Sequence-specific 1H-NMR assignments for the aromatic region of several biologically active, monomeric insulins including native human insulin.

The aromatic region of the 1H-FT-NMR spectrum of the biologically fully-potent, monomeric human insulin mutant, B9 Ser----Asp, B27 Thr----Glu has been investigated in D2O. At 1 to 5 mM concentrations, this mutant insulin is monomeric above pH 7.5. Coupling and amino acid classification of all aromatic signals is established via a combination of homonuclear one- and two-dimensional methods, including COSY, multiple quantum filters, selective spin decoupling and pH titrations. By comparisons with other insulin mutants and with chemically modified native insulins, all resonances in the aromatic region are given sequence-specific assignments without any reliance on the various crystal structures reported for insulin. These comparisons also give the sequence-specific assignments of most of the aromatic resonances of the mutant insulins B16 Tyr----Glu, B27 Thr----Glu and B25 Phe----Asp and the chemically modified species des-(B23-B30) insulin and monoiodo-Tyr A14 insulin. Chemical dispersion of the assigned resonances, ring current perturbations and comparisons at high pH have made possible the assignment of the aromatic resonances of human insulin, and these studies indicate that the major structural features of the human insulin monomer (including those critical to biological function) are also present in the monomeric mutant.

Role of B13 Glu in insulin assembly. The hexamer structure of recombinant mutant (B13 Glu-->Gln) insulin.

The assembly of the insulin hexamer brings the six B13 glutamate side-chains at the centre into close proximity. Their mutual repulsion is unfavourable and zinc co-ordination to B10 histidine is necessary to stabilize the well known zinc-containing hexamers. Since B13 is always a carboxylic acid in all known sequences of hexamer forming insulins, it is likely to be important in the hormone's biology. The mutation of B13 Glu-->Gln leads to a stable zinc-free hexamer with somewhat reduced potency. The structures of the zinc-free B13 Gln hexamer and the 2Zn B13 insulin hexamer have been determined by X-ray analysis and refined with 2.5 A and 2.0 A diffraction data, respectively. Comparisons show that in 2Zn B13 Gln insulin, the hexamer structure (T6) is very like that of the native hormone. On the other hand, the zinc-free hexamer assumes a quaternary structure (T3/R3) seen in the native 4Zn insulin hexamer, and normally associated only with high chloride ion concentrations in the medium. The crystal structures show the B13 Gln side-chains only contact water in contrast to the B13 glutamate in 2Zn insulin. The solvation of the B13 Gln may be associated with this residue favouring helix at B1 to B8. The low potency of the B13 Gln insulin also suggests the residue influences the hormone's conformation.

Absorption of subcutaneously infused insulin: influence of the basal rate pulse interval.

Eight insulin-dependent diabetic patients were given two constant infusions (each 1 IU/h) of 125I-labeled insulin into the abdominal subcutaneous tissue for about 12 h. Insulin was infused in pulses into one side of the abdomen in 6-min intervals (by means of an Auto-Syringe pump) and in the other side of the abdomen, insulin was infused in 1-h intervals (by means of a Medix pump). The size of the subcutaneous depots was continuously measured by counting the radioactivity at the infusion sites. After starting the infusions, the two depots were built up to steady-state levels at the same time and of the same size (approximately 3 IU) and with similar absorption rates. Thus, during basal rate insulin infusion, identical insulin absorption kinetics was achieved, irrespective of a 10-fold difference in the pulse rate.

Absorption kinetics and action profiles of subcutaneously administered insulin analogues (AspB9GluB27, AspB10, AspB28) in healthy subjects.

OBJECTIVE: The subcutaneous absorption and resulting changes in plasma insulin or analogue, glucose, C-peptide, and blood intermediary metabolite concentrations after subcutaneous bolus injection of three soluble human insulin analogues (AspB9GluB27, monomeric; AspB28, mixture of monomers and dimers; and AspB10, dimeric) and soluble human insulin were evaluated. RESEARCH DESIGN AND METHODS: Fasting healthy male volunteers (n = 7) were studied on five occasions 1 wk apart randomly receiving 0.6 nmol.kg-1 s.c. 125I-labeled AspB10 or soluble human insulin (Novolin R, Novo, Copenhagen); 1st study and 0.6 nmol.kg-1 s.c. 125I-labeled AspB28, AspB9GluB27 or soluble human insulin (2nd study). Residual radioactivity at the injection site was measured over 8 h with frequent venous sampling for plasma immunoreactive insulin or analogue, glucose, C-peptide, and blood intermediary metabolite concentrations. RESULTS: The three analogues were absorbed 2-3 times faster than human insulin. The mean +/- SE time to 50% residual radioactivity was 94 +/- 6 min for AspB10 compared with 184 +/- 10 min for human insulin (P less than 0.001), 83 +/- 8 min for AspB28 (P less than 0.005), and 63 +/- 9 min for AspB9GluB27 (P less than 0.001) compared with 182 +/- 21 min for human insulin. delta Peak plasma insulin analogue levels were significantly higher after each analogue than after human insulin (P less than 0.005). With all three analogues, the mean hypoglycemic nadir occurred earlier at 61-65 min postinjection compared with 201-210 min for the reference human insulins (P less than 0.005). The magnitude of the hypoglycemic nadir was greater after AspB9GluB27 (P less than 0.05) and AspB28 (P less than 0.001) compared with human insulin. There was a significantly faster onset and offset of responses in C-peptide and intermediary metabolite levels after the analogues than after human insulin (P less than 0.05). CONCLUSIONS: The rapid absorption and biological actions of these analogues offer potential therapeutic advantages over the current short-acting neutral soluble insulins.

Subcutaneous insulin absorption explained by insulin's physicochemical properties. Evidence from absorption studies of soluble human insulin and insulin analogues in humans.

OBJECTIVE: To study the influence of molecular aggregation on rates of subcutaneous insulin absorption and to attempt to elucidate the mechanism of absorption of conventional soluble human insulin in humans. RESEARCH DESIGN AND METHODS: Seven healthy male volunteers aged 22-43 yr and not receiving any drugs comprised the study. This study consisted of a single-blind randomized comparison of equimolar dosages of 125I-labeled forms of soluble hexameric 2 Zn2+ human insulin and human insulin analogues with differing association states at pharmaceutical concentrations (AspB10, dimeric; AspB28, mixture of monomers and dimers; AspB9, GluB27, monomeric). After an overnight fast and a basal period of 1 h, 0.6 nmol/kg of either 125I-labeled human soluble insulin (Actrapid HM U-100) or 125I-labeled analogue was injected subcutaneously on 4 separate days 1 wk apart. Absorption was assessed by measurement of residual radioactivity at the injection site by external gamma-counting. RESULTS: The mean +/- SE initial fractional disappearance rates for the four preparations were 20.7 +/- 1.9 (hexameric soluble human insulin), 44.4 +/- 2.5 (dimeric analogue AspB10), 50.6 +/- 3.9 (analogue AspB28), and 67.4 +/- 7.4%/h (monomeric analogue AspB9, GluB27). Absorption of the dimeric analogue was significantly faster than that of hexameric human insulin (P less than 0.001); absorption of monomeric insulin analogue AspB9, GluB27 was significantly faster than that of dimeric analogue AspB10 (P less than 0.01). There was an inverse linear correlation between association state and the initial fractional disappearance rates (r = -0.98, P less than 0.02). Analysis of the disappearance data on a log linear scale showed that only the monomeric analogue had a monoexponential course throughout. Two phases in the rates of absorption were identified for the dimer and three for hexameric human insulin. The fractional disappearance rates (%/h) calculated by log linear regression analysis were monomer 73.3 +/- 6.8; dimer 44.4 +/- 2.5 from 0 to 2 h and 68.9 +/- 3.5 from 2.5 h onward; and hexameric insulin 20.7 +/- 1.9 from 0 to 2 h, 45.6 +/- 5.0 from 2.5 to 5 h, and 70.6 +/- 6.3 from 5 h onward. CONCLUSIONS: Association state is a major determinant of rates of absorption of insulin and insulin analogues. The lag phase and the subsequent increasing rate of subcutaneous soluble insulin absorption can be explained by the associated state of native insulin in pharmaceutical formulation and its progressive dissociation into smaller units during the absorption process.

Monomeric insulins and their experimental and clinical implications.

Due to the inherent pharmacokinetic properties of available insulins, normoglycemia is rarely, if ever, achieved in insulin-dependent diabetic patients without compromising their quality of life. Subcutaneous insulin absorption is influenced by many factors, among which the associated state of insulin (hexameric) in pharmaceutical formulation may be of importance. This review describes the development of a series of human insulin analogues with reduced tendency to self-association that, because of more rapid absorption, are better suited to meal-related therapy. DNA technology has made it possible to prepare insulins that remain dimeric or even monomeric at high concentration by introducing one or a few amino acid substitutions into human insulin. These analogues were characterized and used for elucidating the mechanisms involved in subcutaneous absorption and were investigated in preliminary clinical studies. Their relative receptor binding and in vitro potency (free-fat cell assay), ranging from 0.05 to 600% relative to human insulin, were strongly correlated (r = 0.97). In vivo, most of the analogues exhibited approximately 100% activity, explainable by a dominating receptor-mediated clearance. This was confirmed by clamp studies in which correlation between receptor binding and clearance was observed. Thus, an analogue with reduced binding and clearance gives higher circulating concentrations, counterbalancing the reduced potency at the cellular level. Absorption studies in pigs revealed a strong inverse correlation (r = 0.96) between the rate of subcutaneous absorption and the mean association state of the insulin analogues. These studies also demonstrated that monomeric insulins were absorbed three times faster than human insulin. In healthy subjects, rates of disappearance from subcutis were two to three times faster for dimeric and monomeric analogues than for human insulin. Concomitantly, a more rapid rise in plasma insulin concentration and an earlier hypoglycemic response with the analogues were observed. The monomeric insulin had no lag phase and followed a monoexponential course throughout the absorption process. In contrast, two phases in rate of absorption were identified for the dimer and three for the normal hexameric human insulin. The initial lag phase and the subsequent accelerated absorption of soluble insulin can now be explained by the associated state of native insulin in pharmaceutical formulation and its progressive dissociation into smaller units during the absorption process. In the light of these results, the effects of insulin concentration, injected volume, temperature, and massage on the absorption process are now also understood.(ABSTRACT TRUNCATED AT 400 WORDS)

Comparison of subcutaneous soluble human insulin and insulin analogues (AspB9, GluB27; AspB10; AspB28) on meal-related plasma glucose excursions in type I diabetic subjects.

OBJECTIVE: To compare postprandial glucose excursions and plasma free insulin-analogue levels after subcutaneous injection of three novel human insulin analogues (AspB10; AspB9, GluB27; and AspB28) with those after injection of soluble human insulin (Actrapid HM U-100). RESEARCH DESIGN AND METHODS: Six male subjects with insulin-dependent diabetes, at least 1 wk apart and after an overnight fast and basal insulin infusion, received 72 nmol (approximately 12 U) s.c. of soluble human insulin 30 min before, or 72 nmol of each of the three analogues immediately before, a standard 500-kcal meal. RESULTS: Mean basal glucoses were similar on the 4 study days. Compared to human insulin (6.3 +/- 0.8 mM), mean +/- SE peak incremental glucose rises were similar after analogues AspB10 (5.4 +/- 0.8 mM) and AspB9, GluB27 (5.4 +/- 0.7 mM) and significantly lower after analogue AspB28 (3.6 +/- 1.2 mM, P less than 0.02). Relative to soluble human insulin (100% +/- SE21), incremental areas under the glucose curve between 0 and 240 min were 79% +/- 34 (AspB10, NS), 70% +/- 29 (AspB9, GluB27, NS), and 43% +/- 23 (AspB28, P less than 0.02). Basal plasma free insulin levels were similar on the 4 study days. Plasma free insulin-analogue levels rose rapidly to peak 30 min after injection at 308 +/- 44 pM (AspB10); 1231 +/- 190 pM (AspB9, GluB27) and 414 +/- 42 pM (AspB28) and were significantly higher than corresponding (i.e., 30 min postmeal) plasma free insulin levels of 157 +/- 15 pM (P less than 0.02 in each case). CONCLUSIONS: Plasma profiles of the insulin analogues were more physiological than that of human insulin after subcutaneous injection. All three analogues given immediately before the meal are at least as effective as soluble human insulin given 30 min earlier. These analogues are promising potential candidates for short-acting insulins of the future.

Iontophoresis of monomeric insulin analogues in vitro: effects of insulin charge and skin pretreatment.

The aim of this study was to investigate the influence of association state and net charge of human insulin analogues on the rate of iontophoretic transport across hairless mouse skin, and the effect of different skin pretreatments on said transport. No insulin flux was observed with anodal delivery probably because of degradation at the Ag/AgCl anode. The flux during cathodal iontophoresis through intact skin was insignificant for human hexameric insulin, and only low and variable fluxes were observed for monomeric insulins. Using stripped skin on the other hand, the fluxes of monomeric insulins with two extra negative charges were 50-100 times higher than that of hexameric human insulin. Introducing three additional charges led to a further 2-3-fold increase in flux. Wiping the skin gently with absolute alcohol prior to iontophoresis resulted in a 1000-fold increase in transdermal transport of insulin relative to that across untreated skin, i.e. to almost the same level as stripping the skin. The alcohol pretreatment reduced the electrical resistance of the skin, presumably by lipid extraction. In conclusion, monomeric insulin analogues with at least two extra negative charges can be iontophoretically delivered across hairless mouse skin, whereas insignificant flux is observed with human, hexameric insulin. Wiping the skin with absolute alcohol prior to iontophoresis gave substantially improved transdermal transport of monomeric insulins resulting in clinically relevant delivery rates for basal treatment.

Design of insulin analogues for meal-related therapy.

The human insulin in replacement therapy has a hexameric structure. Hexamerization of the insulin molecule facilitates biosynthesis and beta-cell storage of insulin, but is unnecessary for biologic activity and appears to contribute to delayed absorption of exogenous insulin from the subcutis. Insulin analogues with reduced self-association that are produced through recombinant DNA techniques have been shown to have in vivo activity comparable to that of human insulin and absorption kinetics characterized by higher and more constant rates of disappearance from the subcutaneous injection site. In preliminary studies in patients receiving insulin therapy, monomeric insulin analogues have been found to provide glycemic control in the postprandial period that is at least equivalent to that of human insulin. Findings in these studies suggest that the use of such analogues may provide meal-related insulin effects closer to those observed in the physiologic state by limiting excessive postprandial glucose excursions and decreasing the risk of late hypoglycemia. Banting and Best revolutionized diabetes therapy 70 years ago with the extraction of insulin from animal pancreas glands (J Lab Clin Med 7:464-472, 1922). Since that time, many refinements of the therapeutic properties of pharmaceutical preparations of the hormone have been introduced. Until recently, however, such advances have been limited to improvements in insulin purity, insulin species, and adjustment of the composition of the vehicle with respect to auxiliary substances and other additives. With the advent of recombinant DNA techniques, it has become possible to optimize the insulin molecule itself for purposes of replacement therapy.(ABSTRACT TRUNCATED AT 250 WORDS)

Toward understanding insulin fibrillation.

Formation of insulin fibrils is a physical process by which partially unfolded insulin molecules interact with each other to form linear aggregates. Shielding of hydrophobic domains is the main driving force for this process, but formation of intermolecular beta-sheet may further stabilize the fibrillar structure. Conformational displacement of the B-chain C-terminal with exposure of nonpolar, aliphatic core residues, including A2, A3, B11, and B15, plays a crucial role in the fibrillation process. Recent crystal analyses and molecular modeling studies have suggested that when insulin fibrillates this exposed domain interacts with a hydrophobic surface domain formed by the aliphatic residues A13, B6, B14, B17, and B18, normally buried when three insulin dimers form a hexamer. In rabbit immunization experiments, insulin fibrils did not elicit an increased immune response with respect to formation of IgG insulin antibodies when compared with native insulin. In contrast, the IgE response increased with increasing content of insulin in fibrillar form. Strategies and practical approaches to prevent insulin from forming fibrils are reviewed. Stabilization of the insulin hexameric structure and blockage of hydrophobic interfaces by addition of surfactants are the most effective means of counteracting insulin fibrillation.

Studies of the structure of insulin fibrils by Fourier transform infrared (FTIR) spectroscopy and electron microscopy.

Fibril formation (aggregation) of insulin was investigated in acid media by visual inspection, transmission electron microscopy (TEM), and Fourier transform infrared (FTIR) spectroscopy. Insulin fibrillated faster in hydrochloric acid than in acetic acid at elevated temperatures, whereas the fibrillation tendencies were reversed at ambient temperatures. Electron micrographs showed that bovine insulin fibrils consisted of long fibers with a diameter of 5 to 10 nm and lengths of several microns. The fibrils appeared either as helical filaments (in hydrochloric acid) or arranged laterally in bundles (in acetic acid, NaCl). Freeze-thawing cycles broke the fibrils into shorter segments. FTIR spectroscopy showed that the native secondary structure of insulin was identical in hydrochloric acid and acetic acid, whereas the secondary structure of fibrils formed in hydrochloric acid was different from that formed in acetic acid. Fibrils of bovine insulin prepared by heating or agitating an acid solution of insulin showed an increased content of beta-sheet (mostly intermolecular) and a decrease in the intensity of the alpha-helix band. In hydrochloric acid, the frequencies of the beta-sheet bands depended on whether the fibrillation was induced by heating or agitation. This difference was not seen in acetic acid. Freeze-thawing cycles of the fibrils in hydrochloric acid caused an increase in the intensity of the band at 1635 cm(-1) concomitant with reduction of the band at 1622 cm(-1). The results showed that the structure of insulin fibrils is highly dependent on the composition of the acid media and on the treatment.

Recombinant DNA derived monomeric insulin analogue: comparison with soluble human insulin in normal subjects.

OBJECTIVE: To compare the rate of absorption from subcutaneous tissue and the resulting hypoglycaemic effect of iodine-125 labelled soluble human insulin and a monomeric insulin analogue derived by recombinant DNA technology. DESIGN: Single blind randomised comparison of equimolar doses of 125I labelled soluble human insulin and insulin analogue. SETTING: Study in normal people at a diabetes research unit and a university department of medical physics. SUBJECTS: Seven healthy male volunteers aged 20-39 not receiving any other drugs. INTERVENTIONS: After an overnight fast and a basal period of one hour two doses (0.05 and 0.1 U/kg) of 125I labelled soluble human insulin and insulin analogue were injected subcutaneously into the anterior abdominal wall on four separate days. END POINT: To find a fast acting insulin for meal related requirements in insulin dependent diabetics. MEASUREMENTS and main results--Residual radioactivity at the injection site was measured continuously for the first two hours after injection of the 125I labelled preparations and thereafter for five minutes simultaneously with blood sampling. Frequent venous blood samples were obtained over six hours for determination of plasma immunoreactive insulin, insulin analogue, glucose, and glucagon values. Time to 50% of initial radioactivity at the injection site for the insulin analogue compared with soluble insulin was 61 v 135 minutes (p less than 0.05) with 0.05 U/kg and 67 v 145 minutes (p less than 0.001) with 0.1 U/kg. Concentrations in plasma increased faster after the insulin analogue compared with soluble insulin, resulting in higher plasma concentrations between 10 and 150 minutes (0.001 less than p less than 0.05) after 0.05 U/kg and between 40 and 360 minutes (0.001 less than p less than 0.05) after 0.1 U/kg. The hypoglycaemic response to insulin analogue was a plasma glucose nadir at 60 minutes with both doses compared with 90 and 120 minutes with soluble insulin at 0.5 and 0.1 U/kg respectively. The response of glucagon substantiated the earlier and more dramatic hypoglycaemic effect with the insulin analogue. CONCLUSIONS: The much faster absorption from subcutaneous tissue of the disubstituted monomeric insulin analogue compared with soluble insulin suggests that the analogue may be a potential candidate for rapid insulin delivery after subcutaneous bolus injection.

The new era of biotech insulin analogues.

Many of the structural properties of insulin have evolved in response to the requirements of biosynthesis, processing, transport and storage in the pancreatic beta cells, properties that are not necessary for the biological action of the hormone. It is therefore not surprising that wild-type insulin has far from optimal characteristics for replacement therapy. For example, native human insulin self-associates to hexameric units, which limits the possibilities for the absorption of the molecule by various routes. During the last decade new techniques of molecular design have emerged and recombinant DNA technology offers new and exciting opportunities for rational protein drug design. This review describes examples of recent advances in insulin engineering aimed at optimizing the hormone for therapy. Such approaches focus on improvements in the pharmacokinetic properties, storage stability, and feasibility for less intrusive routes of administration.

Chemical stability of insulin. 4. Mechanisms and kinetics of chemical transformations in pharmaceutical formulation.

Insulin decomposes by a multitude of chemical reactions [1-3]. It deamidates at two different residues by entirely different mechanisms. In acid, deamidation at AsnA21 is intramolecularly catalyzed by the protonated C-terminal, whereas above pH 6 an intermediate imide formation at residue AsnB3 leads to isoAsp and Asp derivatives. The imide formation requires a large rotation around the alpha-carbon/peptide carbonyl carbon bond at B3, corresponding to a 10 A movement of the B-chain N-terminal. The main determinant for the rate of B3 deamidation, as well as for the ratio between the two products formed, is the local conformational structure, which is highly influenced by various excipients and the physical state of the insulin. An amazing thermolysin-like, autoproteolytic cleavage of the A-chain takes place in rhombohedral insulin crystals, mediated by a concerted catalytic action by several, inter-hexameric functional groups and Zn2+. Intermolecular, covalent cross-linking of insulin molecules occurs via several mechanisms. The most prominent type of mechanism is aminolysis by the N-terminals, leading to isopeptide linkages with the A-chain side-chain amides of residues GlnA15, AsnA18 and AsnA21. The same type of reaction also leads to covalent cross-linking of the N-terminal in protamine with insulin. Disulfide exchange reactions, initiated by lysis of the A7-B7 disulfide bridge, lead mainly to formation of covalent oligo- and polymers. Activation energy (Ea) for the neutral deamidation and the aminolysis reactions was found to be 80 and 119 KJ/mol, respectively.

Insulin pumps and insulin quality--requirements and problems.

In developing insulin solution suitable for delivery devices the chemical and biological stability, as well as the physical stability, must be taken into consideration. Addition of certain mono- and disaccharides increases the physical stability of neutral insulin solutions, but concurrently the chemical and biological stability decrease to an unacceptable degree. Addition of Ca-ions in low concentrations offers a physiologically acceptable method for stabilizing neutral insulin solutions against heat precipitation without affecting the quality, including the chemical and biological stability.

Chemical stability of insulin. 3. Influence of excipients, formulation, and pH.

The influence of auxiliary substances and pH on the chemical transformations of insulin in pharmaceutical formulation, including various hydrolytic and intermolecular cross-linking reactions, was studied. Bacteriostatic agents had a profound stabilizing effect--phenol > m-cresol > methylparaben--on deamidation as well as on insulin intermolecular cross-linking reactions. Of the isotonicity substances, NaCl generally had a stabilizing effect whereas glycerol and glucose led to increased chemical deterioration. Phenol and sodium chloride exerted their stabilizing effect through independent mechanisms. Zinc ions, in concentrations that promote association of insulin into hexamers, increase the stability, whereas higher zinc content had no further influence. Protamine gave rise to additional formation of covalent protamine-insulin products which increased with increasing protamine concentration. The impact of excipients on the chemical processes seems to be dictated mainly via an influence on the three-dimensional insulin structure. The effect of the physical state of the insulin on the chemical stability was also complex, suggesting an intricate dependence of intermolecular proximity of involved functional groups. At pH values below five and above eight, insulin degrades relatively fast. At acid pH, deamidation at residue A21 and covalent insulin dimerization dominates, whereas disulfide reactions leading to covalent polymerization and formation of A- and B-chains prevailed in alkaline medium. Structure-reactivity relationship is proposed to be a main determinant for the chemical transformation of insulin.

Insulin structure and stability.

Insulin is composed of 51 amino acids in two peptide chains (A and B) linked by two disulfide bonds. The three-dimensional structure of the insulin molecule (insulin monomer), essentially the same in solution and in solid phase, exists in two main conformations. These differ in the extent of helix in the B chain which is governed by the presence of phenol or its derivatives. In acid and neutral solutions, in concentrations relevant for pharmaceutical formulation, the insulin monomer assembles to dimers and at neutral pH, in the presence of zinc ions, further to hexamers. Many crystalline modifications of insulin have been identified but only those with the hexamer as the basic unit are utilized in preparations for therapy. The insulin hexamer forms a relatively stable unit but some flexibility remains within the individual molecules. The intrinsic flexibility at the ends of the B chain plays an important role in governing the physical and chemical stability of insulin. A variety of chemical changes of the primary structure (yielding insulin derivatives), and physical modifications of the secondary to quaternary structures (resulting in "denaturation," aggregation, and precipitation) are known to affect insulin and insulin preparations during storage and use (Fig. 8). The tendency of insulin to undergo structural transformation resulting in aggregation and formation of insoluble insulin fibrils has been one of the most intriguing and widely studied phenomena in relation to insulin stability. Although the exact mechanism of fibril formation is still obscure, it is now clear that the initial step is an exposure of certain hydrophobic residues, normally buried in the three-dimensional structure, to the surface of the insulin monomer. This requires displacement of the COOH-terminal B-chain residues from their normal position which can only be accomplished via monomerization of the insulin. Therefore, most methods stabilizing insulin against fibrillation share the property of being able to counteract associated insulin from being disassembled. Chemical deterioration of insulin during storage of pharmaceutical preparations is mainly due to two categories of chemical reactions, hydrolysis and intermolecular transformation reactions leading to insulin HMWT products. The predominant hydrolysis reaction is deamidation of Asn residues which in acid solution takes place at residue A21, in neutral medium at residue B3. An amazing hydrolytic cleavage of the backbone A chain, presumably autocatalyzed by an adjacent insulin molecule, has been identified in insulin preparations containing rhombohedral crystals in combination with free zinc ions.(ABSTRACT TRUNCATED AT 400 WORDS)

Chemical stability of insulin. 2. Formation of higher molecular weight transformation products during storage of pharmaceutical preparations.

Formation of covalent, higher molecular weight transformation (HMWT) products during storage of insulin preparations at 4-45 degrees C was studied by size exclusion chromatography. The main products are covalent insulin dimers (CID), but in protamine-containing preparations the concurrent formation of covalent insulin-protamine (CIP) products takes place. At temperatures greater than or equal to 25 degrees C parallel or consecutive formation of covalent oligo- and polymers can also be observed. Rate of HMWT is only slightly influenced by species of insulin but varies with composition and formulation, and for isophane (NPH) preparations, also with the strength of preparation. Temperature has a pronounced effect on CID, CIP, and, especially, covalent oligo- and polymer formation. The CIDs are apparently formed between molecules within the hexameric unit common for all types of preparations and rate of formation is generally faster in glycerol-containing preparations. Compared with insulin hydrolysis reactions (see the preceding paper), HMWT is one order of magnitude slower, except for NPH preparations.

Chemical stability of insulin. 5. Isolation, characterization and identification of insulin transformation products.

During storage of insulin formulated for therapy, minor amounts of various degradation and covalent di- and polymerization products are formed [1-3]. The main chemical transformation products were isolated from aged preparations and characterized chemically and biologically. The most prominent products formed in neutral medium were identified as a mixture of deamidation products hydrolyzed at residue B3, namely isoAsp B3 and Asp B3 derivatives. A hydrolysis product formed only in crystals of insulin zinc suspensions containing a surplus of zinc ions in the supernatant was identified as an A8-A9 cleavage product. The small amounts of covalent insulin dimers (CID) formed in all formulations were shown to be a heterogenous mixture of 5-6 different CIDs with a composition dependent on the pharmaceutical formulation. The chemical characteristics of the CIDs indicate that they are formed through a transamidation reaction mainly between the B-chain N-terminal and one of the four amide side-chains of the A chain. GlnA15, AsnA18 and, in particular, AsnA21 participate in the formation of such isopeptide links between two insulin molecules. The covalent insulin-protamine products (CIPP) formed during storage of NPH preparations presumably originate from a similar reaction between the protamine N-terminal with an amide in insulin. Covalent polymerization products, mainly formed during storage of amorphously suspended insulin at higher temperature, were shown to be due to disulfide interactions. Biological in vivo potencies relative to native insulin were less than 2% for the split-(A8-A9)-product and for the covalent disulfide exchange polymers, 4% for the CIPP, approximately 15% for the CIDs, whereas the B3 derivatives exhibited full potency. Rabbit immunization experiments revealed that none of the insulin transformation products had significantly increased immunogenicity in rabbits.