Potential therapeutic effects of mTOR inhibition in atherosclerosis

Abstract

Despite significant improvement in the management of atherosclerosis, this slowly progressing disease continues to affect countless patients around the world. Recently, the mechanistic target of rapamycin (mTOR) has been identified as a pre‐eminent factor in the development of atherosclerosis. mTOR is a constitutively active kinase found in two different multiprotein complexes, mTORC1 and mTORC2. Pharmacological interventions with a class of macrolide immunosuppressive drugs, called rapalogs, have shown undeniable evidence of the value of mTORC1 inhibition to prevent the development of atherosclerotic plaques in several animal models. Rapalog‐eluting stents have also shown extraordinary results in humans, even though the exact mechanism for this anti‐atherosclerotic effect remains elusive. Unfortunately, rapalogs are known to trigger diverse undesirable effects owing to mTORC1 resistance or mTORC2 inhibition. These adverse effects include dyslipidaemia and insulin resistance, both known triggers of atherosclerosis. Several strategies, such as combination therapy with statins and metformin, have been suggested to oppose rapalog‐mediated adverse effects. Statins and metformin are known to inhibit mTORC1 indirectly via 5' adenosine monophosphate‐activated protein kinase (AMPK) activation and may hold the key to exploit the full potential of mTORC1 inhibition in the treatment of atherosclerosis. Intermittent regimens and dose reduction have also been proposed to improve rapalog's mTORC1 selectivity, thereby reducing mTORC2‐related side effects.

Introduction

Atherosclerosis is a chronic inflammatory disease in the arterial wall of large and medium‐sized arteries and the leading cause of mortality and morbidity in modern societies, claiming at least one out of every three deaths 1. The aetiology of this slowly progressing disease is associated with an imbalance in lipid metabolism as well as a chronic inflammation of the arterial wall 2, 3. Endothelial dysfunction following mechanical or chemical stress allows infiltration and retention of modified low‐density lipoprotein particles in the intima as well as the recruitment of monocytes and other inflammatory cells 4, 5, 6. By digesting low‐density lipoprotein (LDL) particles, tissue macrophages within the arterial wall transform into foam cells 5, 7. As cell death is a hallmark of advanced lesions, and clearance of dead cells by macrophages (a process known as efferocytosis) is impaired in advanced plaques, necrotic debris as well as lipids accumulate and lead to the formation of a necrotic core 8, 9. As plaque growth continues asymptomatically, the inner core becomes hypoxic and the vasa vasorum respond by initiating neovascularization to supply the hypoxic areas. Plaque neovascularization is an important risk factor for plaque rupture and clinical symptoms 10, 11, 12.

The use of lipid‐lowering 3‐hydroxy‐3‐methylglutaryl‐coenzyme A (HMG‐CoA) reductase inhibitors (statins) has led to better management of atherosclerosis, thereby reducing morbidity and mortality rates 13, 14. A considerable fracton of patients, however, does not benefit from this therapy 15, 16. For such patients, the novel proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitors may be of great value. These injectable‐only drugs, already approved by the US Food and Drug Administration and available, reduce LDL cholesterol levels by 61% on top of statin therapy and are also able to decrease significantly the onset of cardiovascular events 17, 18. However, while lipid‐lowering drugs are crucial in atherosclerosis management, the inflammatory nature of the disease cannot be ignored. Extensive research has shown that the inhibition of the mechanistic target of rapamycin (mTOR) may offer new avenues leading to stabilization, and even possibly regression, of atherosclerotic plaques 19, 20, 21.

The mechanistic target of rapamycin

The highly conserved, constitutively active serine/threonine kinase mTOR is part of two distinct multiprotein complexes: mTORC1 and mTORC2 (Figure 1) 22. As a result, the two complexes have different cellular functions 23. Under normal conditions, mTORC1 promotes protein synthesis, proliferation, lipogenesis, growth and energy metabolism. These activities are achieved through mTORC1 targets such as S6 kinase 1 (S6K1), 4e‐binding protein‐1 (4eBP‐1), cyclin‐dependent kinases (CDKs) and the hypoxia‐inducible factor 1α (HIF1α), which is involved in the expression of a variety of glycolytic genes 21. Moreover, mTORC1 has been implicated in the activity and biogenesis of mitochondria 24, 25. As such, mTORC1 is very sensitive to nutrient deprivation, intracellular energy status and the availability of growth factors. Loss of mTORC1 signalling results in inhibition of cellular growth, protein synthesis and metabolism as well as the induction of autophagy, a catabolic process for the destruction of organelles and long‐lived proteins aimed at restoring energy levels in the cell 21, 26. Inhibition of mTORC1 is also known to promote longevity in mice 27, 28, whereas over‐activation of mTORC1 is associated with several cancers 29. The less‐studied mTORC2 is known to be responsible for cytoskeletal organization, lipolysis, insulin sensitivity and the activation of several kinases, including protein kinase B (AKT) and the serum/glucocorticoid‐regulated kinase 1 (SGK1) 30. Inhibition of mTORC2 results in insulin resistance and decreases the life span of male mice 31. It also leads to several defects in neutrophil polarization and inhibits cellular migration by impeding cytoskeletal reorganization 32. Recent data suggest that the two mTOR complexes regulate each other's activity through cross‐talk mechanisms 33, 34.

Figure 1.

Schematic representation of the protein complexes mTORC1 and mTORC2. Whereas both complexes feature mTOR, Deptor and mLST8, mTORC1 engages with Raptor, Rheb‐GTP and Pras40 while mTORC2 is associated with Rictor, Protor‐1 and mSin1. Accordingly, both complexes control distinct cellular functions. Rapalogs are able directly to inhibit mTORC1 and, when given chronically, can also disrupt mTORC2 signalling. Other drugs, such as metformin, inhibit mTORC1 via indirect means, such as activation of REDD1 and AMPK, while inhibiting RAG‐GTPase. AMPK, 5' adenosine monophosphate‐activated protein kinase α; mLST8, mammalian lethal with SEC13 protein 8; mTOR, mechanistic target of rapamycin; mTORC, mechanistic target of rapamycin complex; mSin1, mammalian stress‐activated protein kinase‐interacting protein 1; REDD1, regulated in development and DNA damage responses 1; ATP, adenine triphosphate; RAG, ras‐related GTPase; Rheb, ras homolog enriched in brain; Pras40, proline‐rich AKT substrate of 40 kDa; Raptor, regulatory associated protein of mTOR; Deptor, DEP domain containing mTOR‐interacting protein; CDK's, cyclin dependent kinases; 4eBP‐1, eukaryotic initiation factor 4E binding protein 1; Srebp1, sterol regulatory element‐binding protein 1; HIF‐1a, Hypoxia‐inducible factor 1a; VEGF, vascular endothelial growth factor; S6K1, p70 ribosomal protein S6 kinase 1; IRS‐1, insulin receptor substrate 1; AKT, protein kinase B; Protor‐1, protein observed with Rictor 1; Rictor, rapamycin insensitive companion of mTOR; PKCa, protein kinase C a; SGK1, serum and glucocorticoid‐induced protein kinase 1

Pharmacological inhibition of mTOR

The natural anti‐inflammatory macrolide rapamycin, also known as sirolimus, was already being used in the 1970s as an immunosuppressant in autoimmune disorders or following organ transplantations 27, 35. Rapamycin's ability to inhibit mTOR was discovered 20 years later, and, given its potent anti‐proliferative properties, it was rapidly introduced into the field of drug‐eluting stents (DES) and cancer 36, 37. While the drug is known to be a potent inhibitor of mTORC1, its effects on mTORC2 are somewhat controversial and will only appear following chronic administration 38, 39. Rapamycin binds to the intracellular immunophilin 12‐kDa FK‐506 binding protein (FKBP12). The resulting complex physically interacts with mTOR, disturbing its association with raptor and obstructing its signalling 40. Interestingly, the naturally produced macrolide immunosuppressant tacrolimus has a structure that is almost identical to that of rapamycin (Figure 2) and can also form a complex with FKBP12. However, tacrolimus inhibits calcineurin and does not interfere with mTORC1 41. A polyene region in rapamycin's structure appears to be crucial in this context 42, 43. Owing to the poor bioavailability of rapamycin and its long half‐life (up to 70 h) 44, 45, several semi‐derivatives, collectively known as rapalogs, have been synthesized, featuring improved pharmacokinetic and pharmacodynamic properties. These semi‐derivatives include everolimus (RAD‐001), zotarolimus (ABT‐578), ridaforolimus (AP‐23573), umirolimus and the rapamycin's pro‐drug temsirolimus (CCl‐779) 27, 46, 47. Everolimus is widely used in new‐generation DES and in the novel bioresorbable vascular scaffold systems (BVS), with very promising results 48, 49. Due to mTOR over‐activation in many tumours, everolimus has also been approved for the treatment of hormone receptor‐positive breast cancer and, together with temsirolimus, for renal cell carcinoma 50, 51. Clinical trials featuring mTORC1 inhibitors in other cancer types are ongoing, although with mixed results 52. It should be noted that a new class of adenosine triphosphate (ATP)‐competitive mTOR inhibitors that target both mTOR complexes has been synthesized and approved for clinical trials 53, 54. However, these compounds are only intended to treat certain cancer types and are beyond the scope of this review. Yet, it is worthwhile to mention that inhibition of mTORC1 can also be achieved via indirect mechanisms such as activation of the 5' adenosine monophosphate‐activated protein kinase α (AMPK), activation of sirtuin1 and regulation of intracellular calcium concentrations 37, 55. One of the best known indirect mTORC1 inhibitors is the anti‐diabetic drug metformin (Figure 1).

Figure 2.

Chemical structure of sirolimus (or semi‐derivatives thereof) and tacrolimus. (A) Different functional groups can be added to sirolimus, resulting in several semi‐derivatives with improved pharmacokinetics. (B) Unlike sirolimus, tacrolimus has no polyene mTOR‐binding region. Thus, while both drugs form complexes with FKBP12, tacrolimus lacks the ability to influence mTOR signalling. mTOR, mechanistic target of rapamycin

mTORC1 inhibition in atherosclerosis

The introduction of DES coated with rapalogs ushered in a revolution in the field of interventional cardiology 56. Rapalog‐coated DES were superior both to bare metal stents (BMS) 57, 58, 59, 60, 61 and DES coated with the proliferation inhibitor paclitaxel 62, 63. Unlike rapalogs, paclitaxel inhibits smooth muscle cell (SMC) proliferation by stabilizing microtubule chains 64. Inhibitors of mTORC1 are therefore likely to possess beneficial mechanisms reaching beyond proliferation arrest in SMCs. Indeed, implantation of stents eluting everolimus in atherosclerotic arteries of cholesterol‐fed rabbits resulted in a remarkable clearance of macrophages throughout the plaque without altering SMC content 65. This effect can be explained by the fact that plaque macrophages are metabolically very active and thus considerably more sensitive to the inhibition of mTORC1‐driven protein synthesis. Macrophage clearance by everolimus‐eluting stents was associated with massive vacuole formation, a hallmark of autophagy 65. Similarly, a full‐body knockdown of mTOR in ApoE^‐/‐ mice was also found to remove macrophages selectively from the plaque by autophagy 66. Owing to the systemic inflammatory nature of atherosclerosis, it would make more sense to administer mTORC1 inhibitors systemically rather than as a stent coating. The well‐tolerated rapalogs can be delivered in animal models orally, intraperitoneally or subcutaneously via an osmotic minipump. Different studies in ApoE^‐/‐ and low‐density lipoprotein receptor‐deficient (LDLR^‐/‐) mice 67, 68, 69, 70, 71, 72 as well as in rabbits 73, 74, have shown that systemic administration of rapamycin and everolimus is effective in reducing plaque size and complexity (Table 1). Moreover, after 3 years of follow‐up in humans, oral administration of rapamycin during the first 14 days following implantation of a BMS proved to be superior to BMS alone and had only minimal side effects, such as gum sores or diarrhoea 75. Importantly, it was also shown that this combination was as effective as the use of DES 76. As for indirect mTORC1 inhibitors, metformin is known to reduce neointimal formation through inhibition of SMC proliferation and migration as well as via potent induction of autophagy, making it the only anti‐diabetic drug with proven reduction of micro‐ and macrovascular complications in diabetic patients 37, 77, 78.

Table 1.

Overview of animal studies featuring mechanistic target of rapamycin complex 1 (mTORC1) inhibition in atherosclerosis

Treatment	Administration route	Dose (mg kg–1 day–1)	Animal model	Plaque reduction (%)	Serum lipids	Macrophage content	Ref
Rapamycin	Oral	0.01–0.02	ApoE‐/‐ mice	7–48	Cholesterol and TG unchanged	Decreased	67
Rapamycin	Oral	5	ApoE‐/‐ mice	45	Total cholesterol increased	ND	69
Rapamycin	s.c.	1–4	ApoE‐/‐ mice	56–66	Total cholesterol, VLDL, LDL, HDL unchanged	ND. MCP‐1, CCL2 decreased	68
Rapamycin	s.c.	1–8	ApoE‐/‐ mice	45–60	LDL and HDL increased, VLDL and TG unchanged	ND. IFN‐α and IL‐12 decreased	72
Rapamycin	i.p.	3	ApoE‐/‐ mice	ND	Total cholesterol unchanged, TG increased	ND. MMP‐2 decreased, MMP‐9 increased	98
Rapamycin	Oral	0.1–1	LDLR‐/‐ mice	16–62	Unchanged	ND. IL‐6, MCP‐1, IFN‐α, TNF‐α and CD40 decreased	70
Rapamycin	Oral	0.5	Rabbits	72	Unchanged	Decreased. MCP‐1 and MMPs 1, 2, 3, 9, 12 decreased	74
Rapamycin	Oral	0.5	Rabbits	40–50	Unchanged	Decreased	97
Everolimus	s.c.	0.05–1.5	LDLR‐/‐ mice	44–85	VLDL and LDL increased TG unchanged	Increased. IL‐1α, IL‐5, IL‐12 and GM‐CSF decreased	71
Everolimus	Oral	1–5	LDLR‐/‐ mice	0.4–40	VLDL and LDL increased TG unchanged	Decreased	125
Everolimus	Oral	1	Rabbits	38	Total cholesterol unchanged	Decreased	73
mTOR siRNA	–	–	ApoE‐/‐ mice	36	Total cholesterol, LDL and TG decreased	Decreased. MCP‐1 and MMP‐2 decreased	66

GM‐CSF, granulocyte‐monocyte colony‐stimulating factor; HDL, high‐density lipoprotein; IFN, interferon; IL, interleukin; i.p., intraperitoneal; LDL, low‐density lipoprotein; LDLR^‐/‐, low‐density lipoprotein receptor‐deficient; MCP‐1, monocyte chemoattractant protein‐1; MMP, matrix metalloproteinase; ND, not determined; s.c., subcutaneous; siRNA, small interfering RNA; TG, triglycerides; TNF‐α, tumour necrosis factor α; VLDL, very low density lipoprotein.

Pleiotropic anti‐atherosclerotic effects of mTORC1 inhibitors

Apart from the proliferation arrest discussed above, inhibiting mTORC1 initiates a cluster of anti‐atherosclerotic responses, many of which are aimed at decreasing macrophage content in the plaque (Figure 3). Targeting mTORC1 decreases monocyte migration by reducing chemokines such as monocyte chemoattractant protein‐1 (MCP‐1), which is produced by endothelial cells and SMCs to recruit monocytes into the intima 66, 73, 79. Rapamycin also promotes a decline in monocyte adhesion by crippling hyaluronan production in SMCs 80. In addition, everolimus treatment in ApoE^‐/‐ mice reduces the number of circulating LyC6^high proatherogenic monocytes without affecting the less inflammatory LyC6^low subtype (Kurdi A. et al., unpublished). In vivo results in LDLR^‐/‐ mice have shown that everolimus treatment also mediates a decrease in circulating proatherogenic inflammatory cytokines such as interleukin (IL) 1α, IL‐5, IL‐12 and the granulocyte‐monocyte colony‐stimulating factor (GM‐CSF) 71. These results were in contrast to in vitro experiments in primary macrophages showing that mTORC1 inhibition by everolimus causes massive cytokine release and shifts macrophages to a high inflammatory status 81. However, recent data from our laboratory showed that this effect was only seen with supra‐clinical concentrations (10 µmol l^–1), which are unlikely to be achieved in vivo (Kurdi A. et al., unpublished).

Figure 3.

Inhibition of mTORC1 in atherosclerosis. Inhibitors of mTORC1 counteract atherosclerosis on several levels by reducing: (1) chemokine levels; (2) monocyte adhesion and migration; (3) monocyte/macrophage proliferation; (4) endothelial dysfunction; (5) foam cell formation; (6) macrophage inflammatory responses; (7) hypoxia‐inducible factor‐1α (HIF 1α) production; and (8) intraplaque angiogenesis. IL, interleukin; MMP, matrix metalloproteinase; mTORC, mechanistic target of rapamycin complex; TNF‐ α, tumour necrosis factor α

It is known that angiogenesis is induced by hypoxia and associated with the progression and destabilization of atherosclerotic plaques 82, 83, 84, 85. Intraplaque (IP) neovessels contribute to the ever‐increasing monocyte accumulation in advanced atherosclerosis and deliver lipids and inflammatory cells to high‐risk areas in the plaque. Moreover, IP neovessels are fragile and leaky, resulting in IP haemorrhages, iron deposition, macrophage activation and inflammation 86, 87. The ability of mTORC1 to regulate angiogenesis has been widely recognized in ischaemic heart injury and tumour models 88, 89, although recent data seem to indicate a role of mTORC2 in this process 90, 91. Nonetheless, inhibitors of mTORC1 may contribute to plaque stability by restraining IP neovessel formation. Indeed, mTORC1 inhibition blunts the expression of hypoxia‐inducible factor‐1α (HIF‐1α), which is a protein associated with rupture‐prone atherosclerotic plaques 21, as well as its downstream target vascular endothelial growth factor (VEGF), both of which are key regulators of angiogenesis 92, 93. Moreover, macrophages are known to initiate angiogenesis and, by reducing their numbers in the plaque, mTORC1 inhibitors also decrease angiogenesis indirectly 94, 95. Recent evidence indicates that everolimus treatment prevents the formation of neovessels in atherosclerotic plaques of ApoE^‐/‐ Fbn1^C1039G+/‐ mice (Kurdi A. et al., unpublished). This novel mouse model of advanced atherosclerosis and plaque rupture spontaneously develops IP neovessels and haemorrhages after feeding a Western‐type diet 96, features that were only sporadically present in other animal models of atherosclerosis.

With monocyte recruitment brought to a minimum, selective cell death mediated by autophagy finally clears resident macrophages within the plaque as discussed above 65, 97. As such, despite the halt of protein synthesis, the amount of total collagen in the plaque is increased because of diminished matrix metalloproteinase (MMP) activity, thus promoting plaque stability 74, 98. In addition to selective macrophage clearance from the plaque, autophagy is involved in other atheroprotective mechanisms, including anti‐inflammatory responses, stimulation of cholesterol efflux and inhibition of foam cell formation. Indeed, plaques are larger and more inflamed in ApoE^‐/‐ and LDLR^‐/‐ mice bearing autophagy‐deficient macrophages as compared with autophagy‐competent control animals 99, 100. Moreover, autophagy deficiency results in decreased cholesterol efflux out of foam cells 101. In these cells, over‐activated mTORC1 resulted in autophagy inhibition and the use of small interfering RNA to interfere with mTORC1 stimulated autophagy and impaired foam cell formation in vitro 102.

Inhibition of mTORC1 can also affect the haemodynamic system. Indeed, it has been reported that mTORC1 inhibition reduces arterial stiffness, one of the major causes of endothelial dysfunction (ED) 103. However, this was only observed after implantation of DES coated with everolimus and not with rapamycin 104, 105, 106, 107. Nonetheless, systemic rapamycin has been shown to increase endothelial nitric oxide synthase activity specifically in atherosclerosis‐prone regions of low shear stress in vivo, thus counteracting ED and atherosclerosis at its most basic level 108. The role of mTORC1 inhibition in hypertension is somewhat controversial. Although some reports indicate beneficial effects 109, 110, others suggest that the opposite is true (Table 2) 111, 112.

Table 2.

Adverse effects associated with mechanistic target of rapamycin complex 1 mTORC1 inhibition

Frequency	Adverse effect	Management
Very common (10% or more)	Dyslipidaemia	Statin or PCSK9 inhibitor therapy
	Hyperglycaemia	Metformin therapy
	Hypertension	Monitoring
	Anaemia	Anaemia screening every 3 months
		Iron supplementation
		Erythropoiesis stimulation therapy
	Dermatological disorders	Monitoring
		Topical treatment
	Impaired stent endothelialization	Dual anti‐platelet therapy
Common (1–10%)	Proteinurea	ACE inhibitor therapy
		Drug withdrawal
	Pneumonitis	Perform pulmonary function tests
		Rule out infection
		Dose reduction
		Corticosteroid + antibiotic therapy

ACE, angiotensin‐converting enzyme; PCSK9, proprotein convertase subtilisin/kexin type 9. Table modified after Kaplan et al. 111.

Recently, mTORC1 inhibition by rapamycin was found to prevent the osteoplastic differentiation of isolated human SMCs 113. Calcification of the plaque is well known to be associated with advanced atherosclerosis and increased cardiovascular adverse events. However, the size of arterial calcification is critical for its role in the plaque. While macro‐calcification is thought to be the result of a healing process that decreases IP inflammation 114, 115, micro‐calcifications are considered as active plaque destabilizers which increase the occurrence of plaque rupture and cardiovascular events 116. Therefore, inhibition of calcification is a double‐edged sword and more studies are needed to uncover the effect of mTORC1 inhibition on this process.

Clinical suitability

Although rapalogs are well tolerated, mTORC1 inhibition causes several side effects, which is plausible considering its central role in a variety of signalling pathways. Some reactions are mild to moderate, including rash, oedema, delayed wound healing and stomatitis. However, severe adverse effects, such as proteinuria, pneumonitis, thrombocytopenia and anaemia, are not uncommon and may form a serious limitation to their use 27, 117. Many of these side effects can be related to direct effects of rapamycin and were successfully reduced following the introduction of semi‐derivatives with improved pharmacokinetic profiles, such as everolimus 111. Moreover, different approaches, including dose reduction and combination therapy, have been proposed to better manage these adverse effects (Table 2) 111. For example, after the implantation of DES coated with rapalogs, the proliferation arrest mediated by mTORC1 inhibition delays endothelialization of the stented vessel, thereby increasing the risk of thrombosis 118. Systemic administration of rapamycin also results in increased tissue factor activity 119. Therefore, dual anti‐platelet therapy has become a standard practice following the implantation of DES.

Rapalogs are also known to induce metabolic and haemodynamic adverse effects which paradoxically contribute to atherosclerosis. Indeed, hypercholesterolaemia and hypertriglyceridaemia are widely observed in transplant patients receiving everolimus 117. Dyslipidaemia is also observed in animal models of atherosclerosis such as ApoE^‐/‐ and LDLR^‐/‐ mice treated with mTORC1 inhibitors 69, 71.

Inhibition of mTORC1 activity reduces the clearance of circulating lipoproteins by inhibiting: (i) the activity of lipases; (ii) the ability to deposit lipids in tissues; and (iii) the production of bile acids 120, 121, 122, 123. A decrease in hepatic LDL receptors has also been reported 124 but the latter mechanism is less likely to play a major part in rapalog‐induced dyslipidaemia because hypercholesterolaemia occurs in everolimus‐treated LDLR^‐/‐ mice as well 71, 125. However, despite dyslipidaemia, atherosclerotic plaques of treated animals are reduced in size and complexity, and lipid deposition in arteries is significantly decreased 71, 97. Lipid metabolism is controlled by both mTOR complexes. As mTORC1 promotes adipogenesis and lipogenesis 126, 127, mTORC2 positively controls hepatic lipogenesis and prevents lipolysis in white adipose tissue 39. The latter could make mTORC2 inhibition important in rapalog‐mediated dyslipidaemia.

Five per cent to 32% of patients treated with everolimus develop new‐onset diabetes mellitus (NODM) as a result of hyperglycaemia and insulin resistance 111. Physicians are therefore advised to monitor high‐risk NODM patients treated with rapalogs closely. Although no conclusive evidence exists for the role of hyperglycaemia itself in cardiovascular disease 128, 129, diabetes is known to accelerate the development of atherosclerosis through dyslipidaemia, endothelial dysfunction and induced inflammation 130, 131. Incidentally, mTORC1 inhibitors normalize endothelial function and decrease inflammatory cytokines as described above.

Hyperactivation of mTORC1 can cause insulin resistance by increased phosphorylation of the mTORC1 target S6 ribosomal protein (S6rp). S6rp, in turn, inhibits the insulin receptor substrates 1/2 (IRS‐1/2) and AKT 132. Alternatively, both caloric restriction (which results in mTORC1 inhibition) and deletion of the mTORC1 target S6rp are known to increase insulin sensitivity 133. It was therefore tempting to investigate the role of mTORC1 inhibitors in preclinical diabetic settings. Acute treatment with rapamycin (single injection) increased insulin sensitivity and glucose uptake as expected 132, 134, 135, 136. When it was given chronically, however, rapamycin paradoxically led to glucose intolerance in mice, rats and humans 111, 133, 137, 138, 139. This effect was, at least partly, mediated by mTORC2 inhibition following chronic rapalog administration as mTORC2 has recently been identified to be a critical mediator for insulin sensitivity 133, 136; another explanation is mTORC1 resistance following chronic rapalog treatment. In vitro experiments using renal carcinoma cells (RCC) revealed that long‐term everolimus treatment results in hyperphosphorylation of S6rp, a crucial mediator in insulin resistance 140, 141, 142. Along this line, we have obtained in vivo evidence that chronic inhibition of mTORC1 in mice treated with everolimus paradoxically results in over‐activation of mTORC1 as well as in diminished autophagy (Kurdi et al., unpublished results). The ability of rapalogs to shift their actions based on the duration of their administration or dosage is still poorly investigated, despite its importance in the clinic. In anti‐tumour therapy, for example, chronic administration of rapalogs is known to induce resistance to the drugs through different mechanisms 141, 142, 143, 144.

In an attempt to overcome these complications, combination therapy has been proposed to counter rapalog‐induced dyslipidaemia and glucose intolerance 21, 111. Statins and metformin have long been known to possess pleiotropic anti‐atherosclerotic effects beyond their main mechanism of action 145, 146, 147, 148. Both drugs induce AMPK at clinically relevant doses 149, 150, thereby further inhibiting mTORC1 and activating autophagy 37. This could help to reduce the dose of rapalogs and may eliminate several adverse effects. Intriguingly, metformin is also able to reduce plasma LDL levels in diabetic patients, adding to its significance in the prevention of cardiovascular diseases 151. Apart from drug combinations, a strategy based on intermittent dosing regiments of rapalogs could be used to avoid the development of mTORC1 resistance following chronic administration 152, 153. In addition, lower doses of rapamycin can improve its selectivity towards mTORC1 and may deteriorate the onset of drug resistance or mTORC2 inhibition 137.

Concluding remarks

Accumulating evidence suggests that mTOR plays a major role in the pathology of atherosclerosis. However, because of the diverse and complex roles that mTOR fulfils, it is difficult to define which mechanism is responsible for its anti‐atherosclerotic effects. The present review focused on rapalog‐mediated mTORC1 inhibition because rapalogs are extensively investigated in preclinical atherosclerosis models. Even though treatment with rapalogs results in adverse effects, such as dyslipidaemia and insulin resistance, that are known to exacerbate atherosclerosis, the net beneficial effect is indisputable, suggesting a mechanism or a combination of mechanisms powerful enough to counter these unwanted responses.

The choice of a rapalog is critical as side effects triggered by members of this class can vary significantly. Semi‐derivatives of rapamycin, such as everolimus, seem to hold a distinct advantage over the parent compound 111, 154. Moreover, the time span and frequency of administration as well as the chosen dosage can considerably influence the beneficial and adverse effects of a rapalog. Indeed, mTORC2, which was thought to be rapalog‐insensitive, has been shown to be affected by chronic treatment and to be responsible for some serious undesired effects. Chronic treatment may also be responsible for the development of mTORC1 resistance. Future studies should therefore be designed with much more care to avoid such complications.

In the absence of selective, resistance‐free mTORC1 inhibitors, an important clinical aspect is the availability of drug combinations to counteract rapalog‐mediated adverse effects and to boost their anti‐atherosclerotic properties, allowing for dose reduction of rapalogs. Statins and metformin inhibit mTORC1 155 and stimulate autophagy via AMPK activation, and should be considered in this context. Interestingly, in a similar way as rapamycin, these drugs are able to increase longevity in rodent ageing models 21.

Taken together, we are hopeful that focused research can result in clinical guidelines which will allow us to take full advantage of the anti‐atherosclerotic benefits following mTORC1 inhibition while limiting the need to cope with serious adverse effects.

Competing Interests

All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf and declare: no support from any organization for the submitted work; no financial relationships with any organizations that might have an interest in the submitted work in the previous 3 years; no other relationships or activities that could appear to have influenced the submitted work.

This work was supported by the Fund for Scientific Research (FWO)‐Flanders (project G.0160.13 N) and the University of Antwerp (BOF).

Kurdi, A. , De Meyer, G. R. Y. , and Martinet, W. (2016) Potential therapeutic effects of mTOR inhibition in atherosclerosis. Br J Clin Pharmacol, 82: 1267–1279. doi: 10.1111/bcp.12820.