36), also showed significant cardioprotective
`effects [54]. In Wistar – Kyoto rat hearts, GLP-1 increased
`glucose uptake by increasing nitric oxide (NO) production
`and glucose transporter (GLUT)-1 translocation [55]. In the
`same model, GLP-1 also enhanced recovery after a 30-min
`low-flow ischaemia protocol, with significant improvement in
`left ventricular (LV) end-diastolic pressure and LV developed
`pressure, and also showed that the GLP-1-mediated increase in
`glucose uptake was through a non-Akt-dependent mechanism
`distinct from the action of insulin [55].
`In conscious dogs with advanced dilated cardiomyopathy,
`Nikolaïdis et al. [56] showed that a 48-h infusion of GLP-1
`could significantly increase stroke volume and cardiac output,
`while significantly decreasing LV end-diastolic pressure, heart
`rate and systolic vascular resistance. GLP-1 also increased
`myocardial insulin sensitivity and myocardial glucose uptake
`[56]. In rats subjected to ischaemia – reperfusion, GLP-1
`dramatically decreased infarct size, an effect that was abolished
`by a GLP-1 receptor antagonist [57], while data from a study in
`dogs by Moberly et al. [58] indicated that acute intracoronary
`administration of GLP-1 preferentially augments glucose
`metabolism in ischaemic myocardium, independently of
`its effects on cardiac contractile function or coronary blood
`flow. In swine, GLP-1, but not its metabolite GLP-1-(9-36),
`increased cardiac output during ischaemia by increasing
`ventricular preload without changing cardiac inotropy [59].
`In non-diabetic rats, an infusion of native GLP-1, started 10
`min prior to the induction of ischaemia and continued for 24
`h until the end of the reperfusion period, significantly reduced
`infarct size while increasing myocardial glucose uptake in
`the normal heart, and induced metabolic substrate switching
`by increasing the ratio of carbohydrate vs fat oxidation in
`the non-ischaemic myocardium of ischaemic hearts [60]. A
`possible direct effect of GLP-1 on the heart is also suspected,
`as mice with genetic deletion of GLP-1 receptors display
`increased LV thickness, impaired LV contractility and diastolic
`dysfunction after insulin administration, as well as reduced
`LV contractility after epinephrine infusion [61].
`A few studies have analyzed the effects of native GLP-1
`on human hearts in vivo. Thrainsdottir et al. [62] examined six
`diabetic patients with congestive heart failure of ischaemic
`etiology, treated with subcutaneous infusions of 3–4 pmol/
`kg/min of recombinant GLP-1 for 72 h, and reported a trend
`towards myocardial improvement. In one exploratory study,
`a 72-h GLP-1 infusion improved regional and global LV
`function in 10 patients with acute MI and severe diastolic
`dysfunction after successful primary angioplasty, increasing
`the LV ejection fraction (LVEF) from 29 ± 2% to 39 ± 2%
`(p < 0.01). In addition, the in-hospital mortality rate was
`reduced in patients with acute MI and LV dysfunction (27%
`vs 10%, respectively) after successful reperfusion [63]. In
`a pilot study of 20 patients with normal LV function and
`
`278911ANN_DM_NOVO_CS6_PC.indb 7
`
`14/03/2017 13:08:45
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00005
`
`

`

`2S8
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`single-vessel coronary disease, infusion of GLP-1 improved
`recovery of LV systolic and diastolic function 30 min after
`balloon occlusion compared with controls, and reduced LV
`dysfunction after a second balloon occlusion [64]. Sokos et
`al. [65] studied the effects of GLP-1 infusion (1.5 pmol/kg/
`min) before and after coronary artery bypass grafting (CABG)
`in patients with heart disease and preserved LV. Compared
`with controls, these patients needed fewer inotropic and
`vasoactive drug infusions postoperatively to achieve the same
`haemodynamic results, and presented with arrhythmias less
`frequently. These beneficial effects on myocardial function
`were not confirmed in a study of 20 patients without diabetes,
`in those with heart failure and ischaemic heart disease who
`received 48-h GLP-1 (0.7 pmol/kg/min), there was no
`significant increase in either cardiac index or LVEF [66].
`In yet another study, GLP-1 treatments were associated with
`improvements in LV function, functional status and quality
`of life in patients with chronic heart failure, as measured
`by the Minnesota Living with Heart Failure Questionnaire
`quality-of-life score [67].
`
`6.2. Effects of GLP-1 receptor agonists
`
`Treatment with exenatide reduced MI size and improved
`cardiac function in a porcine model [3], but in the same
`animal model, there was no confirmed MI size reduction
`with liraglutide [68]. However, liraglutide was shown to
`confer cardioprotection and other survival advantages in
`mice compared with metformin, with significant reductions
`in MI size despite achieving equivalent glycaemic control;
`the effect lasted for up to 4 days after treatment cessation [4].
`The use of different animal models and study procedures may
`account for such discrepancies among studies.
`Several studies have also extensively analyzed the effects
`of GLP-1 receptor agonists on human hearts. In a randomized
`placebo-controlled study, administration of exenatide over
`3 days in patients with ST-segment elevation MI (STEMI)
`induced significant reductions in plasma levels of creatine
`kinase-MB and troponin I and in infarct size, with a signifi-
`cantly higher LVEF at 6 months compared with a placebo [5].
`In a randomized placebo-controlled study of 172 patients with
`STEMI, administration of exenatide started 15 min before a
`percutaneous coronary intervention increased myocardial
`salvage and reduced infarct size vs placebo [6]; these beneficial
`effects of exenatide on myocardium were independent of
`hyperglycaemia [7].
`In contrast, other clinical studies of GLP-1 receptor
`agonists have shown no improvement in heart failure. In
`one, 12 weeks of treatment with albiglutide vs placebo in
`non-diabetic patients with heart failure (LVEF < 40%) failed
`to improve either LVEF or LV structure or function, and led
`to only a modest increase in peak oxygen consumption [69].
`Similarly, in a study of heart-failure patients with and without
`diabetes (LVEF < 45%), liraglutide (1.8 mg/day) for 24 weeks
`did not significantly improve LVEF [51]. In fact, the absence
`of any effect of GLP-1 receptor agonists in heart failure was
`
`confirmed by the Functional Impact of GLP-1 for Heart
`Failure Treatment (FIGHT) trial, a prospective randomized,
`double-blind, placebo-controlled study performed in patients
`with heart failure, in whom treatment with liraglutide (1.8 mg/
`day) failed to improve time to death, time to hospitalization for
`heart failure or time-averaged changes in levels of N-terminal
`prohormone of brain natriuretic peptide [51]. In the recently
`presented LIVE study, involving 241 patients with chronic
`heart failure, after 24 weeks of treatment, there were no
`significant differences between liraglutide and placebo groups
`in the primary endpoint of change in LVEF [70].
`These studies are reassuring, as they reveal no increased
`risk of heart failure with GLP-1 receptor agonists, although
`a significant increase in heart rate is observed with these
`drugs. These findings also indicate that the CV benefits of
`GLP-1 receptor agonists are not due to any improvement of
`LV function.
`In summary, studies in humans demonstrate significant
`beneficial effects of GLP-1 receptor agonists on ischaemic
`myocardium after very short treatment periods, suggesting
`a direct effect on myocardium. Other evidence to suggest a
`direct effect of GLP-1 on the heart comes from a genome-wide
`association study, which revealed that a variant (Ala316Thr;
`rs10305492) of the GLP-1 gene – associated with lower blood
`glucose and T2DM risk and, therefore, probably increased
`GLP-1 activity – was also associated with a significantly
`lower risk of coronary heart disease [71]. Thus, a direct effect
`of GLP-1 receptor agonists on the ischaemic heart might be
`involved in their beneficial actions on the CV system. However,
`the mechanisms underlying these possible CV and heart effects
`remain unclear, as GLP-1 receptors are found mostly in atrial,
`but not ventricular, cardiomyocytes [23,72,73].
`
`7. GLP-1 and endothelium function
`
`In several in-vitro studies, GLP-1 induced endothelial-
`dependent relaxation [74,75], an effect that is NO-dependent
`[75]. Also, in-vitro GLP-1 decreased reactive oxygen species
`(ROS) generation and subsequently reduced VCAM-1 mRNA
`levels in human umbilical vein endothelial cells (HUVECs)
`exposed to advanced glycation end-products (AGEs) [76],
`while several human studies reported beneficial effects with
`GLP-1 on endothelium function. In healthy non-diabetic
`subjects, GLP-1 infusion enhanced acetylcholine-induced
`forearm blood flow, as measured by venous occlusion
`phlethysmography [77]. In T2DM patients with stable coronary
`artery disease (but not in healthy subjects), GLP-1 vs placebo
`infusion significantly improved endothelial dysfunction by
`increasing flow-mediated vasodilatation in the brachial artery
`(3.1 ± 0.6% vs 6.6 ± 1.0%, respectively; p < 0.05), an effect
`that seems to be independent of insulin sensitivity [78]. In
`T2DM patients, GLP-1 infusion can also further increase
`the flow-mediated dilatation induced by insulin during a
`normoglycaemic – hyperinsulinaemic clamp, while further
`decreasing plasma levels of soluble intercellular adhesion
`molecule (sICAM)-1, plasma PGF2α, nitrotyrosine and IL-6,
`
`278911ANN_DM_NOVO_CS6_PC.indb 8
`
`14/03/2017 13:08:45
`
`Novo Nordisk Exhibit 2110
`Mylan Pharms. Inc. v. Novo Nordisk A/S
`IPR2023-00724
`Page 00006
`
`

`

`
`
`B. Vergès, B. Charbonnel / Diabetes & Metabolism 43 (2017) S3-S12
`
`2S9
`
`thereby demonstrating its vasodilatory, anti-inflammatory and
`antioxidant actions as well [38].
`Whereas the studies performed with native GLP-1 appear
`to indicate an ability to induce endothelial-depen