DRUG EVALUATION

Insulin glulisine: preclinical hallmarks and clinical effi cacy

Insulin glulisine (glulisine) is a new fast-acting human insulin analogue approved for treatment of diabetes Type 1 and 2. Preclinical investigations indicate mitogenic and metabolic potency similar to human insulin, but distinct β-cell protective properties, possibly related to a preferential effect on the insulin receptor substrate-2. Clinical trials with regular human insulin as a comparator consistently revealed superiority of glulisine with regard to glycemic control, which brought post-prandial glucose excursions closer to physiological levels without increasing the risk of hypoglycemia. Experimental trials showed specifi c advantages of glulisine in the treatment of obese and pediatric patients, which requires further evalulation. Certainly, glulisine represents a valuable tool for optimizing glycemic control, and thereby for attenuating micro- and macro-vascular complications.

† KEYWORDS: β-cells euglycemic clamp insulin analogues insulin-like growth Freimut Schliess & factor proliferation obesity systems biology zinc Tim Heise †Author for correspondence: Insulin: physiological action & and two transmembrane β-subunits. Insulin Profi l Ins tut für signaling mechanisms binding triggers autophosphorylation of the Stoff wechselforschung GmbH, Metabolism & gene expression IR-β subunit, which allows recruitment and Hellersbergstrasse 9, D-41460 Insulin shifts cellular metabolism towards an tyrosine phosphorylation of insulin receptor Neuss, Germany anabolic state and is crucial for the systemic substrates (IRS), including IRS-1, IRS-2 and Tel.: +49 2131 4018 225; control of carbohydrate metabolism (FIGURE 1) [1] . the Src homology and collagen protein (Shc). Fax: +49 2131 4018 500; Insulin increases cellular glucose and amino IRS serve as interfaces between the insulin freimut.schliess@ acid uptake, glycolysis and glucose oxidation, as receptor and various signaling pathways lead- profi l-research.de well as the synthesis of glycogen, proteins, fatty ing, for instance, to the activation of PI 3-kinase acids and triglycerides. Simultaneously, insulin and protein kinase B (PKB) and mitogen- inhibits catabolic pathways such as gluconeo- activated protein (MAP) kinases such as the genesis and glycogenolysis, as well as lipolysis and extra cellular signal-regulated kinases Erk-1 and autophagic proteolysis. Insulin stimulates DNA -2 [12] . Cellular internalization of the insulin/IR synthesis and cell proliferation and protects cells complex leads to the degradation of insulin ini- from apoptosis – for instance in the setting of tiated by the insulin-degrading enzyme, which liver regeneration [2]. By preventing postprandial may contribute to the termination of insulin excursions of blood glucose, oxidative stress and signaling [13] . On the other hand, intracellular infl ammatory cytokines, insulin provides protec- insulin/IR complexes may even play an active tion from endothelial dysfunction and vascular part in signal transduction [14] . disease [3,4]. Insulin promotes complex changes As revealed by knockout experiments, insulin- in gene-expression patterns [5,6]. Recent investi- sensitive signal transduction pathways show a gations suggest that insulin action in the brain high degree of redundancy. For example, genetic may signifi cantly contribute to the regulation of deletion of IRS-1 led to the identifi cation of peripheral glucose and fat metabolism and, in IRS-1-independent signaling via IRS-2 [15–17] . concert with leptin, to hunger and satiety [7,8]. Furthermore, the insulin-sensitive signaling Metabolic insulin effects are antagonized by pathway is also addressed by other factors such glucagon, catecholamines and glucocorticoids. as insulin-like growth factor (IGF-1), with func- tional consequences different from those of insu- Signal transduction lin. Insulin and IGF-1 signals cross-talk with A complex signaling network activated by insu- each other by low-affi nity binding of insulin to lin binding to its plasma membrane receptor the IGF-1R, and IGF-1 to the IR, respectively, determines the exact insulin effect [9–11] (FIGURE 1). and the engagement of IR/IGF-1R hybrid recep- The insulin receptor (IR) is a receptor tyrosine tors [18,19] . Among others, the specifi c outcome kinase comprising two extracellular α-subunits of insulin signaling may critically depend on the

10.2217/14750708.6.2.209 © 2009 Future Medicine Ltd Therapy (2009) 6(2), 209–229 ISSN 1475-0708 209 DRUG EVALUATION Schliess & Heise

expression profi le of cell surface receptors and prevents postprandial hyperglycemia, while a downstream operating signaling components, low basal insulin secretion accounts for normo- the cross-talk between signaling pathways and glycemia in the post-absorptive state [1,20,21] . the spatio–temporal compartmentalization of Insulin therapy of Type 1 diabetic patients signaling modules within the cellular matrix. is instituted at the time of diagnosis, and basal-bolus subcutaneous injection regimens Insulin therapy in people or continuous subcutaneous insulin infusion with diabetes (CSII) aim at mimicking meal-related and basal In healthy individuals, nutrient intake provokes a insulin secretion, which is widely absent in this biphasic insulin secretion by the pancreatic β-cells, clinical setting. In Type 2 diabetes, which is with maximum blood insulin concentrations characterized by insulin resistance and pro- achieved within 0.5–1 h, and a return to basal gressive loss of β-cell function, insulin is often levels within another 2–3 h. The meal-related administrated on top of an existing treatment endogenous insulin production almost perfectly with oral anti diabetic drugs (OADs) at later

Insulin, IGF-1, insulin analogues

PI3,4(P2) PI4(P) P-Tyr Tyr-P P-Tyr p110 p85 IRS-1/2 Tyr-P P-Tyr Tyr-P SHC

PDK1 SHP-2 Grb2 Ptdins-3-kinase Grb2 PKB PKCλ PKCζ GLUT4 vesicles β

PKB SOS SOS Ras Leu Raf-1 mTOR GSK3 Foxo MEK MAPK MNK1

pp1G p90rsk

4E-BP1 p70S6K eIF-4E MKP

+ Glucose uptake Metabolic effects + Amino acid uptake + Glycogen synthesis + Protein synthesis + Glycolysis - Proteolysis - Gluconeogenesis + Fatty acid synthesis + DNA synthesis + Triglyceride synthesis + Proliferation - Lipid degradation - Apoptosis

Figure 1. Signaling pathways responsive to insulin, insulin analogues and IGF-1. The scheme depicts part of the cellular signaling network responsive to insulin, IGF-1 and insulin analogues. Insulin and IGF-1 bind with high affi nity to their cognate receptors, and thereby induce distinct, cell-type- and environment-specifi c signaling patterns. In addition, there is low-affi nity binding of insulin (analogues) to the IGF-1 receptor (IGF1-R), and of IGF-1 to the insulin receptor (IR), respectively, and the activation of IR/IGF-1R hybrids induced by ligand binding, autophosphorylation of receptor tyrosine residues generates docking sites for insulin receptor substrates (such as IRS-1, IRS-2 and Shc), which feed signals in different areas of the network. Although signaling pathways can be identifi ed (e.g., the Ras–Raf–MEK–MAPK pathway, the PI 3-kinase–PKB pathway or the intracellular amino acid-dependent signaling via the mammalian target of rapamycin [mTOR]) it should be noted that the activation of specifi c patterns rather than linear pathways mediates the entire metabolic response to insulin, insulin analogues and IGF-1. Further information is provided in the text and in [9–14]. Adapted from [10].

210 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

stages of the disease, with a stepwise intensi- ability of hexameric insulin to dissociate into fi cation of insulin therapy over the following dimers and monomers that can readily permeate years [22,23]. the capillary wall [20,25,27]. At a concentration of approximately 10–3 mol/l in commercially avail- The design of fast-acting able human insulin preparations, approximately insulin analogues 75% of the insulin molecules assemble to the hex- Subcutaneous insulin injections poorly mimic amer formation via trimerization of dimers [26]. the physiological pattern of endogenous meal- Dilution of insulin solutions promotes disso- related and basal insulin secretion. Endogenous, ciation of the hexameric insulin, and the initial but not subcutaneously administrated insulin lag phase of insulin absorption refl ects the time fi rst passes the liver, which is pivotal for estab- required for suffi cient dilution of subcutaneously lishment of the portal–peripheral insulin gra- deposited insulin hexamers around Zn2+, which dient and adequate postprandial glucose con- makes dimeric and monomeric insulin available trol [24]. Furthermore, absorption of regular for absorption. human insulin (RHI) from a subcutaneous Fast-acting human insulin analogues were depot into the circulation is delayed by a lag designed with the objective to provide insu- phase of approximately 3–4 h, which accounts lin-related drugs that are absorbed faster than for a mismatch between the postprandial rise RHI, but at the same time ideally match the of blood glucose concentration and the avail- properties of human insulin regarding recep- ability of insulin. This increases the incidence tor binding, post-receptor signaling, mitogenic of both post-prandial hyperglycemia and late potency and glucodynamic effi cacy. Based on post-absorptive hypoglycemia [20,25]. As Robert x-ray crystallographic investigations and studies Tattersall noted years ago, insulin is given in with mutagenated human insulin, amino acid the wrong place, at the wrong time, and in the residues were identifi ed that critically contribute wrong amounts (not enough for the meal and to multi merization, Zn2+ coordination and bind- too much for the fast). ing of insulin to the IR [20,28–32]. This allowed Insulin is a heterodimetic polypeptide com- modifi cation of the primary insulin structure to prising an A chain and a B chain, covalently decrease self-assembly to hexamers, and thereby connected by a disulfi de bond between CysA20 enhance the rate of absorption and consequently B19 (FIGURE 2) [26] and Cys . In solution there exists improve postprandial glucose disposal after sub- a dynamic equilibrium between monomers, cutaneous injection. For instance, by replacement B10 dimers, tetramers, hexamers and possibly higher of His by Asp (FIGURE 2), a human insulin ana- molecular weight products, which critically logue was created that cannot hexamerize, even depends on insulin concentration, pH and ionic in the presence of Zn2+ [31,33,34]. Accordingly, strength [26]. The presence of Zn2+ promotes euglycemic clamp experiments with the AspB10 the formation of hexamers, which stabilizes the analogue in healthy volunteers revealed a signifi - insulin molecule but delays its absorption fol- cant faster onset of gluco dynamic action as com- lowing subcutaneous injection [26]. Experimental pared with insulin [35,36]. However, in the case and theoretical studies established a correlation of AspB10, a 12-month exposure of rats to high between the rate of insulin absorption and the concentrations of the AspB10 analogue increased

A chain (RHI, AspB10, lispro,

aspart, glulisine) S S

Gly IleVal GluGln Cys Cys Thr Ser Ile Cys SerLeu TyrGln Leu Glu Asn Tyr Cys Asn

S S B chain

S S 3 10 28 29 RHI Phe ValAsn GlnHis Leu Cys Gly Ser His Leu ValGlu AlaLeu Tyr Leu Val Cys Gly Glu ArgGly Phe Phe Tyr Thr Pro Lys Thr

AspB10 Phe Val Asn GlnHis Leu Cys Gly Ser Asp Leu ValGlu AlaLeu Tyr Leu Val Cys Gly Glu ArgGly Phe Phe Tyr Thr Pro Lys Thr

Lispro Phe ValAsn GlnHis Leu Cys Gly Ser His Leu ValGlu AlaLeu Tyr Leu Val Cys Gly Glu ArgGly Phe Phe Tyr Thr Lys Pro Thr

Aspart Phe ValAsn GlnHis Leu Cys Gly Ser His Leu ValGlu AlaLeu Tyr Leu Val Cys Gly Glu ArgGly Phe Phe Tyr Thr Pro Asp Thr Glulisine Phe ValLys GlnHis Leu Cys Gly Ser His Leu ValGlu AlaLeu Tyr Leu Val Cys Gly Glu ArgGly Phe Phe Tyr Thr Pro Glu Thr

Figure 2. Primary structure of regular human insulin and fast-acting insulin analogues.

future science group www.futuremedicine.com 211 DRUG EVALUATION Schliess & Heise

the incidence of mammary gland tumors [37], and was fi rst approved in 2004 by the FDA [201] which prevented clinical development of this and EMEA [202] for the treatment of diabetes ‘superactive insulin’ [33]. mellitus Type 1 and 2 in adults, and in 2008 by The fast-acting insulin analogues currently the EMEA for the treatment of diabetic children marketed are summarized in FIGURE 2. Insulin aged 6 years or older [203]. lispro (lispro) and insulin aspart (aspart) were fi rst approved for the treatment of diabetes mel- Structural considerations & litus in adults and children by the US FDA in formulation of glulisine 1996 and 2000, and by the European Medicines Glulisine was derived from human insulin by Agency (EMEA) in 1996 and 1999, respectively. replacement of AsnB3 by Lys and LysB29 by Glu Both analogues are also available as fi xed mix- (FIGURE 2), which at physiological pH adds a posi- tures of fast- and intermediate-acting (protami- tive charge in position B3 and replaces a positive nated) insulin formulations [38–40]. by a negative charge, at position B29, leading to In lispro, the amino acid sequence ProB28– a formal addition of a negative charge and con- LysB29 of human insulin has been inverted [41], sequently to a slight decrease of the isoelectric whereas in aspart, ProB28 of human insulin was point compared with human insulin [32]. Data on exchanged with aspartic acid [31] (FIGURE 2). These the structural implications of these amino acid amino acid replacements are localized within the replacements were reported in a recent review C-terminal part of the B chain that is critical for [32]. Accordingly, the introduction of Lys at posi- dimerization of insulin molecules, and possibly tion B3 may impair the trimerization of dimers, insulin binding to the IGF-1R, but not involved whereas Glu at position B20 may decrease dimer in the binding to the insulin receptor [28–30]. formation and provide stability to the mono- The replacement of ProB28 strongly impairs self- meric glulisine. In contrast to lispro and aspart, assembly of lispro and aspart molecules, even ProB28 remains preserved in glulisine, which then in the presence of Zn2+, and in case of lispro again may support dimerization [52,53]. self-assembly was further compromized by the A specifi c characteristic of the glulisine formu- additional replacement of LysB29 by Pro [41]. lation is the lack of Zn2+ which is not required for Nonetheless, the marketed formulations of lispro glulisine stabilization [32]. Accordingly, glulisine and aspart contain Zn2+ [32]. This appears to be in the marketed formulation exists in the mono- 2+ a paradox; however, it was shown that Zn in meric and the dimeric form only (FIGURE 3) [32]. concert with the obligatory antimicrobial pheno- It was demonstrated that the addition of Zn2+ lic excipients promotes hexamerization of lispro promotes hexamerization of glulisine, lead- and aspart, and thereby protects the analogues ing to pharmacokinetics and glucodynamic from denaturation, which increases shelf half- time–action profi les that resemble those of RHI life [26,34,41,42]. On the other hand, x-ray crys- [54]. Thus, in the case of glulisine the omission tallographic analysis suggested that, compared of Zn2+ is mandatory to fulfi ll the characteristics with insulin, there is a loss of dimer-stabilizing of a fast-acting insulin analogue. interactions in these hexamers, which may allow In order to achieve suffi cient stability without a more rapid dissociation of the hexamers. This Zn2+, polysorbate 20 is added to the glulisine for- explains the faster absorption and onset of glu- mulation [32], acting as a surfactant that occupies codynamic action of lispro and aspart after sub- hydrophobic interfaces and thereby may confer cutaneous injection (FIGURE 3) [34,41–44]. additional protection of monomeric glulisine In addition to the fast-acting insulin ana- from denaturation and fi brillation [53]. logues, long-acting insulin analogues were designed, which allow a more accurate control of Preclinical investigation: glulisine blood glucose concentration in the post-absorp- effects on signaling, proliferation, tive (inter-prandial) time period than neutral metabolism & β-cell viability protamine hagedorn (NPH) insulin. Excellent The effects of insulin analogues on signal trans- recent reviews have summarized the hallmarks duction, metabolism, cell viability and prolifera- of the currently marketed long-acting insulin tion were examined in vitro, cultured cells (that analogues insulin glargine and insulin detemir express different levels of IR and IGF-1 receptors [22,45–51]. and/or represent relevant insulin target tissues) and in animals in vivo, and usually compared Insulin glulisine: regulatory affairs with the respective effects of RHI, IGF-1 and Insulin glulisine (glulisine) is the most recently the AspB10 analogue. Such investigations address marketed fast-acting insulin analogue (FIGURE 2) safety aspects already at an early stage of drug

212 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

RHI

Lispro, aspart

Glulisine

Figure 3. Self-association properties and absorption of regular human insulin and fast-acting insulin analogues. In aqueous solution insulin and its analogues exist in a dynamic equilibrium between monomers, dimers and hexamers. The protracted absorption of RHI from the subcutaneous depot is determined by the dilution-dependent, and therefore time-consuming, dissociation of hexamers, which are stabilized by Zn2+ (central sphere). Lispro and aspart in the presence of Zn2+ and phenolic excipients form less stable hexamers, which shifts the equilibrium towards the formation of dimers and monomers and thereby accelerates absorption. Based on its structure and marketed formulation glulisine can be stored in the absence of Zn2+, which strongly shifts the equilibrium in favor of dimer and monomer formation. After subcutaneous injection, glulisine becomes immediately absorbed from the depot. RHI: Regular human insulin. Adapted from [51]. development, and use insulin analogues as an concentration achieving half-maximal stimula- exciting tool to examine insulin signaling and tion of glucose uptake or lipogenesis in primary its (patho)-physiological implications. rat adipocytes) and mitogenic potency (i.e., the analogue concentration normalized to the insulin Signaling via IR & IGF-1R: concentration required to stimulate half-maximal metabolic & mitogenic potency of [3H]thymidine incorporation into the DNA of insulin analogues CHO-K1 cells that do not over express the IR) [55]. A panel of insulin analogues different from glu- Remarkably, a disproportionate increase of the lisine was initially used to establish a potential mitogenic potency of analogues was observed at relationship between metabolic and mitogenic Kd values below 40% of the Kd of insulin, which potency on the one hand, and binding to/disso- was refl ected by an exponential increase in the ciation from the IR and the IGF-1R on the other mitogenic potency:metabolic potency ratio [55]. [20,33,42,55,56]. A representative of these studies For instance, the AspB10 analogue displayed a rela- [55] demonstrated with Chinese Hamster Ovary tive Kd of approximately 14% of that of insulin, (CHO) cells that a decreased dissociation from with a metabolic potency of approximately 230% the IR (as estimated in CHO cells overexpress- and a mitogenic potency of approximately 650% ing the human IR) correlates with an increase [55]. The attenuated dissociation of AspB10 from of both relative metabolic potency (i.e., the ana- the IR was associated with a delayed inactiva- logue concentration normalized to the insulin tion of the IRβ kinase and a delayed decline of

future science group www.futuremedicine.com 213 DRUG EVALUATION Schliess & Heise

IRβ and Shc tyrosine phosphorylation follow- was stimulated by AspB10, which was also true for ing withdrawal of the analogue from the culture Shc tyrosine phosphorylation by glulisine, insulin medium, which was suggested to account for the and AspB10 [61]. Dual phosphorylation of the MAP relatively high mitogenicity of the AspB10 ana- kinases Erk-1/Erk-2 by glulisine was weaker com- logue [55]. Studies with other cell types established pared with that induced by insulin and AspB10, a correlation between IGF-1R binding and the while PKB Ser473 and GSK-3α/β Ser21/29 phos- mitogenicity of insulin analogues [56–58]. phorylation by glulisine, insulin and AspB10 was Remarkably, despite the heterogeneity of recep- almost equal in K6 myoblasts [62]. These fi ndings tor binding and dissociation kinetics, the maxi- indicate that the level of receptor occupancy and mal metabolic and proliferative responses to the activation alone may not be suffi cient to defi ne the analogue and the slopes of the respective dose– signaling potency of an insulin analogue. response curves were almost identical, and not dif- Rat-1 fi broblasts overexpressing the human IR ferent from those obtained with RHI in cultured were utilized to study IR-dependent signaling cells [55,56]. Furthermore, equimolar amounts by glulisine. The association kinetics of glulisine of analogues with higher (AspB10) and lower (0.0035 nmol/l) were similar to that of insulin, (AspB9GluB27) affi nity to the insulin receptor and whereas AspB10 displayed a higher affi nity to the RHI showed an almost equal total glucodynamic IR [63]. Consistently, the potency of glulisine and activity in vivo in hyperinsulinemic euglycemic insulin to compete with 125I-labeled AspB10 for IR clamp experiments [59]. This was explained by binding was almost equal and only approximately a higher/lower rate of IR-mediated clearance of 10% of that of unlabeled AspB10 [63]. Recording the high-/low-affi nity analogues, respectively, the release of 125I-labeled insulin/analogue from which may account for equipotent steady state the IR-overexpressing rat-1 fi broblasts revealed concentrations in the circulation [20,59]. similar receptor dissociation kinetics for insulin and glulisine, whereas AspB10 dissociated much IR & IGF-1R signaling by glulisine more slowly from the IR [63]. Accordingly, the Studies with cells overexpressing the IR and/or stimulation of maximal IRβ tyrosine phospho- the IGF-1R in vivo investigated glulisine binding rylation (100%) by either glulisine, insulin or and dissociation kinetics, as well as signal trans- AspB10 (each 1 nmol/l) is followed by a 50% decay duction downstream of these receptors in view in the continued presence of glulisine and insu- of a potentially increased mitogenicity. lin, respectively, whereas a pronounced IRβ tyro- K6 myoblasts derived from rat heart muscle sine phosphorylation persisted in the presence of express high levels of the IGF-1R and only AspB10 [63]. Unfortunately, this presentation does small amounts of the IR [60], and were used to not allow comparison of the potencies of gluli- study IGF-1R-dependent signaling by glulisine. sine and insulin to increase IRβ tyrosine phos- Glulisine and insulin (each 10–8 mol/l) bind- phorylation above basal levels. However, in liver ing to the K6 myoblasts was comparable and and muscle of random-fed mice, in vivo injection approximately half that of AspB10 [61]. Similarly, of glulisine or insulin (2 IU) into the inferior the rate of glulisine and insulin internalization vena cava after 10 min increased IRβ tyrosine was comparable, and approximately 50% that of phosphorylation to comparable levels [64]. AspB10. On the other hand, the extent of glulisine A study with human skeletal muscle cells degradation was similar to that of AspB10, but directly compared the affi nities of glulisine, insu- only approximately half that of insulin [61]. This lin and IGF-1 to both human IR and IGF-1R by indicates that internalization of the analogues measuring the concentrations required for a 50% was not paralleled by their degradation, and sug- displacement of specifi cally bound [125I]insulin gests a prolonged intracellular presence of func- and [125I]IGF-1, respectively [65]. Essentially no tional IGF-1R/glulisine complexes compared differences were observed in IR affi nities of glu- with IGF-1R/insulin. lisine and insulin, and both were equally poor for Glulisine as AspB10 (each 500 nmol/l, 10 min) competing IGF-1 binding to the IGF-1R, indicat- in the K6 myoblasts increased tyrosine phos- ing that affi nity of glulisine to the IGF-1R was low phorylation of the IGF-1R approximately 2.5-fold and comparable with that of RHI [65]. above basal levels, while insulin-induced IGF-1R tyrosine phosphorylation was signifi cantly lower Mitogenicity of glulisine (approximately twofold) [61]. On the other hand, Studies performed in different cell culture sys- glulisine- and insulin-induced recruitment of tems and in animals in vivo suggest a low mito- Shc to the IGF-1R was comparable, and only genic potential of glulisine almost equal to that approximately a quarter of the Shc recruitment of human insulin.

214 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

Glulisine (16 h) and insulin at the concentra- tyrosine phosphorylation of IRS-1, whereas a tion of 500 nmol/l increased the incorporation pronounced tyrosine phosohorylation of both of 5-bromo-2´-desoxyuridine into the DNA of IRS-1 and IRS-2 was found in response to insu- K6 myoblasts to a similar extent, which was lin and AspB10 [61,67]. This was observed in rat strongly surpassed by AspB10 and IGF-1, respec- K6 myoblasts, primary human skeletal muscle tively [61]. Glulisine and insulin (10–100 nmol/l, cells, primary adult rat cardiomyocytes [61] and 16 h) were also equipotent to stimulate [3H]thy- the clonal rat β-cell line INS-1 [61,67]. In spite of midine incorporation into the DNA of mouse- a preferential IRS-2 phosphorylation by glulisine, derived C2C12 myoblasts [64], and glulisine was the metabolic and mitogenic potency of glulisine even less effective than RHI in stimulation of was similar to human insulin in K6 myoblasts [61]. [3H]thymidine incorporation into the DNA of Preferential IRS-2 tyrosine phosphorylation by the nonmalignant human breast cell line MCF10 glulisine was apparently not related to the relative [63]. Furthermore, [3H]thymidine incorporation abundance of IR and IGF-1R. into the DNA in cultured myotubes derived from However, a potential β-cell protective action of nondiabetic and Type 2 diabetic individuals was glulisine was established in INS-1 cells and related comparable with glulisine and insulin (each 1–25 to preferential IRS-2 signaling by this analogue nmol/l, 16 h) and only approximately 10% of [67]. In these cells tyrosine phosphorylation of that observed with IGF-1 in both cell prepara- IRS-1 by glulisine (500 nmol/l) was virtually tions [65]. Similar fi ndings were observed for dual absent, whereas insulin, AspB10 (each 500 nmol/l) phosphorylation of the MAP kinases Erk-1/Erk-2 and IGF-1 (10 nmol/l) produced a pronounced in response to glulisine, insulin and IGF-1 [65]. tyrosine phosphorylation of both IRS-1 and Mammary glands of female rats treated with -2 [67]. Importantly, insulin glulisine (500 nmol/l) human insulin or glulisine (20 and 50 IU/kg was as potent as IGF-1 (10 nmol/l) to prevent body weight twice daily over 12 months) showed caspase-3 activation and DNA fragmentation as no difference in the immunoreactivity of the pro- induced by palmitic acid (250 µmol/l, 24 h) or liferation marker Ki-67 in mammary glands [61]. proinfl ammatory cytokines (4 ng/ml IL-1β plus Unfortunately, no positive control experiment was 10 U/ml IFN-γ, 24 h), whereas RHI and AspB10 included in this study. were less potent [67]. Recording dose–response curves indicated that glulisine at concentrations Glulisine-induced glucose uptake of 1–1000 nmol/l was superior to insulin, lispro Uptake of the nonmetabolizable glucose deriva- and aspart in preventing DNA fragmentation in tive 3-O-methylglucose was used as a surrogate response to palmitic acid and cytokines, respec- marker for metabolic activity. The initial rise of tively [67]. These data suggest that glulisine pro- 3-O-methylglucose uptake by adult rat cardio- tects β-cells from fatty acid- and infl ammatory myocytes exposed to glulisine, insulin or AspB10 cytokine-induced apoptosis. (5 and 500 nmol/l, 10 min) was comparable Studies with a genetic knockout or over- and in line with the almost equal stimulation of expression of IRS-1/IRS-2 provided substantial PKB/GSK-3 phosphorylation by insulin and the evidence for a specifi c role of IRS-2 in β-cell analogues in these cells [61]. Full dose-response function. Different from IRS-1 null mice, IRS-2 curves recorded in fused and differentiated myo- null mice mimic the phenotype of human Type 2 tubes derived from nondiabetic and Type 2 dia- diabetes, with a decreased β-cell mass and a lack betic individuals indicated that the sensitivity of of islet hyperplasia for compensation of insulin 3-O-methylglucose uptake to glulisine and insu- resistance [15–17,68] . A recent investigation of mice lin was essentially equal and slightly less than the with a pancreas-specifi c IRS-2 knockout indi- sensitivity to IGF-1 in these cultures [66]. The cated that pancreatic IRS-2 is not only essential fi ndings were refl ected at the level of PKB Ser473 for β-cell mass, but also for adequate insulin phosphorylation, which was similar with regard secretion in response to glucose [69]. Accordingly, to sensitivity and maximal response to glulisine β-cell-specifi c overexpression of IRS-2 increased and insulin, but higher in response to IGF-1 in β-cell growth, survival and insulin secretion, both cell preparations [66]. and prevented diabetes in IRS-2 knockout mice, obese mice and streptozotocin-treated mice [70,71], Glulisine signaling via IRS-2: while repression of IRS-2 expression by different potential impact on β-cell viability kinds of stress was associated with increased β-cell A remarkable fi nding was that insulin glulisine apoptosis [72]. From this it seems well conceivable produced a pronounced tyrosine phosphorylation that specifi c targeting of IRS-2 by glulisine may of IRS-2, but only a marginal, if at all present, provide β-cell protection.

future science group www.futuremedicine.com 215 DRUG EVALUATION Schliess & Heise

According to another scenario, the lack of frequently used Diabetes Treatment Satisfaction IRS-1 activation may account for the anti- Questionnaire revealed a signifi cant improve- apoptotic effi cacy of glulisine [61,67]. One could ment in favor of the analogues. Overall, the hypothesize that IRS-1 in insulin-stimulated Cochrane analysis suggested only a minor clini- β cells, at least in the face of environmental cal benefi t of fast-acting insulin analogues in the toxins, releases pro-apoptotic signals that are majority of insulin-treated diabetic patients, and absent in β cells exposed to glulisine. Indeed, a recommended a ‘cautious response to the vigor- pro-apoptotic potential of insulin became appar- ous promotion of insulin analogues’ until long- ent in certain experimental settings [73–75], and term effi cacy and safety data would be available. there is evidence that IRS-1 could confer such This analysis was critically commented on in pro-apoptotic signaling. Thus, overexpression a review published in 2007, which also included of IRS-1 was shown to constitutively activate randomized, controlled trials, but more selectively pro-apoptotic pathways in the liver [76]. In a than in the Cochrane review. Only studies with breast cancer mouse model IRS-1-, but not an intervention interval of at least 3 months and IRS-2-defi cient mammary tumor cells, were with a complete change in the insulin regimen to highly invasive and resistant to apoptotic treat- analogues (‘all-analogue’ versus ‘all human insu- ments [77]. Within this framework, glulisine, by lin’ regimens) were included. Data on postpran- leaving IRS-1 signaling at basal levels, would dial and fasting blood glucose concentrations, simply decrease the pro-apoptotic input in β weight changes and within- and between-person cells, and thereby shift the balance in favor of variability of glucodynamic parameters were ana- the IRS-2-mediated survival signals, which may lyzed [48]. According to this analysis, which again support the β cell to cope with toxins. includes one study with glulisine [78], fast-acting analogues in people with Type 1 diabetes gener- Clinical evaluation of glulisine: ally decreased postprandial blood glucose con- pharmacokinetics, glucodynamics, centrations and the incidence of hypoglycemia, safety & tolerability compared with RHI. In people with Type 2 dia- As insulin glulisine is the third fast-acting insulin betes treated with a basal-bolus insulin regimen,

analogue on the market, we will briefl y describe the results regarding HbA1c and hypoglycemia the clinical benefi t of the previously available were heterogeneous, but postprandial plasma analogues that have recently be summarized in glucose concentrations were consistently lower several meta-analyses. We will then put these with the rapid-acting analogues than with RHI data into perspective with those obtained with [48]. Furthermore, premixed fast-acting analogues glulisine in experimental and clinical human were found to be superior over human insulin trials that were reported as full papers in peer- mixes in the control of post-prandial blood reviewed journals registered in the PubMed glucose in Type 2 diabetic people. database (TABLE 1) [204]. Furthermore, where indi- Another recent meta-analysis of randomized, cated, we discuss some information published controlled trials with an intervention interval in abstract form. of at least 4 weeks (including two studies [79,80] with glulisine) concluded that fast-acting insulin

Fast-acting insulin analogues: analogues improved both HbA1c (standardized meta-analysis of clinical trials mean difference: -0.4%, [0.1–0.6%, p = 0.027]) A Cochrane review published in 2006 [49] ana- and postprandial blood glucose in comparison lyzed 49 randomized, controlled trials (includ- with RHI also in people with Type 2 diabetes. ing one with glulisine [78]), with over 8000 par- These improvements in glucose control were

ticipants in total. In terms of % HbA1c, this achieved without signifi cant increase in the risk study calculated a weighted mean difference of severe hypoglycemia [47]. of only -0.1% (95% CI: -0.2 to -0.1) in favor of the fast-acting analogues in Type 1 diabetic Evaluation of glulisine in healthy patients, while no benefi t at all was observed nonobese volunteers in Type 2 diabetes. Although a reliable meta- The metabolic potency of glulisine in com- analysis of hypoglycemic events was compro- parison with that of RHI was investigated in mised by heterogeneous defi nitions of ‘hypo- a single-center, randomized, open-label, two- glycemia’, severe hypoglycemia was stated to way crossover, manual hyperinsulinemic glu- occur less often in the analogue group than in cose clamp study in 16 healthy male individu- the RHI group. Similarly, quality-of-life assess- als using an intravenous infusion at a rate of ment was not standardized; nevertheless, the 0.8 mU·kg-1·min-1 for 2 h [81]. Overall glucose

216 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION [78] [79] [88] [77] Not published (kg) -0.3 (post-meal (post-meal -0.3 GLU vs RHI); (post-meal -0.3 GLU vs pre-meal GLU) 3 IGURE F n.s. n.s. n.s. (end point) estimated from vs RHI); (post-meal -2.57 GLU vs RHI); -0.26 (pre-meal vs post-meal GLU) (pre-mealPrandial: -2.63 GLU vs RHI); -2.21 (post-meal GLU vs RHI); (pre-meal-0.41 GLU vs post-meal GLU) Noct, -5.4% (week 12); Prandial: -1.6 n.s. (pre-meal -2.31 Total: GLU 2 Between-treatment differencesBetween-treatment Ref. paper. paper. IGURE F ) ) estimated * * 4 IGURE F Post test meal: -10.8 (1 h);Post (1 test meal: -10.8 (2 h) -20.9 End point: -10.1 (2 h point:End -10.1 post-breakfast); -6.8 (2 h post-dinner); -5.4 (2 h average daily from -6.6 (2 h post-breakfast); (2 h post-dinner) -10.0 estimated from h, (2 -5.0 post-breakfast); average daily Pre-meal GLU vs RHI, -22.9 -22.9 RHI, vs GLU Pre-meal (2 h post-breakfast); -19.9 (2 h post-dinner) Pre-meal GLU vs post- (2 h meal -13.4 GLU, (2 h post-breakfast); - 11.7 post-dinner) (%) Blood glucose (mg/dl) Hypoglycemia Daily insulin dose (IU) Body weight 1c HbA -0.11 (week 26); -0.11 (end point) -0.11 GLU vs RHI); (pre-meal vs-0.15 post-meal GLU); (post-meal n.s. GLU vs RHI)

1c cant; RHI: Regular human insulin. cacy of glulisine. (%) at baseline at (%) Mean HbA Mean 683, weeks 26 7.59 weeks 26 878, 7.55 weeks 26 n.s.892, 7.54 -0.05 (week 12); n.s. n.s. (2 h -10.1 Week 12: n.s. basal: Total: -1.70 -1.87; n.s. Number of of Number randomized patients, treatment duration 860, 12 weeks860, 12 7.66 (pre-meal -0.13 groups Type 1 diabetes,Type pre-meal GLU vs pre-meal LIS 2 diabetes,Type GLU vs RHI 2 diabetes,Type GLU vs RHI Type 1 diabetes,Type pre-meal GLU vs post-meal GLU RHI vs

et al.

(2005) (2004) (2007) Table 1. Phase 1. IIITable trials evaluating clinical effi Study Treatment Dreyer et al. Dailey et al. Rayman et al. Garg (2005) *Mean of blood glucose excursions after breakfast, lunch and dinner. The table summarizes data from trials published as a full GLU: Insulin glulisine; LIS: Insulin lispro; noct: Nocturnal; n.s.: Not signifi

future science group www.futuremedicine.com 217 DRUG EVALUATION Schliess & Heise

disposal and the onset of glucodynamic action The data indicated that glulisine and lispro were were similar for intravenously infused glulisine absorbed signifi cantly faster than RHI. Mean and RHI, as were the shape of the time–concen- residence time and time to maximum glucose tration profi les [81]. Interpretation of the phar- infusion rate were 50% lower for both lispro and macokinetic results was hampered by diffi cul- glulisine than for RHI. The fast-acting proper- ties with the specifi c assay for glulisine, which ties of glulisine are preserved independent of the produced considerably higher serum concentra- injection site, as demonstrated in a manual glu- tions for glulisine than did the human insulin cose clamp study in 16 healthy male volunteers assay for RHI [81]. Distribution and elimination who received glulisine (0.1 IU/kg subcutaneous) from the systemic circulation of glulisine were into femoral, deltoid or abdominal areas. No sig- comparable to those of RHI [54]. nifi cant differences were observed between the Pharmacokinetics and glucodynamics follow- sites, although a slightly faster absorption and ing subcutaneous injection of glulisine, lispro onset of glucodynamic action was found follow- and RHI (each 0.3 U·kg-1) were compared with a ing abdominal injection [83], as has also been manual hyperinsulinemic, euglycemic clamp in demonstrated for RHI [84]. a randomized, double-blind, three-way crossover The effect of mixing glulisine (0.1 IU·kg-1) single-center study in 16 healthy male volunteers, with NPH (0.2 IU·kg-1) immediately before sub- which was published in an abstract form [82]. cutaneous injection on the time–concentration

160 10 Insulin glulisine Insulin glulisine 140 -1 0.3 U.kg-1 0.3 U.kg -1 8 -1 0.15 U.kg ) 0.15 U.kg

120 -1 ) -1 -1 -1 0.075 U.kg 0.075 U.kg

100 .min 6 -1 80 4 60 Insulin (µU.ml 40 GIR (mg.kg 2 20

0 0

0 2 4 6 810 0 2 4 6 8 10 Time (h) Time (h)

160 10 Human insulin Human insulin 140 -1 0.3 U.kg-1 0.3 U.kg -1 8 -1 0.15 U.kg ) 0.15 U.kg

120 -1 ) -1 -1 -1 0.075 U.kg 0.075 U.kg

100 .min 6 -1 80 4 60 Insulin (µU.ml 40 GIR (mg.kg 2 20

0 0

02468 10 0 2 4 6 8 10 Time (h) Time (h)

Figure 4. Time–concentration and time–action profi les of regular humain insulin and glulisine in individuals with Type 1 diabetes. Time–concentration profi les (A & C) and glucose infusion rates (B & D) after subcutaneous administration of glulisine (A & B) and insulin (C & D) were determined in the course of Biostator-based hyperinsulinemic euglycemic clamp experiments. GIR: Glucose infusion rate. Adapted from [87].

218 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

profi le of glulisine was addressed in a single- dose, randomized, open-label, two-way, cross- 7.7 Glulisine over manual hyperinsulinemic euglycemic clamp Lispro study in 32 healthy male volunteers published in abstract form [85]. In this study mixing caused a decrease of maximum glulisine concentration 7.6

by approximately 27% compared with that after (%) separate injections of glulisine and NPH, while a 1c similar total exposure and only a slight decrease of the time period needed to achieve maxi- 7.5 mum glulisine concentration in the blood was Mean HbA reported [85]. Nevertheless, this study indicates that glulisine should not be mixed with NPH insulin in one syringe, as its fast-acting properties will be blunted. In contrast to the other available 7.4 fast-acting analogues, lispro and aspart, glulisine is not yet available in a premixed formulation Baseline 12 weeks 26 weeks with a free and protamine-retarded component. Timepoint Studies in people with Figure 5. Mean percentage of HbA1c values at baseline and after 12 and Type 1 diabetes 26 weeks of treatment of people with Type 1 diabetes with lispro and End-organ metabolic effects of glulisine were glulisine, respectively. Type 1 diabetic individuals received insulin glargine as compared with those of lispro and RHI in a basal analogue and subcutaneous injections of either glulisine or lispro 0–15 min more detailed prospective single-center, dou- before meals. Both treatment groups had similar decreases in mean percentage of HbA over the study [88]. ble-blind, three-period, cross-over manual 1c hyper insulinemic euglycemic clamp trial with 18 Type 1 diabetic patients [86]. It was demon- Accordingly, glulisine-induced glucose disposal strated that administration of glulisine, lispro in the fi rst 2 h after injection was approximately and RHI by a continuous intravenous infusion twice (p < 0.05) as great as that for RHI, whereas with a stepwise dose increase (0.33, 0.66 and overall glucose disposal was almost equal for glu- 1.0 mU·kg-1·min-1) had comparable effects on lisine and RHI. End-of-action phenomena, such endogenous glucose production and glucose as an increase in blood glucose concentrations uptake. In addition, effects on the blood con- at the end of the glucose clamp experiments, centration of free fatty acids, glycerol and lactate were observed earlier with glulisine, confi rming were comparable, indicating that the potency a shorter glucodynamic action of the analogue. of glulisine is similar to that of lispro and RHI Two large Phase III trials in Type 1 diabetic with regard to both blood-glucose-lowering and patients investigated glulisine as part of a basal- lipolytic effects. bolus regimen in combination with a once-daily As in healthy volunteers, glulisine in Type 1 administration of the basal insulin analogue diabetic patients behaved as a fast-acting insu- glargine (TABLE 1 [78,88]). One of these studies, per- lin analogue with pharmacokinetic and gluco- formed in 683 Type 1 diabetic patients (mean dynamic time–action profi les comparable with baseline HbA1c: 7.59%) compared the effects those of lispro [54]. The dose-proportionality of of glulisine and lispro on the basis of percent- glucose disposal in response to subcutaneous age HbA1c determinations, self-monitored daily injected glulisine and RHI (each 0.075, 0.15 and seven-point blood glucose profi les and registra- 0.3 U/kg) was investigated in a single-center, tion of hypoglycemic episodes [88]. Both insu- randomized, Biostator-supported hyperinsulin- lins resulted in a similar decrease in HbA1c emic euglycemic clamp study in 18 male Type 1 from baseline (-0.14 and -0.13%, respectively, diabetic patients [87]. As shown in FIGURE 4, total FIGURE 5) without signifi cant between-treatment serum exposure, as well as maximum serum differences, neither in the proportion of patients concentrations, were dose-proportional for glu- reaching target HbA1c values nor in self-moni- lisine and RHI. Furthermore, glulisine expo- tored glucose concentrations, or in the rate and sure in the fi rst 2 h was twice (p <0.05) that frequency of symptomatic, severe and nocturnal of RHI, and maximum serum concentrations hypoglycemia. Slight differences were observed were reached approximately twice (p <0.05) concerning the dose of insulin glargine, which as fast with glulisine as with RHI at all doses. was higher in the lispro arm, leading to a very

future science group www.futuremedicine.com 219 DRUG EVALUATION Schliess & Heise

in 20 Type 1 diabetic patients who received a Glulisine post-meal 0.02 standardized 15-min meal covered by either glu- vs RHI p = 0.6698 lisine (0.15 U·kg-1 per injection immediately pre- meal or 15 min post meal timed from the start -0.13 -1 Glulisine pre-meal p = 0.0234 of the meal) or 0.15 U·kg RHI injected 30 min vs RHI or immediately before meal ingestion [89]. No 0.15 differences were observed in the blood glu- p = 0.0062 Glulisine post-meal cose time–concentration profi les between the vs glulisine pre-meal ‘15-min post-meal glulisine’, the ‘immediately pre-meal RHI’ and the ‘30-min pre-meal RHI’ arms. As expected from the glucodynamic prop- erties, with a faster onset of action for glulisine, -0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0.3 0.4 maximum blood glucose concentrations were Δ (HbA ) (%) Mean (98, 33% CI) 1c lower and reached earlier with the pre-meal injection of glulisine versus the ‘immediately Figure 6. Between-treatment differences of percentage HbA1c in Type 1 diabetic individuals treated with glulisine or regular human insulin. pre-meal RHI’ group, resulting in lower blood Individuals with Type 1 diabetes received once-daily insulin glargine as a basal glucose excursions in the fi rst 2 h. Remarkably, analogue and subcutaneous injections of either glulisine (pre-meal or post-meal) or the ‘30-min pre-meal RHI’ administration, in RHI (pre-meal). After 12 weeks, the decrease in percentage HbA was greater in 1c contrast to the ‘immediate pre-meal glulisine’ the pre-prandial glulisine group compared with the RHI group and the treatment, led to an early decrease in blood glu- post-prandial glulisine group, respectively. RHI: Regular human insulin. Data taken from [78]. cose concentrations (FIGURE 7), indicating a risk of pre-meal hypoglycemia associated with this small increase in the total daily insulin dose treatment regimen. compared with glulisine (glulisine: -0.86 IU vs Two additional studies, one clinical and one lispro: +1.01 IU, p = 0.0123). in vitro, compared glulisine and aspart in patients The other Phase III study comparing glu- with CSII. A 12-week, European multi center lisine with RHI over 12 weeks in 860 people study in 59 Type 1 diabetic patients (mean base-

with Type 1 diabetes (mean HbA1c at baseline: line HbA1c: 6.9%) demonstrated a small (nonsig- 7.66%) addressed the potential of glulisine with nifi cant) trend towards fewer catheter occlusions

a more fl exible timing of injection without los- with glulisine, but HbA1c values, daily insulin ing adequate glycemic control [78]. Glulisine was doses, blood glucose profi les and adverse event administered subcutaneously either 0–15 min rates were similar [90]. The in vitro simulated CSII before or immediately after a meal, whereas RHI study proposed a decreased resistance to fi bril- was given 30–45 min before a meal, as recom- lation and a higher rate of soluble high-molec-

mended. ‘Pre-meal glulisine’ reduced HbA1c ular-weight protein formation with glulisine by a signifi cantly greater extent (-0.26%) than compared with aspart [91]. In view of the clinical RHI (-0.13%, p = 0.02) and the ‘postmeal glu- data [90], the relevance of these observations is lisine’ (-0.11%, p = 0.006) group, mainly due to currently unclear. signifi cantly lower postprandial blood glucose levels 2 h post-breakfast and 2 h post-dinner Studies in individuals with (FIGURE 6). The incidence of hypo glycemia was Type 2 diabetes similar in all treatment arms, as were the num- Two large Phase III trials investigated the gly- ber of adverse events. Glycemic control was not cemic control achieved with glulisine or RHI, signifi cantly different between ‘postmeal gluli- both in combination with NPH insulin in sine’ and RHI, but ‘postmeal glulisine’ led to a Type 2 diabetic patients for up to 26 weeks small mean decline in body weight (0.3 kg), in (TABLE 1) [79,80]. contrast to the other treatments that resulted in The fi rst study in 878 relatively well-con-

a gain of 0.3 kg (p = 0.03). Insulin doses, both trolled patients (mean HbA1c: 7.55%) on oral prandial and total, slightly increased in the RHI antidiabetic agents (that were continued in this group in contrast to a small decrease observed in trial) showed improved blood glucose levels at the two glulisine groups (RHI dose: +1.75 IU, all time points with glulisine, reaching statis- glulisine dose: -0.88 [p = 0.0001] and -0.47 IU tical signifi cance 2 h post-breakfast and 2 h [p = 0.0012], respectively). post-dinner, resulting in a small, but statisti-

The potential of postprandial application of cally signifi cant advantage with regard to HbA1c insulin glulisine was confi rmed in a single-dose, improvement (-0.46 vs -0.30%, p = 0.0029) [79]. randomized, four-way, complete cross-over study This seems remarkable, in particular as 78% of

220 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

the patients mixed glulisine with NPH insulin, Glulisine immediately pre-meal which should blunt the advantageous pharmaco- dynamic properties of glulisine by partial prot- RHI 30 min pre-meal aminization of the analogue [51]. The improved 210 glycemic control by glulisine was neither accom- 180 panied by an increased incidence of hypoglyce- mia (overall, nocturnal and severe), nor due to 150 any increase in insulin or OAD doses. No safety concerns were observed for glulisine; adverse (mg/dl) 120 events, laboratory and other safety end points 90 did not show any difference to RHI. Likewise, no statistically signifi cant baseline to end point Blood glucose concentration 60 changes in cross-reactive anti-insulin antibodies -1 0 1 2 3 4 5 6 were detectable. M Time (min) The design of the more recently published Phase III trial in Type 2 diabetic patients was similar to that of the fi rst one, but included blood Glulisine immediately pre-meal glucose estimations related to ingestion of a stan- dardized test meal at week 26 [80]. This study in RHI 30 min pre-meal 100 892 patients (mean HbA1c: 7.54% at baseline) did not show any HbA1c differences between 80 glulisine and RHI, so that noninferiority of glulisine was achieved. In addition, the number 60

of patients with at least one episode of nocturnal (µU/ml) 40 hypoglycemia from month four to the end of treatment was less frequent with glulisine, but it 20 Insulin concentration is unclear if this was a predefi ned end point or a post-hoc analysis. Blood glucose excursions were 0 signifi cantly lower both in the self-monitored -10123456- seven-point blood glucose profi les after breakfast M Time (h) and after dinner, as well as after the test meal. No between-treatment differences with regard to frequency and type of adverse events were Figure 7. Blood glucose concentrations and serum insulin concentrations after administration of glulisine injected immediately pre-meal compared observed in this study. with regular human insulin injected 30 min pre-meal in Type 1 diabetic Two Phase III studies recently published in individuals. Blood glucose (A) and insulin (B) time–concentration profi les. abstract form addressed intensifi cation of insulin Time axis – 0 h: start of the meal; M: End of the meal. therapy of Type 2 diabetic patients by addition Data taken from [89]. of glulisine as ‘bolus insulin’ to existing regi- mens [92,93]. In a randomized, open, multicenter, basal-bolus regimen (meal-time glulisine plus 26-week study with 318 poorly controlled Type 2 once-daily glargine) with a twice-daily injection diabetic patients, glulisine was sub cutaneously of pre-mixed insulin (NPH 70%, 30% RHI or injected either at breakfast or at the meal with aspart) in 310 Type 2 diabetic patients who were greatest glycemic impact (breakfast, lunch or din- poorly controlled with their previous pre-mixed ner) on top of the already existing ‘OAD plus insulins [93]. Compared with the treatment with glargine treatment’ [92]. Glulisine in both arms pre-mixed insulins, the glulisine/glargine regimen produced a signifi cant and almost equal HbA1c provided a signifi cantly improved glycemic con- decrease from baseline to end point, and blood trol in terms of HbA1c, as well as mean daytime glucose concentrations within each arm were and post-prandial blood glucose concentrations reported to decrease for most pre- and post-meal without increasing hypoglycemia. time points. Whereas the glargine dose remained stable, the glulisine dose increased to a similar Impact of obesity on the extent in both arms. Overall, the study dem- pharmacokinetics & the onstrated that adding glulisine at breakfast is glucodynamic effi cacy of glulisine equivalent to adding glulisine at the main meal The time–concentration profile and gluco- with no signifi cant difference in the hypo glycemic dynamic action of subcutaneously injected RHI is event rate. Another 52-week study compared a attenuated and delayed in obese individuals [94–96],

future science group www.futuremedicine.com 221 DRUG EVALUATION Schliess & Heise

200 6 ) -1 )

150 -1 5 114114 minmin 6868 min min .min +7+7 minmin -1 4 pp=0.039 = 0.039 100 +13+13 min min 3 p=0.048p = 0.048

GIR (mg.kg 2 50 +36+36 min min +47+47 min min p<0.05p < 0.05 p<0.05p<0.05 1 Insulin concentration (µU.ml

0 0 -1- 0 1 2 3 -1 0123

Time (h) Time (h)

Figure 8. Pharmacokinetic and glycodynamic profi le of glulisine, lispro and regular human insulin in obese nondiabetic individuals. Glulisine (spaced dotted curve), lispro (close dotted curve) and RHI (solid curve), each 0.3 U·kg-1, was injected subcutaneously in the abdominal area. Time–concentration profi les (A) and glucose infusion rates (B) were captured in the course of manual hyperinsulinemic euglycemic clamp experiments. Filled areas under the curves indicate the time to 20% exposure to RHI or the analogues, respectively, and the time to 20% of overall glucose disposal in course of the clamp experiments. The data indicate a less intense onset of glucodynamic activity for lispro compared with glulisine. GIR: Glucose infusion rate; RHI: Regular human insulin. For further information see [66].

while in case of fast-acting insulin analogues, rate were comparable between the analogues, the the infl uence of obesity may be less pronounced onset of glucose disposal as indicated by the time [97–99]. Recent studies compared the pharmaco- to 10% of overall glucose disposal was slightly, kinetics and glucodynamic time–action profi le but signifi cantly, earlier for glulisine. Remarkably, of glulisine, lispro and RHI in nondiabetic obese the faster onset of glucodynamic action of gluli- individuals [66,100]. In a single dose, randomized sine was consistently observed with both doses double-blind cross-over manual hyperinsulinemic and in all BMI classes, and was confi rmed by the euglycemic clamp study, 18 individuals received pharmacokinetic results that showed a signifi - subcutaneous injections of either glulisine or RHI cantly shorter time to reach 10% of total exposure (each 0.3 U·kg-1) in predetermined sequences [66]. and a signifi cantly higher exposure during the fi rst The study confi rmed an accelerated onset of glu- hour post-dosing for glulisine across BMI ranges. cose disposal of glulisine and lispro versus RHI, Obesity-related differences of pharmaco- but, interestingly, glulisine showed an even shorter kinetics and glucodynamics between glulisine time to achieve 20% of total glucose disposal, and lispro were not confi rmed in an exploratory indicating a slightly more rapid onset of glucose Biostator-supported hyperinsulinemic euglyce- disposal than lispro (FIGURE 8). This was confi rmed mic clamp study in people with Type 2 diabetes in the pharmacokinetic results, which showed a published in abstract form [101], which may be, at faster absorption of glulisine than lispro (FIGURE 8). least partly, due to the incomplete block design These fi ndings were essentially corroborated in used in this study, which leads to high variability a randomized, single-center, double-blind cross- in particular in a heterogeneous population such over study in 80 nondiabetic individuals with a as Type 2 diabetic individuals. A more recent wide range of body mass index (BMI) (from lean randomized, open-label, two-arm, cross-over [<25 kg/m²] to very obese [>35 kg/m²]). The test meal study in 15 individuals with Type 2 pharmacokinetics and glucodynamic time–action diabetes compared the plasma glucose and time– profi les of lispro and glulisine were compared concentration profi les following subcutaneous using two doses (0.2 and 0.4 U·kg-1) in automated injection of either lispro or glulisine (0.15 U·kg-1 hyperinsulinemic euglycemic clamp experiments each) immediately prior to a 500 kcal standard [100] (FIGURE 9). While total glucose disposal and test meal (58% carbohydrate, 20% proteins and the time to reach the maximal glucose disposal 22% fat) which was served as breakfast, lunch

222 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

and dinner, respectively (TABLE 1) [102] . The plasma ratio and skin thickness over the three meals was glucose profi les were virtually identical for lispro mentioned, but data not shown. Both glulisine and glulisine, even though an additional (post and lispro were considered safe and well-tolerated hoc?) analysis revealed approximately 12% lower by the patients included in this study [102] . baseline-corrected maximum glucose excur- sions for glulisine. It was also reported that the Pediatric studies mean blood glulisine concentrations measured The pharmacokinetics and glucodynamic action post-meal were signifi cantly higher than that of of glulisine in children and adolescents with RHI. However, an apparently higher exposure Type 1 diabetes was investigated in a randomized, to glulisine as compared with lispro was also single-dose (0.15 IU·kg-1), double-blind, cross- observed in other trials, and considered to be over study with ten children (aged 5–11 years) an artifact due to different cross-reactivities of and ten adolescents (aged 12–17 years) with a the analogue-specifi c antibodies utilized in the subcutaneous administration of glulisine or RHI radioimmune assays [66,81,100]. Post hoc analysis 2 min before intake of a standardized test meal revealed a signifi cantly faster absorption rate for (TABLE 1) [103] . Maximal blood concentrations and glulisine versus lispro, and some evidence for a initial exposure (during 1 and 2 h after start of linear relationship between the glulisine:lispro the meal) were higher, and the mean residence

0.2 U/kg glulisine 10 10 0.4 U/kg glulisine Lean Lean Overweight Overweight 8 Obese 8 ) ) Obese -1 -1 Severely obese Severely obese

6 .min .min

-1 6 -1

4 4 GIR (mg.kg GIR (mg.kg 2 2

0 0 0 120 240 360 480 600 0 120 240 360 480 600 Time (min) Time (min)

0.4 U/kg lispro 10 0.2 U/kg lispro 10 Lean Lean Overweight Overweight

) 8 ) 8 Obese

-1 Obese -1 Severely obese Severely obese .min .min

-1 6 -1 6

4 4 GIR (mg.kg GIR (mg.kg 2 2

0 0 0 120 240 360 480 600 0 120 240 360 480 600 Time (min) Time (min)

Figure 9. Glucodynamic action of glulisine and lispro in lean, overweight and obese nondiabetic individuals. The indicated doses of glulisine and lispro were injected subcutaneously in lean (BMI <25 kg·m-2), overweight (≥25 kg·m-2, <30 kg·m-2), obese (≥30 kg·m-2, <35 kg·m-2) and severely obese (>35 kg·m-2) nondiabetic individuals. GIR were registered in the course of automated hyperinsulinemic euglycemic clamp experiments. BMI: Body mass index; GIR: Glucose infusion rates. For further information see [100].

future science group www.futuremedicine.com 223 DRUG EVALUATION Schliess & Heise

time was shorter for glulisine, whereas total expo- glulisine, in contrast to lispro, aspart and RHI, sure (during 6 h) was comparable for glulisine increased INS-1 cell protection from apoptotic and RHI. In the case of glulisine, children and toxins [67], which may be related to the shift adolescents presented almost superimposable from IRS-1- to IRS-2-dependent signaling. time–concentration profi les, whereas a higher Unfortunately, IRS-1/IRS-2 signaling by lispro (164% [96% CI: 114–236]) exposure was found and aspart was not investigated in this study and it for RHI in adolescents versus children. Similarly, seems conceivable that IRS-1/IRS-2-independent the maximum blood RHI concentration was mechanisms account for antiapoptotic signaling higher in adolescents for unknown reasons. In by glulisine as well. Further studies are required to line with the pharmacokinetic data, postprandial dissect IRS-1/IRS-2-dependent and -independent glucose excursions were lower with glulisine than contributions, and to prove that glulisine indeed with RHI, and blood glucose tended to increase offers β-cell protection in diabetic animal models again toward the end of the 6-h monitoring in vivo and in individuals with Type 2 diabetes. period. The latter fi nding refl ects the absence of Most of the published preclinical studies basal insulin and is in line with earlier observa- with insulin analogues provide little informa- tions [51,104], which suggested that an adjustment tion regarding biological action beyond activa- of basal insulin may be required when using a new tion of a limited panel of signaling proteins, fast-acting insulin analogue. Glulisine was safe stimulation of DNA synthesis and glucose and well tolerated in the Type 1 diabetic children disposal. Glulisine, similar to RHI and other and adolescents [103] . insulin analogues, may affect cell function at Results of a Phase III study comparing the multiple levels, such as signal transduction, glucodynamic effi cacy and safety of glulisine metabolic pathways, gene expression and stress and lispro as part of a basal-bolus regimen in tolerance. From a system biological point of view 572 Type 1 diabetic children and adolescents these levels represent particular projections of (aged 4–17 years) was published in abstract a highly complex cellular response [106] . As a form [105] . This randomized, parallel-group fi rst approach to increase insight into the entire study reported similar effects for glulisine and biological activity of insulin analogues, gene-

lispro on HbA1c, even though a slightly higher expression profi les (transcriptome and proteome)

fraction of patients achieved age-specifi c HbA1c and post-translational protein modifi cations targets with glulisine (38.4%) than with lispro (e.g., the phosphoproteome) should be recorded (32.0%, p = 0.0251). Interestingly, fasting blood with state-of-the-art technology in human cell glucose concentrations were lower in the glu- lines under standardized experimental condi- lisine group. No differences were observed in tions. Gene-expression profi ling was already hypoglycemia rates, nor in the frequency and applied in the biological characterization of type of adverse events. insulin mimetic peptides [107] . Such a global approach will certainly help to estimate specifi c Conclusion risks and benefi ts beyond mitogenicity and gly- Compared with RHI, insulin glulisine displays cemic control of individual insulin analogues noninferiority with regard to glycemic control in including glulisine. Type 1 and 2 diabetic individuals. Until today no clinical superiority of glulisine over other fast- Glulisine in the treatment of acting analogues was demonstrated. Preclinical Type 1 & Type 2 diabetes mellitus in data and experimental human trials point to adults & children some specifi c characteristics of glulisine. As revealed by hyperinsulinemic euglycemic clamp studies, glulisine in healthy volunteers as Specifi c signaling & cell-protective well as in patients with Type 1 and 2 diabetes dis- properties of glulisine plays pharmacokinetic and glucodynamic profi les Although stimulation of DNA synthesis and that are typical for a fast-acting insulin analogue. glucose uptake was comparable with that of Pharmacokinetic characteristics of glulisine versus RHI [61,63–66], glulisine displayed distinct signal- RHI are closely related to its faster absorption, and ing properties as it preferentially activated IRS-2 include earlier achievement of maximum blood in different cell types including the rat pancreatic concentration and a shorter mean residence time. β-cell line INS-1 [61,67]. This implies that partly This results in a faster onset and a shorter duration different signaling pathways may be involved in of blood glucose-lowering action, which shifts the regulation of cell proliferation and glucose uptake main effect on glucose disposal closer to the time by glulisine and RHI, respectively. Furthermore, point of injection. Thus, as part of a basal-bolus

224 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

regimen, compared with RHI, glulisine decreases postprandial oxidative stress and time fl exibility of the mismatch between insulin absorption and the administration [22,45–51]. However, due to the lack postprandial need for gluco dynamic activity, and of large outcome trials, it still remains unclear if ensures a post-prandial glucose disposal closer to these differences are clinically meaningful – that physiological conditions as observed consistently is, if fast acting analogues attenuate the progres- throughout the clinical studies with Type 1 and sion of diabetes and the incidence of micro- and 2 diabetic patients. This might be of importance, macro-vascular complications better than RHI. as several epidemiological studies showed a direct Experimental trials comparing glulisine and lispro correlation between the extent of postprandial identifi ed some minor advantages of glulisine, in blood glucose excursions, but not of fasting values particular a slightly faster onset of action in obese [66,100] or HbA1c, and the risk for cardiovascular disease people and a small increase in the fraction [108–110] . A recent exploratory study with Type 2 of patients that achieved age-specifi c HbA1c tar- diabetic patients and healthy volunteers suggests gets plus a decrease of fasting blood glucose con- that the oscillations in blood glucose concentra- centrations in children with Type 1 diabetes [105] . tions might be even more deleterious than elevated Preclinical studies point to glulisine-specific mean blood glucose concentrations with regard to action profi les beyond glycemic control (prefer- oxidative stress and endothelial dysfunction [3]. ential targeting of IRS-2, β-cell protection) [61,67]. In this context it is noteworthy that glulisine was From the scientifi c viewpoint, future position- more effective than RHI in attenuating 3´-nitro- ing of glulisine will certainly depend on whether tyrosine and asymmetric dimethylarginine for- large clinical trials will confi rm an additional mation, as well as intact pro-insulin secretion in benefi t of glulisine in glycemic control in obese Type 2 diabetic patients following ingestion of a Type 2 diabetic people and diabetic children. test meal [4]. These observations may imply an In-depth preclinical analysis should focus on improved protection from post-prandial oxida- the evaluation of insulin and insulin analogues tive stress and endothelial dysfunction by the ana- beyond their effects on cellular glucose uptake logue. Again, these fi ndings need to be proven in and DNA synthesis. Such an open-minded large-scale clinical trials. research strategy may unravel a pharmacody- Under the controlled conditions of clinical tri- namic diversity of insulin analogues and fi nally als, glulisine as compared with RHI causes only defi ne a panel of insulin-related products with a minor, if any, additional decay in HbA1c values. distinct action profi les apart from providing However, glulisine might offer increased fl ex- adequate glycemic control. Such a full charac- ibility with regard to injection timing, as glyce- terization may uncover options for individu- mic control, in terms of HbA1c and postprandial ally tailored insulin therapy regimens, and also blood glucose excursions, was maintained with profi le the fast-acting analogues in competition glulisine administered shortly before or even after with upcoming RHI preparations that acceler- ingestion of a meal, while for RHI a 15–30 min ate insulin absorption merely by modifying the interval between injection and meal intake is rec- formulation (e.g., [112]). ommended. Keeping in mind that many diabetic At present, glulisine is undoubtedly a valu- patients ignore this (inconvenient) recommenda- able additional tool in achieving improved tion risking suboptimal glycemic control [111], glu- glycemic control in people with Type 1 and 2 lisine might be able to achieve superior glycemic diabetes, including children and adolescents. control compared to RHI under daily life condi- The preclinical data are promising that a full tions, in particular in diabetic patients with irregu- functional evaluation of this analogue may lar eating habits, such as children, adolescents or uncover additional properties of potential the elderly. therapeutic benefi t.

Future perspective Financial & competing interests disclosure Currently, no human study is available that com- The authors have no relevant affi liations or fi nancial involve- pares the marketed fast-acting insulin analogues ment with any organization or entity with a fi nancial interest head to head. This makes a prognostic position- in or fi nancial confl ict with the subject matter or materials ing of individual analogues within the upcoming discussed in the manuscript. This includes employment, con- 5–10 years diffi cult. There is no doubt regarding sultancies, honoraria, stock ownership or options, expert tes- advantages of the fast-acting insulin analogues timony, grants or patents received or pending, or royalties. over RHI in terms of glycemic control (post- No writing assistance was utilized in the production of prandial blood glucose excursions), suppression of this manuscript.

future science group www.futuremedicine.com 225 DRUG EVALUATION Schliess & Heise

Executive summary Introduction Diabetes mellitus is a worldwide epidemic with increasing incidence in children and adolescents. Insulin glulisine (glulisine) is a new fast-acting insulin analogue approved for the treatment of diabetes mellitus Type 1 and 2. Preclinical evaluation The mitogenic and glucose-metabolizing potency of glulisine is comparable with that of regular human insulin (RHI). Glulisine displays β-cell-protective properties, which may be related to a preferential tasking of the insulin receptor substrate-2 by this analogue. The clinical relevance of this fi nding is currently unclear. Pharmacokinetics & pharmacodynamics Glulisine is absorbed faster than RHI. Therefore, pharmacokinetic hallmarks include shorter time to maximum blood concentration and a shorter mean residence time. This results in a faster onset and a shorter duration of glucodynamic action, which shifts the bulk of glucose disposal closer to the time point of injection. Glulisine showed a faster onset of action compared with other fast-acting insulin analogues in healthy people irrespective of body mass index. The relevance of this fi nding in the treatment of diabetes has not yet been established. Clinical effectiveness Large interventional trials with glulisine as part of a basal-bolus regimen consistently demonstrate that glulisine displays a higher effi cacy in the suppression of postprandial glucose excursions than RHI in people with Type 1 and 2 diabetes. This was related to noninferiority

to RHI in terms of HbA1c levels and blood glucose concentrations. Glulisine achieves good glycemic control in Type 1 diabetic children and adolescents. Glulisine allows a fl exible timing of meal-related injection (including the option of postprandial injections) without marked deteriorations in glycemic control. Safety & tolerability No safety and tolerability fi ndings of concern were observed in any of the clinical studies. First-time use of glulisine as a bolus insulin requires dose adaptation of the basal insulin, as is the case for other fast-acting insulin analogues. Conclusion Preclinical evaluation: insulin exerts pleiotropic effects on cell metabolism and gene expression, in addition to controlling carbohydrate metabolism. In order to get a reliable estimation of all the effects of insulin analogues, we recommend head-to-head comparisons under standardized experimental conditions. This may include recording of gene-expression profi les and post-translational modifi cations in standardized cell lines, which may point to specifi c risks and benefi ts of individual analogues that deserve further investigation (such as β-cell protection through glulisine). Clinical effi cacy: Although long-term experience is lacking, glulisine is as safe and tolerable as RHI. Due to its pharmacokinetic hallmarks, glulisine is superior to cope with post-prandial blood glucose excursions in Type 1 and 2 diabetic patients. Specifi c benefi ts versus other insulin analogues for the treatment of obese and Type 1 diabetic children and adolescents need to be confi rmed. Glulisine represents a valuable tool for optimizing glycemic control, and thereby for attenuating micro- and macro-vascular complications.

Bibliography fl ow and endothelial function in the 11 Zaid H, Antonescu CN, Randhawa VK, postprandial state. Diabetes Care 31, Klip A: Insulin action on glucose Papers of special note have been highlighted as: 1021–1025 (2008). transporters through molecular switches, of interest tracks and tethers. Biochem. J. 413, 201–215 of considerable interest 5 O’Brien RM, Granner DK: Regulation of (2008). 1 Kahn CR: Banting Lecture. Insulin action, gene expression by insulin. Biochem. J. 278, diabetogenes, the cause of Type II diabetes. 609–619 (1991). 12 White MF: The IRS-signalling system: Diabetes 43, 1066–1084 (1994). 6 Schäfer C, Gehrmann T, Richter L et al.: a network of docking proteins that mediate Modulation of gene expression profi les by insulin action. Mol. Cell. Biochem. 182, 3–11 Provides guidance through the molecular hyperosmolarity and insulin. Cell. Physiol. (1998). mechanisms underlying insulin action, Biochem. 20(5) 369–386 (2007). 13 Duckworth WC, Bennett RG, Hamel FG: insulin resistance and the pathogenesis of Insulin degradation: progress and potential. diabetes mellitus. 7 Pliquett RU, Fuhrer D, Falk S, Zysset S, von Cramon DY, Stumvoll M: The effects of Endocr. Rev. 19, 608–624 (1998). 2 Michalopoulos GK, DeFrances MC: Liver insulin on the central nervous system – focus 14 Di Guglielmo GM, Drake PG, Baass PC, regeneration. Science 276, 60–66 (1997). on appetite regulation. Horm. Metab. Res. 38, Authier F, Posner BI, Bergeron JJ: Insulin 3 Ceriello A, Esposito K, Piconi L et al.: 442–446 (2006). receptor internalization and signalling. Mol. Oscillating glucose is more deleterious to 8 Fisher SJ, Kahn CR: Insulin signaling is Cell. Biochem. 182, 59–63 (1998). endothelial function and oxidative stress than required for insulin’s direct and indirect 15 Araki E, Lipes MA, Patti ME et al.: mean glucose in normal and Type 2 diabetic action on hepatic glucose production. J. Clin. Alternative pathway of insulin signalling in patients. Diabetes 57, 1349–1354 (2008). Invest. 111, 463–468 (2003). mice with targeted disruption of the IRS-1 Demonstrates that oscillating blood glucose 9 Shulman GI: Cellular mechanisms of insulin gene. Nature 372, 186–190 (1994). concentrations, rather than a sustained resistance. J. Clin. Invest. 106, 171–176 This paper (similar to [16]) describes the hyperglycemia, may impair endothelial (2000). nondiabetic phenotype of IRS-1 knockout function and redox homeostasis. 10 Schliess F, Häussinger D: Cell volume and mice. A milestone for the understanding of 4 Hohberg C, Forst T, Larbig M et al.: Effect insulin signalling. Int. Rev. Cytol. 225, 187–228 redundancies in insulin signaling. of insulin glulisine on microvascular blood (2003).

226 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

16 Tamemoto H, Kadowaki T, Tobe K et al.: 28 Blundell TL, Cutfi eld JF, Dodson GG, 40 Holman RR, Thorne KI, Farmer AJ et al.: Insulin resistance and growth retardation in Dodson E, Hodgkin DC, Mercola D: Addition of biphasic, prandial, or basal mice lacking insulin receptor substrate-1. The structure and biology of insulin. insulin to oral therapy in Type 2 diabetes. Nature 372, 182–186 (1994). Biochem. J. 125, 50P–51P (1971). N. Engl. J. Med. 357, 1716–1730 (2007).

This paper (similar to [15]) describes the 29 Baker EN, Blundell TL, Cutfi eld JF et al.: 41 Brems DN, Alter LA, Beckage MJ et al.: nondiabetic phenotype of IRS-1 knockout The structure of 2 Zn pig insulin crystals at Altering the association properties of insulin mice. A milestone for the understanding of 1.5 A resolution. Philos. Trans. R. Soc. Lond, by amino acid replacement. Protein Eng. 5, redundancies in insulin signaling. B, Biol. Sci. 319, 369–456 (1988). 527–533 (1992). 17 Withers DJ, Gutierrez JS, Towery H et al.: 30 Fischer WH, Saunders D, Brandenburg D, 42 Bakaysa DL, Radziuk J, Havel HA et al.: Disruption of IRS-2 causes Type 2 diabetes Wollmer A, Zahn H: A shortened insulin Physicochemical basis for the rapid in mice. Nature 391, 900–904 (1998). with full in vitro potency. Biol. Chem. time–action of LysB28ProB29-insulin: Hoppe–Seyler 366, 521–525 (1985). dissociation of a protein-ligand complex. Describes the diabetic phenotype of IRS-2 Protein Sci. 5, 2521–2531 (1996). knockout mice and provides early and 31 Brange J, Ribel U, Hansen JF et al.: profound evidence for the involvement of Monomeric insulins obtained by protein 43 Ciszak E, Beals JM, Frank BH, Baker JC, engineering and their medical implications. Carter ND, Smith GD: Role of C-terminal IRS-2 in β-cell protection. Nature 333, 679–682 (1988). B-chain residues in insulin assembly: the 18 Siddle K, Urso B, Niesler CA et al.: Specifi city structure of hexameric LysB28ProB29-human Broadly introduces the conception of in ligand binding and intracellular signalling by insulin. Structure 3, 615–622 (1995). insulin and insulin-like growth factor receptors. genetically engineered fast-acting insulin 44 Whittingham JL, Edwards DJ, Antson AA, Biochem. Soc. Trans. 29, 513–525 (2001). analogues in the treatment of diabetes. Clarkson JM, Dodson GG: Interactions of 32 Becker RH: Insulin glulisine complementing 19 Entingh-Pearsall A, Kahn CR: Differential phenol and m-cresol in the insulin hexamer, basal insulins: a review of structure and activity. roles of the insulin and insulin-like growth their effect on the association properties of Diabetes Technol. Ther. 9, 109–121 (2007). factor-I (IGF-I) receptors in response to B28 pro –> Asp insulin analogues. insulin and IGF-I. J. Biol. Chem. 279, 33 Schwartz GP, Burke GT, Katsoyannis PG: Biochemistry 37, 11516–11523 (1998). 38016–38024 (2004). A superactive insulin: [B10-aspartic acid] 45 Miles HL, Acerini CL: Insulin analog insulin(human). Proc. Natl Acad. Sci. USA 20 Brange J, Owens DR, Kang S, Volund A: preparations and their use in children and 84, 6408–6411 (1987). Monomeric insulins and their experimental adolescents with Type 1 diabetes mellitus. and clinical implications. Diabetes Care 13, Introduces the ‘superactive’ insulin analogue Paediatr. Drugs 10, 163–176 (2008). 923–954 (1990). AspB10, a prototypal fast-acting analogue 46 Roach P: New insulin analogues and routes of A comprehensive review that outlines the which until today serves as a reference in the delivery : pharmacodynamic and clinical development of insulin analogues. preclinical evaluation of insulin analogues. considerations. Clin. Pharmacokinet. 47, 21 Luzi L, DeFronzo RA: Effect of loss of 34 Richards JP, Stickelmeyer MP, Flora DB, 595–610 (2008). fi rst-phase insulin secretion on hepatic glucose Chance RE, Frank BH, DeFelippis MR: 47 Mannucci E, Monami M, Marchionni N: production and tissue glucose disposal in Self-association properties of monomeric Short-acting insulin analogues vs. regular humans. Am. J. Physiol. 257, E241–E246 insulin analogs under formulation conditions. human insulin in Type 2 diabetes: (1989). Pharm. Res. 15, 1434–1441 (1998). a meta-analysis. Diabetes Obes. Metab.81, 22 Bell DS: Insulin therapy in diabetes 35 Heinemann L, Starke AA, Heding L, 184–189 (2008). mellitus: how can the currently available Jensen I, Berger M: Action profi les of fast 48 Gough SC: A review of human and analogue injectable insulins be most prudently and onset insulin analogues. Diabetologia 33, insulin trials. Diabetes Res. Clin. Pract. 77, effi caciously utilised? Drugs 67, 1813–1827 384–386 (1990). 1–15 (2007). (2007). 36 Nielsen FS, Jorgensen LN, Ipsen M, 49 Siebenhofer A, Plank J, Berghold A et al.: 23 Raccah D: Options for the intensifi cation of Voldsgaard AI, Parving HH: Long-term Short acting insulin analogues versus regular insulin therapy when basal insulin is not comparison of human insulin analogue human insulin in patients with diabetes enough in Type 2 diabetes mellitus. Diabetes B10Asp and soluble human insulin in IDDM mellitus. Cochrane. Database. Syst. Rev. 19, Obes. Metab. 10(Suppl. 2). 76–82 (2008). patients on a basal/bolus insulin regimen. CD003287 (2006). Diabetologia 38, 592–598 (1995). 24 Schade DS, Eaton RP: Insulin delivery: how, 50 Ceriello A, Quagliaro L, Catone B et al.: when, where. N. Engl. J. Med. 312, 1120– 37 Jorgensen LN, Dideriksen L, Drejer K: Role of hyperglycemia in nitrotyrosine 1121 (1985). Carcinogenic effect of the human insulin postprandial generation. Diabetes Care 25, analogue B10 Asp in female rats. Diabetologia 25 Binder C, Lauritzen T, Faber O, Pramming S: 1439–1443 (2002). 35 (1992) (Abstract A3). Insulin pharmacokinetics. Diabetes Care 7, 51 Heise T, Heinemann L: Rapid and long- 188–199 (1984). 38 Heise T, Weyer C, Serwas A et al.: acting analogues as an approach to improve Time–action profi les of novel premixed 26 Brange J: Galenics of Insulin – The Physico- insulin therapy: an evidence-based medicine preparations of insulin lispro and NPL Chemical and Pharmaceutical Aspects of Insulin assessment. Curr. Pharm. Des. 7, 1303–1325 insulin. Diabetes Care 21, 800–803 (1998). and Insulin Preparations. Springer–Verlag, (2001). 39 Bott S, Tusek C, Heinemann L, Friberg HH, Heidelberg, Germany (1987). This comprehensive review applies the Heise T: The pharmacokinetic and 27 Mosekilde E, Jensen KS, Binder C, evidence-based approach to the evaluation of pharmacodynamic properties of biphasic Pramming S, Thorsteinsson B: Modeling insulin analogues and provides in-depth insulin Aspart 70 (BIAsp 70) are signifi cantly absorption kinetics of subcutaneous injected information about the pharmacokinetic and different from those of biphasic insulin soluble insulin. J. Pharmacokinet. Biopharm. glucodynamic characterization of insulin Aspart 30 (BIAsp 30). Exp. Clin. Endocrinol. 17, 67–87 (1989). Diabetes 113, 545–550 (2005). (analogue) preparations.

future science group www.futuremedicine.com 227 DRUG EVALUATION Schliess & Heise

52 Sluzky V, Tamada JA, Klibanov AM, 63 Hennige AM, Strack V, Metzinger E, the ischemic injury and decreases the graft Langer R: Kinetics of insulin aggregation in Seipke G, Haring HU, Kellerer M: Effects of survival rate in rat liver transplantation. aqueous solutions upon agitation in the new insulin analogues HMR1964 (insulin Transplantation 76, 44–49 (2003). presence of hydrophobic surfaces. Proc. Natl glulisine) and HMR1423 on insulin 74 Porras A, Zuluaga S, Valladares A et al.: Acad. Sci. USA 88, 9377–9381 (1991). receptors. Diabetologia 48, 1891–1897 (2005). Long-term treatment with insulin induces 53 Brange J, Andersen L, Laursen ED, Meyn G, 64 Hennige AM, Lehmann R, Weigert C et al.: apoptosis in brown adipocytes: role of Rasmussen E: Toward understanding insulin Insulin glulisine: insulin receptor signaling oxidative stress. Endocrinology 144, fi brillation. J. Pharm. Sci. 86, 517–525 (1997). characteristics in vivo. Diabetes 54, 361–366 5390–5401 (2003). 54 Becker RH, Frick AD: Clinical (2005). 75 Godbout JP, Cengel KA, Cheng SL, pharmacokinetics and pharmacodynamics of 65 Ciaraldi TP, Phillips SA, Carter L, Aroda V, Minshall C, Kelley KW, Freund GG: Insulin insulin glulisine. Clin. Pharmacokinet. 47, Mudaliar S, Henry RR: Effects of the activates caspase-3 by a phosphatidylinositol 7–20 (2008). rapid-acting insulin analog glulisine on 3’-kinase-dependent pathway. Cell Signal. 11, 55 Hansen BF, Danielsen GM, Drejer K et al.: cultured human skeletal muscle cells: 15–23 (1999). Sustained signalling from the insulin receptor comparisons with insulin and insulin-like 76 Wiedmann M, Tamaki S, Silberman R, de la after stimulation with insulin analogues growth factor I. J. Clin. Endocrinol. Metab. Monte SM, Cousens L, Wands JR: exhibiting increased mitogenic potency. 90, 5551–5558 (2005). Constitutive over-expression of the insulin Biochem. J. 315, 271–279 (1996). 66 Becker RH, Frick AD, Burger F, Potgieter JH, receptor substrate-1 causes functional 56 Kurtzhals P, Schaffer L, Sorensen A et al.: Scholtz H: Insulin glulisine, a new rapid- up-regulation of Fas receptor. J. Hepatol. 38, Correlations of receptor binding and acting insulin analogue, displays a rapid 803–810 (2003). metabolic and mitogenic potencies of insulin time-action profi le in obese non-diabetic 77 Nagle JA, Ma Z, Byrne MA, White MF, analogs designed for clinical use. Diabetes 49, subjects. Exp. Clin. Endocrinol. Diabetes 113, Shaw LM: Involvement of insulin receptor 999–1005 (2000). 435–443 (2005). substrate 2 in mammary tumor metastasis. 57 Milazzo G, Sciacca L, Papa V, Goldfi ne ID, 67 Rakatzi I, Seipke G, Eckel J: [LysB3, GluB29] Mol. Cell. Biol. 24, 9726–9735 (2004). Vigneri R: ASPB10 insulin induction of insulin: a novel insulin analog with enhanced 78 Garg SK, Rosenstock J, Ways K: Optimized increased mitogenic responses and phenotypic β-cell protective action. Biochem. Biophys. Res. Basal-bolus insulin regimens in Type 1 changes in human breast epithelial cells: Commun. 310, 852–859 (2003). diabetes: insulin glulisine versus regular human evidence for enhanced interactions with the Provides a head-to-head comparison of insulin in combination with Basal insulin insulin-like growth factor-I receptor. Mol. regular human insulin, lispro, aspart, glargine. Endocr. Pract. 11, 11–17 (2005). Carcinog. 18, 19–25 (1997). B10 glulisine, Asp and IGF-1 with regard to This Phase III trial (similar to [88]) indicates 58 Eckardt K, May C, Koenen M, Eckel J: IGF-1 cell-protective properties. By investigation noninferiority of insulin glulisine with receptor signalling determines the mitogenic of rat INS-1 insulinoma cells, this study regard to glycemic control in Type 1 potency of insulin analogues in human points to a protection of β-cells by glulisine diabetic people. smooth muscle cells and fi broblasts. from fatty acid- and infl ammatory 79 Dailey G, Rosenstock J, Moses RG, Ways K: Diabetologia 50, 2534–2543 (2007). cytokine-induced apoptosis. Insulin glulisine provides improved glycemic 59 Ribel U, Hougaard P, Drejer K, Sorensen AR: 68 Kubota N, Tobe K, Terauchi Y et al.: control in patients with Type 2 diabetes. Equivalent in vivo biological activity of Disruption of insulin receptor substrate 2 Diabetes Care 27, 2363–2368 (2004). insulin analogues and human insulin despite causes Type 2 diabetes because of liver insulin This Phase III trial (similar to [80]) indicates different in vitro potencies. Diabetes 39, resistance and lack of compensatory β-cell noninferiority of insulin glulisine with 1033–1039 (1990). hyperplasia. Diabetes 49, 1880–1889 (2000). regard to glycemic control in Type 2 60 Bahr M, Kolter T, Seipke G, Eckel J: Growth 69 Cantley J, Choudhury AI, sare-Anane H diabetic people. promoting and metabolic activity of the et al.: Pancreatic deletion of insulin receptor 80 Rayman G, Profozic V, Middle M: Insulin human insulin analogue substrate 2 reduces β and α cell mass and glulisine imparts effective glycaemic control [GlyA21,ArgB31,ArgB32]insulin (HOE 901) impairs glucose homeostasis in mice. in patients with Type 2 diabetes. Diabetes Res. in muscle cells. Eur. J. Pharmacol. 320, Diabetologia 50, 1248–1256 (2007). 259–265 (1997). Clin. Pract. 76, 304–312 (2007). 70 Lingohr MK, Dickson LM, Wrede CE, [79] 61 Rakatzi I, Ramrath S, Ledwig D, McCuaig JF, Myers MG Jr, Rhodes CJ: This Phase III trial (similar to ) indicates Dransfeld O, Bartels T, Seipke G, Eckel J: IRS-3 inhibits IRS-2-mediated signaling in noninferiority of insulin glulisine with A novel insulin analog with unique pancreatic β-cells. Mol. Cell. Endocrinol. 204, regard to glycemic control in Type 2 properties: LysB3,GluB29 insulin induces 85–99 (2003). diabetic people. prominent activation of insulin receptor 71 Hennige AM, Burks DJ, Ozcan U et al: 81 Becker RH, Frick AD, Burger F, Scholtz H, substrate 2, but marginal phosphorylation of Upregulation of insulin receptor substrate-2 Potgieter JH: A comparison of the steady-state insulin receptor substrate 1. Diabetes 52, in pancreatic β cells prevents diabetes. J. Clin. pharmacokinetics and pharmacodynamics of 2227–2238 (2003). Invest. 112, 1521–1532 (2003). a novel rapid-acting insulin analog, insulin Provides a comprehensive examination of glulisine, regular human insulin in healthy 72 Li D, Yin X, Zmuda EJ et al.: The repression volunteers using the euglycemic clamp insulin glulisine signal transduction and its of IRS2 gene by ATF3, a stress-inducible technique. Exp. Clin. Endocrinol. Diabetes glucose lowering and mitogenic potency. gene, contributes to pancreatic β-cell 113, 292–297 (2005). Discovery of the preferential activation of apoptosis. Diabetes 57, 635–644 (2008). IRS-2 by glulisine. 82 Becker RH, Frick AD, Wessels DH, 73 Li XL, Man K, Liu YF, Lee TK, Tsui SH, Scholtz HE: Evaluation of the 62 Owens DR, Zinman B, Bolli GB: Insulins Lau CK, Lo CM, Fan ST: Insulin in pharmacodynamic and oharmacokinetic today and beyond. Lancet 358, 739–746 (2001). University of Wisconsin solution exacerbates profi les of insulin glulisine: a novel,

228 Therapy (2009) 6(2) future science group Insulin glulisine: preclinical hallmarks & clinical effi cacy DRUG EVALUATION

rapid-acting human insulin analogue. injection pre-mixed insulin regimen in Type 2 105 Philotheou A, Arslanian S, Blatniczky L, Diabetologia 46, (2003) (Abstract A775). diabetes patients – results of the GINGER Peterkova V, Souhami E, Danne T: Effi cacy 83 Frick AD, Becker RH, Wessels DH, study. Diabetologia 51 (2008) (Abstract A186). and safety of insulin glulisine versus insulin Scholtz HE: Pharmacokinetic and 94 Vora JP, Burch A, Peters JR, Owens DR: lispro as part of a basal-bolus insulin regimen glucodynamic profi les of insulin glulisine: an Relationship between absorption of in children and adolescents with Type 1 evaluation following subcutaneous radiolabeled soluble insulin, subcutaneous diabetes mellitus. Diabetologia 51 (2008) administration at various injection sites. blood fl ow, anthropometry. Diabetes Care 15, (Abstract A950). Diabetologia 46, (2003) (Abstract A776). 1484–1493 (1992). 106 Kitano H: Systems biology: a brief overview. 84 Berger M, Cuppers HJ, Hegner H, Jorgens V, 95 Vora JP, Burch A, Peters JR, Owens DR: Science 295, 1662–1664 (2002). Berchtold P: Absorption kinetics and biologic Absorption of radiolabelled soluble insulin in 107 Jensen M, Palsgaard J, Borup R, De MP, effects of subcutaneously injected insulin Type 1 (insulin-dependent) diabetes: Schaffer L: Activation of the insulin receptor preparations. Diabetes Care 5, 77–91 (1982). infl uence of subcutaneous blood fl ow and (IR) by insulin and a synthetic peptide has 85 Frick AD, Scholtz H, Burger F, Becker RH: anthropometry. Diabet. Med. 10, 736–743 different effects on gene expression in Absorption of insulin glulisine when mixed (1993). IR-transfected L6 myoblasts. Biochem. J. 412, with NPH insulin. Diabetes 53, (2004) 96 Sindelka G, Heinemann L, Berger M, 435–445 (2008). (Abstract A321). Frenck W, Chantelau E: Effect of insulin 108 Ceriello A, Hanefeld M, Leiter L et al.: 86 Horvath K, Bock G, Regittnig W et al.: concentration, subcutaneous fat thickness and Postprandial glucose regulation and diabetic Insulin glulisine, insulin lispro and regular skin temperature on subcutaneous insulin complications. Arch. Intern. Med. 164, human insulin show comparable end-organ absorption in healthy subjects. Diabetologia 2090–2095 (2004). metabolic effects: an exploratory study. 37, 377–380 (1994). 109 No authors listed: Glucose tolerance and Diabetes Obes. Metab. 10, 484–491 (2008). 97 Jehle PM, Fussganger RD, Seibold A, mortality: comparison of WHO and 87 Becker RH, Frick AD, Nosek L, Luttke B, Bohm BO: Pharmacodynamics of American Diabetes Association diagnostic Heinemann L, Rave K: Dose–response insulin Lispro in 2 patients with Type II criteria. The DECODE study group. relationship of insulin glulisine in subjects diabetes mellitus. Int. J. Clin. Pharmacol. European Diabetes Epidemiology Group. with Type 1 diabetes. Diabetes Care 30, Ther. 34, 498–503 (1996). Diabetes Epidemiology: Collaborative 2506–2507 (2007). 98 Holmes G, Galitz L, Hu P, Lyness W: analysis Of Diagnostic criteria in Europe. Lancet 354, 617–621 (1999). 88 Dreyer M, Prager R, Robinson A et al.: Pharmacokinetics of insulin aspart in obesity, Effi cacy and safety of insulin glulisine in renal impairment, or hepatic impairment. 110 Chiasson JL, Josse RG, Gomis R, patients with Type 1 diabetes. Horm. Metab. Br. J. Clin. Pharmacol. 60, 469–476 (2005). Hanefeld M, Karasik A, Laakso M: Acarbose Res. 37, 702–707 (2005). 99 Barnett AH: How well do rapid-acting treatment and the risk of cardiovascular insulins work in obese individuals? Diabetes disease and hypertension in patients with This Phase III trial (similar to [78]) indicates Obes. Metab. 8, 388–395 (2006). impaired glucose tolerance: the STOP- noninferiority of insulin glulisine with regard NIDDM trial. JAMA 290, 486–494 (2003). to glycemic control in Type 1 diabetic people. 100 Heise T, Nosek L, Spitzer H et al.: Insulin glulisine: a faster onset of action compared 111 Heinemann L: Do insulin-treated diabetic 89 Rave K, Klein O, Frick AD, Becker RH: with insulin lispro. Diabetes Obes. Metab. 9, patients use an injection-meal-interval in Advantage of premeal-injected insulin 746–753 (2007). daily life? Diabet. Med. 449–450 (1995). glulisine compared with regular human 112 Steiner S, Hompesch M, Pohl R et al.: A novel insulin in subjects with Type 1 diabetes. 101 Becker RH, Frick AD, Kapitza C, Heise T, insulin formulation with a more rapid onset Diabetes Care 29, 1812–1817 (2006). Rave K: Pharmacodynamics (PD) and pharmacokinetics of insulin glulisine of action. Diabetologia 51, 1602–1606 (2008). This paper provides support for a fl exible compared with insulin lispro (IL) and regular timing of meal-related glulisine injections human insulin (RHI) in patients with Type 2 without loss of proper glycemic control. diabetes. Diabetes 53 (2004) (Abstract 503P). Websites 90 Hoogma RP, Schumicki D: Safety of insulin 102 Luzio S, Peter R, Dunseath GJ, Mustafa L, 201 Food and Drug Administration, Department glulisine when given by continuous Owens DR: A comparison of preprandial of Health and Human Services subcutaneous infusion using an external insulin glulisine versus insulin lispro in people www.fda.gov/cder/foi/ pump in patients with Type 1 diabetes. Horm. with Type 2 diabetes over a 12-h period. appletter/2004/21629ltr.pdf Metab. Res. 38, 429–433 (2006). Diabetes Res. Clin. Pract. 79, 269–275 (2008). 202 Press release: EMEA: Committee for 91 Senstius J, Poulsen C, Hvass A: Comparison 103 Danne T, Becker RH, Heise T, Bittner C, Medicinal Products for Human Use, 1–3 June of in vitro stability for insulin aspart and Frick AD, Rave K: Pharmacokinetics, prandial 2004 insulin glulisine during simulated use in glucose control, safety of insulin glulisine in www.emea.europa.eu/pdfs/human/press/ insulin pumps. Diabetes Technol. Ther. 9, children and adolescents with Type 1 diabetes. pr/1535804en.pdf 517–521 (2007). Diabetes Care 28, 2100–2105 (2005). 203 European Medicines Agency, European 92 Lankisch M, Ferlinz K, Scherbaum W: Indicates that insulin glulisine is safe and Public Assessment Report Adding a single dose of insulin glulisine at guarantees a proper glycemic control in www.emea.europa.eu/humandocs/PDFs/ breakfast or main meal to basal insulin and children and adolescents with Type 1 diabetes. EPAR/apidra/121804en1.pdf oral antidiabetic therapy: which patients of the OPAL study nenefi t most? Diabetologia 51 104 Ciofetta M, Lalli C, Del SP et al.: 204 Pubmed (2008) (Abstract A185). Contribution of postprandial versus www.ncbi.nlm.nih.gov/pubmed/, search: interprandial blood glucose to HbA1c in ‘insulin analogues OR insulin glulisine’, Table 93 Fritsche A, Larbig M, Böhler S, Haring HU: Type 1 diabetes on physiologic intensive Superior effi cacy of basal–bolus regimen with therapy with lispro insulin at mealtime. insulins glargine and glulisine versus a two Diabetes Care 22, 795–800 (1999).

future science group www.futuremedicine.com 229