Derivatisation of Insulin with Polysaccharides

Abstract

The present invention relates to a composition comprising a population of polysaccharide derivatives of a protein, wherein the protein is insulin or an insulin-like protein and the polysaccharide is anionic and comprises between 2 and 125 saccharide units, and wherein the population consists of substantially only N-terminal derivatives of the protein. Typically, the polysaccharide is PSA. The present invention also relates to pharmaceutical compositions comprising the novel compounds, and methods for making the novel compounds.

Claims

1. A composition comprising a population of polysaccharide derivatives of a protein, wherein the protein is insulin or an insulin-like protein and the polysaccharide is anionic and comprises between 2 and 125 saccharide units, and wherein the population consists substantially only of N-terminal derivatives of the protein.

2. The composition according to claim 1 wherein the polysaccharide is selected from polysialic acid, heparin, hyaluronic acid or chondroitin sulphate.

3. The composition according to claim 2 wherein the polysaccharide is polysialic acid, preferably consisting substantially only of sialic acid units.

4. The composition according to any preceding claim wherein the insulin or insulin-like protein is derivatised by the polysaccharide at the reducing terminal unit of the polysaccharide.

5. The composition according to any of claims 1-3 wherein the polysaccharide derivatives are of general formula (I) ##STR00005## wherein m is at least one; HNB is derived from B—NH.sub.2 which is the N-terminus insulin or an insulin-like protein; L is a bond, a linking group, or comprises a polypeptide or a synthetic digomer; GlyO is an anionic saccharide unit; wherein the linking group, if present, is of general formula —Y—C(O)—R.sup.1—C(O)—; wherein Y is NR.sup.2 or NR.sup.2—NR.sup.2; R.sup.1 is a difunctional organic radical selected from the group consisting of alkanediyl, arylene, alkarylene, heteroarylene and alkylheteroarylene, any of which may substituted and/or interrupted by carbonyl, ester, sulfide, ether, amide and/or amine linkages; and R.sup.2 is H or C.sub.1-6 alkyl.

6. The composition according to claim 5 wherein L is a bond or is a group ##STR00006##

7. A composition according to any preceding claim wherein the polysaccharides comprise 10-80 sialic acid units, more preferably 20-60 sialic acid units, most preferably 40-50 sialic acid units.

8. A composition according to any preceding claim wherein the polydispersity of the anionic polysaccharide is less than 1.3, preferably less than 1.2, most preferably less than 1.1.

9. A composition according to any preceding claim which is a pharmaceutical composition and comprises one or more pharmaceutically acceptable excipients.

10. A composition according to any of claims 1-8 for use in therapy.

Description

[0083] The invention is illustrated by Examples 1-6 and by reference to the following drawings:

[0084] FIG. 1 shows the oxidation of colominic acid (alpha-2,8 linked polysialic acid from E. coli) with sodium periodate;

[0085] FIG. 2 shows the selective reduction of the Schiff's base with sodium cyanoborohydride to form a stable irreversible covalent bond with the protein amino group;

[0086] FIG. 3 is an SDS-PAGE of 22 kDa CAO-rh-Insulin conjugates at different temperatures;

[0087] FIG. 4 shows the effects of temperature on the degree of derivatisation;

[0088] FIG. 5 is an SDS-PAGE of 27 kDa CAO-rh-Insulin conjugations at different molar ratios;

[0089] FIG. 6 shows the effects of molar ratio on the degree of derivatisation;

[0090] FIG. 7 is an SDS-PAGE of 8 and 11 kDa CAO-rh-Insulin conjugates;

[0091] FIG. 8 is an SE-HPLC of 8 kDa CAO-rh-Insulin formulations and insulin;

[0092] FIG. 9 shows the in vivo efficacy of 10, 15 and 21.5 kDa CAO-insulin formulations;

[0093] FIG. 10 shows the experimental set-up of preparative HPLC with IEC or HIC for purifying a conjugate;

[0094] FIG. 11 shows the purification of a CAO-insulin conjugate by hydrophobic interaction chromatography over a HiTrap Butyl FF column;

[0095] FIG. 12 shows the purification of a CAO-insulin conjugate from HIC peak 2, with 13 kDa CAO as example, by anion exchange chromatography over a Hitrap Q FF column column;

[0096] FIG. 13 is an Isoelectric focusing (IEF) gel of a 13 kDa and 27 kDa CAO-insulin conjugate;

[0097] FIG. 14 shows In vivo results of CAO-insulin conjugate in mice; and

[0098] FIG. 15 shows Edman amino acid degradation results.

[0099] In one aspect, disclosed herein is a composition comprising a population of polysaccharide derivatives of a protein, wherein the protein is insulin or an insulin-like protein and the polysaccharide is anionic and comprises between 2 and 125 saccharide units, and wherein the population consists substantially only of N-terminal derivatives of the protein.

[0100] In some embodiments, the polysaccharide is selected from polysialic acid, heparin, hyaluronic acid or chondroitin sulphate.

[0101] In some embodiments, the polysaccharide is polysialic acid, preferably consisting substantially only of sialic acid units.

[0102] In some embodiments, disclosed herein is a composition according to any preceding embodiments wherein the insulin or insulin-like protein is derivatised by the polysaccharide at the reducing terminal unit of the polysaccharide.

[0103] In some embodiments, the polysaccharide derivatives are of general formula (I)

##STR00003##

wherein m is at least one;
HNB is derived from B—NH.sub.2 which is the N-terminus insulin or an insulin-like protein;
L is a bond, a linking group, or comprises a polypeptide or a synthetic digomer;
GlyO is an anionic saccharide unit;
wherein the linking group, if present, is of general formula —Y—C(O)—R.sup.1—C(O)—;
wherein Y is NR.sup.2 or NR.sup.2—NR.sup.2; R.sup.1 is a difunctional organic radical selected from the group consisting of alkanediyl, arylene, alkarylene, heteroarylene and alkylheteroarylene, any of which may substituted and/or interrupted by carbonyl, ester, sulfide, ether, amide and/or amine linkages;
and R.sup.2 is H or C.sub.1-6 alkyl.

[0104] In some embodiments, L is a bond or is a group

##STR00004##

[0105] In some embodiments, the composition according to any preceding embodiment wherein the polysaccharides comprise 10-80 sialic acid units, more preferably 20-60 sialic acid units, most preferably 40-50 sialic acid units.

[0106] In some embodiments, wherein the composition according to any preceding embodiment wherein the polydispersity of the anionic polysaccharide is less than 1.3, preferably less than 1.2, most preferably less than 1.1.

[0107] In some embodiments, wherein the composition according to any preceding embodiment which is a pharmaceutical composition and comprises one or more pharmaceutically acceptable excipients.

[0108] In some embodiments, wherein the composition according to any preceding embodiment for use in therapy.

[0109] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein an anionic polysaccharide comprising 2-125 saccharide units, is chemically reacted substantially only at the N-terminal amine of the insulin or insulin-like protein.

[0110] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the anionic polysaccharide has a reactive aldehyde group which reacts with the insulin or insulin-like protein and the derivatisation reaction is carried out under reducing conditions.

[0111] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the reactive aldehyde group is at the non-reducing end of the polysaccharide.

[0112] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the reactive aldehyde is at the reducing end of the polysaccharide and the non-reducing end has been passivated such that it does not react with the insulin or insulin-like protein.

[0113] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the polysaccharide reacts with an amine group of the insulin or insulin-like protein.

[0114] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein where the amine is a terminal amine group.

[0115] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the amine is derived from a lysine amino acid of the insulin or insulin-like protein.

[0116] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the anionic polysaccharide has a reactive aldehyde group which is converted in a preliminary reaction step into an amine, which is then reacted with a bifunctional reagent comprising at least one functional group selected from N-maleimide, vinylsulphone, N-iodoacetamide, orthopyridyl group or N-hydroxysuccinimide, to form a reaction intermediate, wherein the reaction intermediate is reacted with the insulin or insulin-like protein.

[0117] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the anionic polysaccharide or reaction intermediate reacts with a terminal amine group of the insulin or insulin-like protein in a first aqueous solution of acidic pH; and the resultant polysaccharide derivative is purified in a second aqueous solution of higher pH than the first aqueous solution.

[0118] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the pH of the first aqueous solution is in the range 4.0-6.0 and the pH of the second aqueous solution is in the range 6.5-8.5.

[0119] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein which is carried out in the presence of a formulation additive.

[0120] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is selected from one or more buffers, stabilisers, surfactants, salts, polymers, metal ions, sugars, polyols or amino acids.

[0121] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is sorbitol, trehalose or sucrose.

[0122] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is a non-ionic surfactant.

[0123] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is a polymer selected from PSA, PEG or hydroxy-beta-cyclodextrin.

[0124] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is a divalent metal ion, preferably Zn.sup.2+, Ni.sup.2+, Co.sup.2+, Sr.sup.2+, Fe.sup.2+, Ca.sup.2+ or Mg.sup.2+.

[0125] In some embodiments, wherein the method for producing a polysaccharide derivative of insulin or of an insulin-like protein wherein the formulation additive is a buffer and the buffer is sodium phosphate.

28. The polysaccharide derivative of insulin or of an insulin-like protein obtainable by a method according to the preceding embodiments.

EXAMPLES

1. Protein and Colominic Acid Determination

[0126] Quantitative estimation of polysialic acids (as sialic acid) with the resorcinol reagent was carried out by the resorcinol method [Svennerholm, 1957] as described elsewhere [Gregoriadis et al., 1993; Fernandes and Gregoriadis, 1996, 1997]. Protein was measured by the BCA colorimetric method or UV absorbance at 280 nm.

2. Activation of Colominic Acid

[0127] Freshly prepared 0.02 M sodium metaperiodate (NalO4) solution (8 fold molar excess) was mixed with CA at 20° C. and the reaction mixture was stirred magnetically for 15 min in the dark. A two-fold volume of ethylene glycol was then added to the reaction mixture to expend excess NalO.sub.4 and the mixture left to stir at 20° C. for a further 30 min. The oxidised colominic acid (CAO) was dialysed (3.5 kDa molecular weight cut off dialysis tubing) extensively (24 h) against a 0.01% ammonium carbonate buffer (pH 7.4) at 4° C. Ultrafiltration (over molecular weight cut off 3.5 kDa) was used to concentrate the CAO solution from the dialysis tubing. Following concentration to required volume, the filterate was lyophilized and stored at −40° C. until further use. Alternatively, CAO was recovered from the reaction mixture by precipitation (twice) with ethanol.

3. Determination of the Oxidation State of CA and Derivatives

[0128] Qualitative estimation of the degree of colominic acid oxidation was carried out with 2,4 dinitrophenylhydrazine (2,4-DNPH), which yields sparingly soluble 2,4 dinitrophenyl-hydrazones on interaction with carbonyl compounds. Non-oxidised (CA)/oxidised (CAO) were added to the 2,4-DNPH reagent (1.0 ml), the solutions were shaken and then allowed to stand at 37° C. until a crystalline precipitate was observed [Shriner et. al., 1980]. The degree (quantitative) of CA oxidation was measured with a method [Park and Johnson, 1949] based on the reduction of ferricyanide ions in alkaline solution to ferric ferrocyanide (Persian blue), which is then measured at 630 nm. In this instance, glucose was used as a standard.

4. Gel Permeation Chromatography

[0129] Colominic acid samples (CA and CAO) were dissolved in NaNO.sub.3 (0.2M), CH.sub.3CN (10%; 5 mg/ml) and were chromatographed on over 2×GMPWXL columns with detection by refractive index (GPC system: VE1121 GPC solvent pump, VE3580 RI detector and collation with Trisec 3 software (Viscotek Europe Ltd). Samples (5 mg/ml) were filtered over 0.45 μm nylon membrane and run at 0.7 cm/min with 0.2M NaNO.sub.3 and CH.sub.3CN (10%) as the mobile phase.

5. Colominic Acid Stability

[0130] The rules for chemistry of the PEGylation cannot be applied to polysialylation as such because of the difference in the physiochemical properties of these molecules. PSA is an acid labile polymer and is stable for weeks around neutral pH (FIG. 3). The results in FIG. 3 show that at pH 6.0 and 7.4 CAO is stable for 8 days, at pH 5.0 there is slow degradation (after 48 hours 92% of initial MW), and at pH 4.0 there is slow degradation (after 48 hours 70% of initial MW). Polysialic acid is highly hydrophilic whereas PEG is an amphiphilic molecule in nature. When the polysialylation is carried out using conditions used for PEGylation, aggregation and precipitation of the proteins is seen in many cases.

6. Preparation of N-Terminal Protein-CA Conjugates with Formulation Additives

6.1 Preparation of Insulin-CA Conjugates (N-Terminal Method)

[0131] Insulin (5804 Da) was supplied as white solid. The insulin was dissolved by minimum 100 mM HCl, and then adjusted to the required the pH and placed on ice. The amount of CAO to be added for conjugation was calculated based on formula:

[00001]$\mathrm{Weight}\mathrm{of}\mathrm{CAO}=\frac{\mathrm{Amount}\mathrm{of}\mathrm{protein}\left(g\right)}{\left(\mathrm{MW}\mathrm{of}\mathrm{protein}\right)}\times \mathrm{MW}\mathrm{of}\mathrm{CAO}\times \left(\mathrm{Molar}\mathrm{excess}\mathrm{of}\mathrm{CAO}\right)$

[0132] Required amount of CAO was weighed out. CAO was solubilised in 10 mM NaOAc, pH 6.0 gently vortexed the mixture until all the CAO has dissolved and then either filtered into a new container to remove any aggregated/precipitated material. Required amount of insulin protein solution was added to the CAO solution to give a 7.5 molar excess (small scale) and 5 (large scale) of CAO and gently mixed by keeping the reaction mixture on a gentle shaker at 4±1° C. 100 mg/ml NaCNBH.sub.3 solution was added in order to have 8 mg/ml in the final reaction mixture, gently mixed and pH of the final reaction mixture was checked, if necessary adjusted the pH to 6.0 with 0.5 M NaOH/HCl at 4±100. Finally adjusted the volume of the reaction using 10 mM NaOAc, pH 6.0 to give a protein concentration of 1 mg/ml in the reaction mixture. Tube was sealed and stirred at desired temperature (4±1° C.) for 24 hours. The reaction was stopped by an appropriate method (such as tris(hydroxymethyl)aminomethane buffer pH 7.4) and samples were taken out for SDS-PAGE (using 18% Tris glycine gel), SE-HPLC (superose 12 column) and checked the pH of reaction mixture. To eliminate any precipitate the reaction mixture was centrifuged at 13000 rpm for 5 min before SE-HPLC analysis and purification, preferred buffer for SE-HPLC was 0.1 M Na phosphate (pH 6.9).

6.2 Optimisation

[0133] Reductive amination was performed with a range of molecular weights of CAO (10-30 kDa) on insulin for N-terminal and random derivatisation. Range of process variables were studied for conjugation reactions: CAO 10-20 (small scale) and 5-10 (large scale) molar excess; reagent=50-100 mM NaCNBH.sub.3 reaction buffer=10 mM NaOAc pH 5.5-6.5, temperature=4±1° C., time=16-24 hours etc.

[0134] Optimised reaction conditions were found to be as following: CAO=7.5 (small scale) and 5 (large scale) molar excess, reagent=50 mM NaCNBH.sub.3, Reaction buffer=10 mM NaOAc pH 5.5, temperature=4±1° C., time=24 hours.

6.3 Purification and Characterization of Insulin-CA Conjugates (N-Terminal Method)

[0135] To remove free CAO from the mixture, HIC (HiTrap Butyl FF) was used. Prepare loading solution by diluting the insulin reaction mixture with minimum volume using concentrated (NH.sub.4).sub.2SO.sub.4 (e.g. 3 M), 20 mM Na.sub.2HPO.sub.4 (pH 7.4) to give a concentration of 0.8 M (NH.sub.4).sub.2SO.sub.4 in the loading solution. Check pH, should be 7.4 or adjust with 0.5 M HCl/NaOH, the loading solution need be filtered with 0.2 μm membrane filter.

[0136] This solution is then loaded on to the HIC column (rate=0.5 ml/min) previously equilibrated with HIC buffer B (20 mM sodium phosphate+0.8 M (NH.sub.4).sub.2SO.sub.4, pH 7.4). The loading fractions was collected (each fraction 1.5 column volume) and label (L1-Lx) following with washing column with HIC buffer B (at least 5 column volumes; rate=0.5 ml/min; 1.5 column volume fraction) was collected and labelled (W1-Wx). Product with HIC buffer A (10 mM sodium phosphate buffer, pH 7.4) (rate=5 ml/min) was eluted and collected the fractions (1 column volume fraction; 6 column volume) and label (E1-Ex). If two consecutive fractions were absent in protein content (UV280 nm), the next step was carried out. The samples were kept on ice during purification. The protein concentration was analysed by UV (280 nm) (Extinction coefficient of 1 mg/ml of insulin was about 1.043 at 280 nm). The samples were taken for SDS-PAGE and SE-HPLC.

[0137] HIC fractions containing protein fractions are washed with IEC buffer A (20 mM phosphate buffer, pH 7.4). To remove ammonium sulphate if any in Vivaspin 20 (MW: 5 Kd). Check pH and adjust if required to pH 7.4. Load on the IEC column previously equilibrated with IEC buffer A. A gradient system was applied in the following manner:

Loading: 0.25 ml/min of injected sample in IEC buffer A, washing of 3CV
Washing: Gradient system: IEC buffer A: 90%, AEX buffer B (20 mM phosphate buffer+1M NaCl, pH 7.4): 10%, gradient of 5CV & wash of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 68%, IEC buffer B: 32%, gradient of 5 CV & washing of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 35%, 1 EC buffer B: 65%, gradient of 5CV & washing of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 0%, IEC buffer B: 100%, gradient of 5CV & washing of 3CV, flow rate: 0.25 ml/min.

[0138] The IEC fractions containing the purified conjugate are combined, washed to remove salt with buffer change of PBS buffer. Adjust pH after removing salt to 7.4. The solution is then concentrated at 4±1° C. and the protein concentration analysed by UV spectroscopy (280 nm). Conjugates were sterile filtered and samples taken for activity assay and for characterisation by SDS-PAGE and SE-H PLC. If required an aliquot was removed for a protein assay and CA assay. The remainder was stored at 4±1° C. until further use and studied for physical stability by SE-HPLC.

[0139] The effects of various processes affecting the stability of insulin in solution and the degree of derivatization were studied.

6.4 Preparation of Insulin-14 kDaCA Conjugates (Monodisperse)

[0140] Insulin (5808 Da) was supplied as white solid. The insulin was dissolved by adding minimum quantity 100 mM HCl, and then adjusted to the required the pH and placed on ice. The amount of 14 kDa CA to be added for conjugation was calculated based on formula:

[00002]$\mathrm{Weight}\mathrm{of}14\mathrm{kDa}\mathrm{CAO}=\frac{\mathrm{Amount}\mathrm{of}\mathrm{protein}\left(g\right)}{\left(\mathrm{MW}\mathrm{of}\mathrm{protein}\right)}\times \mathrm{MW}\mathrm{of}\mathrm{CAO}\times \left(\mathrm{Molar}\mathrm{excess}\mathrm{of}\mathrm{CAO}\right)$

[0141] Required amount of 14 kDa CAO was weighed out. 14 kDa CAO was solubilised in 10 mM phosphate buffer, pH 6.0 (20% volume of the final reaction volume was used here), gently vortexed the mixture until all the 14 kDa CAO has dissolved and then either filtered into a new container to remove any aggregated/precipitated material. Required amount of insulin protein solution was added to the 14 kDa CAO solution to give a 7.5 molar excess (small scale) and 5 (large scale) of 14 kDa CAO and gently mixed by keeping the reaction mixture on a gentle shaker at 4±1° C. 100 mg/ml NaCNBH.sub.3 solution was added in order to have 63.5 mM or 4 mg/ml in the final reaction mixture, gently mixed and pH of the final reaction mixture was checked, if necessary adjusted the pH to 6.0 with 0.5 M NaOH/HCl at 4±100. Finally adjusted the volume of the reaction using 10 mM NaOAc, pH 6.0 to give a protein concentration of 1 mg/ml in the reaction mixture. Tube was sealed and stirred at desired temperature (4±100) for 24 hours. The reaction was stopped by an appropriate method and samples were taken out for SDS-PAGE (using 18% Tris glycine gel), SE-HPLC (superose 12 column) and checked the pH of reaction mixture. To eliminate any precipitate the reaction mixture was centrifuged at 13000 rpm for 5 min before SE-HPLC analysis and purification, preferred buffer for SE-HPLC was 0.1 M Na phosphate (pH 6.9).

6.5 Optimisation

[0142] Reductive amination was performed with a range of molecular weights of CA (10-30 kDa) on insulin for N-terminal and random derivatisation. Range of process variables were studied for conjugation reactions: CAO 10-20 (small scale) and 5-10 (large scale) molar excess; reagent=50-100 mM NaCNBH.sub.3 reaction buffer=10 mM phosphate buffer, pH=5-7.4; temperature=4-37±1° C., time=16-24 hours etc.

[0143] Optimised reaction conditions were found to be as following: CAO=7.5 (small scale) and 5 (large scale) molar excess, reagent=63.5 mM NaCNBH.sub.3 (4 mg/ml). Reaction buffer=10 mM NaOAc pH 6.0, temperature=4±1° C., time=24 hours.

6.6 Purification and Characterization of Insulin-CA Conjugates (N-Terminal Method)

[0144] To remove free CAO from the mixture, HIC was used. Prepare loading solution by diluting the insulin reaction mixture with minimum volume using concentrated (NH.sub.4).sub.2SO.sub.4 e.g. 3 M), 20 mM Na.sub.2HPO.sub.4, (pH 7.4) to give a concentration of 0.8 M (NH.sub.4).sub.2SO.sub.4 in the loading solution. Check pH, should be 7.4 or adjust with 0.5 M HCl/NaOH, the loading solution need be filtered with 0.2 mm membrane filter.

[0145] This solution is then loaded on to the HIC column (rate=0.5 ml/min) previously equilibrated with HIC buffer B (20 mM sodium phosphate+0.8 M (NH.sub.4).sub.2SO.sub.4, pH 7.4). Collect the loading fractions (each fraction 1.5 column volume) and label (L1-Lx). Wash column with HIC buffer B (at least 5 column volumes; rate=0.5 ml/min; collect 1.5 column volume fraction) collect fractions and label (W1-Wx). Elute the product with HIC buffer A (10 mM sodium phosphate buffer, pH 7.4) (rate=5 ml/min); collect the fractions (1 column volume fraction; 6 column volume) and label (E1-Ex). If two consecutive fractions were absent in protein content (UV280 nm), the next step was carried out. The samples were kept on ice during purification. The protein concentration was analysed by UV (280 nm) (Extinction coefficient of 1 mg/ml of insulin was about 1.043 at 280 nm). The samples were taken for SDS-PAGE and SE-HPLC.

[0146] HIC fractions containing protein fractions are washed with IEC buffer A (20 mM phosphate buffer, pH 7.4). To remove ammonium sulphate if any in Vivaspin 20 (MW: 5 Kd). Check pH and adjust if required to pH 7.4. Load on the IEC column previously equilibrated with IEC buffer A. The gradient system was applied in the following manner:

Loading: 0.25 ml/min of injected sample in IEC buffer A, washing of 3CV
Washing: Gradient system: IEC buffer A: 90%, AEX buffer B (20 mM phosphate buffer+1M NaCl, pH 7.4): 10%, gradient of 5CV & wash of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 68%, IEC buffer B: 32%, gradient of 5 CV & washing of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 35%, 1 EC buffer B: 65%, gradient of 5CV & washing of 3CV, flow rate: 0.25 ml/min
IEC buffer A: 0%, IEC buffer B: 100%, gradient of 5CV & washing of 3CV, flow rate: 0.25 ml/min.

[0147] The IEC fractions containing the purified conjugate are combined, washed to remove salt with buffer change of PBS buffer. The pH is adjusted after removing salt to 7.4. The solution is then concentrated at 4±1° C. and the protein concentration analysed by UV spectroscopy (280 nm). Conjugate were sterile filtered and samples taken for activity assay and for characterisation by SDS-PAGE and SE-HPLC. If required an aliquot was removed for a protein assay and CA assay. The remainder was stored at 4±1° C. until further use and studied for physical stability by SE-HPLC.

[0148] The effects of various processes affecting the stability of insulin in solution and the degree of derivatization were studied.

6.7 SE-HPLC of Insulin Formulations

[0149] HPLC was performed on a Liquid Chromatograph (JASCO) equipped with a Jasco, AS-2057 plus autosampler refrigerated at 4° C., and a Jasco UV-975 UV/VIS detector. Data was recorded by EZchrom Elite software on an IBM/PC. The SEC samples were analysed with an isocratic mobile phase of 0.1 M Na phosphate, pH 6.9; on a Superose 12 column (FIG. 8). FIG. 8 shows just one peak at RT=75.408, which was attributed to insulin.

6.8 SDS Polyacrylamide Gel Electrophoresis & Western Blotting

[0150] SDS-PAGE was performed using 18% triglyine gels. Samples were diluted with either reducing or non reducing buffer and 5.0 ug of protein was loaded into each well. The gels were run on a triglycerine buffer system and was stained with Coomasie Blue (FIGS. 5 & 7). Western blotting was performed using anti PSA antibody. FIG. 4 shows the SDS-PAGE of insulin formulations (site-specific; N-terminal).

6.9 Isoelectric Focusing (IEF) Gel of 27 and 13 kDa CAO-Insulin

[0151] NOVEX® IEF gel was used to determine differences in insulin and CAO-insulin conjugate isoelectric points. Samples was dissolved to a concentration of 0.5 mg/ml. 5 ul sample was diluted with 5 ul NOVEX® IEF Sample Buffer pH 3-10 and then loaded the protein sample on the gel.

6.10 Stability Studies

[0152] Sterile insulin conjugates were stored in PBS buffer; at 4° C. for six weeks. Native-PAGE of the samples was performed every week

6.11 In Vivo Efficacy of Insulin Formulations

[0153] The in vivo efficacy of insulin formulations was studied in female mice CD-1, 7-8 weeks old, 0.3 IU of protein dose (same activity) was injected in mice subcutaneously. Animals were divided into seven groups of four: insulin formulations were given to each animal of each group in the following manner; insulin (0.3 IU/mouse), Lantus (Aventis) insulin-PSA conjugate (14 kDa), PBS. One drop of blood was taken from each animal and was measured blood glucose by ACCU-CHEK Active (Roche Diagnostics).

Results

Activation of CA and Determination of Degree of Oxidation

[0154] Colominic acid (CA) is a linear alpha-2.8-linked homopolymer of N-acetylneuraminic acid (Neu5Ac) residues was used. Exposure of colominic acids to oxidation was carried out for 15 min using 20 mM periodate at room temperature. The integrity of the internal alpha-2.8 linked Neu5Ac residues post periodate treatment was analysed by gel permeation chromatography and the chromatographs obtained for the oxidised (CAO), material was compared with that of native CA. It was found that oxidized and native CA exhibit almost identical elution profiles, with no evidence that the successive oxidation step give rise to significant fragmentation of the polymer chain.

[0155] Quantitative measurement of the oxidation state of CA was performed by ferricyanide ion reduction in alkaline solution to ferrocyanide (Prussian Blue) [Park and Johnson, 1949] using glucose as a standard. It shows that the oxidized colominic acid was found to have a greater than stoichiometric (>100%) amount of reducing agent, i.e. 112 mol % of apparent aldehyde content comprising the combined reducing power of the reducing end hemiketal and the introduced aldehyde (at the other end).

Optimisation of Polysialation Reaction

[0156] The polysialation conditions were optimized using 1 mg rh-insuin with CAO under variable temperatures and reactant molar ratio, and chain length. The results are shown in FIGS. 5-12. 4° C. would appear to be the optimum temperature for CAO and insulin stability during the reaction. However, at a higher temperature, more conjugation can be attained 7.5:1 of molar ratio of CAO to insulin is more effective.

TABLE-US-00001 TABLE 1 shows the effect of molar ratio on polysialylation. Well Number 1 3 5 7 8 11 12 Molar ratio 1:1 1:2 1:4 Ma insulin 1:7.5 1:10 CAO MW 27 kDa CAO with 100% CHO % Insulin (mg) 1 1 1 1 1 CAO (mg) 4.49 8.98 17.96 33.675 44.9 Final NaCNBH.sub.3 4 mg/ml Reaction 4 pH 6

[0157] Under the influence of PBS control, 21.5 kd CAO-insulin has the most significant effect on the decreasing blood glucose in mice among conjugates used in FIG. 9.

TABLE-US-00002 TABLE 2 shows a T-Test (statistical analysis, paired-test) of different chain CAO-insulin conjugates in vivo efficiency. Time (minutes) 0.0 32.0 64.5 99.4 170.8 230.2 303.3 358.0 1354.9 13 kDa ** *** CAO- insulin 21 kDa *** *** * CAO- insulin 27 kDa ** *** ** ** ** CAO- insulin 32 kDa *** *** *** * CAO- insulin insulin *** *** *** Asterisks indicate probability level of the difference between groups against the Tris buffer: *P < 0.05, **P < 0.01, ***P < 0.001

TABLE-US-00003 TABLE 3 Lantus 15 kDa CAO-insulin (Sigma) 0 30 *** ** 60 *** *** 90 ** *** 120 * ** 150 * 210 270 330 Asterisks indicate probability level of the difference between groups against the PBS buffer: * P < 0.05, ** P < 0.01, *** P < 0.001

[0158] The pH effect on in vivo efficiency of 15 kDa CAO-insulin also revealed that pH 6.0 is better than 7.4. Therefore, the afterward experiment was performed at pH 6.0. The data from peptide mapping and Edman degradation further confirmed that the conjugate from pH 6.0 polysialation condition is N-terminally specifically blocked at the B chain of insulin.

Preparation, Purification and Characterisation of Insulin Conjugates

[0159] Monodisperse CAO-insulin can be successfully conjugated and highly pure conjugates were purified by scale-up HIC and IEC. The purification efficiency was improved from the IEC and HIC combination set-up with preparative HPLC instrument as demonstrated in FIG. 10.

[0160] The procedure to prepare and purify colominic acid (CA) conjugates of insulin in an N-terminally selective manner by conducting the reaction at a reduced pH (pH 6.0) and at 4±1° C. is detailed above. This involves conjugation in the presence of sodium cyanoborohydride, followed by purification using hydrophobic interaction chromatography (HIC) to remove free CA (FIG. 11) followed by removal of insulin by ion-exchange chromatography (IEC) (FIG. 12). The low pH was used to favour selective derivatisation at the N-terminus of insulin's B-chain (PheB1), and also in order to minimise aggregation of insulin during the reaction. The composition of the final reaction buffer was 1 mg/ml insulin, 8 mg/ml NaCNBH.sub.3 and 5 molar excess CAO in 10 mM NaOAc at pH 6.0.

[0161] Isoelectric focusing (IEF) gels of 13 kDa and 27 kDa CAO-insulin conjugates in FIG. 13 show the conjugate of polysialylated insulin has no fixed isoelectric point (p1).

[0162] Formation of the insulin-CA conjugates and stability was confirmed by the SE-HPLC (change of retention time of insulin-PSA as compared to insulin; also co-elution of both moieties); ion exchange chromatography (binding of conjugates on to the IEC column) and polyacrylamide gel electrophoresis (SDS-PAGE; shifting of bands with high m.w. species).

[0163] Insulin conjugates used in the in vivo efficiency (on CD-1 mice, average 25 g) showed superior in vivo efficacy (prolongation) and blunt (peak less profile) as compared to insulin. The prolongation of the blood glucose reducing ability of the conjugates was proportional to chain length of polymer used in the formulation as seen in FIG. 14.

[0164] The stability study of 27 kDa CAO-insulin conjugates showed that there was no any degradation during 40 days' 4° C. storage based on the observation from Native-PAGE.

Characterisation of 14 kDa-Insulin Conjugates from Scale-Up Preparation and Purification

[0165] In the example of insulin as 700 mg start reactant, 230 mg 14 kDa CAO-insulin (protein mass weight) was yielded using scale up column purification.

[0166] By comparing chromatograms, changes in amino acid amounts in FIG. 15 reveal the amino acids from N-terminus. The 1 mg/mL aqueous sample in PBS pH 7.4 was diluted×100 with water. 2 μL of this diluted solution was taken for analysis. In conclusion, the sequence G-I-V-E, identifies the A chain of insulin. The absence of amino acid Phe/Val/Asn/Gln indicated that the B chain of insulin was N-terminally blocked.

[0167] The PSA conjugates were found to be active in the in vitro activity assay. In vivo efficacy study shows that PSA-insulin conjugates are vastly superior to insulin.