Synthesis and Evaluation of Insulin−Human Serum Albumin Conjugates

A series of human insulin maleimido derivatives with short and long linkers was ... binding and activation results for this series of insulin-HSA conj...
0 downloads 0 Views 284KB Size
Download PDF
Bioconjugate Chem. 2005, 16, 1000−1008

1000

Synthesis and Evaluation of Insulin-Human Serum Albumin Conjugates Karen Thibaudeau, Roger Le´ger,* Xicai Huang, Martin Robitaille, Omar Quraishi, Chantal Soucy, Nathalie Bousquet-Gagnon, Pieter van Wyk, Ve´ronique Paradis, Jean-Paul Castaigne, and Dominique Bridon ConjuChem Inc., 225 President-Kennedy Avenue, Montreal, QC, H2X 3Y8 Canada. Received April 1, 2005; Revised Manuscript Received May 18, 2005

A series of human insulin maleimido derivatives with short and long linkers was synthesized by exploiting the variations in the pKa values and environment of the three amino groups present in the protein. The syntheses were accomplished in organic solvent because of maleimide’s instability in basic aqueous media. The derivatives thus obtained were conjugated to the free thiol on Cys34 of human serum albumin (HSA) and purified. A structure-activity relationship based on in vitro receptor binding and activation results for this series of insulin-HSA conjugates showed that the best compounds were attached at the B1 position of insulin with either short or long linkers. Two conjugates were administered subcutaneously to streptozotocin-induced diabetic rats and found to possess blood glucose normalizing activity up to 8 h postadministration. The return to diabetic plasma glucose levels was not observed within the time frame of the experiment (48 h). In comparison, the insulin-treated group’s normalization activity lasted 2 h and returned to a diabetic level at 8 h. The onset of the conjugate activities were delayed by 1 h when compared to the activity of human insulin. The study results led to the identification of CJC-1575 as a potent and long lasting human insulin analogue.

INTRODUCTION

Insulin is a vital endocrine hormone that binds to its cellular surface receptor, setting off a cascade of events culminating in glucose absorption from the blood. Low plasma levels of insulin lead to severe disorders such as types I and II diabetes. Type I diabetes is a life threatening disease where the patient must self-administer quick acting insulin for survival. In Type II diabetes, the therapeutic objective is glycemic control to reduce the onset of long-term pathological consequences. The treatment with insulin becomes necessary after the failure of lifestyle changes or glucose lowering drugs. A long lasting form of insulin will lead to better glycemic control and reduce the amount of injections, resulting in increased patient appreciation and compliance (1-3). There are several known “long lasting” insulin drugs in various stages of development or available on the market. Some examples include slow release from the injection site (4), noncovalent binding to blood proteins through lipophilic interactions (5-12), PEGylation (1314), or conjugation to a polymeric support (15). Alternatively, the conjugation of a maleimido derivative of a peptide to Cys34 human serum albumin (HSA) can prolong the presence of the bioactive species in plasma by protecting it against elimination through metabolic or excretion pathways (16-21). We became interested in the application of this methodology to human insulin, 1 (Figure 1), to demonstrate how a small protein can be coupled regioselectively to human serum albumin (HSA) and evaluate the new conjugate’s properties. An insulin-HSA fusion protein prepared through genetic engineering has recently been reported (22). The * To whom correspondence should be addressed. ConjuChem Inc, Research Department, 225 President-Kennedy Ave, Suite 3950, Montreal QC, H2X 3Y8 Canada. Phone: 514-844-5558. FAX: 514-844-1119. E-mail: [email protected].

fusion protein has the insulin linked to the N-terminus of HSA and the two A and B chains tethered together by a peptidic chain. Edman degradation analysis identified the B chain amino acids sequence, indicating that the attachment of the insulin to HSA was at the C-terminus of the A chain. Although the resulting protein was bioactive, there was no further discussion of a structureactivity relationship (SAR) in the study. The product of conjugation of insulin to transferrin was also reported (23). In that study, the insulin was linked to transferrin by a disulfide-containing linking group attached to exposed lysine residues on both molecules. This nonregioselective technique can potentially cause important variations in biological responses due to a possible lack of batch-to-batch reproducibility. Insulin contains two peptide chains with three disulfide bonds and three free amines. Hence, the challenge was to use insulin as a starting material and selectively attach a single maleimido group to each of the available amine sites. We report herein the synthesis of maleimido derivatives of human insulin, their conjugation to HSA, and the in vitro and in vivo biological evaluation of the conjugates. EXPERIMENTAL PROCEDURES

Recombinant human insulin was obtained from ICN Chemicals as a 0.4% zinc salt. Flash column chromatography was carried out by using Biotage “40i flash chromatography” modular system. Unless otherwise specified, semipreparative HPLC purifications were done on a Waters “Breeze” system 1500 series using a Phenomenex luna (RP-18, 10 µ phenyl-hexyl 250 × 21.2 mm) column. The flow rate was 9.5 mL/min, and 9.5 mL fractions were collected. A gradient of acetonitrile (0.1%TFA) in water (0.1%TFA) was used with further details indicated within each compound’s synthetic pro-

10.1021/bc050102k CCC: $30.25 © 2005 American Chemical Society Published on Web 06/25/2005

Synthetic Insulin−HSA Conjugates

Bioconjugate Chem., Vol. 16, No. 4, 2005 1001

Figure 1. Maleimide derivatives of human insulin (1).

cedure. Outflow was monitored by UV absorbance at 214 and 254 nm. Fractions containing the desired product profile were identified by mass spectrometry. Preparative scale purification was done on a Gilson 690 system using a Phenomenex luna column (RP-18, 10 µ phenyl-hexyl 250 × 50.0 mm) using an acetonitrile (0.1% TFA)/water (0.1% TFA) as above. The flow rate was 50 mL/min, and 50 mL fractions were collected. Outflow was monitored by UV absorbance at 214 and 254 nm. Fractions containing the desired product profile were identified by mass spectrometry. LC/MS was performed using an Agilent 1100 series LC-MSD with a single quadrupole and an ESI electrospray source. The liquid chromatography part was done using a Vydac C18 (4.6 × 250 mm) column with a gradient of acetonitrile (0.1%TFA) in water (0.1%TFA) of 20-80% over 20 min. Automated Edman degradation (Applied Biosystems Procise 492 protein sequencer) and high-resolution mass spectroscopy analyses were done at the Sheldon Institute (Genome Quebec) of McGill University (Montreal, QC, Canada). Proton nuclear magnetic resonance spectra were gathered on a Varian 300 MHz Gemini instrument. SDS-PAGE electrophoresis was done on a SE 250 Mighty Small II (Pharmacia/Hoefer) using low range SDS-PAGE Molecular weight standards from BioRad (cat. # 161-0304). N-Succinimidyl 3-Maleimidopropanoate, MPAOSu, 8. To 3-maleimidopropionic acid, 7 (20.5 g), Nhydroxysuccinimide (30.7 g), and N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC, 51.1 g) in dichloromethane (300 mL) was added N-methylmorpholine (133 mL), and the mixture was stirred for 16 h at ambient temperature. The solvent was evaporated and the residue redissolved in dichloromethane (2 L). The organic layer was washed with 0.1 N HCl and brine and dried (Na2SO4). The solution was passed through a glass funnel filled with silica gel (1 kg). The silica gel was rinsed with dichloromethane (2 L) followed by a 1:1 dichloromethane/ ethyl acetate mixture (4 L). The fractions containing

product were combined, and the solvent was removed to give a white powder (23.7 g, 73%) (24). MS (m/z) 267 (M + 1). 1H NMR (300 MHz, CDCl3) δ 2.78 (s, 4H), 2.98 (t, J ) 7 Hz, 2H), 3.89 (t, J ) 7 Hz, 3H), 6.70 (s, 2H). 8-(3-Maleimidopropionamido)octanoic Acid, MPAOA, 9. 8 (6.02 g) was added to a solution of 8-aminooctanoic acid hydrochloride salt (3.95 g) in N,N-dimethylformamide (50 mL) in the presence of N-methylmorpholine (24 mL). The reaction mixture was stirred at ambient temperature for 16 h. The solvent was removed under vacuum. The residue was acidified by addition of 1 N HCl and the product extracted with ethyl acetate (2 × 500 mL). The combined extracts were washed with water and brine and dried (Na2SO4). The solvent was removed under reduced pressure, and the crude product (7.14 g) was used in the next step. 4-Nitrophenyl 8-(3-Maleimidopropionamido)octanoate, MPA-OA-OPNP, 10. Crude 9 (4.56 g) was dissolved in dichloromethane (100 mL). N-Methylmorpholine (3.2 mL) and p-nitrophenyl chloroformate (2.96 g) were added successively, and the reaction was stirred for 45 min at ambient temperature. The solvent was evaporated under reduced pressure, and the residue was taken up into dicholoromethane and divided into three portions. Each portion was injected into a Biotage flash column (40M) eluting with dichloromethane. The fractions containing product were combined and concentrated to give the desired product (4.2 g) as a white solid. Further purification was accomplished by recrystallization from ethyl acetate/hexane to give the desired product (3.14 g, 50%) as a white solid. MS (m/z) 432 (M + 1). Mp 149-150 °C. Thin-layer chromatography Rf ) 0.31 (CH2Cl2/EtOAc 1:1). 1H NMR (300 MHz, CDCl3) δ 1.2-1.5 (m, 8H), 1.64-1.8 (m, 2H), 2.48 (t, J ) 7 Hz, 2H), 2.57 (t, J ) 7 Hz, 2H), 3.18 (m, 2H), 3.81 (t, J ) 7 Hz, 2H), 5.50 (bs, 1H), 6.67 (s, 2H), 7.23 (m, 2H), 8.23 (m, 2H). (A1)-MPA-insulin, 2. 1 (100 mg) was dissolved in N,N-dimethylformamide (2 mL) with trifluoroacetic acid

1002 Bioconjugate Chem., Vol. 16, No. 4, 2005

(100 µL). N-Methylmorpholine (200 µL) and 8 (9.2 mg, 2.5 equiv) were added, and the reaction was stirred for 2 h. The reaction was quenched by addition of water and adjusted to pH 4 with acetic acid. The precipitate was dissolved with water/acetonitrile (3:1, total volume 20 mL). The resulting solution was injected into semipreparative HPLC using a 27 to 31% acetonitrile gradient over 120 min. The fractions at tR ) 36-46 min were combined and lyophilized to give 2 as a white powder (40 mg) along with recovered 1 (23 mg). LC/MS tR ) 15.1 min. (m/z) 5958.0 (M + 4), UV (214 nm) purity >99%. HRMS calculated for C264H388N66O80S6 + 4H+ 5958.696, found 5958.697. (A1B29)-Bis-Boc-insulin, 11. 1 (1.0 g) was dissolved in anhydrous dimethyl sulfoxide (10 mL) and triethylamine (1.0 mL) under sonication. Boc2O (84 mg, 2.2 equivalents) in dimethyl sulfoxide (0.5 mL) was added to the solution, and the reaction was stirred at ambient temperature for 30 min. The reaction was quenched by addition of acetic acid (14 mL) and water (100 mL). The resulting solution was injected into preparative HPLC using a 27 to 40% acetonitrile gradient over 60 min. The fractions at tR ) 38-47 min. were combined and lyophilized to give 11 as a white powder (674 mg, 65%). MS (m/z) 6007.8 (M + 4). HRMS calcd for C267H399N65O81S6 + 3H+ 6006.766, found 6006.804. (B1)-MPA-insulin, 3. 11 (51 mg) in N,N-dimethylformamide (3 mL) was treated with 8 (36 mg) in the presence of triethylamine (30 µL). The reaction was stirred for 2 h at ambient temperature. The solvent was removed under vacuum, and TFA (2 mL) added to the residue and stirred for 10 min. The TFA was evaporated, and the crude product was dissolved in water/acetonitrile (3:1). The resulting solution was injected into semipreparative HPLC using a 27 to 32% acetonitrile gradient over 120 min. The fractions at tR ) 58-59 min were combined and lyophilized to give 3 as a white powder (29 mg, 57%). LC/MS tR ) 15.7 min. (M/z) 5958.4 (M + 4), UV (214 nm) purity >99%. HRMS calcd for C264H388N66O80S6 + 4H+ 5958.696, found 5958.671. (B1)-MPA-OA-insulin, 4. 11 (1.0 g) was dissolved in N,N-dimethylformamide (20 mL) in the presence of N-methylmorpholine (300 µL). 10 (1.0 g) was then added, and the reaction was stirred at ambient temperature for 4 h in the dark. The reaction was quenched by precipitation with Et2O (120 mL) and the precipitate isolated by centrifuge. The residue was treated with TFA (2 mL) for 10 min. The TFA was evaporated, and the crude product was dissolved in acetic acid (20 mL) and water (100 mL). The resulting solution was injected into preparative HPLC using a 27 to 40% acetonitrile gradient over 60 min. The fractions at tR ) 28-37 min were combined and lyophilized to give 4 as a white powder (654 mg, 63%). LC/MS tR ) 16.1 min. (m/z) 6099.6 (M + 4), UV (214 nm) purity >99%. HRMS calcd for C272H403N67O81S6 + 3H+ 6098.803, found 6098.783. (A1B1)-Bis-Boc-insulin, 12. 1 (74 mg) was dissolved in anhydrous dimethyl sulfoxide (2 mL) and acetic acid (46 µL). Boc2O (6.9 mg, 2.5 equiv) was added to the solution, and the reaction was stirred for 5 h at ambient temperature. Water (15 mL) and acetonitrile (5 mL) were added to the reaction mixture, and the resulting solution was injected into semipreparative HPLC using a 27 to 40% acetonitrile gradient over 120 min. The fractions at tR ) 64-71 min were combined and lyophilized to give 12 (30 mg, 40%) as a white powder. MS (m/z) found 6006.8 (M + 4). HRMS calcd for C267H399N65O81S6 + 3H+ 6006.766, found 6006.759.

Thibaudeau et al.

(B29)-MPA-insulin, 5. 12 (30 mg) in N,N-dimethylformamide (2 mL) and N-methylmorpholine (100 µL) was reacted with 8 (10 mg) for 60 min. The solvent was removed under vacuum and the residue treated with TFA for 10 min. The TFA was removed under reduced pressure, the residue was dissolved in water/acetonitrile (3:1) and the solution injected into semipreparative HPLC using a 27 to 32% acetonitrile gradient over 120 min. The fractions at tR ) 39-49 min were combined and lyophilized to give 5 as a white powder (22.2 mg, 75%). LC/MS tR ) 13.1 min. (m/z) 5958.0 (M + 4), UV (214 nm) purity >99%. HRMS calcd for C264H388N66O80S6 + 4H+ 5958.696, found 5958.690. (B29)-MPA-OA-insulin, 6. 12 (20 mg) in N,N-dimethylformamide (1 mL) and N-methylmorpholine (100 µL) was reacted with 10 (20 mg) for 5 h at ambient temperature in the dark. The solvent was removed under vacuum and the residue treated with TFA for 10 min. The TFA was removed under reduced pressure, the residue was dissolved in water and the solution injected into semipreparative HPLC using a 27 to 31% acetonitrile gradient over 120 min. The fractions at tR ) 55-63 min. were combined and lyophilized to give 6 as a white powder (11 mg, 53%). LC/MS tR ) 15.6 min. (m/z) 6099.5 (M + 4), UV (214 nm) purity 97.8%. HRMS calcd for C272H403N67O81S6 + 4H+ 6098.803, found 6098.858. General Procedure for the Hydrolysis of Maleimide. The appropriate MPA-insulin (or MPA-OAinsulin) (20 mg) was dissolved in 10 mL of aqueous sodium hydroxide (0.1 N) and stirred at ambient temperature for 10 min. The reaction mixture was brought to pH 4 by addition of acetic acid. The resulting solution was injected into semipreparative HPLC using a 27 to 32% acetonitrile gradient over 120 min. The fractions at tR ) 45-54 min were combined and lyophilized to give the desired hydrolyzed material. The yields, purity, and mass analysis are described in Supporting Information. Preparation of Conjugates. Each maleimido insulin derivative was solubilized in Nanopure water at a concentration of 10 mM then diluted to 1 mM into a solution of HSA (25%, Cortex-Biochem, San Leandro, CA). The samples were then incubated at 37 °C for 30 min. Prior to their purification, each conjugate solution was diluted to 5% HSA in 20 mM sodium phosphate buffer (pH 7) composed of 5 mM sodium octanoate and 750 mM (NH4)2SO4. Purification of Conjugates. Using an A ¨ KTA purifier (Amersham Biosciences, Uppsala, Sweden), each conjugate was loaded at a flow rate of 2.5 mL/min onto a 50 mL column of butyl sepharose 4 fast flow resin (Amersham Biosciences, Uppsala, Sweden) equilibrated in 20 mM sodium phosphate buffer (pH 7) composed of 5 mM sodium octanoate and 750 mM (NH4)2SO4. When all flow through waste material was removed, a linear gradient of decreasing (NH4)2SO4 concentration (750 to 0 mM) was applied over four column volumes. The fractions containing the conjugate were combined, desalted and concentrated using Amicon ultra centrifugal (30 kDa) filter devices (Millipore Corporation, Bedford, MA). The resulting solution containing purified conjugate was immersed into liquid nitrogen, lyophilized, and stored at -80 °C. The analytical results of the individual conjugates are displayed in Table 2 in the text. SDS-PAGE of Conjugates. A sample (5 µg of protein/ well) of the conjugates was subjected to nonreducing SDS-PAGE using a 10% (w/v) gel. A constant current 20 mA/gel for 1.5 h in separating gel was applied (25). The gels were then stained with 0.1% Coomassie Brilliant Blue R-250 for 30 min at room temperature, followed by

Synthetic Insulin−HSA Conjugates

Bioconjugate Chem., Vol. 16, No. 4, 2005 1003

Scheme 1. Synthesis of Activated Maleimides with Different Chain Lengths

destaining in 10% (v/v) acetic acid and 30% (v/v) methanol in water. Insulin Receptor Binding. Liver membranes from male Wistar rats (175 ( 25 g) were prepared in modified Na-K phosphate pH 7.4 buffer using standard techniques. Three milligram aliquots of membrane were incubated with 30 pM 125I-insulin and the various other compounds for 16 h at 4 °C. Nonspecific binding was estimated in the presence of 1 µM insulin. Each compound was tested at 10-9, 10-8, 3.0 × 10-8, 10-7, 10-6, and 10-5 M, and the insulin control had an extra concentration of 10-10 M. The membranes were filtered and washed three times with PBS buffer, and the filters were counted to determine the 125I-insulin specifically bound. Glucose Uptake Assay. Epidimymal fat tissue (0.03 g/mL) obtained from Wistar rats (175 ( 25 g) was degraded by collagenase in modified HEPES solution pH 7.4 at 37 °C. The test compounds at 10-9, 10-8, 10-7, 10-6, and 10-5 M and control (with an extra concentration of 10-10 M) were incubated with 500 µL aliquots of tissue in modified HEPES buffer pH 7.4 in the presence of D-[33 H]glucose (2.5 µCi/mL) at 37 °C for 2 h. The cells were harvested using GF/B filtermats, washed with PBS buffer, and dried, and the filtermats were counted to determine the cellular uptake of radioactive glucose. Evaluation of Conjugates in StreptozotocinInduced Diabetic Rats. Eight-week old male SpragueDawley rats (Charles River, St-Constant, QC, Canada) were administered a single intravenous injection of streptozotocin (60 mg/kg in 25 mM citrate/0.9% saline, pH4.5) to induce type I diabetes. Diabetic animals had mean blood glucose levels of 25.9 to 29.2 mM while glycemia ranged from 6.0 to 6.4 mM in normal rats. The test compounds (saline control, insulin 0.14 mg/kg, 3-HSA and 4-HSA 8.7 and 21.7 mg/kg) in 0.9% saline were administered as a single subcutaneous bolus injection in the dorsal area 2 days after streptozotocin treatment. Blood sampling was done via the tail tip, and glucose levels were measured using a hand-held glucometer (On Touch Ultra, Lifescan, Canada). The animals were supplied water and standard rat chow (Charles River Diet #5075) ad libitum and monitored throughout the experiment. Blood glucose levels were measured 30 min predose, 1, 2, 3, 4, 6, 8, 10, 11, 24, 30, and 48 h postdose. Pharmacokinetics of 1 and 4-HSA. Subcutaneous pharmacokinetic profiles were done in 7-8 week old male Sprague-Dawley rats with body weights ranging 322347 g. Each treatment group consisted of four rats. 4-HSA was reconstituted in 0.9% saline while 1 was first dissolved in acidified water and further diluted to the required concentration with 0.9% saline. Blood samples were drawn at the following time points: -2, 10, 15, 20, 30, 40 min, 1, -2, 3 h and 2, 30 min, 1, 2, 4, 8, 24, 48, 72

h postadministration for 1 and 4-HSA, respectively. Serum levels of 1 and 4-HSA were determined using a human insulin ELISA kit (Linco Research, cat# EZHI14K). The reference curves were generated with reference standards of known concentrations of 1 and 4-HSA diluted in rat serum. Ethics and Statistical Analyses. The experimental protocols were performed according to the Canadian Council on Animal Protection following approval by the Universite´ du Que´bec a` Montre´al Institutional Committee on Animal Protection. All results are expressed as mean ( SEM or ( SD performed using GraphPad Prism software. RESULTS AND DISCUSSION

Chemistry. The overall objective was to obtain the insulin-HSA conjugate with the best activity profile. In the case of the fatty acid and PEG derivatives, the focus was on the B1 or the B29 positions of 1 as points of attachment. This approach is in agreement with the proposed insulin receptor interaction (26). If direct acylation was to be used as a strategy, in most cases, the A1 position arose as the most abundant product from the mixture except when the C-terminal B30 was removed, which then favored acylation at B29 (27). In our approach, the exploitation of the pKas (approximated as 11.1 for N-Lys B29, 8.4 for NR-Gly A1, and 7.1 for NR-Phe B1 in water) (28) and steric environments of the three amines was proposed as a means to selectively attach a single maleimidopropionamide (MPA) to each of the three individual positions of 1 as illustrated in Figure 1, more specifically A1 (2), B1 (3 and 4), and B29 (5 and 6). Before initiating the synthesis of the insulin derivatives, two activated maleimide containing acylation reagents were prepared as described in Scheme 1. 8 was prepared from 7 using N-hydroxysuccinimide (NHS) and EDC (24). 8 was then used to prepare the longer linker maleimide by direct coupling to 8-aminooctanoic acid to give the intermediate 9 which was used crude in the next step. Attempts to make the succinimidyl ester of 9 led to purification difficulties and a tendency to decompose in aqueous media. 10 was ultimately made under mild conditions using p-nitrophenyl chloroformate (29-30) and was found to be easily purified and had a much improved stability profile. Acylation using aqueous buffers could not be used in this strategy because of the tendency of the maleimido group to hydrolyze leading to a product that loses all ability to conjugate to HSA. This led to the development of buffer like conditions in N,N-dimethylformamide (DMF) with a mixture of trifluoroacetic acid (TFA) and N-methylmorpholine (NMM). These conditions were found to be convenient to solubilize the starting insulin and as a buffering medium. NMM has a lower pKa than the  amine on the B29 lysine. The latter will therefore, remain

1004 Bioconjugate Chem., Vol. 16, No. 4, 2005 Scheme 2. Synthesis of Nr-A1-insulin MPA Derivative 2 Using Bufferlike Conditions in Organic Solvent

protonated while the R amine of the A1 glycine of insulin will have a larger portion of free amine at equilibrium resulting in more acylation product. The B1 phenylalanine is too sterically hindered to react and will require a different strategy. As shown on Scheme 2, the A1 position remained the predominant product from the direct acylation of 1. Compound 2 was obtained in 40% yield in a single step, along with 23% recovered starting insulin after purification. The remainder was a mixture of other positional isomers and bis-acylation products. To obtain the other two B1 and B29 acylation products, a selective protection strategy was required. The bis tertbutyloxycarbamate (bis-Boc) A1 and B29 protection of 1 gave the best improvement in yield of the B1 derivatives. As shown in Scheme 3, intermediate 11 was prepared in 65% yield using a mixture of triethylamine and dimethyl sulfoxide as solvent and 2.5 equiv of di-tert-butyl dicarbonate (Boc2O) (31-34). Compound 11 was then converted to 3 and 4 using either 8 or 10, respectively, under identical coupling conditions. The intermediate product from the acylation reaction was not isolated. Instead, the crude product was treated with TFA and purified. The final compounds 3 and 4 were obtained in similar yields of 57 and 63%, respectively. Obtaining selective A1, B1 bis-Boc insulin 12 involved the addition of acetic acid to the reaction mixture (Scheme 4). In this case, the ammonium/free amine equilibrium in DMSO was such that only the lysine remained protonated throughout the process. 12 was

Thibaudeau et al.

obtained in low yield but enough material could be made to carry on. The completion of the syntheses of 5 and 6 was accomplished as described in Scheme 4. The yield of 5 was similar to the B1 acylation products but the conversion to 6 gave a slightly lower yield. Each synthesis was repeated to give multiple batches of 3, 4, 5, and 6, which were combined after examination of the purity profiles. The demonstration of the position of attachment was done by Edman degradation. The amide bonds on maleimide were labile under high pH aqueous conditions used in routine automated Edman degradation. The study was therefore undertaken on the prehydrolyzed material to minimize any possible ambiguous results. The corresponding structures and compound numbers of the hydrolysis products are shown in Scheme 5, and the Edman degradation results are shown in Table 1. The sequences read during the Edman degradation showed that in the case of 1 both chain amino acids were identified, indicating that no N-terminal amine was acylated. In the case of 13, only the sequence associated with chain B was read, indicating that chain A was acylated. When this logic was applied to compounds 14 through 17, it became possible to confirm that the initial maleimido derivative structures were correctly assigned. Preparation of Insulin-HSA Conjugates. The conjugation of maleimido derivatives to Cys34 of HSA and subsequent purification using hydrophobic chromatography (17), where a purified conjugate is defined as free of starting HSA (mercaptalbumin and cystinylated albumin also termed “capped” HSA) and free of residual maleimido derivative of insulin, has recently become an efficient process. The conjugation step involved mixing the individual derivatives with 25% HSA and incubating for 30 min at 37 °C. The conjugation reactions were found to be very rapid, and there was no maleimide hydrolysis observed. Thereafter, the resulting mixtures were loaded directly onto a column packed with butyl sepharose 4 fast flow resin. The insulin-HSA conjugates adsorbed onto the hydrophobic resin whereas essentially all nonconju-

Scheme 3. Synthesis of the Nr-B1 Insulin MPA Derivatives 3 and 4

Bioconjugate Chem., Vol. 16, No. 4, 2005 1005

Synthetic Insulin−HSA Conjugates Scheme 4. Synthesis of NE-B29 Insulin MPA Derivative 5 and 6

Scheme 5. Insulin

Hydrolysis of Maleimide Derivatives of

Table 1. Edman Degradation of Hydrolyzed Insulin MPA Derivatives Edman degradation results (positions) compound

chain

1

2

3

4

1

A B A B A B A B A B A B

Gly Phe -Phe Gly -Gly -Gly Phe Gly Phe

Ile Val -Val Ile -Ile -Ile Val Ile Val

Val Asn -Asn Val -Val -Val Asn Val Asn

Glu Gln -Gln Glu -Glu -Glu Gln Glu Gln

13 14 15 16 17

gated HSA eluted within the void volume of the column. Each conjugate was further purified from any free (unreacted) maleimido insulin derivatives by applying a linear gradient of decreasing ionic strength. The synthetic results are shown in Table 2. The mass spectral analysis of starting HSA indicated that the MS peaks corresponding to the mercaptalbumin

and capped albumin species were in a ratio of approximately 2:1, respectively. In the unpurified conjugate, the peaks corresponding to mercapatalbumin diminished while the conjugate peak appeared. Once purified, no residual unreacted HSA in either the mercaptalbumin or capped forms were found. The final products were stored as lyophilized powders. SDS-PAGE analysis of the HSA conjugates was done, and the results are shown in Figure 2. As illustrated, all the insulin-HSA conjugates had adequate purity profiles and migrated in the general area corresponding to proteins of molecular weights similar to albumin. A very faint band was observed above each conjugate as well as HSA. This band corresponds to HSA dimer and is often observed when analyzing HSA with gel electrophoresis. In Vitro Evaluation. Insulin is produced in pancreatic Langerhans and acts by binding to the insulin receptor found on the surface of a variety of cell types. The insulin receptor is composed of two extracellular R units and two transmembrane β units. Insulin binds to the extracellular component causing the stimulation of tyrosine kinase activity of the β unit, leading to a cascade of phosphorylations that cause the activation of glycogen synthase as well as the translocation of glucose transporter 4 (GLUT4) to the cell membrane, thus initiating

1006 Bioconjugate Chem., Vol. 16, No. 4, 2005

Thibaudeau et al.

Table 2. Results from Conjugate Synthesis

a

starting material

Mr theoretical

Mr measured

conjugated product

yield, %a

2 (22.8 mg) 3 (29.0 mg) 4 (164.6 mg) 5 (26.1 mg) 6 (15.5 mg)

72403.8 72403.8 72543.2 72403.8 72543.2

72393.0 72391.8 72543.8 72396.5 72536.2

2-HSA (173.8 mg) 3-HSA (246.4 mg) 4-HSA (1382.6 mg) 5-HSA (212.9 mg) 6-HSA (107.9 mg)

71 75 77 72 63

Yield based on the amount of maleimido starting material.

Figure 2. SDS PAGE 10% gel (5 µg/well) analysis under nonreducing conditions, Coomassie Blue staining. Lane A contains reduced molecular weight standards, and the values are expressed in kDa.

Figure 4. Glucose uptake-response curves on rat epididymal fat cells in the presence of insulin-HSA conjugates (n ) 2-4). Table 3. In Vitro Insulin Receptor Binding and Activation Results compound

binding,a IC50 (nM)

glucose uptake,b EC50 (nM)

1 2-HSA 3-HSA 4-HSA 5-HSA 6-HSA

9.1 ( 3.6 2059 ( 226 100 ( 19 89 ( 14 1191 ( 267 747 ( 376

1.3 ( 0.8 117 ( 78 14 ( 2 18 ( 11 78 ( 15 18.9 ( 0.4

a

Figure 3. Binding response curves for insulin-HSA conjugates on insulin receptor expressing rat hepatocytes (n ) 2-4).

the uptake of glucose into the cell (35). The optimum insulin derivative will be able to bind and activate the insulin receptor on the exterior of the cell at the lowest possible concentration, and the response will be the uptake of glucose from the surrounding medium in vitro or the lowering of plasma glucose in vivo. The primary evaluation of conjugated insulin derivatives was done in vitro. The binding to the insulin receptor was accomplished using a homogenized rat hepatocyte membrane assay (36) as shown in Figure 3. The insulin receptor activation was measured in a glucose uptake assay on rat epididymal fat cells (37) as shown in Figure 4. The individual IC50 and EC50 values for receptor binding and glucose uptake, respectively, are shown in Table 3. The structure-activity results from the in vitro evaluation of the five insulin-HSA conjugates indicate that B1 is the best position of attachment on insulin. The short linker on the A1 and B29 positions and the long linker on the B29 position caused large losses in binding affinity as demonstrated in the responses of 2-HSA, 5-HSA, and 6-HSA, respectively. The short and long linker derivatives 3-HSA and 4-HSA, respectively, were essentially equipotent in the binding and glucose uptake assays. An approximately 10-fold loss versus 1 in glucose

Rat hepatocytes. b Rat epididymal fat cells.

uptake EC50 was observed for the two derivatives in this particular assay. Both conjugates were therefore chosen for in vivo evaluation. In Vivo Evaluation. Type I diabetes was induced with STZ in free feeding Sprague-Dawley rats to an average basal blood glucose titer of >25 mmol/L. The blood glucose level was at an average of 5-6 mmol/L in normal nondiabetic rats. Each diabetic rat was administered a single bolus subcutaneous injection of 1, 3-HSA conjugate, 4-HSA conjugate, or saline and were allowed to feed ad libidum. The blood glucose levels for all groups were monitored over a 2-day period. The dose of 1 was limited to 24 nmol/kg per animal to avoid more severe hypoglycemic events observed at higher doses. The experimental results are shown in Figure 5. The glucose lowering effect of 1 was observed at the first time point (1 h). The glucose level proceeded to rise steadily after 2 h and returned to the STZ control group level by 8 h. It should be noted that in another experiment, 1 was injected at 120 nmol/kg and the glycemia returned to a diabetic level after 6-8 h as well (data not shown). 3-HSA and 4-HSA showed dose-dependent decreases in plasma glucose level where the effects reached their maximum between 4 and 6 h postadministration. At the highest doses, both conjugates were able to control the glucose level beyond 48 h. Other parameters such as water and food intake as well as weight gain were monitored over the study period. The results showed that for both 3-HSA and 4-HSA, the three parameters resembled those from the normal rat group for the whole 48-hour data-gathering period.

Synthetic Insulin−HSA Conjugates

Bioconjugate Chem., Vol. 16, No. 4, 2005 1007

Figure 5. Glycemic control by insulin-HSA conjugates in streptozotocin (STZ)-induced type I diabetes in Sprague-Dawley rats. The results of plasma glucose measurements for 3-HSA (A) and 4-HSA (B). The insulin, STZ diabetic, and normal controls are the same in A and B.

Figure 6. Pharmacokinetic profiles of 1 and 4-HSA in normal Sprague-Dawley rats. The test compounds were administered subcutaneously, and plasma levels were monitored using a human insulin ELISA kit. Part A shows the first 4 h and part B the full 72 h.

Alternately, the diabetic control group as well as the one receiving insulin drank more water, ate less, and lost weight. It can be generalized that the STZ-induced diabetic animals returned to a normal feeding and weight gain pattern when administered insulin-HSA conjugates while the animals treated with 1 did not. 4-HSA was chosen for pharmacokinetic evaluation in normal rats using 1 for comparison. The plasma concentration as a function of time is shown in Figure 6. When 1 was injected subcutaneously to rats, its peak plasma concentration was observed at 50 min and was no longer detectable after 3 h. On the other hand, the 4-HSA plasma concentration peaked at 8 h and was still detectable at 72 h postadministration. These results correlate well with the glycemic control experiment where the peak glucose lowering activity was between 1 and 2 h for 1 and 4 and 6 h for 4-HSA. In summary, the three possible direct acylation positional maleimido derivatives of human insulin with either short or long linkers were synthesized. These derivatives were conjugated to HSA in good yield and purified using hydrophobic chromatography. In vitro analysis of the conjugates demonstrated that the B1 linking position of insulin led to the best overall binding and glucose uptake. Further in vivo evaluation identified 4-HSA, also known

as CJC-1575, as a delayed action and long lasting human insulin analogue. CJC-1575 is currently in preclinical development as a potential therapeutic agent. Supporting Information Available: The mass spectral and purity data from the maleimide hydrolysis and Edman degradation results. This material is available free of charge via the Internet at http://pubs.acs.org. ACKNOWLEDGMENT

The authors thank Pierre Pepin from the Sheldon Institute for performing the Edman degradations and Patricia Liscourt for skilled technical assistance in the pharmacokinetic experiment. LITERATURE CITED (1) Hirsch, I. B. (2005) Insulin analogues. N. Engl. J. Med. 352, 174-183. (2) Lindholm, A. (2002) New insulins in the treatment of diabetes mellitus. Best Pract. Res. Clin. Gastroenterol. 16, 475-492. (3) Gerich, J. E. (2002) Novel insulins: expanding options in diabetes management. Am. J. Med. 113, 308-316.

1008 Bioconjugate Chem., Vol. 16, No. 4, 2005 (4) Gerich, J. E. (2004) Insulin glargine: long-acting basal insulin analogue for improved metabolic control. Curr. Med. Res. Opin. 20, 31-37. (5) Barlocco, D. (2003) Insulin detemir. Novo Nordisk. Curr. Opin. Investig. Drugs 4, 449-454. (6) Kurtzhals, P., Havelund, S., Jonassen, I., Kiehr, B., Ribel, U., and Markussen, J. (1996) Albumin binding and time action of acylated insulins in various species. J. Pharm. Sci. 85, 304-308. (7) Kurtzhals, P., Havelund, S., Jonassen, I., Kiehr, B., Larsen, U. D., Ribel, U., and Markussen, J. (1995) Albumin binding of insulins acylated with fatty acids: characterization of the ligand-protein interaction and correlation between binding affinity and timing of the insulin effect in vivo. Biochem. J. 312, 725-731. (8) Dea, M. K., Hamilton-Wessler, M., Ader, M., Moore, D., Scha¨ffer, L., Loftager, M., Vølund, A., and Bergman, R. N. (2002) Albumin binding of acylated insulin (NN304) does not deter action to stimulate glucose uptake. Diabetes 51, 762769. (9) Hamilton-Wessler, M., Ader, M., Dea, M. K., Moore, D., Loftager, M., Markussen, J., and Bergman, R. N. (2002) Mode of transcapillary transport of insulin and insulin analog NN304 in dog hind limb: evidence for passive diffusion. Diabetes 51, 574-582. (10) Tsai, Y. J., Rottero, A., Chow, D. D., Hwang, K. J., Lee, V. H. L., Zhu, G., and Chan, K. K. (1997) Synthesis and purification of NB1-palmitoyl insulin. J. Pharm. Sci. 86, 1264-1268. (11) Hashimoto, M., Takada, K., Kiso, Y., and Muranishi, S. (1989) Synthesis of palmitoyl derivatives of insulin and their biological activities. Pharm. Res. 6, 171-176. (12) Nakashima, K., Miyagi, M., Goto, K., Matsumoto, Y., and Ueoka, R. (2004) Enzymatic and hyperglycemia stability of chemically modified insulins with hydrophobic acyl groups. Bioorg. Med. Chem. Lett. 14, 481-483. (13) Hinds, K. D., and Kim, S. W. (2002) Effects of PEG conjugation on insulin properties. Adv. Drug Delivery Rev. 54, 505-530. (14) Hinds, K., Koh, J. J., Joss, L., Liu, F., Baudysˇ, M., and Kim, S. W. (2000) Synthesis and characterization of poly(ethylene glycol)-insulin conjugates. Bioconjugate Chem. 11, 195-201. (15) Baudysˇ, M., Letourneur, D., Liu, F., Mix, D., Jozefonvicz, J., and Kim, S. W. (1998) Extending insulin action in vivo by conjugation to carboxymethyl dextran. Bioconjugate Chem. 9, 176-183. (16) Chuang, V. T. G., Kragh-Hansen, U., and Otagiri, M. (2002) Pharmaceutical strategies utilizing recombinant human serum albumin. Pharm. Res. 19, 569-577. (17) Le´ger, R., Thibaudeau, K., Robitaille, M., Quraishi, O., van Wyk, P., Bousquet-Gagnon, N., Carette, J., Castaigne, J.-P., and Bridon, D. P. (2004) Identification of CJC-1131-albumin bioconjugate as a stable and bioactive GLP-1(7-36) analog. Bioorg. Med. Chem. Lett. 14, 4395-4398. (18) Le´ger, R., Benquet, C., Huang, X., Quraishi, O., van Wyk, P., and Bridon, D. (2004) Kringle 5 peptide-albumin conjugates with anti-migratory activity. Bioorg. Med. Chem. Lett. 14, 841-845. (19) Le´ger, R., Robitaille, M., Quraishi, O., Denholm, E., Benquet, C., Carette, J., van Wyk, P., Pellerin, I., BousquetGagnon, N., Castaigne, J.-P., and Bridon, D. P. (2003) Synthesis and in vitro analysis of atrial natriuretic peptidealbumin conjugates. Bioorg. Med. Chem. Lett. 13, 3571-3575. (20) Kratz, F., Warnecke, A., Scheuermann, K., Stockmar, C., Schwab, J., Lazar, P., Dru¨ckes, P., Esser, N., Drevs, J., Rognan, D., Bissantz, C., Hinderling, C., Folkers, G., Fichtner, I., and Unger, C. (2002) Probing the cysteine-34 position of endogenous serum albumin with thiol-binding doxorubicin derivatives. Improved efficacy of an acid-sensitive doxorubicin

Thibaudeau et al. derivative with specific albumin-binding properties compared to that of the parent compound. J. Med. Chem. 45, 55235533. (21) Holmes, D. L., Thibaudeau, K., L’Archeveque, B., Milner, P. G., Ezrin, A. M., and Bridon, D. P. (2000) Site specific 1:1 opioid:albumin conjugate with in vitro activity and long in vivo duration. Bioconjugate Chem. 11, 439-444. (22) Duttaroy, A., Kanakaraj, P., Osborn, B. L., Schneider, H., Pickeral, O. K., Chen, C., Zhang, G., Kaithamana, S., Singh, M., Schulingkamp, R., Crossan, D., Bock, J., Kaufman, T. E., Reavey, P., Carey-Barber, M., Krishnan, S. R., Garcia, A., Murphy, K., Siskind, J. K., McLean, M. A., Cheng, S., Ruben, S., Birse, C. E., and Blondel, O. (2005) Development of a longacting insulin analog using albumin fusion technology. Diabetes 54, 251-258. (23) Xia, C. Q., Wang, J., and Shen, W. C. (2000) Hypoglycemic effect of insulin-transferrin conjugate in streptozotocininduced diabetic rats. J. Pharmacol. Exp. Ther. 295, 594600. (24) Keller, O., and Rudinger, J. (1975) Preparation and some properties of maleimido acids and maleoyl derivatives of peptides. Helv. Chim. Acta. 58, 531-541. (25) Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. (26) Huang, K., Xu, B., Hu, S. Q., Chu, Y. C., Hua, Q. X., Qu, Y., Li, B., Wang, S., Wang, R. Y., Nakagawa, S. H., Theede, A. M., Whittaker, J., De Meyts, P., Katsoyannis, P. G., and Weiss, M. A. (2004) How insulin binds: the B-chain alphahelix contacts the L1 beta-helix of the insulin receptor. J. Mol. Biol. 341, 529-550. (27) Hansen, L. B. Selective Acylation Method. US 5,905,140 (May 18, 1999). (28) Gao, J., Mrksich, M., Gomez, F. A., and Whitesides, G. M. (1995) Using capillary electrophoresis to follow the acetylation of the amino groups of insulin and to estimate their basicities. Anal. Chem. 67, 3093-3100. (29) Glattahard, R., and Matter, M. (1963) Neues, mildes verfahren zur herstellung von aktiven carbonsa¨ure-estern und deren verwendung als acylierungsmittel. Helv. Chim. Acta 46, 795-804. (30) Kim, S., Lee, J. I., and Kim, Y. C. (1985) A simple and mild esterification method for carboxylic acids using mixed carboxylic-carbonic anhydrides. J. Org. Chem. 50, 560-565. (31) Liu, F., Song, S. C., Mix, D., Baudys, M., and Kim, S. W. (1997) Glucose-induced release of glycosylpoly(ethylene glycol) insulin bound to a soluble conjugate of concanavalin A. Bioconjugate Chem. 8, 664-672. (32) Markussen, J., Halstrom, J., Wiberg, F. C., and Schaffer, L. (1991) Immobilized insulin for high capacity affinity chromatography of insulin receptors. J. Biol. Chem. 266, 18814-18818. (33) Losse, G., Naumann, W., and Raddatz, H. (1983) Untersuchungen zur tert-butyloxycarbonylferung von insulin. Z. Chem. 23, 22-23. (34) Losse, G., and Richler, B. (1981) Darstellung von B1-Bocund B29-Boc-insulin-B-Ketten-di-S-sulfonat. Z. Chem. 21, 70-71. (35) De Meyts, P., and Whittaker, J. (2002) Structural biology of insulin and IGF1 receptors: implications for drug design. Nat. Rev. Drug Discovery 1, 769-783. (36) Koch, R., and Weber, U. (1981) Partial purification of the solubilized insulin receptor from rat liver membranes by precipitation with concanavalin A. Hoppe Seylers Z. Physiol. Chem. 362, 347-351. (37) Moody, A. J., Stan, M. A., Stan, M., and Gliemann, J. (1974) A simple free fat cell bioassay for insulin. Horm. Metab. Res. 6, 12-16.

BC050102K