Reviews/Commentaries/ADA Statements PERSPECTIVES ON THE NEWS

Diabetes Treatment

ZACHARY T. BLOOMGARDEN, MD -stimulated state the insulin resis- tance of is largely ac- counted for by skeletal muscle insulin his is the third of six articles based the program found a 7.9% prevalence of resistance. Intramyocellular defects in- on presentations at the American diabetic retinopathy among individuals clude impaired glucose transport and de- T Diabetes Association Scientific Ses- with IGT. At the time of diabetes diagno- creased glycogen synthesis. Insulin action sions held 6–10 June 2008 in San Fran- sis, 12.6% had retinopathy, although the begins with insulin receptor auto- cisco, California. mean A1C was only 6.1%. Peripheral phosphorylation, then causing phosphor- neuropathy also was seen in 5–10% of ylation of insulin receptor substrate Type 2 diabetes treatment participants with IGT. DeFronzo con- (IRS)-1, leading to activation of a number approaches cluded that individuals with IGT are near of intracellular processes, with a decrease Ralph DeFronzo (San Antonio, TX) sug- maximally insulin resistant, with de- in the ability of the insulin receptor to gested an interesting set of approaches to creased ␤-cell function and mass, and tyrosine phosphorylate IRS-1 in insulin the treatment of type 2 diabetes. De- with appreciable prevalence of diabetic resistance. At the same time, the mito- Fronzo noted that the natural history of complications. genic insulin response pathway is rela- type 2 diabetes involves a reduction in ␤-cell failure occurs, DeFronzo said, tively increased, with activation of insulin sensitivity during the progression in an age-related fashion and clusters in proinflammatory pathways, abnormali- from lean to obese with normal tolerance, families. The transcription factor 7–like 2 ties which only respond to pharmacologic and that the subsequent progression to gene (TCF7L2) polymorphism is associated intervention with impaired glucose tolerance (IGT) is asso- with reduced insulin secretion, perhaps (TZDs). ciated with a further decrease in insulin from reduced insulin responsiveness to glu- DeFronzo showed fascinating differ- sensitivity and a relative deficiency in insu- cagon-like peptide (GLP)-1. Carriers of the ences between the effects of oral and par- lin secretory function. As IGT progresses to abnormal gene have a 60% increase in dia- enteral glucose. The latter only increases diabetes, insulin secretion decreases betes development. Deficiency of GLP-1 hepatic glucose uptake when plasma glu- without a further worsening in insulin and resistance to glucose-dependent insuli- cose levels increase, even during hyperin- sensitivity. DeFronzo presented studies of notropic peptide (GIP) action also occur in sulinemia. Oral glucose, in contrast, normal glucose-tolerant, impaired glu- type 2 diabetes. Lipotoxicity from increased markedly increases hepatic glucose up- cose-tolerant, and type 2 diabetic individ- plasma free fatty acids (FFAs) is another take in normal individuals, acting to a uals that demonstrated an increase in the factor impairing insulin secretion. A 48-h lesser extent in type 2 diabetic patients, absolute rate of glucose-induced insulin infusion of heparin with triglyceride which suggests an abnormality of a gut secretion during the progression from emulsion elevating FFAs in normal glu- factor. Increased FFAs may play a role in normal to varying degrees of IGT, with cose tolerant offspring of two diabetic inhibiting muscle glucose uptake, in- insulin secretion subsequently decreasing parents led to decreased insulin secretion. creasing hepatic glucose production, and progressively with worsening degrees of Increased glucose levels also impair insu- decreasing insulin secretion. The use of diabetes. Examining the ratio of insulin lin secretion, the phenomenon of gluco- lipid plus heparin infusion to elevate FFA secretion to insulin resistance (the dispo- toxicity; DeFronzo’s studies of phlorizen, in normal individuals decreases hepatic and sition index), DeFronzo showed that the which reduces glucose levels by increas- muscle insulin signaling via a number of ty- logarithm of insulin secretion/insulin re- ing glycosuria, showed improvement in rosine phosphorylation steps and results in sistance is inversely proportional to the ␤-cell function. Increased amyloid a doubling of muscle lipid content. Piogli- log of the 2-h glucose, and that with ad- polypeptide(IAPP) deposition is another taozne increases the expression of perox- vanced degrees of IGT, ϳ80% of insulin factor leading to ␤-cell failure, with rela- isome proliferators–activated receptor secretory capacity is lost, implying that tive islet amyloid area increased in associ- (PPAR)-␥ coactivator (PGC)-1, thereby insulin deficiency begins well before the ation with decreased insulin secretion and reducing intramyocellular lipid and fatty onset of diabetes as currently defined. He increased fasting glucose in a nonhuman acylCoA content, an effect similar to that cited an autopsy study showing that by primate study. with administration of the nicotinic acid the time elevations occur in fasting glu- DeFronzo discussed insulin resis- derivative acipimox to reduce circulating cose, there is a 50% loss of ␤-cell mass, tance in type 2 diabetes, noting potential FFAs. Decreased incretin effect is another with a further decrease in ␤-cell volume differences between the fasting and insu- factor in the pathogenesis of type 2 diabe- with progression to diabetes (1). lin-stimulated states. In type 2 diabetic tes. A 2-week course of in type The Diabetes Prevention Program patients, elevation in basal hepatic glu- 2 diabetic patients showed beneficial ef- further raises concern about the clinical cose production correlates strongly with fects, inlcuding an improved ratio of in- implications of the term “pre-diabetes,” as increase in fasting glucose, while in the sulin secretion to 2-h glucose and ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● increased splanchnic glucose uptake. Ab- Zachary T. Bloomgarden, MD, is a practicing endocrinologist in New York, New York, and is affiliated with normalities of ␣-cell function may be an- the Division of Endocrinology, Mount Sinai School of Medicine, New York, New York. other factor in the pathogenesis of type 2 DOI: 10.2337/dc09-zb03 © 2009 by the American Diabetes Association. Readers may use this article as long as the work is properly diabetes. Marking another benefit of in- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons. cretin treatment approaches, glucagon se- org/licenses/by-nc-nd/3.0/ for details. cretion is increased and correlates with

DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009 e25 Perspectives on the News increased fasting glucose levels, and this effect on AMPK might explain its effect on those receiving combined - further improves after administration of gluconeogenesis. (Abstract numbers refer treatment. Spanheimer et al. somatostatin. In addition, there are cen- to the ADA Scientific Sessions, Diabetes (abstract 299) reported that 33% of pa- tral nervous system effects on glycemia, 57 [Suppl. 2], 2008). tients in the PROspective and the hyperinsulinemia of obesity may This work was supported by Baverel Clinical Trial In macroVascular Events involve central insulin resistance, with et al. (abstract 38), who found a dose- (PROactive) were treated with insulin, evidence of altered hypothalamic func- dependent inhibition by metformin of 39% with a nitrate, and 70% with an tion in obese individuals after glucose gluconeogenesis from lactate in liver angiotensin-directed agent, with no evi- ingestion. slices from Zucker diabetic fatty rats and a dence of these treatments increasing risk Given the variety of pathogenic reduction of cellular ATP levels and of of stroke, myocardial infarction, or mor- abnormalities in type 2 diabetes, its treat- CO2 production from lactate, while lac- tality. Seung Jin Han et al. (abstract 435) ment requires multiple drugs in combina- tate production and ketogenesis nearly administered pioglitazone versus placebo tion. Metformin and TZDs act on the liver, doubled with increased ␤-hydroxybu- to 75 nondiabetic renal allograft recipi- and TZDs act on muscle, the adipocyte, tyrate–acetoacetate ratio, reflecting the ents for 12 months, and found a signifi- and the ␤-cell, suggesting to DeFronzo mitochondrial redox state. Schaefer et al. cant increase versus decrease in insulin that these agents are preferable to met- (abstract 97) treated 19 nondiabetic obese sensitivity and a decrease versus increase formin and to . All long- adults with 850 mg metformin daily for in carotid intima-media thickness. Bao term TZD studies, he said, including one week, then twice daily for three more et al. (abstract 441) studied outcome PERISCOPE, CHICAGO, ADOPT, and weeks, showing a reduction in 24-h en- among 3,713 diabetic patients treated the UKPDS, show that sulfonylureas do ergy expenditure by 3% with carbohy- with metformin alone for at least 12 not give durable glycemic benefit, while drate and fat oxidation increasing 17% months; 29 and 71% of the patients sub- long-term glucose-lowering is seen with and decreasing 33%, respectively. sequently took and a sulfo- TZDs in type 2 diabetic patients and in A number of other agents may regulate nylurea, respectively. On average, the prevention studies such as the DPP, TRI- pathways similar to those affected by met- latter group was 2 years older; the patient POD, PIPOD, DREAM, and ACT-NOW. formin. Van Poelje et al. (abstract 350) groups had a similar sex distribution and “The TZDs and the GLP-1 analogs,” De- found a reduction in glucose production prevalence of hypertension and CVD, and Fronzo concluded, “offer a new therapeu- from lactate in human hepatocytes with baseline resource utilization was similar. tic approach.” This is, he said, preferable the fructose-1,6-bisphosphatase inhibitor Comparing those adding rosiglitazone to the stepwise approach of typically us- MB07803. In diabetic rodents and nonhu- versus sulfonylurea, 74 vs. 69% were ad- ing metformin followed a sulfonylurea man primates, glucose decreased without herent to glucose-lowering therapy, 23 recommended by ADA, which he charac- any change in blood lactate. Motoshima et vs. 27% experienced hospitalizations, 26 terized as “nonphysiological.” He recom- al. (abstract 351) found that the protein vs. 29% had ER visits, and 54 vs. 58% had mended a “pathophysiologic-based kinase C␦ inhibitor rottlerin decreased outpatient visits. There was a 40% greater algorithm” of initial treatment with life- AMPK phosphorylation in adipocytes, adjusted likelihood of adherence to rosigli- style, TZDs, metformin, and exenatide, myocytes, and hepatocytes and increased tazone, and rates of both hospitalizations with an A1C goal Ͻ6%, suggesting that cellular glucose consumption; the latter and ER visits were 20% lower among those this would be durable, would result in effect is not observed with overexpression who added rosiglitazone. ␤-cell preservation, and would not cause of dominant-negative AMPK, which sug- Tint et al. (abstract 303) administered hypoglycemia or weight gain. gests this phenomenon to mediate the rosiglitazone for 16 weeks to 14 type 2 dia- These and many additional ap- glucose-lowering effect observed in ani- betic patients of Chinese and Asian Indian proaches to treatment of type 2 diabetes mal models in vivo. ethnicity. Euglycemic-hyperinsulinemic were explored in studies presented at the clamp insulin sensitivity increased 52 vs. ADA meeting. PPAR␥ agonists 120%, respectively, with somewhat more A number of studies at the ADA meeting weight gain in those of Asian Indian eth- Metformin contributed to the endeavor to unravel nicity. Kritchevsky et al. (abstract 1736) Foretz et al. (abstract 1507) investigated cardiovascular (CV) risks versus benefits administered 30 mg pioglitazone daily the relationship between metformin’s ac- of the TZDs. Bilik et al. (abstract 300) versus placebo to 88 nondiabetic adults tivation of AMP-activated protein kinase compared 8,739 type 2 diabetic patients who had a BMI Ͼ27 kg/m2 and were on a (AMPK) and its inhibition of gluconeo- who were followed from 1999 to 2003 calorie-restricted diet for 4 months; genesis, finding that although hepato- and either received or did not receive a weight loss did not differ between the pio- cytes from mice not expressing AMPK TZD. Mortality among the patients was 5 glitazone and placebo groups, but men had a 30% reduction in gluconeogenesis, vs. 7%, and major CV events occurred in receiving pioglitazone had 3% reduction both in the basal state and in response to 11 vs. 10%, respectively. A total of 817 in percent body fat, while there was a 2% cyclic AMP, metformin reduced glucose took just rosiglitazone and 724 just pio- reduction in the placebo group; there was production to a greater extent in the glitazone, with major CV events in 9 and a greater reduction in visceral fat among knockout hepatocytes than in those from 10%, respectively. However, Wang et al. pioglitazone-treated men. wild-type animals. Mice overexpressing (abstracts 590 and 604) analyzed 11,283 Chou et al. (abstract 304) compared a PGC-1␣, which is distal to AMPK in acti- type 2 diabetic patients receiving either new TZD, rivoglitazone, at 1, 2, and 3 mg vation of gluconeogenesis, continued to metformin or a sulfonylurea alone at doses, with pioglitazone 45 mg daily and respond to metformin. The authors found baseline. They found a 23% greater like- with placebo in a study of 441 type 2 di- that metformin reduced intracellular lihood of a CV event among patients re- abetic patients. A1C decreased by 0, 0.4, ATP, suggesting that this rather than its ceiving add-on rosiglitazone than among 0.5, and 0% and increased 0.6%, re- e26 DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009 Bloomgarden spectively. Triglyceride decreased 10, colesevelam daily versus placebo for 8 line, A1C was7.7–8% and decreased by 15, and 21% with the 1, 2, and 3 mg weeks, finding no effect on the glucose 0.7–0.9% with dapagliflozin, 0.7% with doses and 8% with pioglitazone, while response to a standardized meal tolerance metformin, and 0.2% with placebo, and HDL cholesterol increased 11, 10, 14, test. This finding suggests the effect of the there were 2.7–3.4, 1.7, and 1.2% weight and 8%, respectively. Peripheral edema, agent is not mediated by altered glucose losses, respectively. Adverse events with however, occurred in 14, 17, 24, and absorption. Jialal et al. (abstract 459) an- dapagliflozin included urinary tract infec- 11%, respectively, and weight gain was alyzed the pooled effect of the bile acid tion, nausea, dizziness, headache, fatigue, also more likely to occur at the 2 and 3 mg binding resin colesevelam in 1,081 type 2 back pain, and nasopharyngitis. doses. Truitt et al. (abstract 437) studied diabetic patients receiving insulin, met- Chaudhury et al. (abstract 729), however, 426 patients receiving 0.5, 2, and 5 mg formin, or a sulfonylurea, and found a in an effort to address the question of rivoglitazone, 30 mg pioglitazone, and 0.5% placebo-adjusted reduction in A1C, whether glycosuria is associated with renal placebo. The 2 and 5 mg doses had more a 15 mg/dl reduction in fasting glucose, tubular damage in 106 newly diagnosed potent glycemic effects than pioglitazone, and a 15% reduction in LDL cholesterol untreated type 2 diabetic individuals, although edema occurred in 6 and 16% of but a 7% reduction in non-HDL choles- showed the degree of glycosuria to correlate those receiving the 2 and 5 mg doses but terol, reflecting a 15% increase in triglyc- with a marker of proximal tubular damage. in only 0–1% of those receiving pioglita- eride levels. Guha et al. (abstract 439) A1C was an independent predictor, raising zone. There was also greater weight gain administered an agonist of the gut bile the question of whether a therapeutic ap- with the higher rivoglitazone doses. An acid receptor TGR5 in type 2 diabetic an- proach to increasing glycosuria might have interesting implication is that activation imal models, showing an improvement in adverse renal effects. of PPAR␥ is submaximal with existing glycemia and insulin sensitivity and in- TZDs at recommended dosages, with ad- creased active GLP-1 levels in portal and G protein–coupled receptor ditional glucose lowering possible, al- systemic circulation. Brufau et al. (ab- Fyfe et al. (abstract 297) studied PSN821, though the greater risks of fluid retention stract 1553) reported the cholic acid syn- an agonist of G protein–coupled receptor and weight gain may make the more po- thesis rate to be increased by 70% in type (GPR) 119 expressed in pancreas and gut, tent agents not clinically viable. 2 diabetic patients, with a consequent in- and showed stimulation of both ␤-cell in- Dunn et al. (abstract 499) adminis- crease in deoxycholic acid synthesis, pool sulin and gut GLP-1 secretion in vitro and tered the non-TZD partial PPAR␥ agonist size, and total bile acid synthesis. As bile improved glucose tolerance in type 2 pre- INT131 to 69 type 2 diabetic patients not acids are ligands for nuclear FXR and cell diabetic and diabetic animal models. A1C receiving a glucose-lowering agent. Fast- membrane TGR5 receptors, this may be was lower in the latter, and weight was ing glucose increased from 165 by 8 related to abnormal glycemia in diabetes reduced in a dietary obesity model. Trem- mg/dl with placebo and decreased from and to the beneficial effect of bile acid– blay et al. (abstract 1489) evaluated mice 163 and from 184 by 22 and 46 mg/dl binding resins. not expressing GPR-39, which is nor- with 1 mg and 10 mg doses, respectively. mally expressed in the gastrointestinal Guha et al. (abstract 1478) studied the Sodium-glucose cotransporter 2 tract, adipose tissue, liver, and pancreatic effect of the PPAR␦ agonist KD3010, inhibitors islets. The researchers found reduced se- which exhibits Ͼ1,000-fold selectivity The kidney filters 160 g glucose daily, rum insulin and elevated glucose levels over human PPAR␣ and -␥ and has been with 90% reabsorbed by sodium-glucose associated with a high-fat diet or aging, associated with weight loss, in diabetic cotransporter 2 (SGLT2) and 10% by which suggests that agonists of GPR-39 db/db mice. A1C, fasting insulin, and SGLT1 in the renal tubules. Interestingly, might have glucose-lowering effects. postload glycemia decreased. Multani et in animal models of diabetes and in dia- Zhou et al. (abstract 541) studied activa- al. (abstract 569) administered this agent betic patients, the maximal renal tubular tors of GPR-40 (which is expressed in to normal and obese volunteers, improv- reabsorptive capacity is increased. pancreatic ␤-cells) and found enhanced ing peripheral insulin resistance and re- Wancewicz et al. (abstract 334) adminis- glucose-dependent insulin secretion in ducing fasting insulin levels; no weight tered ISIS 388626, an SGLT2 antisense vitro and improved glucose tolerance in gain or signs of fluid retention or other tox- oligonucleotide designed to specifically type 2 diabetic models. icity were exhibited. Marita (abstract 196) distribute to the kidney, in canine and ro- studied a non-TZD, P1736-05, that does dent diabetic models. Administration of Glucokinase activators not activate human PPAR␥ or -␣ receptors ISIS 388626 resulted in improved glucose Glucokinase (GK) has glucose affinity in but increases adipocyte glucose uptake via a levels and may be an effective treatment the physiologic range of 5–12 mmol/l/l, process involving phosphatidylinositol-3 modality. List et al. (abstracts 329 and allowing it to function as a glucose sensor. kinase and thereby induces translocation of 461) administered 2.5–50 mg of the renal The diabetes variant MODY2 is caused by GLUT4 transporter to the plasma mem- SGLT2 inhibitor dapagliflozin daily, decreased hepatic GK activity, while acti- brane. In a type 2 diabetic model, this pro- 1,500 mg metformin daily, or placebo to vating GK mutations cause hyperinsu- cess reduces glucose and triglyceride levels 389 treatment-naïve type 2 diabetic pa- linemic hypoglycemia of infancy. GK acts and improves muscle insulin-induced glu- tients for 12 weeks, and found dose- in the ␤-cell to form glucose-6-phosphate cose uptake without increasing plasma vol- related 52–85 g/day glycosuria with and increase intracellular ATP, closing the ume at 60-fold the effective dose. dapagliflozin. There was no change in se- ATP-sensitive potassium channel, depo- rum sodium, potassium, or creatinine or larizing the cell, and opening a calcium Bile acid sequestrants in type 2 in serum or urinary calcium. Magnesium channel, thereby leading to insulin secre- diabetes increased 0.1–0.2 mEq/l, urate decreased tion. As such, there has been interest in Schwartz et al. (abstract 440) randomized 1 mg/dl, and serum phosphate increased GK activators as insulin secretagogues. In 35 type 2 diabetic patients to 3.75 g 0.2 mg/dl at the highest doses. At base- the liver, GK is the rate-limiting step for

DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009 e27 Perspectives on the News

haploinsufficiency of the GK gene. In NEWS FROM THE FOOD AND DRUG ADMINISTRATION high-fat-fed mice, glucose tolerance im- proved with the agent, again with and From time to time, new announcements by the FDA pertaining to aspects of diabetes without deletion of one copy of the GK treatment will be highlighted in this section. gene. Bodvarsdottir et al. (abstract 328) An interview with Dr. David Orloff, former director of the division of metabolism studied the liver-specific GK activator and endocrinology products at the FDA, reviewed the recent FDA cardiovascular TTP355, showing increases in vitro in he- risk assessment guidelines for diabetes drugs, pointing out that the guidelines patocyte glucose metabolism, without ef- would increase the cost and time of developing a diabetes drug and suggesting fect on insulin secretion, and showing that several companies are likely to discontinue development of their diabetes improvement in glycemia in a type 2 dia- drugs (interview downloaded January 26, 2009, from http://www.close betic animal model. Bonadonna et al. (ab- concerns.com/). To understand this, it may be useful to review the guidelines, stract 322) reported improved glucose which state, “For completed studies, before submission of the new drug appli- levels and increased insulin secretion in cation (NDA)/biologics license application (BLA), sponsors should compare the 15 mild type 2 diabetic patients receiving incidence of important cardiovascular events occurring with the investigational another GK activator, RO4389620. agent to the incidence of the same types of events occurring with the control group to show that the upper bound of the two-sided 95 percent confidence Dipeptidyl peptidase-4 inhibitor interval for the estimated risk ratio is less than 1.8” (http://www.fda.gov/cder/ treatment guidance/8576fnl.pdf). Hjollund et al. (abstract 1464) measured portal vein active GLP-1 levels in pigs, Consider drugs that are completely neutral with regard to cardiovascular out- finding an increase from 6.6 to 45.1 come. Say that drug A from company A is tested in a population of 2,000 patients, pmol/l after administration of bombesin with 2,000 control subjects, in whom the cardiovascular event rate is 1%. If both (neuromedin C). After dipeptidyl pepti- groups have precisely 20 cardiovascular events, the relative risk (RR) ratio is, of dase-4 (DPP-4) inhibition with vildaglip- course, 1.0. The 95% CI, however, would be 0.5397-1.8528. Now, say company tin, GLP-1 increased from 16.3 to 90.3 B performed the same studies, but in their case there were 20 events among the pmol/l. Portal levels were two to three 2,000 patients receiving drug B but 21 events among the 2,000 control subjects. times greater than peripheral blood levels, The relative risk is now 0.95, with a 95% CI of 0.5179-1.7514 (CIs for RRs and potentially acting on the liver and on va- odds ratios calculated with a statistical calculator downloaded 25 January 2009 gal afferents, which the authors suggest from vl.academicdirect.org/applied_statistics/binomial_distribution/ref/CI might explain the comparable glycemic calculator.xls, according to the methods described in the following source: Ar- effect of DPP-4 inhibitors to those of mitage P, Berry G: Statistical Methods in Medical Research. 3rd ed. London, Black- GLP-1 receptor activators. Peripheral well, 1994, p. 131). Certainly, drug A and drug B have indistinguishable blood GLP-1 receptor activation appears cardiovascular risk. Yet, the language of the FDA document states clearly that to be much lower with DPP-4 inhibition, company A, but not company B, will need to undertake a postmarketing trial “to but portal levels may be comparable. Aul- definitively show that the upper bound of the two-sided 95 percent confidence inger et al. (abstracts 1,545 and 1,551) interval for the estimated risk ratio is less than 1.3.” Company A now needs to reported that although neither GLP-1 nor study a total of 12,000 patients in each group to show that, with 1% cardiovas- reduced food intake given cular event rates, the same RR of 1.0 has a 95% CI of 0.7774-1.2863. It appears, separately in a rat feeding model, com- then, that arbitrary and statistically meaningless differences may lead some com- bined administration was effective. Ex- panies to abandon the development of potentially promising new therapies for enatide showed a more potent and diabetes—certainly an undesirable and hopefully an unintended result of the longer-lasting anorexic effect and, inter- FDA guidance. estingly, the combination of exenatide The FDA issued a Public Health Advisory to alert consumers, patients, health with vildagliptin suppressed food intake care professionals, and caregivers about potentially serious and life-threatening to an even greater extent, suggesting a side effects from the improper use of skin-numbing products such as lidocaine, potential clinical benefit of combined which may be administered to patients with painful diabetic neuropathy. Noting treatment of overweight patients with di- abetes. Two interesting studies suggest that application under occlusion or when skin temperature is increased may lead ␣ to systemic absorption, the advisory stated that the agents may be associated with that some of the effects of -glucosidase arrhythmia, seizures, respiratory difficulty, or decreased mental status. inhibition may be mediated by changes in incretin secretion. Narita et al. (abstract 447) reported effects of on GLP-1 and GIP responses to a mixed meal in nine glucose metabolism and it increases gly- without hyperinsulinemia or weight gain. type 2 diabetic patients, finding a modest cogen formation, so that GK activators In addition to the liver, the ␤-cell, and gut increase in GLP-1 concentrations by ap- could also increase hepatic insulin action. L- and K-cells, GK is expressed in ␣-cells proximately one-third at 60 and 120 min, Archer et al. (abstract 465) studied the and in hypothalamic neurons involved in but a marked reduction in GIP by 60% at small molecule GK activator ARRY-588, physiologic glucose-sensing. Nakamura 30 and 60 min, with a 3-h integrated in- which is capable of increasing glucose- et al. (abstract 493) showed that a small crease in GLP-1 and decrease in GIP by 14 induced ␤-cell insulin secretion as well as molecule GK activator increased glucose- and 47%, respectively. Goto et al. (ab- that of GIP and GLP-1, and of reducing stimulated insulin secretion in islets from stract 470) administered miglitol and the glucose levels in type 2 diabetic models, mice with and without ␤-cell–specific DPP-4 inhibitor SK-0403 in combination e28 DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009 Bloomgarden in a rat model, showing that after a mixed one-half and one-quarter with glomerular tive state, with inhibition of IR-PTP meal the combination increased GLP-1 filtration rates Ͻ50 and Ͻ30 ml/min, re- prolonging the insulin signal. Mice not levels to a greater extent than the DPP-4 spectively, although there is a presumable expressing this enzyme display remark- inhibitor alone. Miglitol alone did not overlap between the 1.7- and 2.1-fold in- able tissue specificity of insulin sensitiv- change GLP-1 levels, and curiously the creases; therefore, the dose might also be ity, with a reduction in fed blood glucose, GLP-1 response to oral sucrose was less reduced in the 51–80 ml/min group. a 50% lowering of insulin levels, and in- with the combination than with SK-0403 Fleck et al. (abstract 479) administered creased tyrosine phosphorylation of the alone. As in the human study, GIP levels 6.25, 12.5, 25, 50, or 100 mg IR in muscle and liver, without effect in after the mixed meal were reduced by ad- daily or placebo for 12 weeks to 265 type adipose tissue, leading to resistance to di- ministration of miglitol. 2 diabetic patients not receiving pharma- et-induced obesity (2). Hormone signal- Williams-Herman et al. (abstract cologic treatment. The researchers found ing in adipocytes involves both insulin 495) and Katzeff et al. (abstract 496) lacebo-adjusted A1C reductions of 0.2, and ␤-adrenergic pathways. Insulin leads found, adjusting for baseline A1C, a 0.5, 0.6, 0.4, and 0.5%, respectively, from to phosphodiesterase 3b phosphorylation greater placebo-controlled reduction in baseline levels of 8–8.2%. Pratley et al. and increases GLUT4 translocation to the A1C by use of among individ- (abstract 478) added alogliptin 12.5 or 25 cell surface, while in the absence of insu- uals in the highest proinsulin/insulin ter- mg or placebo for 26 weeks in 493 type 2 lin, cAMP levels increase thereby leading tile and in the lowest HOMA-␤ tertile in diabetic patients receiving pioglitazone; to activation of protein kinase A (PKA); in four randomized controlled trials of some of the patients were also receiving mice not expressing PTP-1B, PKA activity 1,691 type 2 diabetic patients. Lower metformin or a sulfonylurea. A1C de- is increased in white and brown fat and in ␤-cell function may be associated with creased 0.7% and 0.8% with 12.5 and 25 muscle but not in liver, but there is adi- greater response to sitagliptin. There was mg alogliptin and 0.2% with placebo, ex- pocyte insulin resistance, suggesting that no differential effect by age, sex, or BMI hibiting a greater reduction with a higher in adipocyte PTP-1B is a positive regula- group. Chapell et al. (abstract 512; disclo- baseline A1C level. Nauck et al. (abstract tor of insulin signaling. IR substrate sure: ZB was a coauthor) compared the 477) administered alogliptin 12.5 or 25 (IRS)-1 levels are reduced 40% in mice glucose-lowering effects of sitagliptin, mg or placebo for 26 weeks in 527 type 2 not expressing PTP-1B, and IRS-1 phos- pioglitazone, and rosiglitazone in a meta- diabetic patients receiving metformin, phorylation is decreased, appearing to in- regression analysis of 23 randomized con- finding 0.6%, 0.6%, and 0.1% reduction volve the mTOR pathway. A number of trolled studies, finding weighted mean in A1C and 19, 17, and 0 mg/dl falls in small-molecule PTP-1B inhibitors are be- reduction in A1C of 0.7, 0.9, and 0.5%, fasting glucose. DeFronzo et al. (abstract ing studied, with evidence of improved respectively. Differences in baseline A1C 446) administered alogliptin 12.5 or 25 glycemia in a variety of obese rodent mod- explained most of the apparent difference mg or placebo for 26 weeks to 329 type 2 els. Interestingly, PTP-1B is overex- between the agents, with strong correla- diabetic patients not receiving pharmaco- pressed in breast cancer, and reduced tion between baseline A1C and change in logic treatment, finding 0.6%, 0.6%, and carcinogenesis has been shown in some A1C across studies. no reduction in A1C and a 10 and 16 animal models with PTP-1B inhibitors. Two new DPP-4 inhibitors are under- mg/dl reduction and an 11 mg/dl increase going clinical testing. Rosenstock et al. in fasting glucose, respectively. Pratley et Further therapeutic approaches (abstract 517) administered the DPP-4 in- al. (abstract 445) added 12.5 or 25 mg Santilli et al. (abstract 395) administered hibitor (2.5, 5 and 10 mg alogliptin or placebo for 26 weeks to 500 100 mg three times daily versus daily) to 401 drug-naïve type 2 diabetic type 2 diabetic patients receiving gly- placebo for 20 weeks to 48 type 2 diabetic patients for 24 weeks. The researchers buride, finding 0.4%, 0.5%, and no re- patients with A1C Ͻ7%, finding fasting found a placebo-adjusted reduction in duction in A1C with 5 and 8 mg/dl and 2-h post-test meal glucose to decrease fasting glucose of 21, 15, and 23 mg/dl decreases and a 2 mg/dl increase in fasting from 126 to 117 mg/dl and from 171 to and in A1C of 0.6, 0.6, and 0.7%, respec- glucose, respectively. Rosenstock et al. 139 mg/dl, respectively, with a fall in A1C tively. Adverse events occurring in at least (abstract 444) added 12.5 or 25 mg alo- from 6.7 to 6.3%. Urinary 11-dehydro- 5% of patients included respiratory infec- gliptin or placebo for 26 weeks to 390 thromboxane-B2 and 8-iso-prostaglan- tion, headache, nasopharyngitis, and si- type 2 diabetic patients receiving insulin, din-F2a, markers of platelet activation nusitis, presumably an overlapping alone or with metformin, with 0.6%, and oxidant stress, decreased 40 and complex of diagnoses, and urinary infec- 0.7%, and 0.1% reductions in A1C and a 30%, respectively, and correlated with tion. Karim et al. (abstract 538) adminis- 2 mg/dl increase, a 12 mg/dl decrease, the reduction in postprandial rather than tered the DPP-4 inhibitor alogliptin (50 and a 6 mg/dl increase in fasting glucose, in fasting glucose, potential cardioprotec- mg) to six individuals with a creatinine respectively. tive effects. Hawkins et al. (abstract 344) clearance of 51–80 ml/min, six individu- administered INCB013739, a selective in- als with a creatinine clearance of 30–50 Protein tyrosine phosphatase hibitor of 11␤-hydroxysteroid dehydro- ml/min, six individuals with a creatinine inhibitors genase type 1, to 30 type 2 diabetic clearance of Ͻ30 ml/min (but not on di- Brian Kennedy (Merck Frosst Centre for individuals for 28 days, finding a reduc- alysis), and six individuals with end-stage Therapeutic Research, Pointe-Claire- tion in hepatic glucose production during renal insufficiency. The researchers found Dorval, PQ, Canada) discussed protein hyperinsulinemic, euglycemic, pancre- a 1.7-, 2.1-, 3.2-, and 3.8-fold increased tyrosine phosphatase (PTP)-1B, insulin atic clamp studies, with improved periph- plasma exposure over 5 days when com- sensitivity, and weight control. Insulin re- eral glucose uptake. Fasting glucose pared with six healthy individuals with ceptor (IR) signal transduction involves decreased 18 mg/dl and LDL cholesterol normal renal function. The authors sug- its autophosphorylation. PTP dephos- 20 mg/dl. Plasma ACTH increased 12 pg/ gested that the dose should be reduced to phorylates the IR, returning it to the inac- ml, although morning plasma cortisol lev-

DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009 e29 Perspectives on the News els were unchanged. Huyen et al. nary revascularization, and hospitalization toxicity, with one patient developing liver (abstract 508) studied Gynostemma pen- for angina or congestive heart failure (P ϭ failure. taphyllum, also called jiaogulan or south- 0.036) and a 55% reduction in the com- ern ginseng tea, a traditional Vietnamese bination of myocardial infarction, stroke, herbal treatment. The researchers admin- or death, with benefit seen in subgroups Acknowledgments— Z.T.B. has served on istered 6 g Gynostemma pentaphyllum stratified by A1C (Յ7 vs. Ͼ7%), age, sex, speaker’s bureaus of Merck, NovoNordisk, twice daily to 24 type 2 diabetic patients, or race. Chisholm et al. (abstract 528) Lilly, , Daichi Sankyo, and Glaxo- with a placebo-adjusted 43 mg/dl reduc- randomized 727 type 2 diabetic patients SmithKline; has served on advisory panels for tion in fasting glucose (baseline 180 mg/ to the anti-angina agent ranolazine versus Medtronics, Takeda, Merck, AtherGenics, CV dl) and a 1.8% reduction in A1C, with placebo, and found an A1C reduction Therapeutics, Daichi Sankyo, BMS, and Astra Zeneca; holds stock in Abbott, Bard, evidence of improvement in insulin sen- with active treatment which correlated Medtronic, Merck, Millipore, Novartis, and sitivity. Luo et al. (abstract 333) noted with baseline glucose; there was no rela- Roche; and has served as a consultant for No- that mice not expressing thyrotropin- tionship between glucose and change in vartis, Dainippon Sumitomo Pharma America, releasing hormone (TRH) are hyperglyce- A1C in those receiving placebo. Klug et al. Forest Laboratories, and Nastech. No other mic, and thyroxin does not improve this (abstract 443) and Tardif (abstract 335) potential conflicts of interest relevant to this effect. In a streptozotocin-diabetic model, treated 6,144 patients with acute coro- article were reported. TRH administration markedly reduced nary syndrome with succinobucol, with a the degree of hyperglycemia and main- 19% decrease in the prespecified second- tained normal insulin levels. Normal ani- ary end point of cardiovascular death, car- References mals receiving TRH alone had mild diac arrest, myocardial infarction, and 1. Ritzel RA, Butler AE, Rizza RA, Veldhuis hyperinsulinemia without hypoglycemia. stroke. Of the 2,271 type 2 diabetic pa- JD, Butler PC: Relationship between Several treatment approaches may tients, 1,952 had evaluated A1C data, ␤-cell mass and fasting blood glucose combine glycemic with cardiovascular showing a reduction from 7.2% by 0.5%, concentration in humans. Diabetes Care benefit. Scranton et al. (abstract 331) ad- without an increase in weight, waist cir- 29:717–718, 2006 ministered a rapidly absorbed formula- cumference, or edema. Of those not hav- 2. Elchebly M, Payette P, Michaliszyn E, tion of bromocryptine to increase early ing diabetes, 82 of 1,950 who received Cromlish W, Collins S, Loy AL, Nor- morning dopaminergic activity versus pla- placebo versus 30 of 1,923 who received mandin D, Cheng A, Himms-Hagen J, Chan CC, Ramachandran C, Gresser MJ, cebo for 52 weeks to 3,070 type 2 diabetic succinobucol developed diabetes during Tremblay ML, Kennedy BP: Increased in- patients in the Cycloset Safety Trial (Clini- the period of observation. There was a sulin sensitivity and obesity resistance in calTrials.gov Identifier NCT00377676), trend to increased hospitalization for mice lacking the protein tyrosine phos- showing a 42% reduction in the combina- heart failure, a significant increase in atrial phatase-1B gene. Science 283:1544–8, tion of myocardial infarction, stroke, coro- fibrillation, and the occurrence of hepato- 1999

e30 DIABETES CARE, VOLUME 32, NUMBER 3, MARCH 2009