Reviews/Commentaries/ADA Statements PERSPECTIVES ON THE NEWS

Gut-Derived Incretin Hormones and New Therapeutic Approaches

ZACHARY T. BLOOMGARDEN, MD showing in somatostatin-, insulin-, and glucagon-infused dogs (thus not allowing GLP-1 to change insulin or glucagon lev- els) that peripherally administering the his is the first of a series of articles on from nil to 100%. Conversely, when glu- peptide increased hepatic glucose uptake presentations at the American Dia- cose was infused peripherally versus in 60%, while administration either via the Ϫ Ϫ T betes Association Annual Meeting, the portal vein at 2.5 mg ⅐ kg 1 ⅐ min 1, portal vein or hepatic artery let to a 150% Orlando, Florida, 4–8 June 2004, ad- arterial glucose levels were identical but increase in hepatic glucose uptake, sug- dressing an important theme of the meet- there was a rapid drop from hepatic glu- gesting the effect to be in the liver rather ing: new therapeutic approaches based on cose output to uptake with portal deliv- than specifically requiring GLP-1 levels to gut-derived incretin hormones. ery, whereas a much slower fall was seen increase in the portal venous system. Io- in hepatic glucose output occurring with nut et al. (abstract 1412) similarly showed peripheral delivery, with consequent lack a glucose-lowering effect of portal vein Physiology of glucagon-like peptide of increased peripheral glucose uptake GLP-1 infusion. 1 action following intraportal administration. Pe- Cherrington concluded that the por- Alan Cherrington (Nashville, TN) de- ripheral glucose administration led to tal vein and liver may contain nutrient- scribed a series of studies of peripheral most glucose uptake being peripheral. and hormone-sensing mechanisms that help coordinate the disposition of in- and hepatic arterial and portal vein flow Thus, whole-body glucose clearance did gested nutrients. Portal vein glucose de- measurement and sampling that allowed not change with the method of glucose livery generates a signal that augments the for the assessment of hepatic glucose bal- administration, but the portal signal ap- role of the liver and limits muscle glucose ance and effects of glucagon-like peptide peared to instruct the liver to take up disposition, with portal vein GLP-1 infu- (GLP)-1. In a study comparing peripheral more, and muscle to take up less, glucose. sion studies suggesting an important di- with portal venous glucose delivery dur- This is not due to lower peripheral insulin ing hyperglycemia (glucose increased rect hepatic effect of the hormone, which levels, as these tended to be somewhat mediates at least part of this process. In from 75 to 150 mg/dl), with somatostatin higher after portal glucose administra- to suppress endogenous insulin and glu- response to questions, he suggested that tion, suggesting alternative mechanisms, cagon and administration of glucagon to the molecular mechanism of glucose potentially including the incretin effect replace basal levels and insulin to levels sensing may involve glucokinase and and neurally mediated metabolic signals. approximately four times greater than GLUT2, noted that the effect of hepatic Addressing the incretin effect, GLP-1 basal, there was a similar increase in the ϳ denervation includes decreased response hepatic glucose load, from ϳ25 to 50 mg levels increased from basal levels of 15 to portal and increased response to pe- Ϫ Ϫ ⅐ kg 1 ⅐ min 1, but portal glucose delivery pmol/l to arterial levels of 30 pmol/l and ripheral glucose delivery, and suggested doubled hepatic glucose uptake. Net he- to 60 pmol/l in the portal vein, leading that GLP-1 produces similar augmenta- Cherrington to note that levels above tion of hepatic response with and without patic glucose uptake remained consider- ϳ ably greater with intraportal glucose 100 pmol/l should be considered non- somatostatin, suggesting that this experi- physiological. In a study administering delivery when the hepatic glucose load Ϫ Ϫ mental method did not artifactually alter Ϫ ⅐ 1 ⅐ 1 was varied from 50 to 75 to 100 mg ⅐ kg 1 glucose at a rate of 4 mg kg min via the results of his studies. Ϫ ⅐ min 1. Thus, Cherrington suggested, the portal vein, as well as peripheral glu- In a fascinating study related to this hepatic glucose entry depends on insulin cose to increase plasma glucose levels to topic presented at the meeting, Li and and glucose levels and a “portal signal.” 160 mg/dl, intraportal GLP-1 increased Drucker (abstract 1) studied mice not ex- ϳ Portal glucose delivery, then, can be said arterial and portal GLP-1 to 40 and 85 pressing the GLP-1 receptor, showing to “activate the liver,” which may be in pmol/l and tripled hepatic glucose up- that superimposing transgenic expression part the mechanism of increased hepatic take, in part because of greater increase in of the receptor restricted to pancreatic ␤- glucose uptake following oral glucose de- insulin and suppression of glucagon, but and ductal cells via the PDX-1 gene al- livery. When measuring hepatic sinusoi- peripheral GLP-1 in a similar study had a lowed restoration of normal fasting glu- dal insulin, as levels increase, the net lesser effect in increasing hepatic glucose cose and response to intraperitoneal hepatic glucose uptake increases but uptake, suggesting this to be an important glucose with and without administration there is further uptake with activation of direct influence “without the mediation of of exendin-4, a natural peptide with the portal glucose signal. In further stud- a change in [insulin and glucagon] secre- Ͼ50% sequence homology to GLP-1, but ies varying peripheral and portal vein glu- tion.” In a presentation at the American with greater potency and longer duration cose delivery and maintaining Diabetes Association Annual Meeting on of action, produced in the saliva of the hyperinsulinemia, a fivefold increase in this topic by Cherrington’s group, Dard- Gila monster. This study suggests that al- hepatic glucose uptake was demon- evet et al. (abstract 1410) assessed the po- though the stomach, nervous system, and strated, as portal glucose delivery ranged tential direct effects of GLP-1 on the liver, liver also express the GLP-1 receptor, ac-

2554 DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 Bloomgarden tion at these sites may not be required for duction in body weight, and decelerating mediated insulin secretion, improves the peptide’s systemic glycemic effect. gastric emptying. Although Nauck stated insulin sensitivity, and slows gastric emp- that there is no evidence of effect on insu- tying, either via a central nervous system Potential effects of GLP-1 in type 2 lin sensitivity, the studies discussed by (CNS) effect or by activation of vagal af- diabetes Cherrington suggest the possibility that ferent fibers. An important question, she Michael A. Nauck (bad Laukenberg em the peptide changes hepatic versus pe- noted, is whether albumin-bound GLP-1 Nacht, Germany) discussed GLP-1 and ripheral partitioning of glucose metabo- will be able to access CNS sites. GLP-1 glucose-dependent insulinotropic pep- lism. Initial studies of persons with type 2 also increases ␤-cell mass via stimulation tide (GIP) in type 2 diabetes. GLP-1 is diabetes showed that subcutaneous of neogenesis and inhibition of apoptosis, produced by L-cells in the distal small GLP-1 injection stimulated insulin and again leading to the question of whether bowel, while GIP comes from K-cells in inhibited glucagon secretion, decreased larger albumin-bound complexes will be proximal small intestine. Classically, the gastric emptying, and, with repeated in- able to access ␤-cells to produce trophic incretin effect has been demonstrated by jections, normalized plasma glucose (6). effects. comparing effects of intravenous versus Nauck described a study of 34 persons GLP-1 action controls fasting and oral glucose administered to produce with type 2 diabetes who received a postprandial insulin release in a glucose- similar plasma glucose curves and hence 6-week intravenous infusion of GLP-1 or dependent manner. GLP-1 is rapidly in- “glycemic stimulus.” Under such condi- saline and had baseline HbA1c of 9.4%. activated by dipeptidyl peptidase (DPP)- tions, oral glucose produces a threefold Glucagon levels were suppressed, insulin IV, limiting its therapeutic potential, higher insulin and C-peptide response. In and C-peptide levels were increased, fast- although as discussed above therapeutic persons with type 2 diabetes, however, ing and mean glucose levels were de- effects can be shown when the peptide is the incretin effect is attenuated or even creased by 77 and 99 mg/dl, respectively, administered by continuous infusion. completely lost. One explanation would HbA1c was reduced by 1.3%, and there The development of long-acting safe and be that the relative secretion of hormones was a weight loss of 4 lb (7). Below a efficacious GLP-1 receptor agonists, such mediating the incretin effect is decreased. blood glucose of 120 mg/dl, the insulin as exenatide, allows consideration of Indeed, comparing persons with and secretory and glucagon suppressive ef- practical therapeutic trials. In phase 3 without type 2 diabetes, differences in fects lessened, with levels of both hor- studies completed by Amylin and Lilly, GLP-1 appear ϳ1 h after oral glucose (1), mones returning to the basal range. In a two injections daily are required for ef- while GIP may either be hypo- or hyper- study from Nauck’s group, Fehse et al. fect, with a 1-year open-label study show- secreted (2) and therefore was not felt to (abstract 351) compared 13 persons with ing a 1.2% decrease in HbA1c,an8-lb explain the lack of incretin effect in type 2 type 2 diabetes, using insulin infusion to weight loss, mild to moderate nausea, and diabetes. Studies comparing the effects of attain euglycemia with and without a 5-h hypoglycemia in patients also receiving GIP and GLP-1 in diabetic and nondia- exenatide infusion, and 12 healthy con- sulfonylureas. betic persons show, however, that the re- trol subjects. Both the first- and second- A novel alternative approach involves sponse to GIP is decreased while that to phase insulin response to intravenous linking GLP-1 with albumin. Albumin is a GLP-1 is similar in diabetic and nondia- glucose increased two- to fourfold with 67-kDa protein with 19-day half-life in betic persons (3), suggesting that the lack the GLP-1 analog to the range seen in the humans, and albumin-binding of GLP-1 of GIP effect may indeed be important in nondiabetic group, and glucose disposal slows its absorption following subcutane- type 2 diabetes. Comparing bolus injec- was normalized. However, GLP-1 infu- ous administration, resulting in steady tions of GLP-1 and GIP, persons with type sion increases total GLP-1 to a greater ex- blood concentrations. Liraglutide is a 2 diabetes do show response to GIP as to tent than it affects the biologically active compound developed by Novo Nordisk GLP-1, but longer periods of infusion of molecule and therefore might not be ap- that has completed phase 2 clinical trials. GIP fail to increase insulin levels in the propriate for therapy (8), suggesting the The compound has 97% homology with fashion seen with GLP-1 (4), implying importance of development of incretin- GLP-1, is released slowly from the injec- that the mechanism is complex. In a study mimetic agents (vide ifra). Nauck con- tion site, and has a fatty acid binding moi- of GIP infusion during hyperglycemia, cluded that the incretin effect is reduced ety leading to albumin binding (9). After a approximately half of a group of relatives in persons with type 2 diabetes, explained single injection the half-life of liraglutide of persons with type 2 diabetes had atten- by both abnormal secretion and de- is 11–15 h, with a decrease in glucagon, uation of insulin response. Although this creased response, and that the use of nat- an increase in insulin, and lowering of might seem a likely candidate as an early ural GLP-1 is limited by its rapid 24-h glucose profiles in persons with type marker of type 2 diabetes, there was no degradation and elimination. GLP-1 de- 2 diabetes following once-daily adminis- difference in glucose tolerance between rivatives such as exenatide may therefore tration (10). After 12 weeks of adminis- responders and nonresponders either at be better suited to diabetes therapy. tration, HbA1c levels decrease to a similar baseline or at 4-year follow-up, further extent to that seen with sulfonylureas but indicating the intricacy of these effects Long-acting GLP-1 agonist with weight loss rather than weight gain (5). preparations (11). Baggio noted that an important Effects of GLP-1 in normalizing glu- Laurie Baggio (Toronto, Canada) de- question for study is whether liraglutide cose may go beyond the acute insulin- scribed albumin-bound GLP-1 agonist reaches the same CNS binding sites as stimulatory action of the peptide to effects delivery systems. She stated that GLP-1 is GLP-1. Liraglutide reduces blood glucose increasing ␤-cell mass, decreasing gluca- derived from a larger precursor, stimu- in a glucose-dependent manner, without gon, decreasing food intake leading to re- lates insulin gene expression and glucose- effect in mice not expressing the GLP-1

DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 2555 Perspectives on the News receptor. When injected twice daily in a ing similar effects to those seen with food intake. GLP-1 agonists increase mouse model, the agent stimulates islet exendin-4, although with less potent in heart rate and blood pressure. Further- proliferation, again mimicking biological vitro activation of the GLP-1 receptor. Af- more, GLP-1 exhibits antiapoptotic and actions of GLP-1. Studies reported at the ter intravenous or subcutaneous injec- neurotrophic effects in neuronal cells, meeting suggested that liraglutide will be tion, albugon has a half-life of 11 h in similar to its trophic actions in ␤-cells. an effective therapeutic agent. Nauck et mice and 2–4 days in cynomolgus mon- Perry hypothesized that GLP-1 and al. (abstract 356) administered 2 mg lira- keys. In the mouse model, it mimics the GLP-1 agonists may induce neuronal dif- glutide s.c. daily versus placebo for 5 biological action of GLP-1 in decreasing ferentiation. In vitro, pheochromocytoma- weeks in 144 persons with type 2 diabetes glucose and glucagon and increasing in- derived PC12 cells express a functional receiving 1 g metformin twice daily, lead- sulin levels, and the molecule can be GLP-1 receptor, increasing intracellular ing to a 1.1% fall in HbA1c. Compared shown to act via the pancreatic GLP-1 re- cAMP upon activation. The effects of with metformin plus glimepiride, patients ceptor. Although less potent than ex- GLP-1 are similar to those of nerve growth treated with metformin plus liraglutide endin-4, there is definite effect in factor (NGF), while exendin-4 had some- had a 2.9-kg greater weight decrease and inhibiting gastric emptying in the mouse what different morphologic effects, with a 22-mg/dl greater fall in fasting blood model, with no effect seen in mice lacking the combination of exendin-4 and nerve glucose. In an obese rat model using the GLP-1 receptor. “These studies,” Bag- growth factor leading to an amplified ef- candy feeds, Knudsen et al. (abstract gio stated, “demonstrate that these larger fect. In an ibotenic acid–induced partial 1408) reported that over 12 weeks, body albumin-bound molecules are able to ac- basal nucleus lesion in vivo, cholinester- weight decreased with liraglutide in asso- cess the neuronal pathways.” Intraperito- ase transferase activity was restored by lo- ciation with a decrease in candy, but not neal albugon reduced food intake in the cal GLP-1 infusion, therefore providing chow, intake but did not show significant model, again without effect in mice lack- evidence of benefit in a neurocytotoxic le- change with the DPP-IV inhibitor ing the GLP-1 receptor, with both alb- sion. A series of GLP-1 analogs based on LAF237. ugon and exendin-4 increasing brainstem combined properties of GLP-1 and ex- Two newer products, CJC-1131 and c-Fos in a similar fashion targeting the endin-4 are being developed, some of Albugon, exhibit covalent binding to al- area postrema, further suggesting exhibi- which may exhibit neurotrophic activity. bumin. CJC-1131, which is currently in tion of at least some CNS effects similar to The peptides also protect against apopto- phase 2 clinical trials, was developed by those of GLP-1. In response to a question sis in cultured hippocampal neurons. If the research company ConjuChem based regarding antibody production, Baggio these studies are confirmed, GLP-1 and on their DAC (drug affinity construct) noted that this has been shown in some exendin-4 may be considered neuropro- technology, leading to decreased clear- patients treated with exendin-4, although tective, perhaps involving protein kinase ance. The drug is not prebound to albu- not appearing to affect activity, and has C and cAMP-dependent protein kinase A min before injection but binds covalently not been shown at this point with either pathway activation. In an oxidative insult to endogenous albumin. A modification liraglutide or CJC-1131. model in cultured hippocampal neurons, of the second COOH-terminal amino acid both exendin-4 and GLP-1 showed con- from L-toD-alanine renders the peptide Neural effects of GLP-1 centration-dependent protective effects, resistant to DPP-IV, while not blocking Tracy Ann Perry (Baltimore, MD) dis- with decreased amyloid ␤ protein 1-40 binding to the GLP-1 receptors or reduc- cussed neural actions of incretin peptides. levels, leading Perry to suggest that the ing activation of GLP-1 receptor– GLP-1 receptor expression has been well agent be explored as a treatment for Alz- dependent signal transduction pathways demonstrated in rodent and human heimer’s disease. There is also evidence of in vitro. After a single injection, the half- brain, with circulating GLP-1 entering the a neuroprotective effect in a middle cere- life is ϳ10 days. In CJC-1131–treated brain primarily in the periventricular ar- bral artery ligation stroke model. persons with type 2 diabetes, Baggio eas, although to a lesser extent through- Perry further discussed protective ef- showed evidence of a dose-response re- out the brain. In addition, GLP-1 is fects of GLP-1 and exendin-4 in the pe- duction in a seven-point glycemic profile, synthesized in the nucleus solitarius of ripheral nervous system. In pyridoxine- with evidence of body weight reduction the brainstem. CNS effects of GLP-1 in- induced sensory neuropathy model, in animal models and in the human stud- clude regulation of food intake and body pyridoxine alone led to weight loss, while ies. In a study presented at the meeting, weight consistent with receptor expres- GLP-1 infusion (but not with the receptor Guivarc’h et al. (abstract 535) reported sion in the brain, and CNS GLP-1 signal- antagonist exendin-4 (9-39) protected efficacy of CJC-1131 administered daily ing appears to be due, to large extent, to against the pyridoxine-induced func- to 22 persons with type 2 diabetes for centrally synthesized GLP-1. GLP-1 is tional impairment, as shown by measure- 14–20 days, showing a dose-dependent also a mediator of multiple stress re- ment of inclined screen climbing lowering of fasting and postprandial glu- sponses. Infused into the lateral ventricle, performance. Sciatic nerve morphology cose levels and a weight decrease. Wen et it potentiates the release of corticotropin- showed loss of large-diameter fibers and al. (abstract 634) found no evidence of releasing hormone. Perry noted that the increase in small-diameter fibers, with immunogenicity of CJC-1131 after ad- central nucleus of the amygdala, which both abnormalities improved by GLP-1 ministration of two doses separated by 6 acts as a “fear center,” and the hippocam- and by exendin-4 injection. The degree of weeks. pus, playing a role in memory, also show degeneration was quite marked with pyr- Albugon is a molecule covalently response to GLP-1, with infusion into the idoxine alone but diminished by GLP-1 bound to albumin produced by the com- amygdala leading to anxiety response and and by exendin-4. pany Human Genome Sciences, exhibit- paraventricular infusion suppressing In a study presented at the meeting,

2556 DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 Bloomgarden

Anini et al. (abstract 320) assessed the ef- showed effects in increasing insulin and twice-daily exenatide doses and a signifi- fects on food intake of exendin-4 in mice, decreasing glucagon levels with a conse- cant 1.6-kg weight loss in the 10-␮g showing decreased food intake at doses quent decrease in blood glucose (12). In twice-daily dose group. lower than those decreasing gastric emp- the present study, 113 patients received tying, with capsaicin, which disrupts sen- placebo and the initial dosage for active Clinical studies of DPP-IV inhibitors sory C-type neural pathways, partially treatment was 5 ␮g twice daily; after 4 Bo Ahren (Lund, Sweden) discussed a decreasing the effect of low exendin-4 weeks, 110 of the active treatment pa- 1-year study of the DPP-IV inhibitor doses, suggesting a neural afferent mech- tients continuing this dose and the re- LAF237 in patients with type 2 diabetes anism as well as possible CNS effects at maining 113 participants increased the treated with metformin. He noted that higher doses. Young et al. (abstract 1344) dose to 10 ␮g twice daily. The study was GLP-1 is rapidly eliminated from plasma, noted that the area postrema of the brain completed by 79, 82, and 82% of partic- with t1/2 of 1.5 min, because of inactiva- stem lacks a blood-brain barrier, re- ipants, respectively. From baseline, tion by DPP-IV, a cell surface serine pep- sponds to glucose, as well as to peptides HbA1c levels of 8.2% decreased 0.5 and tidase expressed by endothelial cells and including GLP-1, insulin, amylin, and 0.9% in the 5- and 10-␮g treatment rapidly inactivating biologically active cholecystokinin, and has vagal output, groups at 30 weeks, with 13, 32, and 46% peptides with serine or alanine as the sec- ␮ potentially controlling secretion of these of patients assigned to placebo, 5 g, and ond NH2-terminal peptide. The active ␮ hormones. In a rat model ablating this 10 g, respectively, achieving HbA1c form of DPP-IV exists as a dimmer area, euglycemic clamps following argi- Ͻ7%. A meal tolerance test showed low- (13,14). Thus, Ͼ80% of total circulating nine infusion were associated with the ering of postprandial glucose excursions GLP-1 is inactive due to DPP-IV action doubling of lactate and insulin levels, as well as of fasting glucose in a dose- (8). Animal models not expressing the en- confirming its role in the control of insu- response fashion, with an earlier insulin zyme have been studied and show in- lin secretion. peak and a lower proinsulin-to-insulin ra- creased GLP-1 action (15). Ahre´ n tio in the treatment groups. From a base- reviewed his initial studies of DPP-IV in- Clinical studies of exenatide line BMI of 34 kg/m2, body weight hibition in humans (16). The new DPP-IV Mack et al. (abstract 1717) administered decreased 0.3, 1.6, and 2.8 kg, respec- antagonist LAF237 is similar to the earli- exenatide (the designation for synthetic tively. Nausea occurred in 23, 36, and er-studied compound valine pirolidide, exendin-4 used as a pharmacologic agent) 45% of patients, occurring principally which lacks action on P450 enzymes and to high-fat–fed rats, showing dose-related during the first 2 months, with severe has high enzyme affinity and reversible weight loss over a 28-day period, which at symptoms in 2, 3, and 4% leading to action. Single-dose efficacy studies in per- the highest dose was approximately twice withdrawal of 0, 1, and 2% of patients. sons with type 2 diabetes, as well as as great as that with sibutramine. Hiles et Diarrhea occurred in 8, 12, and 16% and 4-week studies, show decreased fasting, al. (abstract 1585) administered ex- vomiting in 4, 11, and 12% of the respec- prandial, and 24-h glucose levels, with a enatide to mice and rats for 2 years at lev- tive groups. Hypoglycemia occurred in 20-mg/dl decrease in fasting glucose, de- els 6, 25, and 90 times higher (per 5% of participants in each group. In a creased glucagon, increased GLP-1, and kilogram body weight) than the maximal 1-year open-label extension of the study, no change in insulin levels despite lower human dose and showed no increased fre- HbA1c was stable in previously treated pa- glucose levels (17). quency of islet cell hyperplasia, adenoma, tients and decreased by 1.1% in those pre- Ahren presented (see also abstract or carcinoma. viously receiving placebo. Overall, with 354 and abstract 7-LB) a 1-year study of In a human study, Calara et al. (ab- treatment, HbA1c decreased from 8.1 to 107 metformin-treated patients with type stract 508) reported that absorption of ex- 7% and weight decreased from 102 to 98 2 diabetes, 56 of whom received 50 mg enatide was similar after subcutaneous kg at 30 weeks, with further weight loss at LAF237 daily plus metformin at a mean administration in the arm, thigh, and ab- 52 weeks, and multivariate analysis dose of 1.8 g daily and 51 of whom re- domen in 25 persons with type 2 diabe- showed that the weight loss only ex- ceived metformin with placebo. The base- tes. Poon et al. (abstract 588) reported a plained a portion of the improvement in line HbA1c was 7.7% with LAF237 and study of 156 metformin- or diet-treated glycemia. Kendall et al. (abstract 10-LB) 7.8% with placebo. At 3 months, HbA1c persons with type 2 diabetes given pla- reported a similar 30-week study of 733 was 0.7% lower and fasting and mean cebo or 2.5, 5, 7.5, or 10 ␮g exenatide persons receiving sulfonylureas plus met- prandial glucose decreased 22 and 40 mg/ twice daily for 28 days. HbA1c decreased formin, with HbA1c decreasing from 8.5% dl. Forty-two LAF237 and 29 placebo pa- by 0.04, 0.27, 0.37, and 0.49%, respec- by 0.8 and 0.5% with the 10- and 5-␮g tients were followed for 12 months, with tively, from the baseline average of 7.5%, twice-daily exenatide doses, respectively, HbA1c increasing to 8.4% with placebo with weight loss of 0.8, 0.7, 1.4, and with a 1.6-kg weight loss in both groups. but remaining stable at 7.1% with 1.8 kg. Hypoglycemia occurred in 13% of those LAF237, an overall 0.5% decrease vs. Ralph DeFronzo (San Antonio, TX) receiving placebo but in 19 and 28% of 0.6% increase in HbA1c, with 41 vs. 10% ␮ Ͻ reported, at a presentation of late- patients receiving the 5- and 10- g twice- reaching HbA1c 7%. Body weight de- breaking studies and in a poster (abstract daily exenatide doses. Buse et al. (abstract creased similarly by 0.2 kg in both 6-LB) the effects of exenatide treatment 352) reported a final similar 30-week trial groups. In 12 patients treated with 100 for 30 weeks in 336 persons with type 2 of 377 persons with type 2 diabetes re- mg LAF237 daily, HbA1c decreased 0.8% diabetes receiving metformin. He noted ceiving sulfonylureas alone, with a base- at 12 weeks, compatible with a dose- that initial dose-response studies using line HbA1c 8.6%, showing 0.5 and 0.9% related effect. Meal tolerance tests showed ␮ the agent in persons with type 2 diabetes reduction in HbA1c with 5- and 10- g somewhat lower fasting and greater low-

DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 2557 Perspectives on the News ering of postprandial glucose excursion, lated to its degradation of GLP-1. Leiting agent improved glycemia, triglycerides, with increased prandial insulin response. et al. (abstract 6) and Lankas et al. (ab- and free fatty acid levels, with histological No change was reported in homeostasis stract 7) noted that DPP-IV is present as a evidence of increases in insulin-positive model insulin resistance or in lipids. Mild surface marker on activated immune cells islet cells, suggesting that this approach hypoglycemia was seen in three of the pa- (where it is referred to as CD26). Other “may have disease-modifying capacity in tients in the active treatment group, with- peptidases, including QPP, DPP8, and the treatment of type 2 diabetes.” out other adverse events described. Thus, DPP9, may also be involved in immune In a human study, Heins et al. (ab- Ahren characterized the DPP-IV inhibitor regulation, and in an in vitro T-cell acti- stract 539) reported effects of the DPP-IV as a well-tolerated weight-neutral and ef- vation model, they showed that the non- inhibitor P93/01, developed by the com- fective agent achieving glucose lowering selective inhibitor Val-boro-Pro and a pany Probiodrug. Following a 240-mg comparable to GLP-1 agonists while start- DPP8/9 inhibitor inhibited proliferation, oral dose in 16 persons with mild type 2 Ͻ ing from a somewhat lower baseline level whereas there was no T-cell effect in their diabetes, in 8 who had HbA1c 6%, of glycemia. Pratley and Galbreath (ab- model of specific DPP-IV inhibition. GLP-1 and GIP increased, without a stract 355) reported a 12-week study of Other adverse effects of DPP8/9 inhibi- change in glycemia, while those with Ͼ 72 persons with type 2 diabetes not re- tion were, in rats, alopecia, thrombocyto- HbA1c 6% also had a fall in postmeal ceiving other pharmacologic therapy who penia, anemia, splenomegaly, and death glycemia. Herman et al. (abstract 353) ad- were treated with 25 mg LAF237 twice and, in dogs, bloody diarrhea and emesis, ministered 25 or 200 mg MK-0431 to 56 daily. The fall in HbA1c, adjusted for the while QPP inhibition lowered reticulo- persons with type 2 diabetes not receiving response of 28 patients receiving placebo, cyte counts in rats and had no effects in other pharmacologic treatment and was 0.6%, with evidence of a greater fall at dogs. Masur et al. (abstract 49-LB) showed a doubling of GLP-1 levels and a Ͼ higher baseline HbA1c levels at 0.7 and showed that in vitro migration of stimu- 20% increase in insulin with both doses 1.2% for baseline HbA1c 7– 8 and lated human CD8 T-cells was reduced by and 22 and 26% decreases in the blood 8 –9.5%, respectively. Ten percent of ϳ40% by coincubation with GLP-1, sug- glucose increment following oral glucose treated patients had one or more hypogly- gesting physiologic effects on gut and islet and placebo, respectively. cemic episode. immune function but also raising the pos- Ahrens was asked whether, as the du- sibility of untoward effects with exoge- Comparison of GLP-1 with other ration of action of LAF237 appears to be nous administration. pharmacologic agents in diabetes ϳ 12 h, it might be appropriate to use the Larsen et al. (abstract 1413) showed Zander et al. (abstract 1422) compared agent twice daily and noted that this is an evidence of efficacy of another orally ad- glycemic effects of pioglitazone and important area for future study with this ministered DPP-IV inhibitor, NN7201, GLP-1 in eight persons with type 2 diabe- agent. Addressing the lack of weight loss which lowered postload glucose and in- tes, showing a decrease in fasting plasma with the agent, he pointed out that creasing GLP-1 and GIP levels in a glucose from 13.5 mmol/l with neither DPP-IV inhibitors do not achieve the minipig model of mild insulin-deficient drug to 11.7 mmol/l with GLP-1, which same level or duration of GLP-1 action diabetes. Ahren and Hughes (abstract was administered by continuous subcuta- seen with injected GLP-1 agonists, sug- 1406) studied the effect of DPP-IV inhibi- neous infusion, to 11.5 mmol/l with 30 gesting that more potent agents might tion with valine-pyrrolidide in mice ad- mg pioglitazone daily and to 9.5 mmol/l have additional effects. He was asked ministered intravenous glucose alone or with both agents. Insulin levels were about whether other biologically active with GLP-1, GIP, pituitary adenylate cy- higher and glucagon levels lower with peptides regulated by DPP-IV might affect clase–activating polypeptide, or gastrin- GLP-1 compared with pioglitazone, and blood pressure and stated that mice not releasing peptide and showed that the sensation of appetite was reduced with expressing the enzyme do not exhibit insulin response to these agents was in- GLP-1 alone or in combination. Although such abnormality, suggesting safety in creased by 80, 40, 75, and 25%, respec- preliminary, the study suggests that thia- this regard. tively. Thus, the effects of DPP-IV zolidinedione–GLP-1 combinations offer A number of additional reports were inhibitors may be not limited solely to ac- an additional promising treatment ap- presented at the meeting addressing tions on GLP-1. Weber et al. (abstract proach. aspects of DPP-IV inhibition. Bose et al. 633-P) studied the Merck DPP-IV inhibi- (abstract 2) studied potential cardiopro- tor MK-0431 in animal models and tective effects of GLP-1 in an ischemia- showed similar action in decreasing post- References reperfusion model with or without glucose load glycemia in insulin-resistant 1. Toft-Nielsen MB, Damholt MB, Madsbad administration of GLP-1, showing a de- models. Petrov et al., from the same S, Hilsted LM, Hughes TE, Michelsen BK, crease in infarct size both in vivo and in group, used the DPP-IV inhibitor (2S,3S)- Holst JJ: Determinants of the impaired se- vitro in isolated perfused rat hearts, sug- isoleucyl thiazolidide, showing delayed cretion of glucagon-like peptide-1 in type gesting that the effect is not mediated by progression of hyperglycemia and devel- 2 diabetic patients. J Clin Endocrinol Metab increased insulin secretion. Curiously, in opment of insulin deficiency in female 86:3717–3723, 2001 vivo pretreatment with valine pyrolidide, Zucker diabetic fatty rats fed a high-fat 2. Creutzfeldt W, Ebert R, Nauck M, Stock- ␤ mann F: Disturbances of the entero-insu- which inhibits DPP-IV to prevent GLP-1 diet, thereby suggesting prevention of “ - lar axis. Scand J Gastroenterol Suppl 82: breakdown, blocked the cardioprotective cell exhaustion.” Zhang et al. (abstract 58- 111–119, 1983 effect, suggesting either a need for short LB) studied an insulin-deficient mouse 3. Nauck MA, Heimesaat MM, Orskov C, duration of GLP-1 effect or an adverse type 2 diabetic model, showing that a Holst JJ, Ebert R, Creutzfeldt W: Pre- consequence of DPP-IV inhibition unre- 3-month period of treatment with this served incretin activity of glucagon-like

2558 DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 Bloomgarden

peptide 1 [7–36 amide] but not of syn- diabetic patients and in healthy subjects. 3089, 2003 thetic human gastric inhibitory polypep- Diabetes 44:1126–1131, 1995 13. Mentlein R: Dipeptidyl-peptidase IV tide in patients with type-2 diabetes 9. Knudsen LB, Nielsen PF, Huusfeldt PO, (CD26): role in the inactivation of regula- mellitus. J Clin Invest 91:301–307, 1993 Johansen NL, Madsen K, Pedersen FZ, tory peptides. Regul Pept 85:9–24, 1999 4. Vilsboll T, Krarup T, Madsbad S, Holst JJ: Thogersen H, Wilken M, Agerso H: Potent 14. Ludwig K, Yan S, Fan H, Reutter W, Bot- Defective amplification of the late phase derivatives of glucagon-like peptide-1 tcher C: The 3D structure of rat DPPIV/ insulin response to glucose by GIP in with pharmacokinetic properties suitable CD26 as obtained by cryo-TEM and obese type II diabetic patients. Diabetolo- for once daily administration. J Med Chem single particle analysis. Biochem Biophys gia 45:1111–1119, 2002 43:1664–1669, 2000 Res Commun 304:73–77, 2003 5. Meier JJ, Hucking K, Holst JJ, Deacon CF, 10. Degn KB, Juhl CB, Sturis J, Jakobsen G, 15. Marguet D, Baggio L, Kobayashi T, Ber- Schmiegel WH, Nauck MA: Reduced in- Brock B, Chandramouli V, Rungby J, nard AM, Pierres M, Nielsen PF, Ribel U, sulinotropic effect of gastric inhibitory Landau BR, Schmitz O: One week’s treat- Watanabe T, Drucker DJ, Wagtmann N: polypeptide in first-degree relatives of pa- ment with the long-acting glucagon-like Enhanced insulin secretion and improved tients with type 2 diabetes. Diabetes 50: peptide 1 derivative liraglutide (NN2211) glucose tolerance in mice lacking CD26. 2497–504, 2001 markedly improves 24-h glycemia and ␣- Proc Natl Acad SciUSA97:6874–6879, 6. Nauck MA, Wollschlager D, Werner J, and ␤-cell function and reduces endoge- 2000 Holst JJ, Orskov C, Creutzfeldt W, Willms nous glucose release in patients with type 16. Ahren B, Simonsson E, Larsson H, B: Effects of subcutaneous glucagon-like 2 diabetes. Diabetes 53:1187–1194, 2004 peptide 1 (GLP-1 [7–36 amide]) in pa- 11. Juhl CB, Hollingdal M, Sturis J, Jakobsen Landin-Olsson M, Torgeirsson H, Jansson tients with NIDDM. Diabetologia 39: G, Agerso H, Veldhuis J, Porksen N, PA, Sandqvist M, Bavenholm P, Efendic S, 1546–1553, 1996 Schmitz O: Bedtime administration of Eriksson JW, Dickinson S, Holmes D: In- 7. Zander M, Madsbad S, Madsen JL, Holst NN2211, a long-acting GLP-1 derivative, hibition of dipeptidyl peptidase IV im- JJ: Effect of 6-week course of glucagon- substantially reduces fasting and post- proves metabolic control over a 4-week like peptide 1 on glycaemic control, insu- prandial glycemia in type 2 diabetes. Di- study period in type 2 diabetes. Diabetes lin sensitivity, and beta-cell function in abetes 51:424–429, 2002 Care 25:869–875, 2002 type 2 diabetes: a parallel-group study. 12. Kolterman OG, Buse JB, Fineman MS, 17. Ahren B, Landin-Olsson M, Jansson PA, Lancet 359:824–830, 2002 Gaines E, Heintz S, Bicsak TA, Taylor K, Svensson M, Holmes D, Schweizer A: In- 8. Deacon CF, Nauck MA, Toft-Nielsen M, Kim D, Aisporna M, Wang Y, Baron AD: hibition of dipeptidyl peptidase-4 re- Pridal L, Willms B, Holst JJ: Both subcu- Synthetic exendin-4 (exenatide) signifi- duces glycemia, sustains insulin levels, taneously and intravenously administered cantly reduces postprandial and fasting and reduces glucagon levels in type 2 di- glucagon-like peptide I are rapidly de- plasma glucose in subjects with type 2 di- abetes. J Clin Endocrinol Metab 89:2078–

graded from the NH2-terminus in type II abetes. J Clin Endocrinol Metab 88:3082– 2084, 2004

DIABETES CARE, VOLUME 27, NUMBER 10, OCTOBER 2004 2559