GLP-1 receptor agonists in the treatment of type 2 diabetes – state-of-the-art

GLP-1 receptor agonists (GLP-1 RAs) with exenatide b.i.d. first approved to treat type 2 diabetes in 2005 have been further developed to yield effective compounds/preparations that have overcome the original problem of rapid elimination (short half-life), initially necessitating short intervals between injections (twice daily for exenatide b.i.d.).

At present, GLP-1 RAs are injected twice daily (exenatide b.i.d.), once daily (lixisenatide and liraglutide), or once weekly (exenatide once weekly, dulaglutide, albiglutide, and semaglutide). A daily oral preparation of semaglutide, which has demonstrated clinical effectiveness close to the once-weekly subcutaneous preparation, was recently approved. All GLP-1 RAs share common mechanisms of action: augmentation of hyperglycemia-induced insulin secretion, suppression of glucagon secretion at hyper- or euglycemia, deceleration of gastric emptying preventing large post-meal glycemic increments, and a reduction in calorie intake and body weight. Short-acting agents (exenatide b.i.d., lixisenatide) have reduced effectiveness on overnight and fasting plasma glucose, but maintain their effect on gastric emptying during long-term treatment. Long-acting GLP-1 RAs (liraglutide, once-weekly exenatide, dulaglutide, albiglutide, and semaglutide) have more profound effects on overnight and fasting plasma glucose and HbA 1c , both on a background of oral glucose-lowering agents and in combination with basal insulin. Effects on gastric emptying decrease over time (tachyphylaxis). Given a similar, if not superior, effectiveness for HbA 1c reduction with additional weight reduction and no intrinsic risk of hypoglycemic episodes, GLP-1RAs are recommended as the preferred first injectable glucose-lowering therapy for type 2 diabetes, even before insulin treatment. However, GLP-1 RAs can be combined with (basal) insulin in either free- or fixed-dose preparations. More recently developed agents, in particular semaglutide, are characterized by greater efficacy with respect to lowering plasma glucose as well as body weight. Since 2016, several cardiovascular (CV) outcome studies have shown that GLP-1 RAs can effectively prevent CV events such as acute myocardial infarction or stroke and associated mortality. Therefore, guidelines particularly recommend treatment with GLP-1 RAs in patients with pre-existing atherosclerotic vascular disease (for example, previous CV events). The evidence of similar effects in lower-risk subjects is not quite as strong. Since sodium/glucose cotransporter-2 (SGLT-2) inhibitor treatment reduces CV events as well (with the effect mainly driven by a reduction in heart failure complications), the individual risk of ischemic or heart failure complications should guide the choice of treatment. GLP-1 RAs may also help prevent renal complications of type 2 diabetes. Other active research areas in the field of GLP-1 RAs are the definition of subgroups within the type 2 diabetes population who particularly benefit from treatment with GLP-1 RAs. These include pharmacogenomic approaches and the characterization of non-responders. Novel indications for GLP-1 RAs outside type 2 diabetes, such as type 1 diabetes, neurodegenerative diseases, and psoriasis, are being explored. Thus, within 15 years of their initial introduction, GLP-1 RAs have become a well-established class of glucose-lowering agents that has the potential for further development and growing impact for treating type 2 diabetes and potentially other diseases.

Following the approval of exenatide to treat type 2 diabetes (USA: 2005; Europe: 2006), several pharmaceutical companies started diverse developments aiming at GLP-1 receptor stimulation with greater effectiveness and longer duration of action. Exenatide needs to be injected at least twice daily, which mainly provides active circulating concentrations covering two major meals every day, with low levels between the two injections. Liraglutide, approved in 2009, was designed to provide a nearly unchanged amino acid sequence compared to mammalian GLP-1. A free fatty acid side chain was coupled to the peptide, which promotes binding to albumin in plasma and interstitial fluid. Only a minor proportion (estimated 1–2%) of liraglutide circulates in a free (non-albumin-bound) form, ready to diffuse into tissues and bind receptors. The albumin-bound bulk forms a reservoir promoting prolonged action. Overall, the elimination half-life is approximately 13 h, making it a suitable preparation for once-daily injection. The next step was aiming at once-weekly injections of GLP-1 RAs. Exenatide was developed as a novel preparation with the active ingredient slowly released after subcutaneous injection from a matrix dissolving over time. Thus, the onset of action was very much delayed, and a steady state was not reached until 8–10 weeks of treatment [ 17 , 18 ]. Other approaches followed the strategy to couple (modified) GLP-1 to large proteins such as an immunoglobulin Fc fragment (dulaglutide or efpeglenatide) or albumin (albiglutide). These compounds appear to slowly degrade, with half-lives of approximately one week. After subcutaneous injection, they reach effective circulating concentrations relatively early, thus beginning to lower plasma glucose soon after initiating such treatment. Semaglutide is another compound with a structure generally similar to liraglutide (GLP-1 with a free fatty acid side chain) but with a much longer half-life, apparently mediated by even tighter coupling to albumin. Semaglutide is presently available for once-weekly subcutaneous injection. More recently, semaglutide was co-formulated with s odium N -(8-(2-hydroxybenzoyl) a mino) c aprylate (SNAC) for oral treatment. To account for the relatively low bioavailability of semaglutide when absorbed through the gastrointestinal tract, oral semaglutide needs to be administered daily. This is the first GLP-1 RA approved for oral administration. At equivalent doses, subcutaneous and oral semaglutide seem to have similar effects on HbA 1c , body weight, and adverse events [ 19 ]. Details regarding the molecular structures of various GLP-1 RAs and additional pharmacokinetic information were summarized by Nauck and Meier in 2019 [ 20 ].The time between subcutaneous (or oral) administration and the occurrence of peak concentrations is displayed in Figure 1 . Table 1 Characteristics of GLP-1 RAs that have been approved to treat type 2 diabetes as of 2020. Table 1 GLP-1 RA First approved (date) Molecular weight (Da) c Reference amino acid sequence Other important components Elimination half-life Administration schedule Pharmaceutical company Reference For subcutaneous injection Short-acting compounds Exenatide b.i.d. 2005 (USA); 2006 (Europe); Byetta 4186.6 Exendin-4 None 3.3–4.0 h Twice daily AstraZeneca i [ 21 ] Lixisenatide 2013 (Europe); Lyxumia; 2016 (USA); Adlyxin 4858.5 Exendin-4 Poly-lysine tail 2.6 h Once daily Sanofi [ 22 ] Long-acting compounds/preparations Liraglutide 2009 (Europe); 2010 (USA); Victoza 3751.2 Mammalian GLP-1 Free fatty acid e 12.6–14.3 h Once daily Novo Nordisk [ 23 ] Once-weekly exenatide 2012; BYDUREON a 4186.6 Exendin-4 Active ingredient encapsulated in microspheres of poly-( d , l -lactide-co-glycolide) 3.3–4.0 h f Once weekly AstraZeneca i [ 21 ] Dulaglutide 2014; Trulicity 59670.6 Mammalian GLP-1 Immunoglobulin Fc fragment 4.7–5.5 d Once weekly Eli Lilly and Company [ 24 ] Albiglutide 2014 (Europe); Eperzan Tanzeum (USA) b 72971.3 Mammalian GLP-1 Albumin 5.7–6.8 d Once weekly GlaxoSmithKline [ 25 ] Semaglutide 2017 (USA); 2019 (Europe); Ozempic 4113.6 Mammalian GLP-1 Free fatty acid e 5.7–6.7 d Once weekly Novo Nordisk [ 26 ] For oral administration Semaglutide (long-acting) 2020; Rybelsus 4113.6 Mammalian GLP-1 Free fatty acid e 5.7–6.7 d Once daily Novo Nordisk [ 27 ] Fixed-dose combinations With basal insulin (for subcutaneous injection) Liraglutide/insulin degludec (iDegLira) 2014 (Europe); 2016 (USA); Xultophy 3751.2 d Mammalian GLP-1 Basal insulin 12.6–14.3 h Once daily (anytime g ) Novo Nordisk [ 28 ] Lixisenatide/insulin glargine (iGlarLixi) 2016 (USA); Soliqua 100/33; 2017 (Europe); Suliqua 4858.5 d Exendin-4 Basal Insulin 2.6 h Once daily h Sanofi [ 29 ] a Improved once-weekly auto-injector BYDUREON BCise was approved in 2018. b Marketing was discontinued in 2018. c Mammalian GLP-1: 3297.7. d For the GLP-1 RA component only. e Promoting binding to albumin. f Identical to the short-acting preparation. g Approx

Since the parent compound of GLP-1 RAs, GLP-1, has a very short elimination half-life that precluded its clinical use outside settings characterized by continuous administration, compounds/preparations with longer intervals between injections have been developed over time ( Table 1 ). While this at first was thought to be mainly relevant with respect to the injection frequency, thus representing a convenience issue, essential pharmacological differences were later identified that suggested that both short- and long-acting GLP-1 RAs may have specific advantages and indications [ 31 ]. By definition, short-acting GLP-1 RAs (exenatide b.i.d. and lixisenatide) are characterized by short-lived peaks in plasma drug concentrations following each injection, with intermittent periods of near-zero concentrations. Thus, the time–action profile changes between periods (lasting a few hours) during which patients are exposed to effective circulating drug concentrations, and “resting” periods, during which GLP-1 receptors are not activated. In contrast, long-acting GLP-1 RAs, once at a steady state, are characterized by constantly elevated drug concentrations in a range leading to substantial GLP-1 receptor stimulation and only minor fluctuations between injections (e.g., a 24-h period for liraglutide and a week-long period for semaglutide). Of note, this definition does not rest on the injection frequency alone but on the pharmacological kinetics. Consequently, once-daily lixisenatide is a short-acting compound, whereas once-daily liraglutide is a long-acting GLP-1 RA ( Table 1 ). One obvious consequence of the different temporal patterns of short- and long-acting GLP-1 RAs with reduced exposure during the night in short-acting compounds is the ability of long-acting GLP-1 RAs to more profoundly lower fasting plasma glucose than short-acting GLP-1 RAs. This was best exemplified by a study comparing un-retarded (b.i.d.) and long-acting release (once-weekly) exenatide [ 18 ], although the differences were valid for the comparison of any short- and long-acting GLP-1 RA ( Figure 4 ). Figure 4 Comparison of approved GLP-1 RAs with respect to their effectiveness in reducing HbA 1C (A), fasting plasma glucose (B), and body weight (C). A linear regression analysis relating reductions in fasting plasma glucose to reductions in HbA 1c is shown in panel D. A comparison of the reported coefficients of variation for reducing HbA 1c and body weight is displayed in panel E. All data are from clinical trials reporting head-to-head comparisons between various GLP-1 RAs (exenatide b.i.d. vs lixisenatide [ 36 ], exenatide b.i.d. vs liraglutide [ 37 ], lixisenatide vs liraglutide [ 38 ], exenatide once-weekly vs liraglutide [ 39 ], albiglutide vs liraglutide [ 40 ], dulaglutide vs liraglutide [ 41 ], subcutaneous semaglutide vs dulaglutide [ 42 ], and oral semaglutide vs liraglutide [ 43 ]) on a background of oral glucose-lowering agents. Data concerning the same GLP-1 RA were pooled using conventional equations to calculate common means and their standard deviations. Figure 4 Another peculiarity relates to the effectiveness of GLP-1 RAs to slow gastric emptying in light of tachyphylaxis: while intermittent stimulation of GLP-1 receptors (short-acting GLP-1 RAs) is associated with preserved effects on gastric motility, even long-term continuous stimulation leads to desensitization, which probably begins early (within 4–24 h) and reaches its full expression after several weeks or months [ 32 ]. Since the velocity of gastric emptying is tightly coupled to the absorption of nutrient carbohydrates, slowed gastric emptying means reduced and/or delayed glycemic increases after meals. For short-acting GLP-1 RAs, delayed gastric emptying is the main mechanism for post-meal reductions in plasma glucose rises [ 33 ]. It has been claimed that short-acting GLP-1 RAs act preferentially on post-meal glycemic rises through their effect on gastric emptying, which are preserved over time [ 18 , 33 , 34 ], while there is substantial tachyphylaxis for long-acting compounds [ 18 , 34 ]. First, long-acting GLP-1 RAs reduce post-prandial glucose as well, mainly through increasing insulin and suppressing glucagon [ 31 ]. The effect on gastric emptying relates only to meals, before which the short-acting GLP-1 RA has been administered (once daily with lixisenatide and twice daily with exenatide b.i.d.), with minor effects at most for other meals [ 33 ]. Whether this translates into a net advantage is far from clear. In a recent meta-analysis comparing short- and long-acting GLP-1 RAs on a basal insulin background, post-prandial glucose increases were not significantly different [ 35 ]. Conditions under which a reduction in post-meal glycemic excursions through a lasting deceleration of gastric emptying cause an obvious advantage of short-over long-acting GLP-1 RAs still need to be defined. The effectiveness of short- and long-acting GLP-1 RAs for controlling fasti

Therapy with basal insulin may fail because it may be successful in controlling fasting plasma glucose but does not sufficiently limit post-prandial glycemic excursions. Treatment intensification can mean adding one to three prandial insulin injections per day or adding a GLP-1 RA to ongoing insulin treatment. Nevertheless, GLP-1 RA therapy with a background of oral glucose-lowering medications may fail to achieve glycemic targets as well. In this case, combining it with insulin (mainly basal insulin) is a well-documented method of improving fasting, post-prandial, and overall (HbA 1c ) glycemic control [ [51] , [52] , [53] , [54] , [55] , [56] , [57] ]. The combination of (basal) insulin with a GLP-1 RA is a highly effective treatment even for advanced stages of type 2 diabetes. It should only be used in patients needing a combination of two injectable treatments, especially considering the costs of such a combination. When a GLP-1 RA is added to (basal) insulin, the combination is as effective as an intensified (basal bolus) insulin regime in terms of HbA 1c control, but with a much lower risk of hypoglycemia and weight gain [ 58 ]. When insulin is added to a GLP-1 RA, it helps control fasting plasma glucose. In combination with post-prandial effects of GLP-1 RAs (through decelerating gastric emptying, stimulating insulin, or suppressing glucagon secretion [ 31 ]), this provides excellent chances to achieve the target ranges for fasting, post-prandial, and overall (HbA 1c ) glycemic control. In studies comparing basal insulin and GLP-1 RAs alone and in combination with each other, the combination achieved the lowest HbA 1c or highest HbA 1c reduction and a body weight transformation in between GLP-1 RA alone (lowest) and insulin alone (highest) [ 59 ]. There is a risk of hypoglycemic episodes with this combination, which is higher than treating with GLP-1 RAs alone, but lower compared to insulin treatment alone [ 59 ]. The fact that a combination of a GLP-1 RA with basal insulin is a highly efficacious glucose-lowering treatment regime for advanced stages of type 2 diabetes has led to the development of fixed-dose combinations. GLP-1 RAs that are usually injected once daily (liraglutide or lixisenatide) were combined with basal insulin designed for once-daily injection (insulin degludec or insulin glargine), resulting in the fixed-dose combinations iDegLira [ 28 , 59 ] and iGlarLixi [ 60 , 61 ]. Since insulin must be titrated slowly as part of the dose-finding process, the GLP-1 RA component of these fixed-dose combinations is titrated slowly as well. This approach for introducing GLP-1 RA therapy has resulted in fewer problems with nausea, vomiting, or diarrhea. Apparently smaller steps of increasing GLP-1 RA exposure better support an adaptation process increasing patients’ tolerance to such adverse reactions. It has been postulated that short-acting GLP-1 RAs are particularly suited for combination with basal insulin because the strength of long-acting compounds, a greater effect on fasting plasma glucose, is not needed in this combination since the role of basal insulin would be to control fasting plasma glucose. However, the effect of slowing gastric emptying leading to slower absorption of nutrients (which is preserved over time with short-acting GLP-1 RAs) is a mechanism limiting post-meal glycemic excursion [ 31 ]. Notably, with short-acting GLP-1 RAs, this effect only applies to the meal before which the agent has been injected. A recent meta-analysis described the advantages of combining long-acting GLP-1 RAs (compared to short-acting GLP-1 RAs) with basal insulin [ 50 ]. This applied to free combinations (dosage determined separately for the GLP-1 RAs and basal insulin) as well as fixed-dose combinations [ 50 ]. As depicted in Figure 5 , not only HbA 1c was lowered significantly more and HbA 1c targets were achieved in a higher proportion of patients, but also fasting plasma glucose concentrations and body weight were controlled better with long-acting GLP-1 RAs [ 50 ]. In addition, the risk of hypoglycemic episodes and gastrointestinal side effects was slightly, but significantly lower with long-acting GLP-1 RAs [ 50 ].

Side effects most reported with GLP-1 RAs are nausea, vomiting, and diarrhea, often summarized as gastrointestinal adverse events. They are typically most prominent when initiating treatment with (any) GLP-1 RA or after increasing the dose (e.g., during recommended up-titration regimens). Since these symptoms can occur in fasting subjects, they are probably not related to the effects of GLP-1 RA treatment on gastrointestinal functions (e.g., deceleration of gastric emptying) but instead are caused by direct interactions with CNS GLP-1 receptors ( Figure 6 ) most likely located in the brain stem (area postrema). Nausea is typically reported in up to 25% and vomiting or diarrhea in up to 10% of subjects treated with GLP-1 RAs [ 20 , 49 ]. For most patients, these are short, self-limited episodes that cease spontaneously, even with continued treatment. The time point of occurrence is probably related to the time characterized by maximum drug concentrations typically occurring at T max following several hours to days after each injection ( Figure 1 ). The probability of these side effects varies with sudden incremental exposure to GLP-1 RAs. An often-used recommendation to avoid these adverse events is a standardized, slowly increased exposure through up-titration regimens ( Figure 2 ), which have been shown to mitigate gastrointestinal side effects. Experience with fixed-dose combinations with basal insulin (which must be titrated much more slowly) underscore the effectiveness of this approach. Summarizing adverse event reporting from clinical trials examining GLP-1 RAs discloses subtle differences in the risk of these side effects depending on the short- (worse) vs long-acting nature (better) background medication (worse in combination with metformin or insulin) that are also related to the individual compound/preparation [ 49 ]. In part related to adverse events, patients randomized to GLP-1 RA treatment often discontinue this medication. Table 4 shows reported figures from cardiovascular (CV) outcome trials with GLP-1 RAs, the largest trials available reporting the longest durations of exposure to GLP-1 RAs. The proportions of patients reporting adverse events were not generally different from shorter clinical trials [ 49 ]. This indicates that while the frequency and severity of side effects can be successfully modulated through optimized up-titration regimens, a certain percentage of patients does not tolerate this treatment with the current regimens of initiating GLP-1 RA. Interestingly, in a recent study allowing individual titration of oral semaglutide, most patients discontinuing this treatment did so after exposure to the lowest (initial) dose of 3 mg per day [ 93 ]. This may indicate that the sensitivity of patients toward developing gastrointestinal adverse events is considerably heterogeneous, such that some patients fail to tolerate low doses, while for others, higher doses than currently used may offer better effectivity without increasing side effects. Along these lines, higher doses of some GLP-1 RAs are being explored, especially to further reduce body weight [ 65 , 67 , 68 , 71 , 86 ]. The reported nausea, vomiting, and diarrhea rates are generally lower in Japanese than Caucasian populations, suggesting that the cultural background and eating behaviors may also have an impact on the induction of nausea with GLP-1 RAs [ 94 ]. Table 4 Proportions of patients randomized to GLP-1 RA treatment in CV outcome trials discontinuing study drug treatment, proportion of the follow-up period during which patients were exposed to the study drug, and proportions discontinuing due to adverse events. Table 4 GLP-1 RA Proportion of patients permanently discontinuing the study drug [%] a Proportion of follow-up period during which the study drug was taken [%] Proportion of patients discontinuing the study drug because of adverse events [%] Trial/reference Lixisenatide 27.5 90.5 11.4 ELIXA [ 95 ] Liraglutide n.p. 84 9.5 LEADER [ 96 ] Once-weekly exenatide 43.0 76 4.5 c EXSCEL [ 97 ] Dulaglutide 26.8 82.2 9.1 REWIND [ 98 ] Albiglutide 24.5 87 8.6 HARMONY Outcomes [ 99 ] Semaglutide s.c. 21.3 86.5 13.2 SUSTAIN-6 [ 100 ] Oral semaglutide 15.3 n.p. b 11.6 PIONEER-6 [ 101 ] n.p.: Not presented. a Not counting transient “drug holidays.” b 75% received the study medication for more than 1 year (total follow-up of 15.3 months). c Counting only gastrointestinal adverse events. No CV outcome trial has been reported for exenatide b.i.d. (approved before these studies became mandatory). When GLP-1 RAs were introduced as novel agents to treat type 2 diabetes, there was uncertainty about several potential adverse effects such as acute pancreatitis, pancreatic cancer, and thyroid cancer [ 102 , 103 ]. The availability of large databases from randomized CV outcome studies that defined pancreatitis, pancreatic cancer, and thyroid cancer as “adverse events of special interest” with protocols carefully adjudicating suspected cas

As previously demonstrated in detail [ 115 ], GLP-1 RAs modify a number of risk factors for cardiovascular complications, including body weight reduction, lower systolic blood pressure, reduced plasma LDL cholesterol and triglyceride concentrations, and improved glycemic control (reductions in fasting and post-meal plasma glucose resulting in lower HbA 1c ; see Figure 4 ). Thus, a reduction in the incidence of ischemic events could be the consequence of a more beneficial risk profile under treatment with GLP-1 RAs. Mediation analysis is an approach to identify potential mediators that might explain the findings observed in terms of endpoints. While several mathematical approaches have been developed, their common aim is to show that taking into account the changes in a potential mediator reduces the effect size with respect to the endpoint of interest. Potential mediators are variables measured in the trial that are differentially affected by active drugs and placebo. For example, GLP-1 RAs reduce systolic blood pressure by 2–4 mmHg compared to placebo treatment [ 115 ]. Considering this reduction in systolic blood pressure, if the difference in MACE outcomes is reduced, it can be concluded that a reduction in systolic blood pressure mediates the prevention of MACE. If the effect is nullified, this mechanism is responsible for 100% of the effect, but partial mediation is also possible. Using slightly different approaches, mediation analyses have been published on the effects of liraglutide in the LEADER trial [ 116 ] and the effects of dulaglutide in the REWIND study [ 117 ]. Interestingly, both analyses concluded that HbA 1c reduction was a potential mediator, responsible for up to 82% of the total effect. A reduction in urinary albumin excretion was found to be another potential mediator in the LEADER trial (responsible for up to 33% of the total effect). Of note, any potential mechanism that does not leave a measurable trace or has not been assessed in a given trial will never be identified as a potential mediator using this approach. This applies to intravascular changes associated with the progression of atherosclerosis unless they are accompanied by, for example, inflammatory responses, which can be identified by measuring C-reactive protein or inflammatory cytokines (which was not done in any of the CV outcome trials of GLP-1 RAs). Hence, identifying HbA 1c reduction as a potential mediator of CV benefits induced by GLP-1 RAs leaves a number of open questions, especially since it has been difficult to establish a relationship of HbA 1c reduction with cardiovascular benefits in other glucose-lowering medications [ 118 ]. For CV outcome studies of GLP-1 RAs, a relationship that links the magnitude of the HbA 1c reduction achieved (versus placebo) to the hazard ratio for major adverse CV events was suggested by Caruso et al. [ 119 ]. In particular, they identified a relationship between the mean reduction in HbA 1c in individual trials and the corresponding hazard ratio for stroke [ 119 ]. Figure 8 presents a similar analysis, however, including CV outcome studies with DPP-4 and SGLT-2 inhibitors. Remarkably, these additional data points were positioned along the same regression lines, and the relationship between the reduction in HbA 1c and MACE ( Figure 8 A) or stroke ( Figure 8 C) remained significant. Similar trends of non-fatal acute myocardial infarction and CV death were non-significant. A significant reduction in hospitalizations for heart failure was restricted to SGLT-2 inhibitors and independent from reductions in HbA 1c ( Figure 8 F). Such an analysis may not only confirm the relationship initially observed by Caruso et al. [ 119 ], but may also explain why DPP-4 inhibitors and SGLT-2 inhibitors did not consistently reduce MACE [ 110 , 111 , 115 ]. Given that all of the CV outcome trials aimed at glycemic equipoise (similar if not identical glycemic control for active drug and placebo treatment), only trials with potent glucose-lowering medications such as GLP-1 RAs, which failed to achieve glycemic equipoise, underscored the potential for a CV benefit. Figure 8 Regression analysis of differences achieved in HbA 1c concentrations between patients treated with placebo and active drug vs hazard ratios for major adverse cardiovascular outcomes (MACE; A), cardiovascular death (B), non-fatal stroke (C), non-fatal myocardial infarction (D), and hospitalization for heart failure (E) reported from cardiovascular outcome studies with GLP-1 receptor agonists (red), SGLT-2 inhibitors (blue), and DPP-4 inhibitors (green). Significant associations are shown for MACE (A) and non-fatal stroke (C) with similar slopes of the regression lines, while for cardiovascular death (B) and non-fatal myocardial infarction (D), a less prominent, non-significant correlation resulted from the analysis. Regarding hospitalization for heart failure (E), hazard ratios did not vary with HbA 1c reduction. Analyzing GLP-1 recep

LDL cholesterol is transported across the intima layer of arterial blood vessels and in part oxidized to oxidized LDL particles (oxLDL) through reactive oxygen species (ROS). Contact of monocytes and macrophages with oxLDL and ROS promotes further infiltration of monocytes by secreting adhesion molecules such as vascular cell adhesion protein 1 (VCAM-1), monocyte chemoattractant protein 1 (MCP-1), intercellular adhesion molecule 1 (ICAM-1), and E-selectin. Stimulated by oxLDL, monocytes transform into macrophages. M1 macrophages produce pro-inflammatory cytokines such as tumor necrosis factor α (TNF-α), interleukin (IL)-6, and IL-1β. M2 macrophages take up lipid particles through phagocytosis and suppress the formation of Krüppel-like factor 2 (KLF-2), which in turn suppresses endothelial NO synthase (eNOS), leading to lower NO production and preventing vasodilation through NO-mediated vascular smooth muscle relaxation. In an environment dominated by ROS and oxLDL, M2 macrophages transform into foam cells that can undergo apoptosis and release their lipid content into the lipid core of nascent atherosclerotic plaques. Stable plaques are characterized by a dense fibrous cap mainly composed of collagen that helps prevent rupture. However, as atherogenesis progresses, larger necrotic areas form, endothelial cells (EC) undergo apoptosis, and matrix metalloproteinases (MMP) proteolytically destroy the fibrous cap. This results in plaque rupture, thrombus formation, and bleeding into necrotic plaque areas.

The discovery of beneficial renal effects using GLP-1 RAs is a recent achievement mainly based on the observations that GLP-1 RAs prevented new-onset macroalbuminuria [ 99 , 164 , 165 ], reduced urinary albumin excretion [ 164 , 165 ], or slowed the decline in the estimated glomerular filtration rate (eGFR) over time [ [164] , [165] , [166] ]. The mechanisms leading to these renal benefits are largely unknown. While significant reductions in achieving renal composite outcomes were reported [ 99 , 164 , 165 ], they heavily relied on dominating effects preventing new-onset persistent macro-albuminuria. Clinical events indicating progression to end-stage renal disease (doubling in serum creatinine, major reduction by 30–50% in eGFR, achieving eGFR below 15 ml/min per 1.73 m 2 , necessary to initiate renal replacement therapy or perform renal transplantation, or death due to renal causes) have rarely been reported in numbers allowing a meaningful analysis. This is due in part to the fact that the populations studied had fairly good renal function at baseline. Studies of selected patients with prominent or advanced renal disease are lacking. A dedicated trial studying the effects of semaglutide on renal outcomes in type 2 diabetic patients with chronic kidney disease is underway to clarify these issues: FLOW (ClinicalTrials.gov NCT03819153 ). Since most GLP-1 RAs can be used in chronic kidney disease, while SGLT-2 inhibitors lose some of their glucose-lowering efficiency with reduced glomerular filtration rates, further studies appear to be needed, especially since patients with reduced eGFR at baseline seem to benefit most in terms of preventing rapid declines in eGFR [ 164 ]. For the time being, more robust effects have been reported for SGLT-2 inhibitors, which are preferred glucose-lowering medications interfering with the progression of diabetic renal disease even in patients with moderately reduced eGFR [ 110 , 111 , 167 ].

As presented in Table 4 , some patients randomized to GLP-1 RA treatment discontinued the assigned medication. The proportion withdrawing from GLP-1 RA treatment in CV outcome trials ranged from 15% (oral semaglutide) to approximately 25%; an exceptionally high withdrawal rate was observed with once-weekly exenatide (43%), possibly related to the less comfortable pen injection device requiring resuspension of the active ingredient in buffer ( Figure 3 ) or the occurrence of subcutaneous nodules at injection sites [ 175 ]. Approximately one-half of the discontinuations were reported to be associated with adverse events ( Table 4 ). In the trials reporting discontinuation because of any adverse events and those specifically due to gastrointestinal side effects, the latter were responsible for approximately one-half of the withdrawals. Another potential reason contributing to withdrawals was a perception of ineffective glycemic and body weight control achieved (including a suspicion to have been randomized to placebo), perhaps as a consequence of the progression of the type 2 diabetes mellitus disease process [ 176 ]. Whether or not GLP-1 RA treatment counters this progression (e.g., through β cell-preserving effects [ 177 ]) remains an open question. In rodents, these effects are restricted to earlier periods in life [ 178 ] when β cells have a propensity to proliferate, which they lose in adult animals [ 179 ]. Overall, randomized controlled clinical trials showed that high persistence regarding GLP-1 RA treatment could be maintained for periods up to 5 years, which contrasts with data from observational studies (as previously described). Efforts to encourage persistent use of GLP-1 RA, as successful in clinical trials, may be necessary to achieve better persistence in clinical practice as well.

Since 2005, when exenatide was first approved, rapid development began that has yielded progress with respect to GLP-1 RAs pharmacokinetics, with the obvious consequence that instead of 2 (or more) injections per day, now once-weekly injections are available. While advances making GLP-1 RA treatment more comfortable are welcome, it should not be overlooked that the effectiveness of GLP-1 RAs has increased in large steps (e.g., going from exenatide to liraglutide, the first long-acting GLP-1 RA, but also advancing to semaglutide, which clearly has superior efficacy than other GLP-1 RAs, especially with respect to body weight reduction as depicted in Figure 4 ). These significant advances, occurring in substantial leaps, suggest that this development has not yet come to an end. 8.1 Oral administration of GLP-1 RAs One development worth noting is that, despite the peptide nature of all of the GLP-1 RAs, semaglutide is now available for oral administration. An absorption enhancer molecule (SNAC; see Section 2 ) must be part of the oral preparation to promote absorption through the gastric mucosa. Low bioavailability after oral administration makes daily administration of a semaglutide tablet necessary to avoid wide fluctuations in drug exposure. It must be taken on an empty stomach, and for 30 min after taking oral semaglutide, no other food, drink, or medication should be administered to allow undisturbed absorption. With these precautions, in principle, quantitatively similar effects can be achieved with respect to glycemic control and lowering body weight [ 19 ]. The phase 3 PIONEER program, however, was conducted with somewhat lower doses (maximum, 14 mg/d) than would be necessary to match the effectiveness of subcutaneous semaglutide at 0.5 or 1.0 mg/week [ 43 ]. In addition to developing peptide-based GLP-1 RAs for oral administration, some reports described small molecules with GLP-1 receptor agonist properties that should be suitable for oral administration without additives and/or sensitive procedures. To date, the binding affinities of these compounds has been too low to support further development as clinically effective drugs [ [186] , [187] , [188] , [189] ].

Cluster analysis was applied to define subgroups within the type 2 diabetes population that differ with respect to insulin secretory capacity, insulin sensitivity, age at diagnosis, and the presence of autoimmune markers [ [198] , [199] , [200] ]. These subgroups display significant differences in the development of CV and renal complications [ [198] , [199] , [200] ]. Thus, given the beneficial actions of GLP-1 RAs on preventing CV events (and on the progression of nephropathy), they may turn out to be particularly effective in those presenting a high a priori risk of these complications. Identifying a central pathophysiological defect (e.g., reduced insulin secretory capacity) may also help select a specific therapy addressing this point (e.g., GLP-1 RAs augmenting insulin secretion triggered by hyperglycemia). While prospective studies comparing various therapies in type 2 diabetes patients belonging to different subgroups are lacking, this sub-classification promises to be a helpful tool assisting in a more individualized approach toward selecting glucose-lowering medications for a given patient.

One avenue of further increasing the potency of GLP-1 RAs is developing molecules that address not only GLP-1 receptors, but a second (co-agonist) or even a third (tri-agonist) receptor (choosing from glucagon, glucose-dependent insulinotropic polypeptide [GIP], or peptide YY [PYY] receptors). Preliminary findings suggest that highly effective compounds (e.g., tirzepatide [ 210 ]) can be developed this way, in particular providing weight loss far exceeding that reported with pure GLP-1 RAs. These developments will be the focus of another manuscript on the present supplement volume (Baggio et al. [ 211 ]).

Interest in using GLP-1 RAs to treat neurodegenerative diseases emerged from preclinical studies showing that GLP-1 receptor signaling is involved in cognitive functions [ 220 ] and GLP-1 RAs can induce neuronal growth and synaptic plasticity and reduce apoptosis and oxidative stress [ 221 ]. In Alzheimer's disease, the most prevalent form of dementia, animal studies have shown the positive effects of GLP-1 RAs on cognitive impairment [ 222 , 223 ]. In a clinical trial of patients with prediabetes and type 2 diabetes, memory function improved after 4 months of liraglutide administration [ 224 ]. However, there was no placebo control. In another trial with non-diabetes subjects at increased risk of Alzheimer's disease, administration of liraglutide for 3 months improved brain region connectivity (assessed by functional MRI), but cognitive functions did not improve [ 225 ]. Another trial in non-diabetes patients with Alzheimer's disease found that 6 months of liraglutide treatment prevented a further decline in brain glucose uptake (assessed by positron emission tomography), but did not change cognitive function tests [ 226 ]. A larger clinical trial is currently ongoing that is investigating the effects of liraglutide on mild Alzheimer's disease using comprehensive neurological and cognitive assessment [ 227 ]. Parkinson's disease is another neurodegenerative disease for which GLP-1 RAs are being explored as treatment options [ 221 , 228 ]. In a mouse model of Parkinson's disease, a novel GLP-1 RA protected dopaminergic neurons and ameliorated behavioral deficits, most likely by blocking the formation of a neurotoxic astrocyte variant [ 229 ]. In clinical trials, exenatide improved the MDS-UPDRS score (a standardized assessment scale for patients with Parkinson's disease) [ 230 , 231 ]. Nevertheless, a recent systematic Cochrane Database review declared the evidence of improved motor impairment in GLP-1 RA-treated patients with Parkinson's disease as “low certainty” [ 232 ]. In an animal model of Huntington's disease, mice treated with exendin-4 for 9 weeks presented with reduced huntingtin protein aggregates in the cortex compared to placebo and had longer life spans [ 233 ]. We are not aware of any clinical trials in human patients with Huntington's disease. Psoriasis is associated with type 2 diabetes [ 234 ]. Two case reports generated interest in using GLP-1 RAs as a potentially novel treatment option for psoriasis [ 235 , 236 ]. Two subsequent prospective studies found positive effects on psoriasis severity scores in type 2 diabetes patients treated with GLP-1 RAs [ 237 , 238 ], a finding that could not be confirmed in non-diabetes subjects [ 239 ]. This possibly suggests a clinical effectiveness that may differ in diabetes and non-diabetes patients.

M.A.N. has been a member of advisory boards or consulted with AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, GlaxoSmithKline, Menarini/Berlin-Chemie, Merck, Sharp & Dohme, and Novo Nordisk. He has received grant support from 10.13039/100004325 AstraZeneca , 10.13039/100004312 Eli Lilly and Company , Menarini/Berlin-Chemie , 10.13039/100004334 Merck , Sharp & Dohme , and Novo Nordisk. He has also served on the speakers’ bureau of AstraZeneca, Boehringer Ingelheim, Eli Lilly and Company, Menarini/Berlin-Chemie, Merck, Sharp & Dohme, and Novo Nordisk. J.J.M. has received consulting and speaker honoraria from AstraZeneca, Eli Lilly and Company, Merck, Sharp & Dohme, Novo Nordisk, Sanofi, and Sevier. He has received research support from Boehringer Ingelheim, Eli Lilly and Company, Merck, Sharp & Dohme, Novo Nordisk, and Sanofi. J.W. and D.Q. have nothing to declare.