AU2013237740B2 - Insulin analogues containing penta-fluora-phenyalanine at position B24 - Google Patents

Abstract

Abstract of the disclosure An insulin analogue comprises a B-chain polypeptide incorporating a halogenated phenylalanine at position B24, B25 or B26. The halogenated phenylalanine may be ortho-monofluoro-phenylalanine, ortho-monobromo-phenylalanine, ortho-monochloro-phenylalanine, or para-monochloro phenylalanine or penta-fluoro-Phenylalanine at position B24 alone or in combination with an acidic substitution (Aspartic Acid or Glutamic Acid) at position B 10. The analogue may be of a mammalian insulin, such as human insulin. A nucleic acid encodes such an insulin analogue. The halogenated insulin analogues retain significant activity. A method of treating a patient comprises administering a physiologically effective amount of the insulin analogue or a physiologically acceptable salt thereof to a patient. Halogen substitution-based stabilization of insulin may enhance the treatment of diabetes mellitus in regions of the developing world lacking refrigeration. PRIOR ART CcC

Description

INSULIN ANALOGUES CONTAINING ΡΕΝΤΑ -FLOURO-PHENYALANINE AT POSITON B24

Statement regarding federally sponsored research or development [0001] This invention was made with government support under cooperative agreements awarded by the United States National Institutes of Health under grant numbers DK40949 and DK074176. The U.S. government may have certain rights to the invention.

Background of the invention [0002] This invention relates to polypeptide hormone analogues that exhibits enhanced pharmaceutical properties, such as increased thermodynamic stability, augmented resistance to thermal fibrillation above room temperature, and decreased mitogenicity. More particularly, this invention relates to insulin analogues that are modified by substitution of non-standard Phenylalanine at position 24 of the B chain (designated herein position B24). Such non-standard sequences may optionally contain standard amino-acid substitutions at other sites in the A or B chains of an insulin analogue. Even more particularly, this invention relates to insulin analogues that are characterized by the incorporation of the element fluorine into phynylalanine at position B24 of the insulin analogue. Fluorine is distinguished from the ordinary constituents of proteins by its atomic radius, electronegativity, stereoelectronic distribution of partial charges, transmitted effects on the stereoelectronic properties of neighboring atoms in a molecule.

[0003] The engineering of non-standard proteins, including therapeutic agents and vaccines, may have broad medical and societal benefits. Naturally occurring proteins—as encoded in the genomes of human beings, other mammals, vertebrate organisms, invertebrate organisms, or eukaryotic cells in general—often confer multiple biological activities. A benefit of non-standard proteins would be to achieve selective activity, such as action in one organ, tissue, or cell type and not in another. Yet another benefit would be decreased risk of an unintended and unfavorable side effect, such as promotion of the growth of cancer cells. An example of a therapeutic protein is provided by insulin. Wild-type human insulin and insulin molecules encoded in the genomes of other mammals bind to insulin receptors is multiple organs and diverse types of cells, irrespective of the receptor isoform generated by alternative modes of RNA splicing or by alternative patterns of post-translational glycosylation. Such ubiquitous binding leads to opposing biological signals, such as an anabolic effect of insulin signaling in the liver, muscle, and fat but a catabolic effect of insulin signaling in the mammalian brain as mediated by specific insulin-stimulated and insulin-modulated circuits in the hypothalamus.

[0004] An example of a further medical benefit would be optimization of the stability of a protein toward unfolding or degradation. Such a societal benefit would be enhanced by the engineering of proteins more refractory than standard proteins with respect to degradation at or above room temperature for use in regions of the developing world where electricity and refrigeration are not consistently available. Analogues of insulin containing non-standard amino-acid substitutions may in principle exhibit superior properties with respect to resistance to thermal degradation or mitogenicity. The challenge posed by its physical degradation is deepened by the pending epidemic of diabetes mellitus in Africa and Asia. Because fibrillation poses the major route of degradation above room temperature, the design of fibrillation-resistant formulations may enhance the safety and efficacy of insulin replacement therapy in such challenged regions.

[0005] The utility of some halogen substitutions in amino acids is well established in medicinal chemistry. Indeed, fluorinated functional groups are critical to the efficacy of such widely prescribed small molecules as atorvastatin (Liptor™), an inhibitor of cholesterol biosynthesis, and fluoxetine hydrochloride (Prozac™), a selective serotonin reuptake inhibitor used in the treatment of depression and other affective disorders. Although the atomic radius of fluorine is similar to that of hydrogen, its large inductive effects modify the stereo-electronic properties of these drugs, in turn enhancing their biological activities. Such observations have motivated the study of fluorinated amino acids in proteins. Similar considerations of physical organic chemistry pertain to the incorporation of larger halogen atoms, such as chlorine and bromine. The small molecule montelukast sodium (Singulair™) is a leukotriene inhibitor whose pharmaceutical properties are enhanced by covalent incorporation of a chlorine atom.

[0006] Attention has previously focused on the use of multiply fluorinated aliphatic side chains (such as trifluoro-y-CFs-Val, trifluoro-6-CF3-Val, trifluoro- 6-CF3-Ile, hexafluoro-Yi.2-CF3-Val, and hexafluoro-6|.2-CF3-Leu) to maximize the gain in hydrophobicity associated with this modification. An example is provided by the stabilization of a model α-helical motif, the homodimeric coiled coil. Its interfacial aliphatic chains were simultaneously substituted by trifluoro-analogs, creating a fluorous core whose stability is enhanced by 0.3 - 2.0 kcal/mole. The degree of stabilization per fluorine atom is < 0.1 kcal/mole. More marked stabilization per fluorine atom has been achieved in an unrelated α-helical domain by substitution of a single internal Phe by pentafluoro-Phe (F^-Phe)1 (AAGU 0.6 kcal/mole per five fluorine atoms). Stabilization occurs only at one specific position in the protein, suggesting that its mechanism requires a particular spatial environment. The structure of the Fj-Phe-modified domain is identical to that of the unmodified domain. Structural stabilization however, is only a portion of the requirements for a biologically active polypeptide. At least a significant portion of the activity must also be maintained.

[0007] An extensive literature describes the use of fluorine labels in proteins as 19F-NMR probe. Whereas such labels are widely regarded as non-perturbing, applications of fluorinated amino acids in protein engineering seek to exploit their altered physicochemical properties. Studies of a model α-helical fold (the villin headpiece subdomain) containing single F^-Phe substitutions have demonstrated that whether and to what extent such a modification may affect protein stability depends in detail on structural environment. Indeed, the stability of this model fold was enhanced by an F^-Phe substitution only at one of the seven sites tested in the core. It should be noted however, that this stabilizing effect was only demonstrated in a nonstandard polypeptide analogue containing a sulfur atom in the main chain near the site of fluorination; in our hands the self-same Fs-Phe substitution in villin headpiece subdomain with native polypeptide main chain has no effect on its stability. Thus, to our knowledge, bona fide data demonstrating enhancement of protein stability due to halogenation of an aromatic residue in an otherwise native protein has not previously been reported. These observations suggest that a generally hydrophobic environment does not in itself assure that the modification will be stabilizing. Because protein interiors are often stabilized by aromatic-aromatic interactions with a specific distance- and angular dependence, a subset of stabilizing Fs-Phe substitutions may adopt particularly favorable perfluroaryl/aryl geometries. Such interactions arise from the asymmetric distribution of partial charges in these aromatic systems. Modulation of the chemical, physical, and biological properties of proteins by the site-specific incorporation of chlorine or bromine atoms into modified amino acids are less well characterized in the scientific literature than are the above effects of incorporation of fluorine atoms.

[0008] Aromatic side chains may engage in a variety of weakly polar interactions, involving not only neighboring aromatic rings but also other sources of positive- or negative electrostatic potential. Examples include main-chain carbonyl- and amide groups in peptide bonds.

[0009] Administration of insulin has long been established as a treatment for diabetes mellitus. Insulin is a small globular protein that plays a central role in metabolism in vertebrates. Insulin contains two chains, an A chain, containing 21 residues and a B chain containing 30 residues. The hormone is stored in the pancreatic β-cell as a Zn2+-stabilized hexamer, but functions as a Zn2+-free monomer in the bloodstream. Insulin is the product of a single-chain precursor, proinsulin, in which a connecting region (35 residues) links the C-terminal residue of B chain (residue B30) to the N-terminal residue of the A chain (Fig. 1A). Although the structure of proinsulin has not been determined, a variety of evidence indicates that it consists of an insulin-like core and disordered connecting peptide (Fig. IB). Formation of three specific disulfide bridges (A6-A11, A7-B7, and A20-B19; Figs. 1A and IB) is thought to be coupled to oxidative folding of proinsulin in the rough endoplasmic reticulum (ER). Proinsulin assembles to form soluble Zn2+-coordinated hexamers shortly after export from ER to the Golgi apparatus. Endoproteolytic digestion and conversion to insulin occurs in immature secretory granules followed by morphological condensation. Crystalline arrays of zinc insulin hexamers within mature storage granules have been visualized by electron microscopy (EM). The sequence of insulin is shown in schematic form in Figure 1C. Individual residues are indicated by the identity of the amino acid (typically using a standard three-letter code), the chain and sequence position (typically as a superscript). Pertinent to the present invention is the designation of individual hydrogen atoms attached to carbon atoms in the aromatic ring of Phenylalanine, designated ring positions 2-6.

[0010] Insulin functions in the bloodstream as a monomer, and yet it is the monomer that is believed to be most susceptible to fibrillation and most forms of chemical degradation. The structure of an insulin monomer, characterized in solution by NMR, is shown in Figure ID. The A-chain consists of an N-terminal α-helix (residues A1-A8), non-canonical turn (A9-A12), second a-helix (A12-A18), and C-terminal extension (A19-A21). The B chain contains an N-terminal arm (B1-B6), β-turn (B7-B10), central α-helix (B9-B19), β-turn (B20-B23), β-strand (B24-B28), and flexible C-terminal residues B29-B30. The two chains pack to form a compact globular domain stabilized by three disulfide bridges (cystines A6-A11, A7-B7, and A20-B19).

[0011] Absorption of regular insulin is limited by the kinetic lifetime of the Zn-insulin hexamer, whose disassembly to smaller dimers and monomers is required to enable transit through the endothelial lining of capillaries. The essential idea underlying the design of Humalog® and Novolog® is to accelerate disassembly. This is accomplished by destabilization of the classical dimer-forming surface (the C-terminal anti-parallel β-sheet). Humalog® contains substitutions ProB28—"Lys and LysB29—>Pro, an inversion that mimics the sequence of IGF-I. Novolog® contains the substitution ProB28—>Asp. Although the substitutions impair dimerization, the analogs are competent for assembly of a phenol- or meta-cresol-stabilized zinc hexamer. This assembly protects the analog from fibrillation in the vial, but following subcutaneous injection, the hexamer rapidly dissociates as the phenol (or m-cresol) and zinc ions diffuse away. The instability of these analogs underlies their reduced shelf life on dilution by the patient or health-care provider. It would be useful for an insulin analogue to augment the intrinsic stability of the insulin monomer while retaining the variant dimer-related β-sheet of Humalog®.

[0012] Use of zinc insulin hexamers during storage is known and represents a classical strategy to retard physical degradation and chemical degradation of a formulation in the vial or in the reservoir of a pump. Because the zinc insulin hexamer is too large for immediate passage into capillaries, the rate of absorption of insulin after subcutaneous injection is limited by the time required for dissociation of hexamers into smaller dimers and monomer units. Therefore, it would advantageous for an insulin analogue to be both (a) competent to permit hexamer assembly at high protein concentration (as in a vial or pump) and yet (b) sufficiently destabilized at the dimer interface to exhibit accelerated disassembly - hence predicting ultra-rapid absorption from the subcutaneous depot. These structural goals walk a fine line between stability (during storage) and instability (following injection).

[0013] Amino-acid substitutions in insulin have been investigated for effects on thermodynamic stability and biological activity. No consistent relationship has been observed between stability and activity. Whereas some substitutions that enhance thermodynamic stability also enhance binding to the insulin receptor, other substitutions that enhance stability impede such binding. The effects of

AO substitution of Thr by several other amino acids has been investigated in wild-type human insulin and in the context of an engineered insulin monomer containing three unrelated substitutions in the B-chain (His —>Asp, Pro —>Lys, and Lys —>Pro) have been reported. Examples are also known in the art of substitutions that accelerate or delay the time course of absorption. Such

πλο R9R R9Q substitutions (such as Asp in Novalog® and [Lys , Pro ] in Humalog®) can be and often are associated with more rapid fibrillation and poorer physical stability. Indeed, a series of ten analogs of human insulin have been tested for susceptibility to fibrillation, including Asp -insulin and Asp -insulin. All ten were found to be more susceptible to fibrillation at pH 7.4 and 37 °C than is human insulin. The ten substitutions were located at diverse sites in the insulin molecule and are likely to be associated with a wide variation of changes in classical thermodynamic stability. Although a range of effects has been observed, no correlation exists between activity and thermodynamic stability.

[0014] Insulin is a small globular protein that is highly amenable to chemical synthesis and semi-synthesis, which facilitates the incorporation of nonstandard side chains. Insulin contains three phenylalanine residues (positions Bl, B24, and B25). Conserved among vertebrate insulins and insulin-like growth factors, the aromatic ring of Phe packs against (but not within) the hydrophobic core to stabilize the super-secondary structure of the B-chain. Phe lies at the classical receptor-binding surface and has been proposed to direct a change in conformation on receptor binding.

[0015] The present theory of protein fibrillation posits that the mechanism of fibrillation proceeds via a partially folded intermediate state, which in turn aggregates to form an amyloidogenic nucleus. In this theory, it is possible that amino-acid substitutions that stabilize the native state may or may not stabilize the partially folded intermediate state and may or may not increase (or decrease) the free-energy barrier between the native state and the intermediate state. Therefore, the current theory indicates that the tendency of a given amino-acid substitution in the insulin molecule to increase or decrease the risk of fibrillation is highly unpredictable.

[0016] Fibrillation, which is a serious concern in the manufacture, storage and use of insulin and insulin analogues for diabetes treatment, is enhanced with higher temperature, lower pH, agitation, or the presence of urea, guanidine, ethanol co-solvent, or hydrophobic surfaces. Current US drug regulations demand that insulin be discarded if fibrillation occurs at a level of one percent or more. Because fibrillation is enhanced at higher temperatures, diabetic individuals optimally must keep insulin refrigerated prior to use. Fibrillation of insulin or an insulin analogue can be a particular concern for diabetic patients utilizing an external insulin pump, in which small amounts of insulin or insulin analogue are injected into the patient’s body at regular intervals. In such a usage, the insulin or insulin analogue is not kept refrigerated within the pump apparatus and fibrillation of insulin can result in blockage of the catheter used to inject insulin or insulin analogue into the body, potentially resulting in unpredictable blood glucose level fluctuations or even dangerous hyperglycemia. At least one recent report has indicated that lispro insulin (an analogue in which residues B28 and B29 are interchanged relative to their positions in wild-type human insulin; trade name Humalog®) may be particularly susceptible to fibrillation and resulting obstruction of insulin pump catheters. Insulin exhibits an increase in degradation rate of 10-fold or more for each 10° C increment in temperature above 25° C; accordingly, guidelines call for storage at temperatures < 30° C and preferably with refrigeration. The propensity of current products Humalog®, Novalog®, and Apidra® to form fibrils at or above room temperature is exacerbated on dilution, as may be used in the treatment of Type I diabetes mellitus in children or adults with low body mass. Accordingly, the shelf life of such diluted pharmaceutical formulations in use at room temperature is reduced in accordance with instructions contained in package inserts regarding disposal of diluted insulin analogue formulations. Fibrillation of basal insulin analogues formulated as soluble solutions at pH less than 5 (such as Lantus® (Sanofi-Aventis), which contains an unbuffered solution of insulin glargine and zinc ions at pH 4.0) also can limit their shelf lives due to physical degradation at or above room temperature; the acidic conditions employed in such formulations impairs insulin self-assembly and weakens the binding of zinc ions, reducing the extent to which the insulin analogues can be protected by sequestration within zinc-protein assemblies.

[0017] Insulin fibrillation is an even greater concern in implantable insulin pumps, where the insulin would be contained within the implant for 1-3 months at high concentration and at physiological temperature (i.e. 37 °C), rather than at ambient temperature as with an external pump. Additionally, the agitation caused by normal movement would also tend to accelerate fibrillation of insulin. In spite of the increased potential for insulin fibrillation, implantable insulin pumps are still the subject of research efforts, due to the potential advantages of such systems. These advantages include intraperitoneal delivery of insulin to the portal circulatory system, which mimics normal physiological delivery of insulin more closely than subcutaneous injection, which provides insulin to the patient via the systemic circulatory system. Intraperitoneal delivery provides more rapid and consistent absorption of insulin compared to subcutaneous injection, which can provide variable absorption and degradation from one injection site to another. Administration of insulin via an implantable pump also potentially provides increased patient convenience. Whereas efforts to prevent fibrillation, such as by addition of a surfactant to the reservoir, have provided some improvement, these improvements have heretofore been considered insufficient to allow reliable usage of an implanted insulin pump in diabetic patients outside of strictly monitored clinical trials.

As noted above, the developing world faces a challenge regarding the safe storage, delivery, and use of drugs and vaccines. This challenge complicates the use of temperature-sensitive insulin formulations in regions of Africa and Asia lacking consistent access to electricity and refrigeration, a challenge likely to be deepened by the pending epidemic of diabetes in the developing world. Insulin exhibits an increase in degradation rate of 10-fold or more for each 10 °C increment in temperature above 25 °C, and guidelines call for storage at temperatures < 30 °C and preferably with refrigeration. At higher temperatures insulin undergoes both chemical degradation (changes in covalent structure such as formation of iso-aspartic acid, rearrangement of disulfide bridges, and formation of covalent polymers) and physical degradation (non-native aggregation and fibrillation. A major goal of conventional insulin replacement therapy in patients with diabetes mellitus is tight control of the blood glucose concentration to prevent its excursion above or below the normal range characteristic of healthy human subjects. Excursions below the normal range are associated with immediate adrenergic or neuroglycopenic symptoms, which in severe episodes lead to convulsions, coma, and death. Excursions above the normal range are associated with increased long-term risk of microvascular disease, including retinapathy, blindness, and renal failure. The classical paradigm of insulin action has focused on organ-specific functions of adipocytes (where insulin regulates storage of fuels in the form of tryglyceride droplets), the liver (where insulin regulates the production of glucose via gluconeogenesis and regulates the storage of fuel in the form of glycogen) and muscle (where insulin regulates the influx of glucose from the bloodstream via trafficking to the plasma membrane of GLUT4) as the target tissues of the hormone. Recent research has revealed, however, that insulin has physiological roles in other organs and tissues, such as in the hypothalamus of the brain, wherein insulin-responsive neural circuitry influences hepatic metabolism, appetite, satiety, and possibly the set point for ideal body weight. In addition to the brain as a site of hormonal regulation of overall metabolism and feeding behaviors, resistance of the brain to the action of insulin is proposed to contribute to the decline of cognitive function in Alzheimer’s Disease and may be a molecular mechanism contributing to its pathogenesis. Although the human genome contains a single gene encoding the insulin receptor, a transmembrane protein containing a cytoplasmic tyrosine-kinase domain, its pre-messanger RNA undergoes alterative splicing to yield distinct A and B isoforms, whose fractional distribution may differ from organ to organ and whose signaling functions may differ within the same cells. The A and B isoforms (designated IR-A and IR-B) differ in affinity for insulin, and only IR-A (lacking a peptide domain in the alpha subunit encoded by exon 11) binds IGF-II with high affinity. Insulin analogues have been described by investigators at Novo-Nordisk that differ from wild-type insulin in the ratio of respective affinities for IR-A and IR-B. In addition, the insulin receptor undergoes a complex pattern of N- and O-linked glycosylation, which may vary from organ to organ or from tissue to tissue; it is not known whether such potential variation can influence the relative binding affinities and signaling potencies of insulin analogues. A further source of organ- or cell-type-specific variation in response to insulin or an insulin analogue could be the co-expression of other membrane proteins or receptors that structurally or functionally interact with the insulin receptor to modulate its signaling effects; such an interacting receptor has been described in the liver due to functional interactions between IR and the MET proto-oncoprotein, the receptor for hepatic growth factor. It is also possible that kinetic and thermodynamic features of how insulin binds to the insulin receptor affect the extent and efficiency of insulin signaling differently in different organs. Together, the complex biochemistry and cell biology of the insulin receptor make it difficult to predict the physiological effects of an insulin analogue based solely on its receptor-binding affinity in vitro.

The relative contributions of organ-specific insulin pathways may differ among mammals. In rodents, for example, insulin regulates hepatic gluconeogenesis in large part by an indirect mechanism involving insulin-responsive neural circuits in the hypothalamus, which modulate vagal inputs to the liver. By contrast, in dogs the major mechanism by which insulin regulates hepatic gluconeogenesis is through the binding of insulin to insulin receptors expressed on the surface of hepatocytes. Whereas the amino-acid sequence of the rodent insulin receptor in rodents is almost identical to the amino-acid sequence of the canine insulin receptor, their physiological mode of regulation of the liver (direct or indirect) is not equivalent. In both animals insulin is nonetheless thought to function in the hypothalamus to regulate appetite and satiety in conjunction with leptin, melanocortin, and neuropeptide Y. Comparison of the regulatory pathways of insulin in these two mammals raises the possibility that a putative organic-specific insulin analogue, should such a molecule be found, could exhibit different integrated hypoglycemic potencies in different animals and different integrated hypoglycemic potencies in the same animal depending on its physiological status in the period following subcutaneous injection (i.e., depending on the extent to which hepatic glucose production and output is contributing to the maintenance or elevation of blood glucose concentration).

[0018] Non-classical functions of insulin extrinsic to adipose tissue, liver, and muscle may have beneficial or undesired effects. An example of a beneficial effect is provided by the action of insulin in the brain to suppress appetite. An example of an undesired effect of insulin is its ability to stimulate the proliferation of cancer cells. Wild-type insulin exhibits mitogenic activity, especially at the high concentrations needed to treat patients with diabetes mellitus who also exhibit marked insulin resistance. Indeed, hyperinsulinemia (whether endogenous due to beta-cell hypersecretion or as a consequence of insulin replacement therapy) is associated on an epidemiological level with an increased risk of several common tumors, including breast cancer and colon cancer. Although molecular mechanisms of such mitogenic signaling are not understood in detail, proliferation of human cancer cell lines is increased by insulin analogs that exhibit higher affinity for IR-A and the homologous Type 1 insulin-like growth factor receptor or that exhibit prolonged residence times on these receptors. It would be desirable, therefore, to invent an insulin analogue with negligible mitogenicity that nonetheless retains at least a portion of the glucoselowering effect of wild-type insulin. More generally, there is a need for an insulin analogue that displays increased thermodynamic stability and increased resistance to fibrillation above room temperature while exhibiting at least a subset of the multiple organ-specific biological activities of the corresponding wild-type insulin.

[0019] Amino-acid substitutions have been described in insulin that stabilize the protein but augment its binding to the insulin receptor (IR) and its cross-binding to the homologous receptor for insulin-like growth factors (IGFR) in such a way as to confer a risk of carcinogenesis. An example known in the art is provided by the substitution of His by aspartic acid. Although Asp -insulin exhibits favorable pharmaceutical properties with respect to stability and pharmacokinetics, its enhanced receptor-binding properties were associated with tumorigenesis in Sprague-Dawley rats. Although there are many potential substitutions in the A- or B chains that can be introduced into Asp -insulin or related analogues to reduce its binding to IR and IGFR to levels similar to that of human insulin, such substitutions generally impair the stability of insulin (or insulin analogues) and increase its susceptibility to chemical and physical degradation. It would be desirable to discover a method of modification of insulin and of insulin analogues that enabled “tuning” of receptorbinding affinities while at the same time enhancing stability and resistance to fibrillation. Such applications would require a set of stabilizing modifications that reduce binding to IR and IGFR to varying extent so as to offset the potential carcinogenicity of analogues that are super-active in their receptor-binding properties.

[0020] Therefore, there is a need for alternative insulin analogues, including those that are stabilized during storage while maintaining at least a portion of the biological activity of the analogue.

Summary of the invention [0021] It is, therefore, an aspect of the present invention to provide an insulin analogue that provides greater stability by halogen substitution in an amino acid, where the analogue than maintains at least a portion of biological activity of the corresponding non-halogenated insulin or insulin analogue. It is also an aspect of the present invention to provide insulin analogues that provide augmented stability, increased resistance to fibrillation above room temperature, and decreased mitogenicity while retaining at least a portion of the glucose-lowering activity of wild-type insulin in rodents following subcutaneous injection. The present invention addresses the utility of insulin analogues that, as a seeming paradox, exhibit marked decrements in affinity for the insulin receptor as expressed in non-neuronal cell lines and activity in non-neuronal tissues.

[0022] In general, the present invention provides an insulin analogue comprising a modified B- chain polypeptide that contains pento-fluoro-Phenylalanine instead of Phe at position B24. In one example the B-chain also contains substitutions Pro —>Lys and Lys —>Pro (a framework designated KP-insulin); the latter substitutions are known in the art to confer rapid action as embodied in the product Humalog® (Eli Lilly and Co.). In another example the KP-insulin framework was further modified by the substitution His —>Asp, yielding a penta-fluoro-Phe analog in a framework designated DKP-insulin. In another example the insulin analogue contains penta-f\uoro-Phenylalanine at position B24 in combination with Aspartic Acid at position B10 and Ornithine at position B29. In yet another example the DKP B-chain framework is substituted by an “EKP” framework containing Glu instead of Asp . In one particular set of embodiments, the B-chain polypeptide comprises an amino-acid sequence selected from the group consisting of SEQ ID NOS. 7-15 and polypeptides having three or fewer additional amino-acid substitutions thereof.

[0023] The invention also provides a method of treating a patient. The method comprises administering a physiologically effective amount of an insulin analogue or a physiologically acceptable salt thereof to the patient, wherein the insulin analogue or a physiologically acceptable salt thereof contains a B-chain polypeptide incorporating a penta-fluoro-phenylalanine at position B24. In still another embodiment, the insulin analogue may be a mammalian insulin analogue, such as an analogue of human insulin. In one particular set of embodiments, the B-chain polypeptide comprises an amino acid sequence selected from the group consisting of SEQ. ID. NOS. 7-15 and polypeptides having three or fewer additional amino acid substitutions thereof.

[0024] The present invention pertains to the substitution of Phenylalanine at position B24 (designated Phe ) by penta-fluoro-Phe in combination with standard or non-standard substitutions elsewhere in the A-chain or B-chain or in combination with C-terminal extension of the B-chain as a novel strategy to modify and improve distinct properties of insulin. During the past decade specific chemical modifications to the insulin molecule have been described that selectively modify one or another particular property of the protein to facilitate an application of interest. Whereas at the beginning of the recombinant DNA era (1980) wild-type human insulin was envisaged as being optimal for use in diverse therapeutic contexts, the broad clinical use of insulin analogues in the past decade suggests that a suite of analogues, each tailored to address a specific unmet clinical need, would provide significant medical and societal benefits. A current unmet clinical need is posed by the absence of insulin analogs known in the art to confer greater signaling activity in the hypothalamus than in the rest of the body. This need is important because insulin replacement therapy is often accompanied by weight gain as a consequence of its predominant anabolic signaling in the liver, muscle, and adipose tissue. Brain-selective insulin analogues may also provide new approaches to the treatment of obesity in general and to the arrest or slowing of the rate of cognitive decline in Alzheimer’s Disease. Despite such compelling needs, the engineering of brain-selective insulin analogues has seemed impossible because the human genome contains a single insulin-receptor gene and because any differences in its expressed product (due to RNA splicing or post-translational modification) do not significantly affect the binding of wild-type human insulin.

Brief description of the several views of the drawings [0025] FIG. 1A is a schematic representation of the sequence of human proinsulin including the A- and B-chains and the connecting region shown with flanking dibasic cleavage sites (filled circles) and C-peptide (open circles).

[0026] FIG. IB is a structural model of proinsulin, consisting of an insulin-like moiety and a disordered connecting peptide (dashed line).

[0027] FIG. 1C is a schematic representation of the sequence of human insulin indicating the position of residue B24 in the B-chain.

[0028] FIG. ID is a ribbon model of an insulin monomer showing aromatic residues in relation B15 B24 to the three disulfide bridges. The adjoining side chains of Leu (arrow) and Phe are shown. The A- and B-chain chains are otherwise shown in light and dark gray, respectively, and the sulfur atoms of cysteines as circles. B24 [0029] FIG. IE is a space-filling model of insulin showing the Phe side chain within a pocket at the edge of the hydrophobic core.

[0030] FIG. 2A is a representation of pentafluoro-phenylalanine (F^-Phe).

[0031] FIG. 2B is a representation of o/t/zo-monofluoro-phenylalanine (2F-Phe).

[0032] FIG. 2C is a representation of meta-monofluoro-phenyl alanine (3F-Phe).

[0033] FIG. 2D is a representation of pc/ra-monofluoro-phenylalanine (4F-Phe).

[0034] FIG. 2E is a representation of ort/zo-monochlorinated-phenylalanine (2C1-Phe).

[0035] FIG. 2F is a representation of meta-monochlorinated-phenylalanine (3C1-Phe).

[0036] FIG. 2G is a representation of pc/ra-monochlorinated-phenylalanine (4C1-Phe).

[0037] FIG. 3 is a set of four circular dichroism (CD) spectra for the far-UV wavelengths. Panel A: DKP-insulin (solid black line) and 2F-PheB24-DKP (□); Panel B: DKP-insulin (solid black line) and 3F-PheB24-DKP (A); Panel C: DKP-insulin (solid black line) and 4F-PheB24-DKP (▼); Panel D: DKP-insulin (solid black line) and F5-PheB24-DKP (·).

[0038] FIG. 4 is a set of four graphs for CD-detected guanidine denaturation studies. Panel A: DKP-insulin (solid black line) and 2F-PheB24-DKP (□); Panel B: DKP-insulin (solid black line) and 3F-PheB24-DKP (A); Panel C: DKP-insulin (solid black line) and 4F-PheB24-DKP (▼); Panel D: DKP-insulin (solid black line) and F^-PheB24-DKP (·).

[0039] FIG. 5 is a graph showing the results of receptor-binding studies of insulin analogs. 125

Relative activities are determined by competitive binding assay in which receptor-bound I-labeled human insulin is displaced by increasing concentrations of DKP-insulin () or its analogs: pentafluoro-PheB24 (A), 2F-PheB24 (▼), 3F-PheB24 (♦), and 4F-PheB24 (·).

[0040] FIG. 6 is a graph showing the results of receptor-binding studies of insulin analogs. Relative activities are determined by competitive binding assay in which receptor-bound I-labeled human insulin is displaced by increasing concentrations of KP-insulin () or its analogs: 2Br-PheB24-KP (A), 2Cl-PheB24-KP (▼), 2F-PheB24-KP (♦), and 4F-PheB24-KP (·). Assay employed the B-isoform of the insulin receptor and 125I-TyrA14-human insulin as tracer.

[0041] FIG. 7A is a graph for CD-detected guanidine denaturation studies of human insulin (solid line), KP-insulin (dashed line; "parent"), KP analogues containing monofluorous substitution of 2F-PheB24-KP (A), 3F-PheB24-KP (), and 4F-PheB24-KP (o).

[0042] FIG. 7B is a graph for CD-detected guanidine denaturation studies of human insulin (solid line), KP-insulin (dashed line; "parent"), 2-Cl-PheB24-KP (▼) and 2-Br-PheB24-KP (A).

[0043] FIG. 8A is a NMR spectra for DKP analogs comparing fluoro-substitutions at positions 2, 3, and 4 of PheB24 in relation to DKP-insulin, recorded at 700 MHz at 32 °C and pD 7.0.

[0044] FIG. 8B is a NMR spectra for KP analogs comparing fluoro-, bromo- and chloro-substitutions at position 2 of PheB24 in relation to KP-insulin, recorded at 700 MHz at 32 °C and pD 7.0.

[0045] FIG. 9A is a graph showing the results of in vitro receptor-binding assays using isolated B24 insulin receptor (isoform B): human insulin (triangles), KP-insulin (squares), and 4C1-Phe -KP-insulin (inverted triangles).

[0046] FIG. 9B is a graph showing the results of in vitro receptor-binding assays employing B24 IGF-1R: human insulin (triangles), KP-insulin (squares), 4C1-Phe -KP-insulin (inverted triangles), and native IGF-I (circles).

[0047] FIG. 9C is a graph comparing the results of in vitro receptor-binding assays using isolated insulin receptor (isoform B): human insulin (solid line), KP-insulin (dashed line), 4C1-PheB24-KP-insulin (triangles) 4F-PheB24-KP-insulin (squares).

[0048] FIG. 9D is a graph comparing the results of in vitro receptor-binding assays using isolated insulin receptor (isoform B): human insulin (solid line), KP-insulin (dashed line), 4C1-PheB26-KP-insulin (triangles) 4F-PheB26-KP-insulin (squares). B24 [0049] FIG. 10 is a graph showing the hypoglycemic action of subcutaneous of 4C1-Phe -KP-insulin in STZ induced diabetic Lewis rats over time (inverted triangles) relative to diluent alone (circles), human insulin (crosses), and KP-insulin (squares).

[0050] FIGS. 11A-C are graphs showing averaged traces of insulin cobalt solutions showing characteristic spectral profiles from 400-750 nm before and after addition of 2 mM EDTA. Samples were dissolved in 50 mM Tris (pH 7.4), 50 mM phenol, and 0.2 mM C0CI2. NaSCN was then added to a final concentration of 1 mM. Solid lines show data pre-EDTA extraction. Dashed lines show data post-EDTA extraction. Panel A: wild type insulin; Panel B: KP-insulin; Panel C; 4Cl-PheB24-KP-insulin.

[0051] FIG. 1 ID is a graph showing the kinetics of hexamer dissociation after addition of 2 mM EDTA as monitored at 574 nm (25° C and pH 7.4). Data were normalized to time zero for each B24 sample: wild type (solid line), KP-insulin (dashed line), and 4C1-Phe -KP-insulin (dotted line).

[0052] FIG. 12 is a graph showing a plot of the mean filtered glucose infusion rate versus time B24 after insulin dose for KP-insulin (Lispro insulin) and 4C1-Phe -KP-insulin (4-Cl-Lispro insulin) at a dosage of 0.2 Units per kilogram of body weight.

[0053] FIG. 13 is a bar graph summarizing 20 pharmacodynamic studies in pigs demonstrating B24 significant improvement in Vi T-max late in 4C1-Phe -KP-insulin over KP-insulin at five different dosing levels.

[0054] FIG. 14 is a bar graph summarizing 14 pharmacodynamic studies in pigs suggesting B24 improvement in Vi T-max early in 4C1-Phe -KP-insulin over KP-insulin at three different dosing levels.

[0055] FIG. 15 is a summary of ten, matched pharmacodynamics studies comparing the relative B24 potencies 4C1-Phe -KP-insulin with that of KP insulin as measured by area under the curve (AUC) in which the slightly reduced average potency for 4-C1-KP was found not to be statistically significant (p=0.22).

[0056] FIG. 16 provides ribbon depictions (stereo pairs) of the crystal structure of wild-type B24 insulin and penta-fluoro-Phe -KP-insulin as R-state protomers: (A) wild-type insulin and (B) analogue. In each panel the A-chain ribbon is shown in light gray and the B-chain ribbon is shown in dark gray.

[0057] FIG. 17 provides ribbon depictions of the crystal structure of wild-type insulin and B24 penta-fluoro-Phe -KP-insulin as R-state dimers: (A) wild-type insulin and (B) analogue. In each panel the A-chain ribbon is shown in light gray and the B-chain ribbon is shown in dark gray.

[0058] FIG. 18 provides ribbon depictions (stereo pairs) of the crystal structure of wild-type insulin and penta-fluoro-Phe -KP-insulin as R-state hexamers: (A) wild-type insulin and (B) analogue. In each panel the A-chain ribbon is shown in light gray and the B-chain ribbon is shown in dark gray. Axial zinc ions are shown in each panel as overlying spheres coordinated by the six imidiazole side chains of His . B24 [0059] FIG. 19 depicts the structural environment of Phe in a wild-type insulin R-state B24 protomer versus the structural environment of penta-fluoro-Phe : (A) wild-type insulin and (B) analogue.

[0060] FIG. 20 is a graph showing the results of receptor-binding studies of insulin analogues. B24

Studies of pmm-fluoro-Phe -DKP-insulin versus DKP-insulin and wild-type human insulin were performed by competitive displacement of 125I-TyrA14-labeled human insulin bound to isoform B of the human insulin receptor. Symbols: penta-fluoro-PheB24-DKP-insulin (| |): human insulin (|); and DKP-insulin (A).

[0061] FIG. 21 is a graph showing the results of receptor-binding studies of insulin analogues. B24

Studies of penta-fluoro-Phe -DKP-insulin versus DKP-insulin and wild-type human insulin were performed by competitive displacement of I-Tyr31-labeled human IGF-I bound to the Type 1 IGF receptor (IGF-1R). Symbols: penta-f\uoro-PheB24-DKP-insu 1 in (| |): human insulin (|); and DKP-insulin (A).

[0062] FIG. 22 provides the results of biological testing of medium-dose insulin analogues in rats (20 μg per rat) rendered diabetic by steptozotocin. The following analogues were tested: KP-insulin, DKP-insulin (AspB10-KP-insulin), EKP-insulin (GluB10-KP-insulin), penta-fluoro-PheB24-DKP-insulin, and penta-fluoro-Phe -EKP-insulin. Symbols are defined within the panel. The potency and pharmacokinetics of KP-insulin in the STZ Sprague-Dawley rat model are indistinguishable from those of wild-type human insulin. Whereas a dose of 5 μg of wild-type human insulin per rat corresponds to ca. 50% of maximal potency, a dose of 20 μg of wild-type human insulin represents ca 80-90% of maximal potency.

[0063] FIG. 23 provides a histogram summarizing results of a soft-agar mitogenicity assay employing human breast-cancer cell line MCF-7 at a ligand concentration of 100 nM. “Basal” represents effect of buffer alone whereas “HI” represents stimulation of growth due to addition of wild-type human insulin. No significant mitogenic activity (relative to buffer alone) was observed in comparative testing of j7ento-fluoro-PheB24-AspB10-OrnB29-insulin (“5F-POT”), penta-fluoro-Phe -DKP-insulin (“5F-DKP”), human proinsulin (“HPI”), or diabetes-associated inactive clinical variant LeuA3-insulin (insulin Wakayama', “A3Leu”).

Detailed description of the invention [0064] The present invention is directed an insulin analogue that provides greater stability by halogen substitution in an amino acid, where the analogue than maintains at least a portion of biological activity of the corresponding non-halogenated insulin or insulin analogue. Particularly, the present invention provides insulin analogues that provides greater stability by substitution of a single halogen in an amino acid, where the analogue than maintains at least a portion of biological activity of the corresponding non-halogenated insulin or insulin analogue. In one example, the present invention provides an insulin analogue that provides greater stability by fluorine, chlorine or bromine substitution in an amino acid, where the analogue than maintains at least a portion of biological activity of the corresponding non-halogenated insulin or insulin analogue. One potential application is to augment the chemical and physical stability of an insulin analogue while retaining a portion of its biological activity; another application is to calibrate the receptor-binding properties of an insulin analogue so not to exceed that of human insulin. To these ends, the present invention provides insulin analogues that contain a halogenated phenylalanine (Phe) residue substitution at position at B24, B25 or B26. While substitutions of a halogenated phenylalanine at B24 or B25 do not change the basic amino acid sequence of insulin or an insulin analogue, a halogenated phenylalanine at position B26 of insulin must substitute for tyrosine (Tyr) found in the wild type sequence of insulin. In one particular embodiment, the present invention provides insulin analogues that contain a pc/ra-monochloro-phenylalanine (4C1-Phe ) residue as a substitution for wild type phenylalanine at position B24.

[0065] The present invention is also directed toward an insulin analogue that provides resistance to fibrillation above room temperature and increased thermodynamic activity due to substitution of Phe by penta-fluoro-Phe. The chemical structure of penta-fluoro-Phe is shown in Figure 2. This modification causes a marked reduction in the affinity of insulin analogues for the insulin receptor or mitogenic Type 1 IGF receptor (IGF-1R). Such a reduction enables the simultaneous incorporation of an acidic residue (Aspartic Acid or Glutamic Acid) at position BIO, which provides further resistance to fibrillation and a further increase in thermodynamic stability. The combination of these modifications attenuates or blocks the undesirable mitogenicity otherwise associated with acidic substitutions at position BIO of insulin. Because Asp or Glu block the binding of zinc ions to the classical axial metal-ion binding sites of the insulin hexamer and destabilize the trimer-related self-association surface, thereby reducing self-assembly of such insulin analogues, application to rapid-acting insulin analogues provides a next generation of such analogues where the analogue maintains at least a portion of biological activity of the corresponding unmodified insulin or insulin analogue. It is a surprising aspect of the present invention that penta-fluoro-Phe -containing analogues exhibiting low receptor-binding affinities in vitro nonetheless retain significant biological activity as tested by an in vivo assay of hypoglycemic potency in male Sprague-Dawley rats rendered diabetic by steptozotocin.The present invention pertains to modification of Phe by penta-fluoro-Phe, an aromatic ring whose hydrophobicity is greater than that of the unmodified aromatic ring of Phenylalanine. The electronegative character of the fluorine atom leads to an inversion of the sign of the electrostatic quadrapole moment of the aromatic ring with negligible change in its small electrostatic dipole moment. While not wishing to be constrained by theory, the change in the quadrapolar electrostatic field would be expected to stabilize or destabilize a protein or protein-protein interface depending on the details of its structural environment, i.e., the positions of any neighboring electrostatic charges, distribution of solvent water molecules, and the positions and orientations of neighboring electrostatic moments, including dipole moments of amide and carbonyl functions in the main-chain of a polypeptide.

[0066] The present invention is not limited to human insulin and its analogues however. It is also envisioned that these substitutions may also be made in animal insulins such as porcine, bovine, equine, and canine insulins, by way of non-limiting examples.

[0067] Furthermore, in view of the similarity between human and animal insulins, and use in the past of animal insulins in human diabetic patients, it is also envisioned that other minor modifications in the sequence of insulin may be introduced, especially those substitutions considered “conservative” substitutions. For example, additional substitutions of amino acids may be made within groups of amino acids with similar side chains, without departing from the present invention. These include the neutral hydrophobic amino acids: Alanine (Ala or A), Valine (Val or V), Leucine (Leu or L), Isoleucine (lie or I), Proline (Pro or P), Tryptophan (Trp or W), Phenylalanine (Phe or F) and Methionine (Met or M). Likewise, the neutral polar amino acids may be substituted for each other within their group of Glycine (Gly or G), Serine(Ser or S), Threonine (Thr or T), Tyrosine (Tyr or Y), Cysteine (Cys or C), Glutamine (Glu or Q), and Asparagine (Asn or N). Basic amino acids are considered to include Lysine (Lys or K), Arginine (Arg or R) and Histidine (His or H). Acidic amino acids are Aspartic acid (Asp or D) and Glutamic acid (Glu or E). Unless noted otherwise or wherever obvious from the context, the amino acids noted herein should be considered to be L-amino acids. In one example, the insulin analogue of the present invention contains three or fewer conservative substitutions other than the halogenated-Phe substitutions of the present invention. In another example, the insulin analogue of the present invention contains one or fewer conservative substitutions other than the halogenated-Phe substitutions of the present invention.

[0068] As used in this specification and the claims, various amino acids in insulin or an insulin analogue may be noted by the amino acid residue in question, followed by the position of the amino acid, optionally in superscript. The position of the amino acid in question includes the A or B chain of insulin where the substitution is located. Thus, Phe denotes a phenylalanine at the twenty B25 fourth amino acid of the B chain of insulin, while Phe denotes a phenylalanine at the twenty fifth amino acid of the B chain of insulin and Phe denotes a phenylalanine substitution for tyrosine at the twenty sixth amino acid of the B chain of insulin. A fluorinated amino acid may be indicated with the prefix “F-,” a brominated amino acid may be indicated with the prefix “Br-“ and a chlorinated amino acid may be indicated with the prefix “C1-.” Therefore, fluorinated phenylalanine may be abbreviated “F-Phe,” chlorinated phenylalanine may be abbreviated “Cl-Phe” and brominated phenylalanine may be abbreviated “Br-Phe.” In the case of phenylalanine, the position of the fluorine substituents or substituents on the phenyl side chain may be further indicated by the number of the carbon to which the fluorine is attached. Therefore, ortho-monofluoro-phenylalanine (shown in Fig. 2B) is abbreviated “2F-Phe,” meta-monofluoro-phenylalanine (shown in Fig. 2C) is abbreviated “3F-Phe” and pc/ra-monofluoro-phenylalanine (shown in Fig. 2D) is abbreviated 4F-Phe. Pentafluoro-phenylalanine (shown in Fig. 2A) is abbreviated as “Fs-Phe.” Similarly, ort/zo-monobromo-phenylalanine may be abbreviated ‘^Br-Phe,” ortho-monochloro-phenylalanine may (shown in Fig. 2E) be abbreviated “2C1-Phe,” meta-monochloro-phenylalanine (shown in Fig. 2F) is abbreviated “3C1-Phe” and pc/ra-monochloro-phenylalanine (shown in Fig. 2G) is abbreviated “4C1-Phe.” [0069] The phenylalanine at B24 is an invariant amino acid in functional insulin and contains B24 an aromatic side chain. The biological importance of Phe in insulin is indicated by a clinical B24 mutation (Ser ) causing human diabetes mellitus. As illustrated in Figs. ID and IE, and while not wishing to be bound by theory, Phe is believed to pack at the edge of a hydrophobic core at the classical receptor binding surface. The models are based on a crystallographic protomer (2-Zn molecule 1; Protein Databank identifier 4INS). Lying within the C-terminal β-strand of the B-chain (residues B24-B28), Phe adjoins the central α-helix (residues B9-B19). One face and edge of the aromatic ring sit within a shallow pocket defined by Leu and Cys ; the other face and edge are exposed to solvent (Fig. IE). This pocket is in part surrounded by main-chain carbonyl and amide groups and so creates a complex and asymmetric electrostatic environment. The effects of the nearsymmetric substituent tetrafluoro-phenylalanine (F^-Phe ) (with its enhanced general hydrophobicity and fluoraryl-aryl interactions;) (Fig. 2A) with those of “unbalanced” analogs containing single ortho-, meta-, or pc/ra-fluorous substitutions (designated 2F-Phe, 3F-Phe, or 4F-Phe, Figs. 2B-2D, respectively) on insulin stability are provided. B25 [0070] Phe is thought to lie at the surface of the insulin monomer and in solution structures projects into solvent with less structural organization than that of Phe . Tyr , like Phe , lies at the edge of the hydrophobic core where it packs near nonpolar residues IleA2, ValA3, and ValB12. Photo-cross-linking studies suggest that the side chains of residues B24, B25, and B26 each contact the insulin receptor. In dimers and hexamers of insulin residues, B24-B26 are believed to participate in an intermolecular anti-parallel β-sheet. Aromatic side chains at these positions are thought to stabilize packing between monomers at this β-sheet interface.

[0071] It is envisioned that the substitutions of the present invention may be made in any of a number of existing insulin analogues. For example, the halogenated phenylalanine (H-Phe) substitutions provided herein may be made in insulin analogues such as Lispro (KP) insulin, insulin Aspart, other modified insulins or insulin analogues, or within various pharmaceutical formulations, such as regular insulin, NPH insulin, lente insulin or ultralente insulin, in addition to human insulin. Aspart insulin contains an Asp substitution and is sold as Novalog® whereas Lispro insulin contains Lys and Pro substitutions and is known as and sold under the name Humalog®. These analogues are described in US Pat. Nos. 5,149,777 and 5,474,978, the disclosures of which are hereby incorporated by reference herein. Both of these analogues are known as fast-acting insulins.

[0072] A halogenated-Phe substitution at B24, B25 and/or B26 may also be introduced into analogues of human insulin that, while not previously clinically used, are still useful experimentally, such as DKP insulin, described more fully below, or miniproinsulin, a proinsulin analogue containing a dipeptide (Ala-Lys) linker between the A chain and B chain portions of insulin in place of the normal 35 amino acid connecting region between the C-terminal residue of the B chain and the N-terminal residue of the A chain (see Fig. IB). Incorporation of halogenated aromatic residues at positions B24, B25, or B26 in DKP-insulin (or other insulin analogues that contain Asp or that exhibit receptor-binding affinities higher than that of human insulin) can reduce their receptor-binding affinities to be similar to or below that of human insulin and so potentially enable clinical use. In this manner the cross-binding of insulin analogues to the mitogenic IGFR may also be reduced.

[0073] The amino-acid sequence of human proinsulin is provided, for comparative purposes, as SEQ. ID. NO. 1. SEQ. ID. NO. 1 (proinsulin)

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Tyr-Thr-Pro-Lys-Thr-Arg-Arg-Glu-Ala-Glu-Asp-Leu-Gln-Val-Gly-Gln-Val-

Glu-Leu-Gly-Gly-Gly-Pro-Gly-Ala-Gly-Ser-Leu-Gln-Pro-Leu-Ala-Leu-Glu-Gly-Ser-Leu-Gln-Lys-

Arg-Gly-Ile-Val-Glu-Gln-Cys-Cys-Thr-Ser-Ile-Cys-Ser-Leu-Tyr-Gln-Leu-Glu-Asn-Tyr-Cys-Asn [0074] The amino acid sequence of the A chain of human insulin is provided as SEQ. ID. NO. 2. SEP. ID. NO. 2 (A chain)

Gly-Ile-Val-Glu-Gln-Cys-Cys-Thr-Ser-Ile-Cys-Ser-Leu-Tyr-Gln-Leu-Glu-Asn-Tyr-Cys-Asn [0075] The amino acid sequence of the B chain of human insulin is provided as SEQ. ID. NO. 3. SEP. ID. NO. 3 (B chain)

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Tyr-Thr-Pro-Lys-Thr [0076] The amino acid sequence of a B chain of human insulin may be modified with a substitution of a halogenated-Phe at position B24, B25, or B26. An example of such a sequence is provided As SEQ. ID. NO 4, where Xaa may be Tyr or Phe. The halogen used in any of these substitutions may be fluorine, chlorine or bromine, for example. SEP. ID. NO. 4

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Xaa-Thr-Pro-Lys-Thr [Xaa is Tyr or Phe] [0077] Further combinations of other substitutions are also within the scope of the present invention. It is also envisioned that the substitutions of the present invention may also be combined with substitutions of prior known insulin analogues. For example, the amino acid sequence of an analogue of the B chain of human insulin containing the Lys Pro substitutions of lispro insulin (Humalog®), in which one of the halogenated Phe substitutions may also be introduced, is provided as SEQ. ID. NO. 5. Likewise, the amino acid sequence of an analogue of the B chain of human insulin containing the Asp substitution of aspart insulin, in which a F-Phe or a Cl-Phe substitution may also be introduced, is provided as SEQ. ID. NO. 6. SEP. ID. NO. 5

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Xaa-Thr-Lys-Pro-Thr [Xaa is Tyr or Phe] SEP. ID. NO. 6

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Xaa-Thr-Asp-Lys-Thr [Xaa is Tyr or Phe] B24 B24 [0078] The F-Phe or Cl-Phe substitution may also be introduced in combination with other insulin analogue substitutions such as analogues of human insulin containing His substitutions at residues A4, A8 and/or B1 as described more fully in co-pending International Application No. PCT/US07/00320 and US Application Ser. No. 12/160,187, the disclosures of which are incorporated by reference herein. For example, a F-Phe or Cl-Phe substitution may be present with [HisA4, HisA8], and/or HisB1 substitutions in an insulin analogue or proinsulin analogue having the amino acid sequence represented by SEQ. ID. NO. 7,

Rl-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-His-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-Arg-Gly-Phe-Phe-R2-Thr-R3-R4-Thr- Xaao_35- Gly-Ile-Val-R5-Gln-Cys-Cys-R6-Ser-Ile-Cys-Ser-Leu-Tyr-Gln-Leu-Glu-Asn-Tyr-Cys-Asn; (SEQ. ID. NO. 7) wherein R1 is His or Phe; wherein R2 is Tyr or Phe, R3 is Pro, Lys, or Asp; wherein R4 is Lys or Pro; wherein R5 is His or Glu; wherein R6 is His or Thr; and wherein Xaao_35 is 0-35 of any amino acid or a break in the amino acid chain; and further wherein at least one substitution selected from the group of the following amino acid substitutions is present: R1 is His; and R6 is His; and R5 and R6 together are His.

[0079] A halogenated-Phe substitution at B24, B25, or B26 may also be introduced into a single chain insulin analogue as disclosed in co-pending U.S. Patent Application No. 12/419,169, the disclosure of which is incorporated by reference herein. B24 [0080] It is also envisioned that the chlorinated-Phe substitution may be introduced into an insulin analogue containing substitutions in the insulin A-chain. For example, an insulin analogue may additionally contain a lysine, histidine or arginine substitution at position A8, as shown in SEQ. ID. NO. 27, instead of the wild type threonine at position A8 (see SEQ. ID. NO. 2). SEP. ID. NO. 27

Gly-Ile-Val-Glu-Gln-Cys-Cys-Xaa-Ser-Ile-Cys-Ser-Leu-Tyr-Gln-Leu-Glu-Asn-Tyr-Cys-Asn [Xaa = His, Arg, or Lys] [0081] Fluorine or chlorine substitutions such as those at B24 may be introduced within an engineered insulin monomer of high activity, designated DKP-insulin, which contains the substitutions AspB1° (D), LysB28 (K) and ProB29 (P). These three substitutions on the surface of the B-chain are believed to impede formation of dimers and hexamers. Use of an engineered monomer circumvents confounding effects of self-assembly on stability assays. The structure of DKP-insulin closely resembles a crystallographic protomer. The sequence of the B-chain polypeptide for DKP insulin is provided as SEQ. ID. NO. 8, where Xaa is Tyr or Phe. SEP. ID. NO. 8 (DKP B-Chain Sequence)

Phe-Val-Asn-Gln-His-Leu-Cys-Gly-Ser-Asp-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-

Arg-Gly-Phe-Phe-Xaa-Thr-Lys-Pro-Thr

RID

[0079] The Asp substitution of a variant B chain is combined with Ornithine at position B29 as provided in SEQ ID NO: 28. SEP ID NO: 28

Phe-Val-Glu-Gln-His-Leu-Cys-Gly-Ser-Asp-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-

Gly-Glu-Arg-Gly-Phe-Phe-Try-Thr-Pro-Orn-Thr

Still other envisioned variants are provided as SEQ ID NOS: 29-38 SEP ID NO: 29

Phe-Val-Glu-Gln-His-Leu-Cys-Gly-Ser-Xaal-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-

Glv-Glu-Arg-Glv-Xaa2-Phe-Trv-Thr-Pro-Lvs-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-ΐΐuoro-Phe. SEP ID NO: 30

Phe-Val- Glu-Gln-His-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Lys-Pro-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-ΐΐuoro-Phe. SEP ID NO: 31

Phe-V al- Glu-Gln-His-Leu-Cys-Gly-Ser-Xaai-Leu-V al-Glu-Ala-Leu-T yr-Leu-V al-Cys-Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Pro-Asp-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-ΐΐuoro-Phe. SEP ID NO: 32

Phe-Val- Glu-Gln-Asp-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Pro-Lys-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-ΐΐuoro-Phe. SEP ID NO: 33

Phe-Val-Lys-Gln-Glu-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-

Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Pro-Glu-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-ΐΐuoro-Phe. SEP ID NO: 34

Phe-Val- Glu-Gln-Asp-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Pro-Orn-Thr

Where Xaai indicates His, Asp, or Glu; and where Xaa2 indicates penta-fluoro-Phe. SEP ID NO: 35

Phe-Val-Glu-Gln-Asp-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-

Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Xaa3-Pro-Thr

Where Xaai indicates His, Asp, or Glu; where Xaa2 indicates penta-ΐΐuoro-Phe; and where Xaa3 indictes Ala, Ser, Asp, Glu, Ser, Asn, Gin, or diamino-propionic acid, diamino-butyric acid, or norleucine. SEP ID NO: 36

Phe-Val-Glu-Gln-Asp-Leu-Cys-Gly-Ser-Xaai-Leu-Val-Glu-Ala-Leu-Tyr-Leu-Val-Cys-

Gly-Glu-Arg-Gly-Xaa2-Phe-Try-Thr-Pro-Xaa3-Thr

Where Xaai indictes His, Asp, or Glu; where Xaa2 indicates penta-ΐΐuoro-Phe; and where Xaa3 indictes Ala, Ser, Asp, Glu, Ser, Asn, Gin, or diamino-propionic acid, diamino-butyric acid, or norleucine.

Peptides used for semi-synthesis techniques include SEQ ID NOS: 37-42 SEP ID NO: 37

Gly-Xaai-Phe-Tyr-Thr-Lys-Pro-Thr

Where Xaai indicates penta-fluoro-Phe. SEP ID NO: 38

Gly-Xaai-Phe-Tyr-Thr-Xaa2-Pro-Thr

Where Xaai indicates penta-fluoro-Phe; and where Xaa2 indicates Ala, Ser, Asp, Glu, Ser, Asn, Gin, or diamino-propionic acid, diamino-butyric acid, or norleucine. SEP ID NO: 39

Gly-Xaai-Phe-Tyr-Thr-Asp-Lys-Thr

Where Xaai indicates pm/r/-fluoro-Phe. SEP ID NO: 40

Gly-Xaai-Phe-Tyr-Thr-Asp-Glu-Thr Where Xaai indicates penta-fluoro-Phe. SEP ID NO: 41

Gly-Xaai-Phe-Tyr-Thr-Pro-Orn-Thr Where Xaai indicates penta-fluoro-Phe. SEP ID NO: 42

Gly-Xaai-Phe-Tyr-Thr-Pro-Xaa2-Thr

Where Xaai indicates penta-fluoro-Phe; and where indicates Xaa2 indicates Ala, Ser, Asp, Glu, Ser, Asn, Gin, or diamino-propionic acid, diamino-butyric acid, or norleucine.

[0082] Analogues of DKP-insulin were prepared by trypsin-catalyzed semi-synthesis and purified by high-performance liquid chromatography (Mirmira, R.G., and Tager, H.S., 1989. J. Biol. Chem. 264: 6349-6354.) This protocol employs (i) a synthetic octapeptide representing residues (N)-GF*FYTKPT (including modified residue (F*) and “KP” substitutions (underlined); SEQ. ID. NO. 9) and (ii) truncated analogue Je.v-octapeptide[B23-B30]-AspB10-insulin (SEQ. ID. NO. 16). Because the octapeptide differs from the wild-type B23-B30 sequence (GFFYTPAT; SEQ. ID. NO. 10) by interchange of Pro and Lys (italics), protection of the lysine f -amino group is not required during trypsin treatment. In brief, i/es-octapeptide (15 mg) and octapeptide (15 mg) were dissolved in a mixture of dimethylacetamide/l,4-butandiol/0.2 M Tris acetate (pH 8) containing 10 mM calcium acetate and 1 mM ethylene diamine tetra-acetic acid (EDTA) (35:35:30, v/v, 0.4 mL). The final pH was adjusted to 7.0 with 10 μΕ of A-methylmorpholine. The solution was cooled to 12 °C, and 1.5 mg of TPCK-trypsin was added and incubated for 2 days at 12 °C. An additional 1.5 mg of trypsin was added after 24 hr. The reaction was acidified with 0.1% trifluoroacetic acid and purified by preparative reverse-phase HPLC (C4). Mass spectrometry using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF; Applied Biosystems,

Foster City, CA) in each case gave expected values (not shown). The general protocol for solid-phase synthesis is as described (Merrifield et al., 1982. Biochemistry 21: 5020-5031). 9-fluoren-9-yl-methoxy-carbonyl (F-moc)-protected phenylalanine analogs were purchased from Chem-Impex International (Wood Dale, IL).

[0083] Circular dichroism (CD) spectra were obtained at 4 °C using an Aviv spectropolarimeter (Weiss et al., Biochemistry 39: 15429-15440). Samples contained ca. 25 μΜ DKP-insulin or analogs in 50 mM potassium phosphate (pH 7.4); samples were diluted to 5 μΜ for guanidine-induced denaturation studies at 25°C. To extract free energies of unfolding, denaturation transitions were fitted by non-linear least squares to a two-state model as described by Sosnick et al., Methods

Enzymol. 317: 393-409. In brief, CD data , where x indicates the concentration of denaturant, are fitted by a nonlinear least-squares program according to

[0084] where x is the concentration of guanidine and where Θα and ΘΒ are baseline values in the native and unfolded states. Baselines are approximated by pre- and post-transition lines

[0085] Relative activity is defined as the ratio of analogue to wild-type human insulin required 125 to displace 50 percent of specifically bound I-human insulin. A human placental membrane preparation containing the insulin receptor (IR) was employed as known in the art. Membrane fragments (0.025 mg protein/tube) were incubated with I-labeled insulin (ca. 30,000 cpm) in the presence of selected concentrations of unlabelled analogue for 18 hours at 4 °C in a final volume of 0.25 ml of 0.05 M Tris-HCl and 0.25 percent (w/v) bovine serum albumin at pH 8. Subsequent to incubation, mixtures are diluted with 1 ml of ice-cold buffer and centrifuged (10,000g) for 5 min at 4°C. The supernatant is then removed by aspiration, and the membrane pellet counted for radioactivity. Data is corrected for nonspecific binding (amount of radioactivity remaining membrane associated in the presence of 1 μΜ human insulin. In all assays the percentage of tracer bound in the absence of competing ligand was less than 15% to avoid ligand-depletion artifacts. An additional assay to monitor changes in activity during the course of incubation of the single-chain analogue at 37 °C was performed using a microtiter plate antibody capture as known in the art. Microtiter strip plates (Nunc Maxisorb) were incubated overnight at 4 °C with AU5 IgG (100 μΐ/well of 40 mg/ml in phosphate-buffered saline). Binding data were analyzed by a two-site sequential model. A corresponding microtiter plate antibody assay using the IGF Type I receptor was employed to assess cross-binding to this homologous receptor.

[0086] The far-ultraviolet circular dichroism (CD) spectra of the singly fluorinated analogs are essentially identical to that of the parent analogue (Fig. 3); a slight distortion is observed in the spectrum of the F^-Phe analogue (Fig. 3, Panel D). The stabilities and receptor-binding activities of the analogs are provided in Table 1. Modified B24 residues were introduced within the context of DKP-insulin. Activity values shown are based on ratio of hormone-receptor dissociation constants relative to human insulin; the activity of human insulin is thus 1.0 by definition. Standard errors in the activity values were in general less than 25%. Free energies of unfolding (AGU) at 25 °C were estimated based on a two-state model as extrapolated to zero denaturant concentration. Lag time indicates time (in days) required for initiation of protein fibrillation on gentle agitation at 37 °C in zinc-free phosphate-buffered saline (pH 7.4).

Table 1

Activities and Stabilities of Insulin Analogs

[0087] Substitution of PheB24 by F^-Phe augments the stability of DKP-insulin by 0.8 ±0.1 kcal/mole at 25 °C (Fig. 4, Panel D). Despite such a favorable effect on stability, the activity of the analogue is negligible (< 1% receptor-binding affinity relative to the parent analog; Table 1 and Fig. 5). This impairment is more marked than that ordinarily observed on standard single-amino-acid substitution. By contrast, the singly substituted analogs each retain significant activity: ortho (37% relative to human insulin), meta (100%), and para (43%). Such activities are each above the threshold of 10% (corresponding to a hormone-receptor dissociation constant of < 1 nM) generally predictive of therapeutic efficacy. Further, whereas the 3F-Phe and 4F-Phe analogs are no more stable (and possibly slightly less stable) than the parent analogue (Fig. 4 and Table 1), 2F-PheB24-DKP-insulin is at least as stable as the F^-Phe analogue (AAGU 0.9 ±0.1 kcal/mole). The physical stabilities of the analogues were evaluated in triplicate during incubation in 60 μΜ phosphate-buffered saline (PBS) at pH 7.4 at 37 °C under gentle agitation. The samples were observed for 20 days or until signs of precipitation or frosting of the glass vial were observed. F5-, 2F- and 3F-PheB24-DKP-insulin analogues (but not 4F- PheB24-DKP-insulin) each exhibit an increase in lag time (relative to the parent analogue) prior to onset of protein fibrillation (Table 1). Substitution of a single fluorine atom is thus able to augment the stability of insulin with retention of reduced but clinically significant activity; the modification can also extend the lag time prior to protein fibrillation on gentle agitation at pH 7.4 and 37 °C. B24 [0088] Whereas the F5- Phe substitution is highly stabilizing (similar to the most favorable of the core substitutions in the villin subdomain), this insulin analogue is essentially without biological B24 activity. Remarkably, the regiospecific modification 2F-Phe is equally stabilizing while retaining substantial (although reduced) affinity for the insulin receptor.

[0089] Fluorine substitutions were also introduced essentially as described above with the exception of ortho-, meta-, para-, and penta-fluorinated phenylalanine being introduced at positions B24, B25 and B26 in lispro insulin (that is, an insulin analogue also containing the substitutions Lys , Pro (sold under the name Humalog®)). Analogues containing ortho-, meta-, para-, and penta-fluorinated phenylalanine substitutions at position B24 are designated 2F-Phe -KP, 3F-PheB24-KP, 4F-PheB24-KP, Fs-PheB24-KP, respectively. Analogues containing ortho-, meta-, para-, and penta-fluorinated phenylalanine substitutions at position B25 are designated 2F-Phe -KP, 3F-PheB25-KP, 4F-PheB25-KP, F5-PheB25-KP, respectively. Analogues containing ortho-, meta-, para-, and penta-fluorinated phenylalanine substitutions at position B26 (substituting for tyrosine) are designated 2F-PheB26-KP, 3F-PheB26-KP, 4F-PheB26-KP, F5-PheB26-KP, respectively. Additionally, ortho-bromo and ortho-chloro phenylalanine was substituted into lispro insulin at position B24. These analogues are designated 2Br-PheB24-KP and 2Cl-PheB24-KP, respectively.

[0090] More specifically, for halogenated phenylalanine substitutions at B24, a synthetic octapeptide representing residues (N)-GF*FYTKPT (halogenated phenylalanine residue indicated as “F*” and “KP” substitutions (underlined); SEQ ID NO. 9) and truncated analogue des-octapeptide[B23-B30]-insulin (wild type at position B10, SEQ ID NO. 17) were used. For fluorinated phenylalanine substitutions at B25, a synthetic octapeptide representing residues (N)-GFF*YTKPT (fluorinated phenylalanine indicated again as “F*” and “KP” substitutions (underlined); SEQ ID NO. 9) and truncated analogue c/e.v-oc tapeptide[B23-B30J-insu 1 in (wild type at position B10; SEQ. ID. NO 17) were used. For fluorinated phenylalanine substitutions at B26, a synthetic octapeptide representing residues (N)-GFFF*TKPT (fluorinated phenylalanine indicated again as “F*” and “KP” substitutions (underlined); SEQ ID NO. 18) and truncated analogue des-octapeptide[B23-B30]-insulin (wild type at position B10; SEQ. ID. NO 17) were used, des-octapeptide (15 mg) and octapeptide (15 mg) were dissolved in a mixture of dimethylacetamide/l,4-butandiol/0.2 M Tris acetate (pH 8) containing 10 mM calcium acetate and 1 mM ethylene diamine tetra-acetic acid (EDTA) (35:35:30, v/v, 0.4 mF). The final pH was adjusted to 7.0 with 10 μΕ of A-methylmorpholine. The solution was cooled to 12 °C, and 1.5 mg of TPCK-trypsin was added and incubated for 2 days at 12 °C. An additional 1.5 mg of trypsin was added after 24 hr. The reaction was acidified with 0.1% trifluoroacetic acid and purified by preparative reverse-phase HPLC (C4). Mass spectrometry using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF; Applied Biosystems, Foster City, CA) in each case gave expected values. Relative activity and dissociation constants were determined as described above.

[0091] The resulting data is presented in Table 2 for substitution of a fluorinated phenylalanine at position B24 of insulin. For comparison purposes, wild type insulin and 2F-Phe -DKP were tested under the same conditions at 37 °C. Data for testing at 37 °C are presented below in Table 3. Data for substitutions of fluorinated phenylalanine at positions B24, B25 and B26 and bromo- and chloro- substituted phenylalanine in a lispro insulin analogue background are presented below in Table 4.

Table 2

Stability and Activity of PheB24-DKP

Table 3

Stability and Activity of Insulin and PheB24-DKP [37 °C]

Table 4

Stability And Activity Of Halogenated-Phe Analogue Of Lispro Insulin

[0092] Receptor-binding data for analogues of lispro insulin by competitive displacement are also presented in Fig. 6. The assay employed the B-isoform of the insulin receptor and 125I-TyrA14-human insulin as tracer.

[0093] Lispro insulin analogues containing either chloro- or bromo-substituted phenylalanine at B24 (2Cl-PheB24-KP and 2Br-PheB24-KP, respectively) were tested for the ability to lower blood sugar in diabetic rats in comparison to non-halogenated lispro (KP) insulin. Male Lewis rats (-250 g body weight) were rendered diabetic with streptozotocin. Lispro insulin analogue, 2C1-Phe -KP and 2Br-Phe -KP were purified by HPLC, dried to powder, and dissolved in insulin diluent (Eli Lilly Corp). Rats were injected subcutaneously at time = 0 with 20 pg or 6.7 pg insulin analogue in 100 pi of diluent. Blood was obtained from clipped tip of the tail at time 0 and every 10 minutes up to 90 min. Blood glucose was measured using a Hypoguard Advance Micro-Draw meter. Blood glucose concentration changes (in milligrams (mg) per decaliter (dL) per hour (h)) are listed in

Tables 5 and 6 (these studies were performed with different sets of rats and so exhibit differences in baseline pharmacodynamic response to control injections of KP-insulin; top lines in each table).

Table 5

Blood Glucose Levels In Response To Halogenated Analogues Of Lispro Insulin

Table 6

Blood Glucose Levels In Response To Halogenated Analogues

[0094] As seen in Tables 5 and 6, the halogenated-phenylalanine containing insulin analogues B24 were equal or greater in potency than lispro (Humalog®) insulin with the exception of 2F-Phe - B24 KP-insulin. The potency of a 2F-Phe insulin analog can be restored to that of KP-insulin by inclusion of the Asp substitution in the 2F-Phe -DKP-insulin analogue (line 2 of Table 6). Additionally, the presence of a halogenated phenylalanine at position B24 increases fibrillation lag time between three-fold and four-fold; thus, whereas KP-insulin under the above experimental conditions exhibits a lag time of about 3 days, substitution of Phe by either 2F-Phe , 2C1-Phe , or 2Br-Phe extends the lag time to between 10-12 days. These modifications also increase the thermodynamic stability (as probed by guanidine denaturation) by between 0.5 and 1 kcal/mol (Table 4).

[0095] Dissociation constants (Kd) were determined as described by Whittaker and Whittaker (2005. J. Biol. Chem. 280: 20932-20936), by a competitive displacement assay using 125I-TyrA14-insulin (kindly provided by Novo-Nordisk) and the purified and solubilized insulin receptor (isoform B) in a microtiter plate antibody capture assay with minor modification; transfected receptors were tagged at their C-terminus by a triple repeat of the FLAG epitope (DYKDDDDK; SEQ. ID. NO 11) and microtiter plates were coated by anti-FLAG M2 monoclonal antibody (Sigma). The percentage of tracer bound in the absence of competing ligand was less than 15% to avoid ligand-depletion artifacts. Binding data were analyzed by non-linear regression using a heterologous competition model (Wang, 1995, FEBS Lett. 360: 111-114) to obtain dissociation constants.

Table 7

Cross-Binding of Insulin Analogues to the IGF Receptor

[0096] Measurements of cross-binding of selected insulin analogues to the IGF receptor are summarized in Table 7. Human insulin under these conditions binds to IGFR 260-fold less tightly than does IGF-I. Cross-binding of Asp -insulin is increased by twofold whereas cross-binding by an analogue in wide clinical use, Arg , Arg , Gly -insulin (insulin glargine ; Fantus ), is increased by threefold. Such augmentation of IGFR cross-binding is undesirable as treatment of Sprague-Dawley rats with Asp -insulin is associated with an increased incidence of mammary tumors whereas recent clinical studies suggest the possibility that human use of Fantus confers a dose-dependent increase in the risk of diverse cancers. (The sequence of IGF-I contains Glu at R1 Π position BIO, similar in negative charge to Asp . While not wishing to be bound by theory, it is believed that cross-binding by Lantus is augmented by the two basic residues at the C-terminus of the B-chain. Because IGF-I also contains Lys at position B28 and Pro at position B29, it is possible that KP-insulin may exhibit a small increase in cross-binding but the above data are not sufficiently precise to establish this.) DKP-insulin exhibits a six-fold increase in IGFR cross-binding. Significantly, this increase is counter-balanced by an ortho-monofluoro substitution at B24: 2F-Phe -DKP-insulin exhibits the same low level of cross-binding as wild-type human insulin. Its hypoglycemic potency in diabetic rats is also the same as that of human insulin (Table 6). Because 2F-Phe -DKP-insulin is monomeric even at protein concentrations > 0.5 mM, this analog is therefore of potential utility as an ultra-fast acting and ultra-stable insulin formulation for clinical use. Cross-binding affinities of 2C1-Phe -KP-insulin and 2Br-Phe -KP-insulin are also at a low level similar to that of human insulin. These analogs are also fully active in diabetic rats (Table 5).

[0100] CD spectra were determined for human insulin, KP-insulin (lispro) and halogenated Phe insulin analogues as described above. CD-detected protein denaturation is provided as a function of the concentration of guanidine hydrochloride in Figs. 7A and 7B. Fig. 7A provides a comparison of human insulin (solid line) and KP-insulin (dashed line; "parent") with KP analogues containing monofluorous substitution of PheB24 at position 2, 3, or 4 (filled triangles, filled squares, and open circles, respectively). Fig. 7B provides a comparison of human insulin (solid line) and KP-insulin (dashed line; "parent") with analogues containing 2-C1 or 2-Br modification of PheB24 (filled inverted triangles and open triangles, respectively). Inferred thermodynamic parameters are provided above in Table 4. Fractional unfolding was monitored by mean residue ellipticity at 222 nm at 25 °C.

[0101] NMR structures of selected analogs have been obtained to demonstrate that halogen substitutions are well accommodated in insulin and do not cause transmitted conformational perturbations. See Figs. 8A and 8B. The NMR structure of 2C1-Phe -KP-insulin as a monomer in solution is thus similar to that of KP-insulin; similarly, the NMR structure of 2F-Phe -DKP-insulin (also as a monomer in solution) is similar to that of DKP-insulin. Qualitative analysis of NMR spectra of 2F-PheB24-KP-insulin and 2Br-PheB24-KP-insulin likewise indicate native-like structures.

[0102] Crystal structures have been determined of 2F-PheB24-KP-insulin and 4F-PheB24-KP-insulin as zinc-stabilized hexamers in the presence of the preservative phenol (data not shown).

The structures are similar to those previously reported for KP-insulin as a zinc-stabilized hexamer £· in the presence of phenol. In each case the conformation of the hexamer is T3R 3 containing two axial zinc ions and three bound phenol molecules. Electron density for the halogen atoms is readily observed. The structure of the dimer- and trimer-forming surfaces are preserved in the halogenated analogues, suggesting that pharmaceutical formulations may be obtained similar to those employed for KP-insulin and other fast-acting insulin analogues.

[0103] Modified residues were introduced within the context of KP-insulin. Activity values shown are based on ratio of hormone-receptor dissociation constants relative to human insulin; the activity of human insulin is thus 1.0 by definition. Standard errors in the activity values were in general less than 25%. Free energies of unfolding (AGU) at 25° C were estimated based on a two-state model as extrapolated to zero denaturant concentration. Lag time indicates time (in days) required for initiation of protein fibrillation on gentle agitation at 37° C in zinc-free phosphate-buffered saline (pH 7.4). B24 [0104] The chlorinated-Phe substitution provided herein may be made in insulin analogues such as lispro insulin (that is, an insulin analogue also containing the substitutions Lys , Pro (sold under the name Humalog®)). Such an insulin analogue is designated chlorinated-PheB24-KP-insulin. For comparative purposes, fluorine substitutions were also introduced, essentially as described above, with the exception of para-fluorinated phenylalanine being introduced at positions B24 and B26 in lispro insulin. Analogues containing pc/ra-fluorinated phenylalanine substitutions at position B24 are designated 4F-Phe -KP-insulin. Analogues containing para -fluorinated phenylalanine substitutions at position B26 (substituting for tyrosine) are designated 4F-Phe -KP-insulin. Designations of respective analogues of KP-insulin and DKP-insulin may be abbreviated in tables and figures as KP and DKP with “insulin” omitted for brevity.

[0105] For chlorinated phenylalanine substitutions at B24, a synthetic octapeptide representing residues (N)-GF*FYTKPT (chlorinated phenylalanine residue indicated as “F*” and “KP” substitutions (underlined); SEQ. ID. NO. 9) and truncated analogue ί/c.v-octapeptide[B23-B30J-insulin (wild type at position B10, SEQ. ID. NO. 17) were used. For fluorinated phenylalanine substitutions at B26, a synthetic octapeptide representing residues (N)-GFFF*TKPT (fluorinated phenylalanine indicated again as “F*” and “KP” substitutions (underlined); SEQ. ID. NO. 18) and truncated analogue Je.v-octapeptide[B23-B30]-insulin (SEQ. ID. NO. 17) were used.

[0106] The resulting data for substitution of a halogenated phenylalanine at position B24 or B26 in a lispro insulin analogue background are presented below in Table 8.

Table 8

Stability And Activity Of Halogenated-Phe Analogues Of Lispro Insulin

B24 [0107] Fibrillation Assays. The physical stability of 4C1-Phe -KP-insulin was evaluated in triplicate during incubation in zinc-free phosphate-buffered saline (PBS) at pH 7.4 at 37° C under gentle agitation in glass vials. The samples were observed for 12 days at a protein concentration of 60 μΜ for visual appearance of cloudiness; twice daily aliquots were withdrawn for analysis of thioflavin-T (ThT) fluorescence. Because ThT fluorescence is negligible in the absence of amyloid but is markedly enhanced on onset of fibrillation, this assay probes for the lag time. Respective lag times for human insulin, KP-insulin, and 4C1-Phe -KP-insulin are 5±1 days, 3±1 days, and more than 12 days. 4C1-Phe -KP-insulin is therefore at least 4-fold more resistant to fibrillation under these conditions than is KP-insulin and at least 2-fold more resistant than human insulin. While not wishing to condition patentability on theory, it is envisioned that increased fibrillation resistance of 4C1-Phe insulin analogues will allow them to be formulated in a zinc-free formulation to enhance the fast-acting nature of the insulin analogue without significantly shortening the storage time of a sample of the analogue, either before or after an individual sample has begun to be used.

[0108] Thermodynamic Stability. We measured the free energy of unfolding of 4C1-Phe -KP-insulin relative to KP-insulin and LysA8-KP-insulin in a zinc-free buffer at pH 7.4 and 25° C (10 mM potassium phosphate and 50 mM KC1). This assay utilized CD detection of guanidine-induced denaturation as probed at helix-sensitive wavelength 222 nm. Values of AGu were extrapolated to zero denaturant concentration to obtain estimates by the free energy of unfolding on the basis of a two-state model. Whereas the substitution ThrA8—«-Lys augmented thermodynamic stability by 0.6 ± 0.2 kcal/mole, the 4Cl-PheB24 modification decreased stability by 0.4 ± 0.2 kcal/mole.

[0109] The affinity of 4Cl-PheB24-KP-insulin-A8T for the detergent-solubilized and lectin purified insulin receptor (isoform B) is similar to that of human insulin. A competitive displacement assay using I-labeled human insulin as tracer is shown in Figure 9A using isolated insulin receptor (isoform B): human insulin (triangles), KP-insulin (squares), 4C1-Phe -KP-insulin (inverted triangles). All three curves are closely aligned, indicating similar receptor-binding affinities. The affinity of 4C1-Phe -KP-insulin for the insulin receptor was indistinguishable from that of KP-human insulin, in each case slightly lower than the affinity of wild-type insulin.

[0110] Fig. 9B shows results of corresponding assays employing Insulin-like Growth Factor I Receptor (IGF-1R), probed by competitive displacement using I-labeled IGF-I as tracer. Symbols are the same with the addition of native IGF-I (circles). The rightward shift of the 4C1-Phe -KP-insulin curve indicates decreased cross-binding to IGF-1R. The cross-binding of 4C1-Phe -KP-insulin to IGF-1R is reduced by approximately 3-fold relative to that of KP-insulin or wild-type insulin.

[0111] A similar comparison between human insulin (solid black line), KP-insulin (dashed B24 B24 line), 4C1-Phe -KP-insulin (triangles), 4F-Phe -KP-insulin (squares) using isolated insulin B24 receptor (isoform B), is shown in Fig. 9C. The rightward shift of the 4F-Phe -KP-insulin curve relative to 4C1-Phe -KP-insulin, wild type human insulin and lispro insulin shows decreased receptor-binding affinity with the use of a different halogen at the same position on the phenylalanine ring in comparison to a para-chloro substitution of the phenylalanine at B24. In contrast, the insulin receptor-binding affinity of 4C1-Phe -KP-insulin is similar to that of wild type human insulin and lispro (KP) insulin.

[0112] A further comparison between human insulin (solid line), KP-insulin (dashed line), 4C1-Phe -KP-insulin (triangles) 4F-Phe -KP-insulin (squares) using isolated insulin receptor (isoform B), is shown in Fig. 9D. Both 4Cl-PheB26-KP-insulin and 4F-PheB26-KP-insulin show decreased insulin receptor-binding affinity in comparison to wild type and lispro insulins. As stated above, 4C1-Phe -KP-insulin does not show a similar decrease in receptor-binding affinity. B24 [0113] The in vivo potency of 4C1-Phe -KP-insulin in diabetic rats is similar to that of KP-insulin. To enable characterization of biological activity, male Lewis rats (-300 g body weight) were rendered diabetic with streptozotocin. Human insulin, KP-insulin, and 4C1-Phe -KP-insulin were purified by HPLC, dried to powder, and dissolved in insulin diluent (Eli Lilly Corp). Rats were injected subcutaneously at time = 0 with either 20 pg or 6.7 pg of KP-insulin or 4C1-Phe -KP-insulin in 100 pi of diluent; the higher dose is at the plateau of the wild-type insulin dose-response curve whereas the lower dose corresponds to 50-70% maximal initial rate of glucose disposal. Injection of diluent alone was performed as a negative control. 8 rats were studies in each group. Blood was obtained from clipped tip of the tail at time 0 and at successive intervals up to 120 min. Blood glucose was measured using a Hypoguard Advance Micro-Draw meter. Blood glucose concentrations were observed to decrease as shown in Figure 10. The initial rate of fall of the blood glucose concentration during the first 24 min after injection are similar on comparison of 4Cl-PheB24-KP-insulin (-225 ± 29 mg/dl/h), KP-insulin (-256 ± 35 mg/dl/h), and human insulin (255 ± 35 mg/dl/h). Any differences in initial rate are not statistically significant. The duration of action of 4C1-Phe -KP-insulin over the next 60 min appears shorter, however, than the durations of human insulin or KP-insulin. B24 [0114] Given the native receptor-binding affinity of 4C1-Phe -KP-insulin, it would be unusual for its potency to be less than that of human insulin. Indeed, insulin analogs with relative affinities in the range 30-100% relative to wild-type typically exhibit native potencies in vivo. It is formally possible, however, that the biological potency of 4C1-Phe -KP-insulin is somewhat lower (on a molar basis) than the potencies of human insulin or KP-insulin. If so, we note that any such decrease would be within the threefold range of the molar activities of current insulin products in clinical use (by convention respective international units (IU) are redefined to reflect extent of glucose lowering, leading to product-to-product differences in the number of milligrams or nanomoles per unit). It should be noted that the slow decline in blood glucose concentration on control injection of protein-free diluent (brown dashed line in Fig. 10) reflects diurnal fasting of the animals following injection.

[0115] A surrogate marker for the pharmacokinetics of insulin hexamer disassembly (designated the EDTA sequestration assay) employs cobalt ions (Co2+) rather than zinc ions (Zn2+) to mediate hexamer assembly. Although Co2+ and Zn2+ hexamers are similar in structure, the cobalt ion provides a convenient spectroscopic probe due to its unfilled d-electronic shell.

[0116] The principle of the assay is as follows. Solutions of Rr, phenol-stabilized Co2+ insulin hexamers are blue due to tetrahedral Co2+ coordination; on disassembly the protein solution is colorless as octahedral Co2+ coordination by water or EDTA (ethylene-diamine-tetra-acetic acid; a strong chelator of metal ions) lacks optical transitions at visible wavelengths as a consequence of ligand field theory. The EDTA sequestration assay exploits these spectroscopic features as follows. At time t=0 a molar excess of EDTA is added to a solution of Rr, insulin hexamers or insulin analog hexamers. Although EDTA does not itself attack the hexamer to strip it of metal ions, any Co2+ ions released in the course of transient hexamer disassembly become trapped by the chelator and thus unavailable for reassembly. The rate of disappearance of the blue color (the tetrahedral d-d optical transition at 574 nm of the R-specific insulin-bound Co2+) thus provides an optical signature of the kinetics of hexamer disassembly.

[0117] Averaged traces of insulin cobalt solutions showing characteristic spectral profiles from 400-750 nm were determined before and after addition of 2 mM EDTA (Figs. 11A-C). Samples were dissolved in 50 mM Tris (pH 7.4), 50 mM phenol, and 0.2 mM C0CI2. NaSCN was then added to a final concentration of 1 mM. The kinetics of hexamer dissociation after addition of 2 mM EDTA as monitored at 574 nm (25° C and pH 7.4) are also shown. The spectra of the analogues before EDTA extraction are shown as solid lines. Post-EDTA extraction, the spectra are displayed as dashed lines. Wild type is shown in Panel A, KP-insulin in Panel B, and 4C1-Phe -KP-insulin as in Panel C.Data were normalized to time zero for each sample.

[0118] On the one hand, the baseline optical absorption spectra of the hexameric cobalt complexes at t=0 are similar among wild-type insulin hexamers, KP insulin hexamers, and 4C1-Phe -KP-insulin hexamers (see Figs. 11A-11C). The similar shapes and magnitudes of these respective d-d electronic transitions imply that the metal ions are in similar R^-specific tetrahedral coordination sites in wild-type and variant hexamers. This result is significant as it implies that 4C1-Phe -KP-insulin remains competent for metal-ion-mediated assembly and hence a zinc-based formulation.

[0119] The kinetics of hexamer dissociation after addition of 2 mM EDTA as monitored at 574 nm (25° C and pH 7.4) shows that the wild-type and variant hexamers exhibit marked differences in rates of EDTA-mediated Co2+ sequestration. As expected, the wild-type hexamer exhibits the greatest kinetic stability (solid line in Fig. 1 ID), followed by KP-insulin (dashed-dotted line in Fig. 11D), and 4-Cl-PheB24-KP-insulin (dotted line in Fig. 11D). Respective half-lives are 481 sec (wild type), 363 sec (KP-insulin), and 66 sec (4C1-Phe -KP-insulin). The extent of acceleration induced by the pc/ra-chloro-aromatic substitution is thus more profound than that associated with the “KP switch” of Lispro insulin (Humalog™). Because diffusion of zinc ions from the site of subcutaneous injection is analogous to the in vitro sequestration of cobalt ions in the EDTA Sequestration assay, these findings predict that 4C1-Phe -KP-insulin will exhibit a marked acceleration of absorption.

[0120] The pharmacokinetic (PK) and pharacodynamic (PD) properties and potency of 4-C1-Phe -KP-insulin were investigated in relation to wild-type insulin (Humulin™; Eli Lilly and Co.) and KP-insulin (Humalog™) in adolescent Yorkshire farm pigs (weight 25-45 kg). The wild type and KP-insulin were used as provided by the vendor (Eli Lilly and Co.) in U-100 strength. The 4-Cl-Phe -KP-insulin was formulated in Lilly diluent with a ratio of protein to zinc ions similar to that of the wild type and KP-insulin products; its strength was U-87. On the day of study, each animal underwent anesthesia induction with Telazol and then general anesthesia with isoflurane. Each animal was endotreacheally intubated, and oxygen saturation and end-tidal expired CO2 were continuously monitored. To block endogenous pancreatic a- and β-cell secretion, pigs were given a subcutaneous injection of octreotide acetate (44 pg/kg) approximately 30 min before beginning the clamp study and every 2 h thereafter. After IV catheters were placed and baseline euglycemia was established with 10% dextrose infusion, an IV injection of the insulin was given through the catheter. In order to quantify peripheral insulin-mediated glucose uptake, a variable-rate glucose infusion was given to maintain a blood glucose concentration of approximately 85 mg/dl. Such a glucose infusion was typically required for 5-8 h until the glucose infusion rate returned to the preinsulin baseline. Glucose concentrations were measured with a Hemocue 201 portable glucose analyzer every 10 min (instrument error rate: 1.9%). The computerized protocol for glucose clamping was as described by Matthews et al. 2-ml blood samples for insulin assay was also obtained according to the following schedule: from 0-40 min after insulin delivery: 5-minute intervals; from 50 - 140 min: 10-minute intervals, and from 160 min - to the point when GIR is back to baseline: 20-min intervals. For analysis of PK/PD, a 20-min moving mean curve fit and filter was applied. PD was measured as time to early half-maximal effect (½ Tmax Early), time to late half-maximal effect (½ Tms Late), time to maximal effect, and area-under-the-curve (AUC) over baseline. For each of these analyses, the fitted curve, not the raw data, was used. Each of 3 pigs underwent 3 studies. The results of these studies are provided in Figs. 12-15.

[0121] 4 -Cl-PheB24-KP-insulin (abbreviated in Fig. 12 as 4-Cl-Lispro Insulin) was found to exhibit a significantly less prolonged late “tail” than KP-insulin or wild-type insulin. The improved turn-off of insulin action suggests a potential clinical benefit with regard to late post-prandial hypoglycemia.

[0122] Fig. 13 summarizes 20 pharmacodynamic studies in pigs demonstrating significant improvement in Vi T-max late in 4C1-Phe -KP-insulin over KP-insulin at five different dosing levels, 0.05 U/kg, 0.1 U/kg, 0.2 U/kg, 0.5 U/kg, and 1 U/kg.

[0123] Fig. 14 summarizes 14 pharmacodynamic studies in pigs suggesting improvement in Vi B24 T-max early in 4C1-Phe -KP-insulin over KP-insulin at three different dosing levels, 0.05 U/kg, 0.1 U/kg, 0.2 U/kg.

[0124] Fig. 15 summarizes ten, matched pharmacodynamics studies comparing the relative potencies 4C1-Phe -KP-insulin with that of KP insulin as measured by area under the curve (AUC) in which the slightly reduced average potency for 4-C1-KP was found not to be statistically significant (p=0.22). The pharmacokinetic (PK) and pharmacodynamic (PD) properties of 4-C1-Phe -KP-insulin in relation to wild-type insulin and KP-insulin (Lispro-insulin) under similar formulation conditions (zinc insulin hexamers or zinc insulin analog hexamers stabilized by phenol and meta-cresol) show that the potency of 4-Cl-Phe -KP-insulin, as measured by area-under-the-curve (AUC) method, was similar to those of wild-type insulin and KP-insulin.

[0125] A method for treating a patient comprises administering a halogen-substituted insulin analogue to the patient. In one example, the halogen-substituted insulin analogue is an insulin analogue containing halogen-substituted phenylalanine, such as a fluoro-, chloro- or bromo-substituted phenylalanine. In one particular example the halogen substituted phenylalanine is 2F-PheB24, 2F-PheB25, 2F-PheB26 or 4Cl-PheB24. In another example, the halogen-substituted insulin analogue additionally contains one or more substitutions elsewhere in the insulin molecule designed to alter the rate of action of the analogue in the body. The insulin analogue may optionally contain a histidine, lysine or arginine substitution at position A8. In still another example, the insulin analogue is administered by an external or implantable insulin pump. An insulin analogue of the present invention may also contain other modifications, such as a tether between the C-terminus of the B-chain and the N-terminus of the A-chain as described more fully in co-pending International Application No. PCT/US07/080467 (U.S. Patent Application No. 12/419,169), the disclosure of which is incorporated by reference herein.

[0126] A pharamaceutical composition may comprise such insulin analogues and which may optionally include zinc. Zinc ions may be included in such a composition at a level of a molar ratio of between 2.2 and 3.0 per hexamer of the insulin analogue. In such a formulation, the concentration of the insulin analogue would typically be between about 0.1 and about 3 mM; concentrations up to 3 mM may be used in the reservoir of an insulin pump. Modifications of meal-time insulin analogues may be formulated as described for (a) “regular” formulations of Humulin™ (Eli Lilly and Co.), Humalog™ (Eli Lilly and Co.), Novalin™ (Novo-Nordisk), and Novalog™ (Novo-Nordisk) and other rapid-acting insulin formulations currently approved for human use, (b) “NPH” formulations of the above and other insulin analogues, and (c) mixtures of such formulations. As mentioned above, it is believed that the increased resistance to fibrillation will permit 4C1-Phe -containing insulin analogues to be formulated without the presence of zinc to maximize the fast acting nature of the analogue. However, it is also believed that even in the presence of zinc, the 4C1-Phe -containing insulin analogues will dissociate from hexamers into dimers and monomers sufficiently quickly as to also be considered a fast-acting insulin analogue formulation.

[0127] Excipients may include glycerol, glycine, other buffers and salts, and anti-microbial preservatives such as phenol and meta-cresol; the latter preservatives are known to enhance the stability of the insulin hexamer. Such a pharmaceutical composition may be used to treat a patient having diabetes mellitus or other medical condition by administering a physiologically effective amount of the composition to the patient.

[0128] A nucleic acid comprising a sequence that encodes a polypeptide encoding an insulin analogue containing a sequence encoding at least a B-chain of insulin with a fluorinated phenylalanine at position B24, B25 or B26 is also envisioned. This can be accomplished through the introduction of a stop codon (such as the amber codon, TAG) at position B24 in conjunction with a suppressor tRNA (an amber suppressor when an amber codon is used) and a corresponding tRNA synthetase, which incorporates a non-standard amino acid into a polypeptide in response to the stop codon, as previously described (Furter, 1998, Protein Sci. 7:419-426; Xie et al., 2005,

Methods. 36: 227-238). The particular sequence may depend on the preferred codon usage of a species in which the nucleic acid sequence will be introduced. The nucleic acid may also encode other modifications of wild-type insulin. The nucleic acid sequence may encode a modified A- or B-chain sequence containing an unrelated substitution or extension elsewhere in the polypeptide or modified proinsulin analogues. The nucleic acid may also be a portion of an expression vector, and that vector may be inserted into a host cell such as a prokaryotic host cell like an E. coli cell line, or a eukaryotic cell line such as S. cereviciae or Pischia pastoris strain or cell line.

[0129] For example, it is envisioned that synthetic genes may be synthesized to direct the expression of a B-chain polypeptide in yeast Piscia pastoris and other microorganisms. The nucleotide sequence of a B-chain polypeptide utilizing a stop codon at position B24 for the purpose of incorporating a halogenated phenylalanine at that position may be either of the following or variants thereof: (a) with Human Codon Preferences: TTTGTGAACC AACACCTGTGCGGCTCAC ACCTGGTGGAAGCTCTCT ACCTAGTGTGCGG GG A ACG AGGCT AGTTCT AC AC ACCC A AG ACC (SEQ. ID. NO. 12) (b) with Pichia Codon Preferences: TTTGTT AACCAAC ATTTGTGTGGTTCTCATTTGGTTG AAGCTTTGT ACTTGGTTTGTGGT GAAAGAGGTTAGTTTTACACTCCAAAGACT (SEQ. ID. NO. 13) [0130] Similarly, a full length pro-insulin cDNA having human codon preferences and utilizing a stop codon at position B24 for the purpose of incorporating a halogenated phenylalanine at that position may have the sequence of SEQ. ID NO. 14. TTTGTGAACC AACACCTGTG CGGCTCACAC CTGGTGGAAG CTCTCTACCT AGTGTGCGGG GAACGAGGCT AGTTCTAC AC ACCCAAGACC CGCCGGGAGG CAGAGGACCT GCAGGTGGGG CAGGTGGAGC TGGGCGGCGG CCCTGGTGCA GGCAGCCTGC AGCCCTTGGC CCTGGAGGGG TCCCTGCAGA AGCGTGGCAT TGTGGAACAA TGCTGTACCA GCATCTGCTC CCTCTACCAG CTGGAGAACT ACTGCAACTA G (SEQ.ID. N0.14) [0131] Likewise, a full length human pro-insulin cDNA utilizing a stop codon at position B24 for the purpose of incorporating a halogenated phenylalanine at that position and having codons preferred by P. pastoris may have the sequence of SEQ. ID. NO 15.

TTTGTTAACC AACATTTGTG TGGTTCTCAT TTGGTTGAAG CTTTGTACTT

GGTTTGTGGT GAAAGAGGTT AGTTTTACAC TCCAAAGACT AGAAGAGAAG

CTGAAGATTT GCAAGTTGGT CAAGTTGAAT TGGGTGGTGG TCCAGGTGCT

GGTTCTTTGC AACCATTGGC TTTGGAAGGT TCTTTGCAAA AGAGAGGTAT

TGTTGAACAA TGTTGTACTT CTATTTGTTC TTTGTACCAA TTGGAAAACT ACTGTAACTA A (SEQ. ID. NO. 15) [0132] Similarly, a nucleotide sequence of a B-chain polypeptide utilizing a stop codon at position B25 for the purpose of incorporating a halogenated phenylalanine at that position may be either of the following or variants thereof: (a) with Human Codon Preferences: TTTGTGAACC AACACCTGTGCGGCTCAC ACCTGGTGGAAGCTCTCT ACCTAGTGTGCGG GG A ACG AGGCTTCT AGT AC AC ACCC A AG ACC (SEQ. ID. NO. 19) (b) with Pichia Codon Preferences: TTTGTT AACCAAC ATTTGTGTGGTTCTCATTTGGTTG AAGCTTTGT ACTTGGTTTGTGGT GAAAGAGGTTTTT AGT AC ACTCCAAAGACT (SEQ. ID. NO.20) [0133] A full length pro-insulin cDNA having human codon preferences and utilizing a stop codon at position B25 for the purpose of incorporating a halogenated phenylalanine such as fluorinated phenylalanine at that position may have the sequence of SEQ. ID NO. 21. TTTGTGAACC AACACCTGTG CGGCTCACAC CTGGTGGAAG CTCTCTACCT AGTGTGCGGG GAACGAGGCT TCTAGTACAC ACCCAAGACC CGCCGGGAGG CAGAGGACCT GCAGGTGGGG CAGGTGGAGC TGGGCGGCGG CCCTGGTGCA GGCAGCCTGC AGCCCTTGGC CCTGGAGGGG TCCCTGCAGA AGCGTGGCAT TGTGGAACAA TGCTGTACCA GCATCTGCTC CCTCTACCAG CTGGAGAACT ACTGCAACTA G (SEQ.ID. N0.21) [0134] Likewise, a full length human pro-insulin cDNA utilizing a stop codon at position B25 for the purpose of incorporating a halogenated phenylalanine such as a fluorinated phenylalanine at that position and having codons preferred by P. pastoris may have the sequence of SEQ. ID. NO 22. TTTGTTAACC AACATTTGTG TGGTTCTCAT TTGGTTGAAG CTTTGTACTT GGTTTGTGGT GAAAGAGGTT TTTAGTACAC TCCAAAGACT AGAAGAGAAG CTGAAGATTT GCAAGTTGGT CAAGTTGAAT TGGGTGGTGG TCCAGGTGCT GGTTCTTTGC AACCATTGGC TTTGGAAGGT TCTTTGCAAA AGAGAGGTAT TGTTGAACAA TGTTGTACTT CTATTTGTTC TTTGTACCAA TTGGAAAACT ACTGTAACTA A (SEQ. ID. NO. 22) [0135] Likewise, a nucleotide sequence of a B-chain polypeptide utilizing a stop codon at position B26 for the purpose of incorporating a halogenated phenylalanine at that position may be either of the following or variants thereof: (a) with Human Codon Preferences: TTTGTGAACC AACACCTGTGCGGCTCAC ACCTGGTGGAAGCTCTCT ACCTAGTGTGCGG GGAACGAGGCTTCTTCTAGACACCCAAGACC (SEQ. ID. NO. 23) (b) with Pichia Codon Preferences: TTTGTT AACCAAC ATTTGTGTGGTTCTCATTTGGTTG AAGCTTTGT ACTTGGTTTGTGGT GAAAGAGGTTTTTTTTAGACTCCAAAGACT (SEQ. ID. NO. 24) [0136] A full length pro-insulin cDNA having human codon preferences and utilizing a stop codon at position B26 for the purpose of incorporating a halogenated phenylalanine at that position may have the sequence of SEQ. ID NO. 25.

TTTGTGAACC AACACCTGTG CGGCTCACAC CTGGTGGAAG CTCTCTACCT AGTGTGCGGG GAACGAGGCT TCTTCTAGAC ACCCAAGACC CGCCGGGAGG

CAGAGGACCT GCAGGTGGGG CAGGTGGAGC TGGGCGGCGG CCCTGGTGCA

GGCAGCCTGC AGCCCTTGGC CCTGGAGGGG TCCCTGCAGA AGCGTGGCAT

TGTGGAACAA TGCTGTACCA GCATCTGCTC CCTCTACCAG CTGGAGAACT ACTGCAACTA G (SEQ.ID. N0.25) [0137] Likewise, a full length human pro-insulin cDNA utilizing a stop codon at position B26 for the purpose of incorporating a halogenated phenylalanine at that position and having codons preferred by P. pastoris may have the sequence of SEQ. ID. NO 26.

TTTGTTAACC AACATTTGTG TGGTTCTCAT TTGGTTGAAG CTTTGTACTT

GGTTTGTGGT GAAAGAGGTT TTTTTTAGAC TCCAAAGACT AGAAGAGAAG

CTGAAGATTT GCAAGTTGGT CAAGTTGAAT TGGGTGGTGG TCCAGGTGCT

GGTTCTTTGC AACCATTGGC TTTGGAAGGT TCTTTGCAAA AGAGAGGTAT

TGTTGAACAA TGTTGTACTT CTATTTGTTC TTTGTACCAA TTGGAAAACT ACTGTAACTA A (SEQ. ID. NO. 26) [0138] Other variants of these sequences, encoding the same polypeptide sequence, are possible, given the synonyms in the genetic code.

[0139] Analogues of insulin containing penta-fluoro-Phe were prepared by trypsin-mediated semi-synthesis. The protocol for semi-synthesis employed a c/e.v-oc tapeptide[B23-B30J fragment of R1Π R| η human insulin or insulin analogue (such as analogues containing Asp or Glu ) together with a synthetic octapeptide containing an N-terminal Glycine. The <:/<?.v-octapeptide[B23-B30] fragment or analogue thereof contains the three native disulfide bridges of wild-type insulin; the protocol including purification of the fragment, peptide, and product by high-performance liquid chromatography was a modification of that described (Mirmira, R.G., and Tager, H.S., 1989. J. Biol. Chem. 264: 6349-6354.) This protocol employs a synthetic peptide containing a monosaccaride pyranoside adduct (SEQ ID NO: 39-41). With respect to non-standard synthetic peptide sequences containing an N-terminal Glycine, sequence design exploited the inability of trypsin does not cleave after Orn, Norleucine, diaminobutyric acid, or diaminopropinic acid and further exploited the fact that trypsin cleaves Lys-Pro steps inefficiently (as in KP-insulin and KP-insulin analogues; residues B28-B29) in contrast to the efficient cleavage of Pro-Lys-Xaa steps (as in wild-type human, pork, and beef insulin; residues B28-B30). Synthetic peptides were prepared by solid-phase peptide synthesis using 9-fluoren-9-yl-methoxy-carbonyl (F-moc) protected precursors.

[0140] In brief, i/es-octapeptide (15 mg) and octapeptide (15 mg) were dissolved in a mixture of dimethylacetamide/l,4-butandiol/0.2 M Tris acetate (pH 8) containing 10 mM calcium acetate and 1 mM ethylene diamine tetra-acetic acid (EDTA) (35:35:30, v/v, 0.4 mL). The final pH was adjusted to 7.0 with 10 μΕ of X-methylmorpholine. The solution was cooled to 12 °C, and 1.5 mg of TPCK-trypsin was added and incubated for 2 days at 12 °C. An additional 1.5 mg of trypsin was added after 24 hr. The reaction was acidified with 0.1% trifluoroacetic acid and purified by preparative reverse-phase HPLC (C4). Mass spectrometry using matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF; Applied Biosystems, Foster City, CA) in each case gave expected values (not shown). The general protocol for solid-phase synthesis is as described (Merrifield et al., 1982. Biochemistry 21: 5020-5031). 9-fluoren-9-yl-methoxy-carbonyl (F-moc)-protected phenylalanine analogues were purchased from Chem-Impex International (Wood Dale, IL).

[0141] The above protocol was employed to prepare analogues of human insulin containing B29 penta-fluoro-Phe at position B24 in the context of Orn -insulin, KP-insulin, DKP-insulin (containing AspB1°), AspB10-OrnB29-insulin, and EKP-insulin (containing GluB1°). The crystal structure of penta-fluoro-Phe -Orn -insulin was determined at a resolution of 2.5 A as a zinc-and phenol-stabilized insulin analogue hexamer with conformation (Figures 16-19).

Comparison of this crystal structure with the corresponding crystal structure of wild-type human insulin revealed that the modified aromatic ring incurs no steric clash and is readily accommodated within a native-like dimer interface (Figure 17). No significant local or transmitted structural perturbations were observed in the mode of hexamer assembly (Figure 18). The conformations of neighboring side chains are similar to those in the wild-type structure (Figure 19).

[0142] Circular dichroism (CD) spectra were obtained at 4° C and/or 25° C using an Aviv spectropolarimeter (Weiss et al., Biochemistry 39: 15429-15440). Samples contained ca. 25 μΜ DKP-insulin or analogues in 50 mM potassium phosphate (pH 7.4); samples were diluted to 5 μΜ for guanidine-induced denaturation studies at 25° C. To extract free energies of unfolding, denaturation transitions were fitted by non-linear least squares to a two-state model as described by

Sosnick et al., Methods Enzymol. 317: 393-409. In brief, CD data where x indicates the concentration of denaturant, were fitted by a nonlinear least-squares program wherein baselines were approximated by pre- and post-transition lines.

[0143] Relative activity is defined as the ratio of the hormone-receptor dissociation constants of 125 analogue to wild-type human insulin, as measured by a competitive displacement assay using I-human insulin. Microtiter strip plates (Nunc Maxisorb) were incubated overnight at 4° C with AU5 IgG (100 μΐ/well of 40 mg/ml in phosphate-buffered saline). Binding data were analyzed by a two-site sequential model. Data were corrected for nonspecific binding (amount of radioactivity remaining membrane associated in the presence of 1 μΜ human insulin. In all assays the percentage of tracer bound in the absence of competing ligand was less than 15% to avoid ligand-depletion artifacts. Representative data are provided in Figure 20; corresponding assays conducted with the Type I IGF receptor (IGF-1R) are shown in Figure 21. Dissociation constants (Kd) were determined by fitting to a mathematic model as described by Whittaker and Whittaker (2005. J. Biol. Chem. 280: 20932-20936); the model employed non-linear regression with the assumption of heterologous competition (Wang, 1995, FEBS Lett. 360: 111-114). Results are summarized in Table 8. In each case the insulin analogues modified by penta-fluoro-Phe exhibit a marked reduction in the affinity of binding to the insulin receptor or to the IGF receptor.

Table 8. In Vitro Receptor-Binding Affinities

Footnote: Experiments 1 and 2 were performed on different dates and employed different preparations of receptors. Abbrevations: IR-B, splicing isoform B of the insulin receptor; IGF-1R, Type 1 IGF receptor. “KP” indicates substitutions FysB28 and ProB29. “DKP” indicates additional substitution Asp whereas “EKP” indicates additional substitution Glu . B24 [0144] The far-ultraviolet circular dichroism (CD) spectra of penta-fluoro-Phe derivatives were found to be greater than those of the respective parent analogues. Free energies of unfolding (AGU) at 25° C were estimated based on a two-state model as extrapolated to zero denaturant concentration; the denaturant was guanidine hydrochloride. The resulting estimates of AGU were in each case greater than those of the parent analogues lacking a modified B24 ring. In the context of KP-insulin the gain in stability (AAGU) was 1.0 ± 0.2 kcal/mole; in the context of DKP-insulin the gain in stability was 1.5 ± 0.2 kcal/mole.

[0145] Physical stability was probed by measurement of lag times prior to onset of fibrillation as detected by enhancement of Thioflavin T fluorescence and visible cloudiness of the samples. Fag times thus indicate the time (in days) required for initiation of protein fibrillation on gentle agitation at 37° C in zinc-free phosphate-buffered saline (pH 7.4). Substitution of penta-fluoro-Phe in the context of KP-insulin or DKP-insulin was found to extend the lag time prior to fibrillation by at least 1.5-fold relative to that of the corresponding parent analog (KP-insulin or DKP-insulin, respectively).

[0146] Mitogenicity was evaluated in soft-agar tissue-culture studies of MCF-7 human breast-cancer cells as grown in 6-well plates. In this protocol 1500 cells were typically seeded in triplicate in a top layer of 0.3% agar in MEM growth medium ±100 nM ligand (human insulin or insulin analogue) on top of a base layer of 0.6% agar in MEM growth medium. Growth medium ±100 nM ligand was overlayed, and this medium changed 3X/week for 21 days. Colonies >150 microns were counted under a microscope. The results, summarized in Figure 23, demonstrate that penta-fluoro-Phe mitogates the enhanced mitogenicity otherwise associated with the Asp substitution. In this assay the mitogenic activity of [Asp , penta-fluoro-Phe (-analogues of insulin was no greater than that of buffer alone and substantially less than wild-type human insulin.

[0147] To evaluate the biological activity and potency of the analogues in an animal model, male Sprague-Dawley rats (mean body mass -300 grams) were rendered diabetic by treatment with streptozotocin (STZ). Protein solutions containing KP-insulin or analogues of KP-insulin containing an acidic substitution at B10 (Asp or Glu) were prepared using a protein-free sterile diluent (obtained from Eli Lilly and Co.) composed of 16 mg glycerin, 1.6 mg meta-cresol, 0.65 mg phenol, and 3.8 mg sodium phosphate pH 7.4. The activity of penta-fluoro-Phe analogs was evaluated in relation to that of KP-insulin (Figure 22). These formulations of KP-insulin or KP-insulin analogues were injected subcutaneously, and resulting changes in blood glucose concentration were monitored by serial measurements using a clinical glucometer (Hypoguard Advance Micro-Draw meter). To ensure uniformity of formulation, insulins were each re-purified by reverse-phase HPLC, dried to powder, dissolved in diluent at the same maximum protein concentration (300 μg/mL) and re-quantified by analytical C4 rp-HPLC; dilutions were made using the above buffer. Rats were injected subcutaneously at time t = 0 with 20 μg insulin or insulin analogs in 100 μΐ of buffer per 300 g rat. Blood was obtained from the clipped tip of the tail at time 0 and every 10 minutes up to 360 min. The penta-fluoro-Phe analogues of the present invention were found, in combination with an acidic substitution at position B10, to retain a substantial proportion of the biological activity of KP-insulin.

[0148] While not wishing to be constrained by theory, we ascribe the seeming paradox of low in vitro receptor-binding affinities and substantial biological activity in rodents to non-classical effects of signaling by insulin or insulin analogs in the brain. In particular, binding of insulin or insulin analogues to neurons in the hypothalamus in the rodent brain is known in the art to contribute to suppression of hepatic gluconeogenesis and hence hepatic glucose output, the predominant source of the high circulating blood glucose concentrations in the streptozotocin rat model in the fasted state.

[0149] A method for treating a patient with diabetes mellitus comprises administering an insulin analogue containing a monosaccaride adduct as known in the art or described herein. It is another aspect of the present invention that insulin analogues containing penta-fluoro-Phe may readily be obtained by semi-synthesis. This protocol employs tryptic digestion of a two-chain or single-chain precursor polypeptide following its folding to achieve native disulfide pairing, yielding a des-octapeptide[B23-B30] fragment of insulin or an insulin analogue containing additional substitutions in the A- and B chains. It is yet another aspect of the present invention that use of non-standard amino-acid substitutions enables a rapid and efficient method of preparation of pento-fluoro-Phe - modified insulin analogues by trypsin-mediated semi-synthesis using unprotected octapeptides. We further envision the analogues of the present invention providing a method for the treatment of diabetes mellitus or of obesity in the context of prediabetes, in the context of the metabolic syndrome, or in the absence of metabolic abnormalities. For this application, the route of delivery of the insulin analogue may alternatively be nasal, buccal, transdermal, or subcutaneous.

[0150] In still another example, the insulin analogue is administered by an external or implantable insulin pump. An insulin analogue of the present invention may also contain other modifications, such as a halogen atom at positions B24, B25, or B26 as described more fully in copending U.S. Patent Application No. 13/018,011, the disclosure of which is incorporated by reference herein. An insulin analogue of the present invention may also contain a foreshortened B-chain due to deletion of residues B1-B3 as described more fully in co-pending U.S. Provisional Patent Application 61/589,012.

[0151] A pharmaceutical composition may comprise such insulin analogues and which may optionally include zinc. Zinc ions may be included. In such a formulation, the concentration of the insulin analogue would typically be between about 0.1 and about 3 mM; concentrations up to 3 mM may be used in the reservoir of an insulin pump. Modifications of meal-time insulin analogues may be formulated as described for (a) “regular” formulations of Humulin® (Eli Lilly and Co.), Humalog® (Eli Lilly and Co.), Novalin® (Novo-Nordisk), and Novalog® (Novo-Nordisk) and other rapid-acting insulin formulations currently approved for human use, (b) “NPH” formulations of the above and other insulin analogues, and (c) mixtures of such formulations. Analogues of insulin lacking residues B1-B3 and containing Glu may also be formulated in the absence of zinc ions as known in the art for the formulation of insulin glulisine.

[0152] Excipients may include glycerol, glycine, arginine, Tris, other buffers and salts, and antimicrobial preservatives such as phenol and meta-cresol; the latter preservatives are known to enhance the stability of the insulin hexamer. Such a pharmaceutical composition may be used to treat a patient having diabetes mellitus or other medical condition by administering a physiologically effective amount of the composition to the patient.

[0153] Based upon the foregoing disclosure, it should now be apparent that halogen-substituted insulin analogues will carry out the objects set forth hereinabove. Namely, these insulin analogues exhibit enhanced thermodynamic stability, resistance to fibrillation and potency in reducing blood glucose levels. The halogen substituted phenylalanine-containing insulin analogues also have reduced cross-reactivity to insulin-like growth factor (IGFR). It is, therefore, to be understood that any variations evident fall within the scope of the claimed invention and thus, the selection of specific component elements can be determined without departing from the spirit of the invention herein disclosed and described.

[0154] The word ‘comprising’ and forms of the word ‘comprising’ as used in this description and in the claims does not limit the invention claimed to exclude any variants or additions.

[0155] In this specification where a document, act or item of knowledge is referred to or discussed, this reference or discussion is not an admission that the document, act or item of knowledge or any combination thereof was at the priority date publicly available, known to the public, part of the common general knowledge or known to be relevant to an attempt to solve any problem with which this specification is concerned.

[0156] The following literature is cited to demonstrate that the testing and assay methods described herein would be understood by one of ordinary skill in the art.

Furter, R., 1998. Expansion of the genetic code: Site-directed p-fluoro-phenylalanine incorporation in Escherichia coli. Protein Sci. 7:419-426.

Merrifield, R.B., Vizioli, F.D., and Boman, H.G. 1982. Synthesis of the antibacterial peptide cecropin A (1-33). Biochemistry 21: 5020-5031.

Mirmira, R.G., and Tager, H.S. 1989. Role of the phenylalanine B24 side chain in directing insulin interaction with its receptor: Importance of main chain conformation. J. Biol. Chem. 264: 63496354.

Sosnick, T.R., Fang, X., and Shelton, V.M. 2000. Application of circular dichroism to study RNA folding transitions. Methods Enzymol. 317: 393-409.

Wang, Z.X. 1995. An exact mathematical expression for describing competitive biding of two different ligands to a protein molecule FEBS Lett. 360: 111-114.

Weiss, M.A., Hua, Q.X., Jia, W., Chu, Y.C., Wang, R.Y., and Katsoyannis, P.G. 2000. Hierarchiacal protein "un-design": insulin's intrachain disulfide bridge tethers a recognition a-helix. Biochemistry 39: 15429-15440.

Whittaker, J., and Whittaker, L. 2005. Characterization of the functional insulin binding epitopes of the full length insulin receptor. J. Biol. Chem. 280: 20932-20936.

Xie, J. and Schultz, P.G. 2005. An expanding genetic code. Methods. 36: 227-238.

Claims (12)

1. Claims What is claimed is: 1 An insulin analogue comprising an insulin B-chain polypeptide modified by penta-fluoro-Phenylalanine at position B24.
2. 2 An insulin analogue of claim 2 that is further modified by an acidic substitution (Aspartic Acid or Glutamic Acid) at position BIO.
3. 3 An insulin analogue of claim 1 that is further modified by paired substitutions LysB28 and ProB29.
4. 4 An insulin analogue of claim 1 that is further modified by paired substitutions LysB3 and GluB29.
5. 5 An insulin analogue of claim 1 that is further modified by substitution AspB28.
6. 6 An insulin analogue of any one of claim 3, 4, or 5 that is further modified by by an acidic substitution (Aspartic Acid or Glutamic Acid) at position BIO.
7. 7 An insulin analogue of claim 6 that is further modified by removal of residues B1-B3.
8. 8 An insulin analogue of claim 1 that is further modified according to SEQ. ID 35.
9. 9 An insulin analogue of claim 1 that is further modified according to SEQ. ID 36.
10. 10 An insulin analogue of any one of claims 1-9, wherein in addition residue A8 contains one of the following nitrogen-containing non-beta-branched side chains: His, Lys, Arg, or Trp.
11. 11 An insulin analogue of any one of claims 1-9, wherein the analogue is an analogue of a mammalian insulin.
12. 12 A method of treating a patient comprising administering a physiologically effective amount of an insulin analogue or a physiologically acceptable salt thereof to the patient, wherein the insulin analogue is modified by penta-fluoro-Phe at position B24.