| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Maxygen (S.P., B.v.d.H., J.C., S.G.-N., C.B.J., K.V.A., T.H., S.O., H.T.S.), DK-2970 Hoersholm; and Department of Gynecology and Obstetrics (S.P.), Hvidovre Hospital, 2650 Hvidovre, Copenhagen, Denmark
Address all correspondence and requests for reprints to: Dr. Hans T. Schambye, Maxygen, Agern Allé 1, DK-2970 Hoersholm, Denmark. E-mail: hs{at}maxygen.dk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-chain (FSH1208) was efficiently glycosylated at both new sites. This resulted in a molecule with increased size and charge, factors known to reduce renal clearance of proteins. FSH1208 was found to have a 3- to 4-fold increased serum half-life, compared with wild-type recombinant FSH. Furthermore, in spite of a lower in vitro activity, FSH1208 had a markedly increased in vivo potency, as shown by increased ability to augment the ovarian weight and stimulate the serum estradiol levels in rats. These characteristics make FSH1208 a possible candidate for improved infertility treatment. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
FSH is a glycoprotein composed of two subunits, termed and ß. The -subunit of FSH is identical to the -subunit of other glycoprotein hormones [i.e. chorionic gonadotropin (CG), LH, and TSH], whereas the ß-subunit is unique. The -chain consists of 92 amino acid residues and has two N-glycosylation sites located at asparagine residues N52 and N78. It is noncovalently linked to the FSH ß-chain composed of 111 amino acids. The ß-chain also contains two N-glycosylation sites, located at residues N7 and N24 (7).
An important mechanism for elimination of FSH is through glomerular filtration (8). It is well known that FSH exists in many different isoforms because of differences in the glycosylation pattern and, hence, in the sialic acid content (7, 9). Isoforms with high content of sialic acid have a low isoelectric point, are negatively charged at physiological pH, and are not as easily filtered by the glomeruli. Thus, the isoforms with a high sialic acid content remain longer in circulation and have been shown to have a higher in vivo activity (10, 11, 12).
Previously a chimeric FSH-hCG molecule created by fusion of the FSH ß-chain with a sequence of the C terminus of hCG has been constructed. This molecule, which has a higher molecular weight than wild-type FSH because of the presence of O-glycosylation in the extension, was shown to have increased half-life and in vivo activity (13, 14, 15, 16).
The goal of the present study was to develop an FSH-based molecule with retained in vitro and in vivo action and prolonged in vivo half-life, which would allow reduction of the injection frequency and, thus, increase patient convenience and compliance. The strategy was to increase the molecular weight and the charge of the molecule by introduction of additional glycosylation and, thus, reduce glomerular filtration.
Two different approaches were used for introduction of extra glycosylation sites. One was based on identification of areas within the native structure of the FSH molecule in which glycosylation sites potentially could be introduced without distortion of receptor binding and the overall tertiary structure. The second approach was based on N-terminal elongation using sequences encompassing glycosylation sites. Several molecules with increased glycosylation were identified through both strategies and one of these, termed FSH1208, was selected for further in vitro and in vivo studies to substantiate that it could form the basis for a new, improved infertility treatment.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A series of 50 models of the three-dimensional structure of hFSH was built based on two available hCG structures (17, 18) using the program Modeler 98 (MSI Inc., San Diego, CA). In the -chain four N-terminal residues (A1, P2, D3, and V4) and three C-terminal residues (H90, K91, and S92) were not modeled because they are not identified in the hCG structures. All of the hFSH-ß-chain was modeled, even the part that has no homologous residues in the hCG structures. The accessible surface area was calculated for each of the 50 model structures using the computer program Access, version 2 (Yale University, New Haven, CT). Although this work was in progress, a structure of an hFSH variant was published (19). The accessible surface area based on this structure generally confirmed the calculation based on the model structures.
Construction of plasmids for expression of FSH
Genes encoding the human FSH - and ß-subunits were constructed by assembly of synthetic oligonucleotides using PCR. The native signal sequences were maintained, except for a mutation in the ß-subunit at position 2 (Lys to Glu). The codon usage of the genes was optimized for high expression in mammalian cells using information on Homo sapiens codon usage from http://www.kazusa.or.jp/codon/ (20), at the time based on 12,512 coding sequences.
The pcDNA3.1(+)/Hygro (Invitrogen, Carlsbad, CA) and pcDNA3.1(+)/Zeo (Invitrogen) were used as expression vectors. An intron from pCI-Neo (Stratagene, La Jolla, CA) was added to further optimize the protein production. The intron was amplified from pCI-Neo using 5'-CCGTCAGATCCTAGGCTAGCTTATTGCGGTAGTTTATCAC-3' and 5'-GAGCTCGGTACCAAGCTTTTAAGAGCTGTAAT-3' as primers. The PCR product was subcloned into pcDNA3.1(+)/Hygro using NheI and HindIII, yielding PF033. The intron was transferred to plasmid pcDNA3.1(+)/Zeo by replacing a 758-bp MluI-XhoI fragment in the plasmid with the intron-containing MluI-XhoI fragment from PF033, resulting in pBvdH957. Using BamHI and XbaI, the synthetic -gene was subcloned into PF033 and the ß-gene into pBvdH957, resulting in pBvdH977 and pBvdH1022, respectively. Mutant forms of FSH- and -ß were generated by PCR-based site-directed mutagenesis using pBvdH977 and pBvdH1022 as templates. The sequences of all constructs were verified by DNA sequencing. Plasmids containing both the - and ß-encoding genes were generated by subcloning FSH- containing NruI-PvuII fragments from pBvdH977 (or derivatives) into pBvdH1022 (or derivatives) linearized with NruI.
Expression of FSH in CHO cells
FSH was expressed in Chinese hamster ovary (CHO) K1 cells (CCL-61; ATCC, Manassas, VA). Cells were grown in media (catalog no. 32571-028 or 31330-038; Life Technologies, Inc., Carlsbad, CA) containing 1:10 fetal bovine serum (FBS) (catalog no. 02-701F; BioWhittaker, Inc., Walkersville, MD) and 100 U/ml penicillin and 100 µg/ml streptomycin (pen-strep) at 37 C, 5% CO2. Plasmids were transfected using Lipofectamine 2000 (Life Technologies, Inc.). Twenty-four to 48 h after transfection, culture media were collected for Western blotting. Stable clones expressing FSH were isolated by growth under selection pressure. The expression levels of individual clones were determined by ELISA. High producers were selected for large-scale FSH production: Cells were grown until confluence in 1700-cm2 roller bottles containing 300 ml media (catalog no. 31330-038; Life Technologies, Inc.) with the same additives. The media were then changed to UltraCHO with L-glutamine (catalog no. 12-724Q; BioWhittaker, Inc.) with addition of 1:500 EX-CYTE VLE (catalog no. 81-129, Serological Proteins Inc., Kankakee, IL) and pen-strep. The media were replaced with fresh media after 48 h of growth. After another 48 h, the media was replaced with serum-free media (catalog no. 21041-025, with the addition of 1:500 ITS-A; catalog no. 51300-044, 1:500 EX-CYTE VLE and pen-strep; Life Technologies, Inc./BRL). Subsequently, every 24 h, culture media were harvested and replaced with 300 ml fresh serum-free media. Growth in roller bottles with daily harvests was continued for up to 10 d.
Purification and characterization of FSH
Harvests from the CHO cells were concentrated by ultrafiltration (10-kDa cut-off membrane), pH was adjusted to 8.0, and the conductivity lowered to 1–2 mS/cm by dilution with distilled water. A diethylaminoethyl Sepharose (Pharmacia, Stockholm, Sweden) column was equilibrated with 16 mM ammonium acetate buffer (pH 8). Subsequently the sample was applied and FSH was eluted with 0.16 M ammonium acetate (pH 8). After addition of ammonium sulfate to a concentration of 1.5 M and adjustment of the pH to 7.0, the sample was applied on a butyl Sepharose (Pharmacia) column. After application, the column was washed with 1.5 M (NH4)2SO4, 20 mM ammonium acetate (pH 7), and FSH was eluted with 20 mM ammonium acetate buffer (pH 7). Eluated FSH was dialyzed against 50 mM sodium phosphate, 150 mM NaCl (pH 7.2), using 12–14 kDa cut-off dialysis tubing. For the final step, an anti-FSH-ß monoclonal antibody (RDI-FSH909, Research Diagnostics) was immobilized to CNBr-activated Sepharose (Pharmacia). One milligram antibody was coupled per milliliter resin. The immunoaffinity resin was equilibrated with 50 mM sodium phosphate, 150 mM NaCl (pH 7.2). After application of the sample, the column was washed with 50 mM sodium phosphate, 1 M NaCl (pH 7.2), followed by equilibration with 50 mM sodium phosphate, 150 mM NaCl (pH 7.2). Elution from the column was performed using 1 M NH3, 40% (vol/vol) isopropanol, and the pH in the eluted fractions was immediately neutralized with 2 M acetic acid to pH 6–8. The purified FSH was concentrated and diafiltrated using Vivaspin 20 modules, 10-kDa cut-off membrane (Vivascience, Hannover, Germany), to 50 mM sodium phosphate, 150 mM NaCl (pH 7.2), and microfiltrated using a 0.22-µm filter. The purified protein was analyzed by SDS-PAGE under nondissociating conditions (without boiling). For storage at -80 C, BSA was added to 0.1% (wt/vol).
N-terminal sequencing was carried out following SDS-PAGE and electroblotting onto polyvinyl difluoride membranes using a 494 protein sequencer (PE Applied Biosystems, Foster City, CA). For amino acid analysis, lyophilized protein samples were hydrolyzed for 16 h at 110 C in 6 M HCl containing 1% phenol in N2-blanketed glass vials before quantification of the liberated amino acids using the AccQTag system (Waters, Milford, MA).
Western blotting was performed using rabbit antihuman FSH (AHP519, Serotec) or mouse antihuman FSH-ß (MCA338, Serotec, Raleigh, NC) as primary antibody, and an ImmunoPure Ultra Sensitive ABC peroxidase staining kit (Pierce Chemical Co., Rockford, IL) for detection.
Samples were separated on pH 3–7 isoelectric focusing (IEF) gels (Novex) for analysis of pH at the isoelectric point (pI). After electrophoresis, proteins were blotted onto Immobilon-P (Millipore Corp., Billerica, MA) membranes, and a Western blot was performed.
In vitro activity assay, ELISA, and specific activity
FSH in vitro activity was measured using a reporter cell line expressing the hFSH receptor and a reporter gene (21, 22). This cell line was obtained as follows: CHO-K1 cells were transfected with a pcDNA3 (Invitrogen) derivative containing the FSH receptor cDNA (kindly supplied by Dr. A. Hsueh, Stanford University, Palo Alto, CA). Stable clones were isolated, and a clone that responded to FSH stimulation was found using a cAMP-SPA RIA (Amersham, Cardiff, UK). Subsequently a plasmid containing 6 cAMP-responsive elements upstream of the Firefly luciferase reporter gene was stably introduced into this clone. For measurement of FSH activity, the reporter cells were suspended in DMEM/F-12 (without phenol red) containing 1.25% FBS and seeded in white 96-well culture plates at a density of 15,000 cells/well (100 µl). After incubation overnight at 37 C, 5% CO2, 25 µl sample or FSH standard (follitropin-ß), diluted in media containing 10% FBS was added to each well. Subsequently the plates were incubated for 3 h, followed by addition of 125 µl LucLite substrate (Packard Bioscience, Groningen, the Netherlands). Finally, the luminescence was measured on a TopCount luminometer (Packard Bioscience) in single photon counting mode.
FSH concentrations were determined by ELISA (FSH EIA, DRG Instruments GmbH, Marburg, Germany). This ELISA kit has a reported minimal detection level of 1 mIU/ml (
0.1 ng/ml) and a coefficient of variance of less than 6% within the measured range. A dilution series of follitropin-ß (Puregon, Organon, Oss, The Netherlands) in PBS with 0.1% BSA was used as reference. A follitropin-ß batch containing 400 IU [relative to international reference IS 70/45 (23)] per milliliter was used. It was assumed that the specific activity of this batch was 10,000 IU/mg, i.e. that the batch contained 40 µg FSH per ml (23). An amino acid analysis of an FSH1208 aliquot determined a protein concentration of 0.34 mg/ml, corresponding exactly to the concentration measured (0.34 mg/ml) in the ELISA assay. This demonstrated that the ELISA set-up was suitable for determination of concentrations of FSH1208.
The specific activity of FSH1208 was determined by measurement of the in vitro bioactivity using the luciferase assay, and measurement of the FSH content of the samples was determined by ELISA. The specific activities were assessed using follitropin-ß as standard and expressed in international units (see above).
Animal studies
Female Sprague Dawley rats were obtained from the breeder at a weight of 200–220 g for the pharmacokinetic studies and at an age of 21–22 d for the pharmacodynamic studies. They were housed in a temperature-controlled room at 23 C, with a 12-h light period from 0600 h to 1800 h and free access to food and water. All experiments were conducted in accordance with accepted standards of humane animal care, reviewed and accepted by the Animal Experimentation Inspectorate in accordance with the Animal Testing Act set by the Danish Ministry of Justice.
Pharmacokinetics were studied by injection of 10 µg protein either iv, in an injection volume of 250 µl, or sc in an injection volume of 500 µl. Preparations were formulated in a DPBS buffer (catalog no. BE 17-512F; BioWhittaker, Inc.) containing 0,1% (wt/vol) BSA (product no. A3059; Sigma, St. Louis, MO). In each experiment, three rats in the reference group received follitropin-ß (Puregon; Organon), and three other rats received FSH1208. Blood samples were taken in a time span over 72 h and incubated at room temperature for 90 min. Serum was isolated by centrifugation (5000 rpm for 30 min) and stored at -80 C. Serum samples were analyzed for FSH content by ELISA.
Pharmacodynamics were determined by the ovarian weight augmentation assay (24) and measurement of estradiol levels in serum. All animals received a daily dosage of 13.3 U hCG (catalog no. C-1063; Sigma) and either FSH or a buffer vehicle. Injections had a volume of 200 µl and were given sc. The hCG was formulated in 0.01 M sodium phosphate (pH 7.2) containing 30 mg mannitol/ml, and FSH was formulated as described above. The first group of animals received vehicle on d 1 and 2; the second group received 0.25 µg (2.5 IU) follitropin-ß on d 1 and vehicle on d 2; the third group received 0.25 µg follitropin-ß both days; the fourth group received 0.25 µg (equivalent to about 0.6 IU) FSH1208 on d 1 and vehicle on d 2; the fifth group received 2.5 IU FSH1208 on d 1 and vehicle on d 2. In the 72-h studies, the groups and dosages were the same, but the rats in group 3 were stimulated for an extra day with follitropin-ß, and all others received vehicle on d 3. Blood samples were withdrawn every 12 h in the 48-h setup. Serum samples were isolated, stored as described above, and analyzed for 17ß-estradiol level using an ELISA kit (catalog no. RE-52041; IBL-Hamburg). The minimum detectable concentration of estradiol was reported by the provider to be 10 pg/ml, and the coefficient of variance was maximally 10% in the measured range. The animals were killed by cervical dislocation, and the ovarian weights were determined after decapsulation.
The results were analyzed using the computer program PRISM (version 3.01, 1999; GraphPad Software, Inc., San Diego, CA). Statistical significance of differences was assessed by ANOVA tests either one way or for repeated measures followed by Bonferroni’s multiple comparison test. Probabilities less than 0.05 were considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Two different approaches of introduction of additional glycosylation sites were applied in the current study. In one approach the structural model of FSH was used to identify suitable locations for introduction of mutations. The other approach used the model to extend the protein with sequences housing possible glycosylation sites.
The structural analysis of FSH was performed in an attempt to identify sites in the molecule in which N-linked glycosylation sites (Asn-X-Thr/Ser) could be introduced without significant distortion of the tertiary structure or the receptor binding of the molecule. Consequently, the structural analysis was focused on identifying sites that were located on the surface of the molecule distant from the presumed receptor-binding area (7, 25). Such possible N-glycosylation sites were chosen through identification of amino acid residues in which more than 50% of the side chain was exposed at the surface of the molecule, calculated as an average over the series of the 50 models of the three-dimensional structure of hFSH that were built. Furthermore, residues that had a proline residue at position +1 or + 3 were excluded because it is known that proline residues in these positions reduce the glycosylation efficacy of the asparagine residues (26). About 50 sites were identified in each of the two subunits in the series of models that were built. Subsequently residues that had a threonine or serine residue located at position +2 were identified because these could be transformed into glycosylation sites through exchange of only one amino acid residue. The -chain housed four such possible N-glycosylation sites [E9()N, F17()N, R67()N, and H90()N], whereas five sites were identified in the ß-chain [Y58(ß)N, L73(ß)N, S89(ß)N, D90(ß)N, and Y103(ß)N]. Only three of the four possible exchanges were constructed in the -chain because it is known that residues in the C-terminal part of the subunit, including H90(), are important for receptor binding and signal transduction, and thus, mutation of this residue would most likely affect the function of FSH (25, 27). Similarly, S89 and D90 of the ß-chain were excluded because it is believed that this part of the ß-chain is also involved in the receptor binding and activation (7, 28).
An alternative approach was based on introduction of extra glycosylation sites by extension of the two N termini because they are located distant from the presumed receptor-binding site and because it is unlikely that an extension results in distortion of the tertiary structure. The extensions comprised four to nine amino acid residues including one or two Asn-X-Thr sites. Three different extensions were used: ANIT, ANITV, and ANITVNITV, all of which included amino acid residues with relatively small, nonpolar side chains (Ile, Val, AL) to minimize steric hindrance and 1–2 amino acids conserved in the extreme N terminus, compared with the wild-type chains.
All muteins were produced in CHO cells in different combinations, and Western blot analyses were performed to assess the degree of glycosylation. Examples of different degrees of glycosylation are shown in a Western blot in Fig. 1B
. Lanes 2 and 4 show the muteins Y103(ß)N and Y58(ß)N, respectively. It is seen that only a small percentage of the Y103(ß)N molecules has a higher molecular weight than the wild-type recombinant FSH (rFSH) in lane 1, indicating that only a minor fraction of the molecules was glycosylated at the introduced asparagine residue. In contrast, almost all molecules of the extension variant in lane 3, which has two additional glycosylation sites added to the N-terminal part of the ß-chain between residue 2 and 3 (NS NITVNITV CEL), were glycosylated, at least at one position in the extension.
|
-chain, was selected for further studies because this mutein was found to be glycosylated to a higher extent than any of the other variants. Characterization of FSH1208
Expression and purification.
Stable CHO clones coexpressing the mutated -subunits with the wild-type ß-subunits were isolated and used for large-scale production. Expression yields were about 0.2–0.25 mg/l. FSH1208 was purified as described, resulting in 90–95% purity as judged by SDS-PAGE (Fig. 1A
). SDS-PAGE, run under nondissociating conditions (Fig. 1A), showed wild-type, rFSH (follitropin-ß) migrating as a band with an apparent mass of 45 kDa (lane 2), slightly diffuse because of heterogeneity in the attached carbohydrates. In comparison, FSH1208 migrated as a band with an apparent mass of 55 kDa (lane 3). The ß-extension mutein (NSNITVNITV) seen in lane 4 has a molecular weight in between the weight of follitropin-ß and FSH1208, indicating that this extension mutein is not as efficiently glycosylated as FSH1208.
The charge of the purified protein was analyzed by IEF. Recombinantly produced wild-type FSH (follitropin-ß) was found to consist of a series of isoforms with a pI between 4.0 and 5.2, in accordance with the literature (29). In contrast, FSH1208 consisted of isoforms with a pI between 3 and 4.8 with the majority of the isoforms in the lowest part of the range (Fig. 2).
|
N-terminal sequencing of FSH1208 demonstrated the presence of the N-terminal extension on the -chain. The sequence of the N-terminal ß-chain was also identical to the known sequence of the wild-type ß-FSH (30), although the first two residues were absent. This truncation has previously been found for both pituitary and rFSH (31, 32, 33). N-terminal amino acid sequence determination of three different purification batches of FSH1208 indicated that approximately 80% of N2 was glycosylated and more than 95% of N6 was glycosylated.
In vitro bioactivity assay showed that the specific activity of FSH1208, 2.3 ± 0.5 IU/µg (n = 11), was lower than that of rFSH, follitropin-ß, 10 IU/µg (23).
Pharmacokinetic studies
The pharmacokinetic profile of FSH1208 was determined either sc or iv by administration of a single dose of 10 µg/rat. Figure 3A shows the serum concentration of wild-type rFSH (follitropin-ß) and FSH1208 after iv administration as measured by ELISA. The concentration of both proteins declined rapidly in the initial distribution phase, followed by a constant elimination at a slower rate. FSH1208 had a significantly longer in vivo half-life than follitropin-ß. The terminal half-life for FSH1208 after iv administration was calculated to 22 h, and the terminal half-life for follitropin-ß was calculated to 6 h, in accordance with previous reports (23). It was not possible to detect follitropin-ß beyond the first 24 h.
|
). The maximal level of follitropin-ß was reached after 12 h, whereas the concentration of FSH1208 continued to rise until 24 h after the injection. The half-life of follitropin-ß after sc administration was equal to the half-life measured after iv administration, i.e. 6 h, whereas FSH1208 had a half-life of 18 h after sc administration. The area under the concentration curve was 4 to 6 times larger for FSH1208 than for wild-type rFSH (follitropin-ß), depending of the route of administration (Table 1). The studies indicate that the bioavailability of FSH1208 was about 80% and, thus, higher than of follitropin-ß of about 50%. In vitro bioactivity measurement showed that there was still FSH activity in the 48-h samples (data not shown).
|
The in vivo potency of the FSH forms was tested by analysis of their ability to augment ovarian weight in rats (24) on 48 or 72 h of stimulation. Furthermore, the estradiol level was measured in serum over a 48-h period. To test whether the increased half-life of FSH1208 translated into higher in vivo activity, FSH1208 was dosed only once, but follitropin-ß was dosed every day.
Ovarian weight augmentation experiments
The effect of FSH1208 during 48 h of stimulation was studied in an experiment in which all animals received a daily dosage of hCG and FSH or vehicle. Two doses of FSH1208 were studied: One group of animals was given 0.25 µg (approximately 0.6 IU), i.e. equal to the amount of follitropin-ß given to the rFSH group, and the other group was given 2.5 IU, equal to the amount of in vitro bioactivity given to the follitropin-ß group. Blood samples were drawn every 12 h for estradiol analysis, and the weight of the ovaries was measured after 48 h. In another experiment, the animals were stimulated for 72 h before they were euthanized, and the ovaries were decapsulated and weighed. The groups and the dosages were the same as in the previous experiment except the rats were treated for an extra day with either follitropin-ß or vehicle.
The ovarian weights after 48 h of stimulation are shown in Fig. 4A. All groups stimulated with either of the FSH formulations were significantly different from the control group stimulated with hCG only (P < 0.001–0.01). The group that received 2.5 IU of FSH1208 had a mean ovary weight on 51.8 ± 3.7 mg, compared with 34.9 ± 4.7 mg in the group that received follitropin-ß once (P < 0.001). In the 72-h stimulation experiment, the effect of FSH1208 was even more pronounced (Fig. 4B). Any dose of FSH1208, e.g. equal amount of protein or bioactivity, respectively, was markedly more potent than 1x follitropin-ß. Interestingly, 0.25 µg FSH1208 given once was at least as efficient in augmenting the ovary weight as 0.25 µg follitropin-ß given once daily for 3 d, despite the fact that the specific activity for FSH1208 is only about one fourth of the specific activity of follitropin-ß.
|
Blood samples were drawn during the 48-h stimulation assay for analysis of the estradiol level in serum. Figure 5 shows the ELISA measurements of the estradiol level as a function of time. The curves suggest a circadian variation in the estradiol concentration after FSH stimulation, as previously described (34). It was not possible to detect any increase in the estradiol level during the first 12 h. After 24 h, a significantly higher content of estradiol (P < 0.05) was detected in the serum of the rats that were stimulated with either a dose of FSH1208, compared with the control group, or FSH1208 dosed in equal in vitro bioactivity (u), compared with 1x follitropin-ß. The peak values of estradiol were measured after 24 h for all the groups with levels at a range from about 20 pg/ml for the control group to about 200 pg/ml for the FSH1208(u) group (Fig. 5).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2.5 IU) for 2–3 d. FSH1208 was consistently more potent in rats than follitropin-ß when equal amounts of in vitro bioactivity were injected. This demonstrates that the increased in vivo half-life translates to a more potent in vivo action. In this study it has been shown that the prolonged half-life of FSH1208 is of major importance for increasing the in vivo activity. Recent studies suggest that apart from the metabolic clearance rate/terminal half-life, the binding affinity for the receptor and the efficiency in signal transduction also contribute to the net in vivo activity of a glycovariant (38, 39). The pharmacokinetic profile of FSH1208 results in a relatively stable serum concentration, which appears to be at least as effective as the high peaks and troughs obtained through daily injections of a molecule with a relatively short in vivo half-life such as rFSH, follitropin-ß. This is in agreement with previous findings showing that several injections of small amounts of wild-type FSH, giving rise to a relatively constant concentration level, is more efficient than the total dose given as a single bolus injection (14).
In principle, endogenous rat FSH may have influenced the serum concentration measurements. However, if any such interference occurred, its effect on the results was only minor. Endogenous FSH levels in rats have been found to reach a maximum of about 10 ng/ml during the estrous cycle (40). In our experiments, the measured FSH levels are initially well below 10 ng/ml (0.01 µg/ml) (Fig. 3B) and drop well below this level at the last time points (Fig. 3, A and B). This indicates that the endogenous FSH levels must have been lower than 0.01 µg/ml or that our ELISA does not efficiently detect rat FSH. In any event, even if the endogenous rat FSH levels would have been maximal throughout our experiment, the concentration curves lie for the most part well above the 0.01 µg/ml level. Furthermore, it is likely that the FSH treatment leads to reduction in endogenous FSH levels through pituitary feedback inhibition.
N-linked glycosylation has previously been used to alter the half-life of erythropoietin (EPO). EPO is a glycoprotein with three naturally occurring glycosylation sites and a molecular mass of approximately 30 kDa, thus fairly similar to FSH. A variant of EPO with two additional glycosylation sites (darbepoetin-) was recently approved for the treatment of anemia. The molecular mass of darbepoetin- is 8 kDa higher than EPO, and the in vivo terminal half-life was increased approximately 3-fold in rats, compared with EPO (41). These results are, thus, similar to what was observed in the present study. The terminal half-life in humans of darbepoetin- was increased to the same extent, approximately 3-fold, and darbepoetin- was able to stimulate erythropoiesis with less frequent dosing as effectively as recombinant human EPO (42, 43). The molecular size and the clearance mechanisms of FSH and EPO are very similar, and it seems reasonable to predict that the 3-fold increase in terminal half-life in rats for FSH1208 would be reflected in humans as well, just as it was observed for darbepoetin-. FSH1208 may, thus, form the basis for an improved treatment of infertility with a reduced frequency of injections. This will naturally need to be verified in clinical studies.
Additional carbohydrate moieties may reduce the elimination of proteins because of the increase in size and charge. Sialic acid bears a net negative charge at physiological pH and makes a glycoprotein hormone more metabolically stable by decreasing the glomerular filtration (44, 45) and through protection against clearance by hepatic asialo-glycoprotein receptors (46, 47). Microheterogeneity in the protein structure is seen for all glycoproteins and is caused by variation in the posttranslational glycosylation process in terms of differences in the degree of sialylation and different branching of the carbohydrate moieties. For FSH this gives rise to about 20 naturally occurring isoforms with different pI, biological activity and terminal serum half-life (48). The degree of sialylation of FSH changes both during the menstrual cycle and through the various stages of life. Acidic isoforms are mainly present in prepubertal and postmenopausal life and during the follicular phase of the cycle, whereas the majority of the isoforms are basic during puberty and during the midcyclic surge of FSH (49, 50). These changes in the secreted hormone from the pituitary gland are influenced by the endocrine milieu surrounding the gland and the current need/stage in the body (51, 52, 53, 54, 55). The differences in the degree of sialylation and metabolic stability of the different isoforms are reflected in the time the isoforms stay in circulation and, thus, the time they have to exert their effect. Sialic acid is, thus, of major importance in determining the half-life and, thus, the in vivo activity of a glycoprotein. Thus, a likely explanation for the changes in the in vivo characteristics of FSH1208, compared with recombinant wild-type FSH, follitropin-ß, is a higher content of sialic acid, rather than just the amino acid extension or the presence of the remaining carbohydrate structures.
In the present study, two different approaches were used for introduction of new glycosylation sites: structure-aided, site-directed introduction of sites within the FSH molecule and addition of N-terminal extensions. A total of six sites identified within the structure were tested, and none of these were glycosylated to a large extent. All the changes in the native FSH molecule were made on the assumptions of the tertiary structure based on computer models and structural analysis. The single-point mutations were meant to be introduced on the surface of the molecule at sites believed to be unimportant for receptor and interchain interaction. The results suggest that the influence of the microenvironment on the glycosylation process is fairly unpredictable. The fact that only a few mutations were introduced and only at sites in which a threonine or serine already was located at position +2 together with the fact that the glycosylation process was not optimized by exchanging residues at important positions surrounding the glycosylation site (56, 57) minimized a priori the chances of identifying highly glycosylated muteins by simple point mutations. In contrast, most of the extension muteins were glycosylated and in the case of FSH1208 to a very high extent.
It has previously been shown that it is possible to make an extension of FSH without loosing the activity. A chimeric FSH-hCG molecule was created by fusion of the FSH ß-chain with a sequence from the C terminus of hCG encompassing four possible O-linked glycosylation sites. The C terminus of the hCG molecule is responsible for the longer circulatory half-life of hCG, compared with the other glycoprotein hormones in the family and deletion of that region in hCG, has been shown to decrease the in vivo activity (58). The changed pharmacokinetic and dynamic profile of the chimeric FSH-hCG molecule are, thus, explained by the properties of the C terminus of hCG and probably by the sialic acids coupled to the four O-linked glycan structures. This chimeric molecule has recently been tested in clinical trials. The half-life in humans was prolonged and FSH-hCG induced multiple follicular growth in a dose-dependent manner. Furthermore, no safety problems were encountered (15, 59). The N-terminally N-glycosylated variant FSH1208 described here resembles the FSH-hCG chimera in the animal studies. However, because the two molecules have important structural differences, they may turn out to differ in clinical efficacy and/or appropriateness for large-scale manufacturing.
A consensus sequence is known for N-linked glycosylation (Asn-X-Thr or Asn-X-Ser), whereas no such clear motif is known for O-linked glycosylation (60, 61, 62). Thus, even though this study showed that N-linked glycosylation sites are not always used, the availability of the consensus sequence allows, with a reasonable change of success, to tailor-make protein extensions with any desired number of additional N-linked glycosylations. Thus, this approach is generally applicable and attractive for many other proteins (patent application WO 02/02597).
The change of the structure of FSH by introduction of a nonnative nine amino acid extension sequence might trigger an immune response in the body. However, the extension has a high extend of glycosylation, compared with the size of the extension itself, and it is likely that a possible antigenic site in the amino acid core will be shielded by the carbohydrate moieties (63, 64). Furthermore, it has been shown that the introduction of two additional glycosylated sites in darbepoetin- does not cause an antibody response (42). It is possible that FSH1208 does not cause an immune response, but this remains to be demonstrated in clinical trials.
Differently charged forms of a glycoprotein have in addition to different half-lives also different biopotency. It is well established for many proteins, including FSH, that acidic isoforms are less potent in vitro than the isoforms with higher pI (10, 12, 23, 49, 65, 66). In line with this, FSH1208 with a lower pI profile has a reduced specific activity, compared with follitropin-ß. The reduced specific activity of the charged glycoproteins is believed to be caused by reduced receptor-binding affinity, possibly because of charge-based repulsion and, thus, reduced signal transduction (67, 68).
In conclusion, introduction of additional glycosylation sites through addition of an N-terminal extension has been successfully used to create an FSH molecule with an increased in vivo half-life and prolonged in vivo activity. This molecule may form the basis for an improved treatment of infertility in the future.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
S.P. and B.v.d.H. contributed equally and should be considered co-first authors.
Present address for C.B.J.: Molecular Pharmacology, Novo Nordisk, DK-2760 Maaloev, Denmark; for B.v.d.H.: Høiberg A/S, DK-1264 Copenhagen K, Denmark; and for S.G.-N.: Lundbeck, DK-2500 Valby, Denmark.
Abbreviations: CG, Chorionic gonadotropin; CHO, Chinese hamster ovary; EPO, erythropoietin; FBS, fetal bovine serum; IEF, isoelectric focusing; h, human; rFSH, recombinant FSH.
Received July 30, 2002.
Accepted March 18, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B.C.J.M. Fauser, B.M.J.L. Mannaerts, P. Devroey, A. Leader, I. Boime, and D.T. Baird Advances in recombinant DNA technology: corifollitropin alfa, a hybrid molecule with sustained follicle-stimulating activity and reduced injection frequency Hum. Reprod. Update, May 1, 2009; 15(3): 309 - 321. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 The Corifollitropin Alfa Dose-finding Study Group A randomized dose-response trial of a single injection of corifollitropin alfa to sustain multifollicular growth during controlled ovarian stimulation Hum. Reprod., November 1, 2008; 23(11): 2484 - 2492. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Stork, K. A. Zettlitz, D. Muller, M. Rether, F.-G. Hanisch, and R. E. Kontermann N-Glycosylation as Novel Strategy to Improve Pharmacokinetic Properties of Bispecific Single-chain Diabodies J. Biol. Chem., March 21, 2008; 283(12): 7804 - 7812. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Antenos, M. Stemler, I. Boime, and T. K. Woodruff N-Linked Oligosaccharides Direct the Differential Assembly and Secretion of Inhibin {alpha}- and {beta}A-Subunit Dimers Mol. Endocrinol., July 1, 2007; 21(7): 1670 - 1684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. J. Brown, G. C. Miller, N. Griffon, C. J. Long, and D. J. Rader Glycosylation of endothelial lipase at asparagine-116 reduces activity and the hydrolysis of native lipoproteins in vitro and in vivo J. Lipid Res., May 1, 2007; 48(5): 1132 - 1139. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Perlman, T. Bouquin, B. van den Hazel, T.H. Jensen, H.T. Schambye, S. Knudsen, and J.S. Okkels Transcriptome analysis of FSH and FSH variant stimulation in granulosa cells from IVM patients reveals novel regulated genes Mol. Hum. Reprod., March 1, 2006; 12(3): 135 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Ge, G. Yang, X. Yu, T. Pourbahrami, and C. Li Oligomerization state-dependent hyperlipidemic effect of angiopoietin-like protein 4 J. Lipid Res., November 1, 2004; 45(11): 2071 - 2079. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Weenen, J. E. Pena, S. V. Pollak, J. Klein, L. Lobel, R. K. Trousdale, S. Palmer, E. G. Lustbader, R. T. Ogden, and J. W. Lustbader Long-Acting Follicle-Stimulating Hormone Analogs Containing N-Linked Glycosylation Exhibited Increased Bioactivity Compared with O-Linked Analogs in Female Rats J. Clin. Endocrinol. Metab., October 1, 2004; 89(10): 5204 - 5212. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. O. Elmagbari, R. D. Egleton, M. M. Palian, J. J. Lowery, W. R. Schmid, P. Davis, E. Navratilova, M. Dhanasekaran, C. M. Keyari, H. I. Yamamura, et al. Antinociceptive Structure-Activity Studies with Enkephalin-Based Opioid Glycopeptides J. Pharmacol. Exp. Ther., October 1, 2004; 311(1): 290 - 297. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Devroey, B. C. Fauser, P. Platteau, N. G. Beckers, M. Dhont, and B. M. Mannaerts Induction of Multiple Follicular Development by a Single Dose of Long-Acting Recombinant Follicle-Stimulating Hormone (FSH-CTP, Corifollitropin Alfa) for Controlled Ovarian Stimulation before in Vitro Fertilization J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2062 - 2070. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
) in rats. A single dose of 10 µg FSH was administrated either iv (A) or sc (B).
