Glucagon-like receptor 1 agonists and DPP-4 inhibitors: potential therapies for the treatment of stroke

Abstract

During the past decades, candidate drugs that have shown neuroprotective efficacy in the preclinical setting have failed in clinical stroke trials. As a result, no treatment for stroke based on neuroprotection is available today. The activation of the glucagon-like peptide 1 receptor (GLP-1) for reducing stroke damage is a relatively novel concept that has shown neuroprotective effects in animal models. In addition, clinical studies are currently ongoing. Herein, we review this emerging research field and discuss the next milestones to be achieved to develop a novel antistroke therapy.

Introduction

Stroke is the primary cause of severe disability and the second most common cause of death according to the World Health Organization. In all, 85% of all strokes result from any major cerebral artery occlusion leading to cerebral ischemia, brain damage, and consequent neurologic impairments and disability.

Recombinant tissue plasminogen activator is the only FDA-approved pharmacological treatment for ischemic stroke. However, this treatment is not available for the majority of stroke patients due to short effective therapeutic window (up to 4.5 hours from stroke symptom onset) and increased risk of cerebral hemorrhage.

The ischemic cerebral region can be divided in two subregions: ischemic core and penumbra. The core is the area of infarction where neurons die extremely quickly after a stroke. The surrounding penumbra maintains residual blood flow and neurons could potentially be salvaged by timely intervention. However, if untreated, the penumbra will slowly progress into an infarction. So far, preclinically efficacious neuroprotective drugs targeting the penumbra have failed clinically.¹ The reasons for this failure are multiple and they range from differences in dosage, routes of administration, and timing of drug administration from the onset of stroke. In addition, preclinical efficacy experiments have often been performed in rodent models without typical comorbidities of stroke patients such as diabetes and hypertension, thus not mimicking the likely conditions of stroke patients.

Recent research has showed neuroprotective properties against stroke by drugs targeting the glucagon-like peptide-1 receptor (GLP-1R). Some of the stroke efficacy data has been achieved under preclinical conditions of clinical relevance. Furthermore, these substances are already in clinical use for the treatment of type 2 diabetes (T2D) and present a good safety profile and minimal side effects. Therefore, the potential repositioning of GLP-1R activating drugs into antistroke treatments seems promising.

The glucagon-like peptide-1 receptor

Glucagon-like peptide-1 receptor is a G-protein-coupled receptor that is expressed in a wide range of tissues including pancreas, heart, and brain.² It is activated by GLP-1; a small peptide hormone released from intestinal L cells and exerting numerous pleiotropic effects. The best-characterized property of GLP-1 is its incretin effect, e.g., enhancing meal-stimulated insulin secretion from pancreatic β cells in a glucose-dependent manner.² This effect accounts for the largest part of the postprandial insulin secretion in healthy subjects. Glucagon-like peptide-1 also decreases glucagon secretion from the pancreas. Since these effects are glucose dependent, activation of the GLP-1R carries a low risk of hypoglycemia. The signal transduction pathway of GLP-1 and its analogs has been characterized in pancreatic β cell and it mainly occurs via adenylate cyclase and the cAMP/PKA pathways.²

Endogenous GLP-1 is rapidly degraded by the enzyme dipeptidyl-peptidase 4 (DPP-4). Thus, despite its glycemic regulatory properties GLP-1 as such could not be developed clinically for the treatment of diabetes. However, there are todays several stable synthetic GLP-1R agonists resistant to DPP-4 degradation and DPP-4 inhibitors that are used clinically to treat T2D.

Glucagon-like peptide-1 receptor activation for the treatment of stroke

In the past few years, neuroprotection via GLP-1R activation has been shown in several animal models of Alzheimer's, Parkinson's, Huntington's, and traumatic brain injury. This research field is rapidly growing and several excellent reviews are available (e.g.³Holscher³). Herein, we specifically focus on data showing GLP-1R-mediated neuroprotection against stroke.

Glucagon-Like Peptide-1 Receptor Agonists

Exendin-4

Exendin-4 (Ex-4) is the first GLP-1R agonist that was developed by Amylin Pharmaceuticals (San Diego, CA, USA) as exenatide for the treatment of T2D. It shares a 53% amino-acid sequence homology with GLP-1. The safety profile of Ex-4 is good, although patients can present side effects such as nausea.²

Besides its metabolic properties, GLP-1Rs are expressed throughout the brain and Li et al⁴ showed that intracerebral administration of Ex-4 enhanced neuroprotection and locomotor activity after stroke in the rat. The authors also showed that the effect was mediated by GLP-1R since Ex-4 was ineffective in GLP-1R knockout mice. The work represents proof-of-concept for neuroprotection against stroke by GLP-1R activation. However, the findings have low clinical relevance since Ex-4 was given at stroke onset and via a route of administration not suitable for clinical applications.

Exendin-4 can pass the blood brain barrier and data have recently shown that peripheral administration of this compound leads to neuroprotection against stroke. Teramoto et al⁵ showed that Ex-4 is neuroprotective and improves neurologic deficit after stroke induced by transient middle cerebral artery occlusion (MCAO) after intravenous injections in mice.⁵ Efficacy was achieved when Ex-4 was given acutely at stroke onset or 1 hour later. At 3 hours, the neuroprotective effect was lost. The neuroprotective effect of Ex-4 was independent of glycemic effects and correlated with increased cAMP levels and decreased oxidative stress and inflammation. This study proved the concept for the efficacy of Ex-4 via peripheral administration. The work has also clinical relevance if the data are interpreted together with a recent work in our group by Darsalia et al⁶ showing that Ex-4-mediated neuroprotection could be achieved even if the treatment started 3 hours after MCAO. Although both works could not show the decreased stroke volume when Ex-4 was administered 3 hours after stroke, in the latter study, neuroprotection was detected by stereological counts of surviving neurons in cortex and striatum. The neuroprotective effect correlated with the increase of M2 reparative-microglia markers. This method appears to be more sensitive than stroke volume measurements in 'unmasking' neuroprotection.^{6, 7, 8} This work also showed that Ex-4 was neuroprotective up to 3 hours after stroke in aged obese/T2D mice. In conclusion, these results suggest that if administered rapidly after stroke Ex-4 could have neuroprotective potential in clinical perspective, in both normal and T2D patients.

Type 2 diabetes is a major risk factor for stroke. Furthermore, T2D and stroke are major causes of morbidity and mortality.⁹ To simulate the clinical situation of a diabetic patient experiencing a stroke, we showed that 4 weeks intraperitoneal pretreatment with Ex-4 before MCAO followed by another 4 weeks of Ex-4 treatment decreases brain damage in T2D rats.⁸ The effect was already significant at the clinically T2D dose (0.1 μg/kg) and occurred independently of the regulation of glycemia. In clinical perspective, the finding has the advantage that T2D patients could receive a therapy based on GLP-1R activation primarily against their diabetes (i.e., antihyperglycemic), while at the same time improving the stroke prognosis. Briyal et al¹⁰ used a similar strategy although the studies were performed in non-diabetic rats. The protective effect of Ex-4 via pretreatment could also be extended to hippocampus by using a cerebral ischemia model targeting the CA1 region in gerbils.¹¹ As for the studies where Ex-4 was administered after stroke, also in these studies antiinflammatory action by Ex-4 on microglia cells have been shown. In one recent study, Jin et al¹² also showed that hypoxia-inducible factor-1α regulation by Ex-4 may be involved in neuroprotection.¹²

Liraglutide

Liraglutide is a stable GLP-1 analog developed by Novonordisk with ~97% homology to GLP-1.² Similarly to Ex-4, liraglutide is used clinically against T2D, and the side effects are similar as for Ex-4.²

Interestingly, Sato et al¹³ showed that liraglutide administered intraperitoneally 2.5 hours after stroke onset induced neuroprotection in the rat in correlation with VEGF upregulation.¹³ As for the works of Teramoto et al and Darsalia et al described above, the data are clinically relevant since efficacy could be achieved in a poststroke setting compatible with the clinical situation. In another recent study, Briyal et al¹⁴ showed that pretreatment with liraglutide in normal and diabetic rats is neuroprotective against stroke in correlation with decreased apoptosis and oxidative stress. These studies could have clinical relevance since the liraglutide chronic regime of administration before stroke could improve stroke prognosis in T2D patients while exerting antidiabetic effects.

See Table 1A for a summary of the studies above.

Table 1A. Bibliometric overviews of preclinical studies and ongoing trials with GLP-1R agonists and DPP-4 inhibitors in stroke (A. Preclinical studies).

Author	Substance	Stroke model	Occlusion time (min)	Species	Administration route	No. of animals	Comorbidity	Main outcome measuresa	Ref. no.
Li, Y	Exendin-4	MCAO	60/90b	Rat/Mouseb	Iv. 15 minutes before MCAO	14/28b	None	SV, BT	4
Lee, CH	Exendin-4	Bil carotid occlusion	5	Gerbil	Ip. 2 hours before surgery and 1 hour after reperfusion	142	None	CC, GLP-1R Western Blot, BT	11
Teramoto, S	Exendin-4	MCAO	60	Mouse	Iv.. 0, 1 or 3 hours after reperfusion	81	None	SV, BT	5
Briyal, S	Exendin-4	MCAO	Permanent	Rat	Ip. 7 days before MCAO	30	None	SV, BT	10
Darsalia, V	Exendin-4	MCAO	90	Rat	Ip. 4 weeks before and after MCAO	42	Diabetes	SV, CC	8
Darsalia, V	Linagliptin	MCAO	30	Mouse	Oral 4 weeks before and 3 weeks after MCAO	42	Diabetes, Obese	SV, CC	7
Sato, K	Liraglutide	MCAO	90	Rat	Ip.	44	None	SV, BT	13
Yang, D	Alogliptin	3VO	15	Mouse	Oral 3 weeks before 3VO	40	None	SV, BT	20
Briyal, S	Liraglutide	MCAO	Permanent	Rat	Sc. 2 weeks before MCAO	30	None	SV, BT	14
Darsalia, V	Exendin-4	MCAO	30	Mouse	Ip. after MCAO	128	Diabetes, Obese, old	SV, CC, Inflammation	6

Abbreviations: BT, behavioral/motor tests; CC, cell counting; DPP-4, dipeptidyl-peptidase 4; GLP-1R, glucagon-like peptide-1 receptor; Ip., intraperitoneal; Iv., intravenous; Ivt., intraventricular; MCAO, middle cerebral artery occlusion; Sc., subcutaneous; SV, stroke volume; 3VO, three vessel occlusion.

PubMed searched 2015 January 08 for preclinical GLP-1 or DPP-4 studies examining stroke. Articles appear in the order of publication, with Li being the first.

^aMost studies report multiple outcome measures, here are only the main measures listed.

^bThe study by Li Y has two arms with rat or mouse. First value is mouse, second rat.

Clinical studies with glucagon-like peptide-1 receptor agonists

Retrospective database analysis on 39,275 patients with T2D has showed that exenatide-treated patients had decreased incidence of cardiovascular events, including stroke.¹⁵ A number of clinical trials with GLP-1R agonists are ongoing where the effect on stroke is one of the outcome measures (see Table 1B). While most of these studies are performed in diabetic patients focusing on combined cardiovascular outcome measures, a few of them are directly focusing on stroke patients. Our institution is running a clinical study where patients with suspected stroke are given exenatide en route in the ambulance (see Table 1B #1). Interestingly, a recent clinical feasibility study in 11 stroke patients with diabetes showed that exenatide treatment was safe and did not cause any serious adverse events, but mild nausea.¹⁶ To assess the cardiovascular safety of liraglutide, a phase 3B, multicenter, randomized, double-blinded, placebo-controlled clinical trial with long-term follow-up (LEADER) was initiated in 2010 and it is ongoing¹⁷ (see also Table 1B #6).

Table 1B. Bibliometric overviews of preclinical studies and ongoing trials with GLP-1R agonists and DPP-4 inhibitors in stroke (B. Ongoing clinical trials).

Substance	EU CTR Identifier no.	CT.gov Identifier no.	Study	Planned size	Primarily stroke trial	#
Exenatide (GLP-1R)	2011-002780-16	—	Prehospital study, hyperglycemic patients with suspected stroke	100	*	1
Exenatide (GLP-1R)	2013-001558-87	—	Exenatide in patients receiving thrombolytic therapy for stroke	40	*	2
Exenatide (GLP-1R)	2010-021069-63	NCT01144338	EXSCEL trial. Diabetics given exenatide as add on, CV events evaluated	14,000		3
Exenatide (GLP-1R)	2012-002219-25	NCT01455896	Diabetics, exenatide given with s.c. pump, CV events evaluated	3,000		4
Liraglutide (GLP-1R)	2011-003572-36	—	Nil by mouth stroke patients. Stroke volume measured by MRI	40	*	5
Liraglutide (GLP-1R)	2009-012201-19	NCT01179048	LEADER trial. Diabetics, time to CV event primary outcome. Effect and safety	8,754		6
Saxagliptin (DPP-4)			No ongoing studies in both databases			7
Alogliptin (DPP-4)			No ongoing studies in both databases			8
Sitagliptin (DPP-4)	2008-006719-20	NCT00790205	TECOS trial. Diabetics, time to CV events evaluated	14,757		9
Linagliptin (DPP-4)	2011-004148-23	NCT01897532	CARMELINA trial. Diabetics, time to CV events evaluated	10,790		10
Linagliptin (DPP-4)	2009-013157-15	NCT01243424	CAROLINA trial. Diabetics, time to CV event evaluated	6,000		11

Abbreviations: CV, cardiovascular; DPP-4, dipeptidyl-peptidase 4; GLP-1R, glucagon-like peptide-1 receptor; MRI, magnetic resonance imaging.

EU Clinical Trial Register (EU CTR) and ClinicalTrials.gov (CT.gov) databases searched with query ‘stroke'+'substance' on 2014 November 8. Each row constitutes a unique study. Retrospective noninterventional studies excluded.

Dipeptidyl-Peptidase 4 Inhibitors

Dipeptidyl-peptidase 4 is a serine aminopeptidase enzyme inactivating the incretins GLP-1 and glucose-dependent insulinotropic polypeptide through a dipeptide cleavage of the penultimate N-terminal amino acid.² The DPP-4 inhibition increases incretins short (1 to 2 minutes) half-life and due to this property DPP-4 inhibitors are clinically used to treat T2D. The DPP-4 inhibition also modulates the activity of other factors with potential neuroprotective properties.¹⁸ In contrast to Ex-4 and liraglutide, which are injectable drugs, DPP-4 inhibitors are orally administered.

Recent data have shown that intracerebral administration of the DPP-4 inhibitor sitagliptin reduced cortical lesions after MCAO in the rat.¹⁹ Although the work has low clinical relevance due to the employed route of administration, it was the first to show neuroprotection by DPP-4 inhibition. We recently showed that 4 weeks pretreatment with clinical doses of the DPP-4 inhibitor linagliptin (orally) in both normal and T2D/obese mice reduced neuronal loss after MCAO.⁷ As expected, the effects of linagliptin correlated with increased plasma GLP-1. However, the neuroprotective effect appeared to be independent of glycemic control. In agreement with this finding, Yang et al²⁰ showed that 3 weeks treatment with the DPP-4 inhibitor alogliptin before stroke induced by the 3-vessel occlusion technique in normal mice reduced stroke volume. The neuroprotective effect correlated with increased brain derived neurotrophic factor.

Although these results are interesting, the responsible mechanisms remain unclear. Unlike GLP-1R agonists, DPP-4 inhibitors do not cross the blood brain barrier. Furthermore, DPP-4 inhibition does not induce an increase in plasma GLP-1 equivalent to pharmacological doses of Ex-4 or liraglutide. Finally, opposite to what has been observed for GLP-1R agonists, antistroke efficacy by alogliptin was lost when the treatment started after stroke onset.²⁰ We also achieved similar results to those with alogliptin by using linagliptin (unpublished). In conclusion, these data indicate that the biology at the basis of neuroprotection mediated by DPP-4 inhibitors is complex and still largely unexplored.

See Table 1A for a summary of the studies above.

Clinical studies with dipeptidyl-peptidase 4 inhibitors

A recent meta-analysis study showed lower incidence of cardiovascular diseases (CVD) in diabetic patients receiving DPP-4 inhibitors.²¹ A larger 2-year efficacy/safety study comparing linagliptin with glimepiride in T2D patients suggests that linagliptin has a beneficial action on CVD.²² Recently, a small study including 378 patients showed that DPP4 inhibitors treatment before stroke showed a trend for less severe stroke compared with untreated patients.²³ However, no effect of DPP-4 inhibition on CVD end points, including stroke, was shown in two recently published large cardiovascular outcome trials for saxagliptin²⁴ and alogliptin.²⁵ Further, analysis of US commercial insurance including 79,000 patients did not show differences between untreated or DPP-4 inhbitor-treated patients regarding CVD end points including stroke.²⁶Table 1B lists ongoing trials with DPP-4 inhibitors. It remains to be shown in the coming trials of TECOS (sitagliptin, see Table 1B #9), CARMELINA (linagliptin, see Table 1B #10), and CAROLINA (linagliptin, see Table 1B #11) if DPP-4 inhibitors lower stroke incidence and/or improve stroke outcome. In particular, CARMELINA contains a poststroke functional substudy using the modified RANKIN score to assess stroke disability approximately one week after stroke and at ~3 months after stoke onset.

Of note is that no DPP-4 inhibitor studies are primarily stroke trials.

Potential mechanisms and future directions

During the past few years, experimental evidence has shown that GLP-1R activation is efficacious against stroke in various rodent models. However, the molecular/cellular mechanisms at the basis of GLP-1R-mediated neuroprotection are still largely unknown. Altogether the results point out direct, antiapoptotic and neuroprotective mechanisms and indirect antiinflammatory mechanisms involving microglia regulation (Figure 1). These data are also supported by the fact that GLP-1R has been found on both neurons²⁷ and microglia.²⁸ Since neuroprotective efficacy by GLP-1R activation can be achieved both after a chronic pretreatment before stroke and acutely after stroke, it will be important to determine whether the protective mechanisms are the same. Interesting, recent literature by Gejl et al has shown that in response to plasma glucose pathologic alterations, GLP-1 reduces brain glucose levels fluctuation by regulating the brain vasculature.²⁹ Hyperglycemia worsens the stroke outcome and this mechanism may prove to be neuroprotective during hyperglycemia. Also, it is not known whether neuroprotection mediated by GLP-1R agonists and DPP-4 inhibitors occur through the same mechanism of action and whether neuroprotection by DPP-4 inhibition is indeed GLP-1R dependent. These questions will need to be answered by performing efficacy studies using GLP-1R antagonists and mice lacking the glp-1r.

Figure 1.

Neuroprotective mechanisms against stroke induced by glucagon-like peptide-1 receptor (GLP-1R) agonists and dipeptidyl-peptidase 4 (DPP-4) inhibitors. GLP-1R agonists and DPP-4 inhibitors could exert neuroprotection via common or independent mechanisms. Evidence is accumulating that direct antiapoptotic survival effects on neurons and indirect effects through the regulation of microglia could be at the basis of acute neuroprotection by GLP-1R agonists. In addition to stimulate neuroprotection by increased plasma GLP-1 levels, DPP-4 inhibitors could also modulate the activity of other neuroprotective factors that will need to be identified. Other potential mechanisms ranging from effects on adult neurogenesis and on synaptic plasticity will need to be investigated to understand whether the effect of both GLP-1R agonists and DPP-4 inhibitors can have a role in late stroke recovery.

The GLP-1R activation has been reported to increase neurologic recovery and to improve learning and memory in different animal models of neurodegenerative disorders.³ In addition, GLP-1R activation promotes synaptic plasticity and neurite outgrow³ and can stimulate adult neurogenesis³⁰ (Figure 1).³⁰ The findings could form the bases for potential regenerative therapy for chronic stroke patients. At this regard, experimental data and future work are highly needed.

Conclusions

Preclinical efficacy studies and clinical results indicate that an antistroke therapy based on GLP-1R activation might be translated into the clinical practice. A number of these molecules are already in clinical use for the treatment of T2D, making this potential translation easier since these molecules are rather safe. If translation is possible, then the preclinical data indicate that diabetic/high-risk stroke patients receiving GLP-1R therapy before stroke would benefit mostly from this type of treatment. Whether a poststroke acute treatment can lead to neuroprotection in a clinical relevant manner remains to be investigated since the protective effect of GLP-1R agonists in rodents decreases quickly after stroke onset. If so, it would require that the drug administration occurs soon after a stroke event; possibly already during the ambulance transport. Finally, safety/feasibility studies assessing whether this type of therapeutic strategy is possible will need to be completed and analyzed in the near future.

Work in CP's laboratory is supported by Boehringer Ingelheim Pharma GmbH & Co. TK is an employee of Boehringer Ingelheim Pharma GmbH & Co. TN is on the national advisory board of Eli Lilly, Novonordisk, and Sanofi.