Bifunctional Gonadotropin-Releasing Hormone Antagonist-Progesterone Analogs with Increased Efficacy and Duration of Action
http://www.100md.com
《内分泌学杂志》
Medical Research Council Human Reproductive Sciences Unit (K.E.R., H.M.F., R.S., R.P.M.), The Queen’s Medical Research Institute, Edinburgh EH16 4TJ, Scotland, United Kingdom
Peptide Biology Laboratory (J.R.), Salk Institute, La Jolla, California 92037
Department of Medical Biochemistry (R.P.M.), University of Cape Town, Observatory 7925, South Africa
Abstract
GnRH peptide analogs are widely used to treat diverse clinical conditions. However, they have poor oral activity and exhibit rapid metabolic clearance, thus requiring injection and depot formulation. Because steroid hormones are bound to plasma proteins, we explored the possibility of conjugating hydroxylated progesterones to GnRH analogs to reduce metabolic clearance of the peptides. Conjugation of [D-Lys6]GnRH agonist to the 11-hydroxyl of 11-hydroxyl progesterone via a hemi-succinate bridge increased the plasma half-life after iv injection in rabbits by 3.6-fold while retaining high binding affinity, thus providing proof of concept. Five GnRH antagonists were then synthesized with 21-hydroxyprogesterone conjugated via C21-hydroxyl to positions six (conjugates A and B) and position seven (conjugates C and D) of GnRH antagonists. In the fifth compound the NH2 terminus of a GnRH antagonist lacking the first two amino acids was conjugated via the C21-hydroxyl to 21-hydroxyprogesterone (conjugate E). All five analogs bound to guinea pig progesterone binding globulin with relatively high affinities (264–1020 nM). Moreover, all five conjugates retained high progestogenic activity in stimulating a progesterone-response-element-driven chloramphenicol acetyltransferase reporter gene in the T47D breast cancer cell line. Conjugation via the -amino function of D-Lys6 (conjugates A and B) produced compounds with high binding affinity for the human GnRH receptor (15 and 7 nM) comparable to that of the unconjugated GnRH antagonists (4 and 26 nM). Conjugation via the -amino function of Lys7 (conjugates C and D) or the NH2 terminus of an N-terminally truncated antagonist (conjugate E) produced compounds of low binding affinity. Conjugates A and B also exhibited high functional antagonism of GnRH stimulation of inositol phosphate production in COS-7 cells expressing the human GnRH receptor (2.6 and 16 nM) compared with the unconjugated antagonists (1.3 and 122 nM). In accordance with their poor receptor binding affinity, conjugates C, D, and E had poor functional antagonism. Preliminary dose-finding studies in female marmosets showed transitory progesterone inhibition by 0.25 mg and prolonged suppression of 12 and 17 d by 0.5- and 1.0-mg doses. Injection of conjugate A in adult male marmosets (0.5 mg sc) rapidly suppressed plasma testosterone levels, which remained suppressed for at least 3 d. In contrast, the unconjugated parent antagonist alone or with progesterone suppressed testosterone for only 8 h to 1 d. The findings demonstrate that conjugation of progesterone to GnRH antagonists conveys plasma binding and progestogenic properties and increases their efficacy and duration of action in vivo. These new GnRH antagonists show promise as therapeutic agents for hormone-dependent diseases and as contraceptives.
Introduction
GnRH ANALOGS ARE extensively used in the treatment of sex hormone-dependent diseases such as prostate cancer, endometriosis, and leiomyoma, and for precocious puberty (1, 2, 3, 4, 5) and show promise as contraceptives for men and women (6, 7, 8). Although GnRH peptide agonists and antagonists are most efficacious in these applications, they have limitations in poor oral bioavailability and rapid metabolic clearance (9, 10). Consequently they are administered as injectable slow release formulations (11). Although this approach overcomes some of the problems, it limits the ability to vary dosages and does not allow withdrawal of treatment when desired. As a result, a number of pharmaceutical companies are developing nonpeptide orally active small molecule GnRH antagonists (12).
Alternative approaches are to conjugate moieties to GnRH peptide antagonists, to convey plasma binding capacity, to reduce metabolic clearance, and to convey active absorption from the gastro-intestinal tract. The majority of steroid hormones are bound to plasma binding proteins (>90%) and exhibit reduced metabolic clearance (13, 14). Therefore, we have conjugated GnRH antagonists to steroid hormones and demonstrated that this reduced their metabolic clearance and extended their duration of action. Moreover, steroid hormone bioactivity was maintained, thereby creating bifunctional molecules with potentially improved efficacy.
Materials and Methods
Synthesis of GnRH analog-steroid conjugates
The conjugation method was adapted from that of Mattox et al. (15) and Rajkowski and Cittanova (16). All chemicals were obtained from Sigma-Aldrich Co. Ltd. (Poole, Dorset, UK) with the exception of radiochemicals purchased from Amersham Pharmacia Biotech UK Limited (Little Chalfont, Buckinghamshire, UK) or unless otherwise stated. 11-Hydroxyprogesterone or 21 hydroxyprogesterone (1.5 M) in N-dimethyl formamide (DMF) was added slowly with stirring to 5 M succinic anhydride in DMF, and an alkaline pH was maintained with tributylamine. The mixture was placed on ice for 2 h and then stood for 12 h at room temperature. Forty-eight hours after the addition of an equal volume of water, the crystalline product was removed and subjected to spectrophotomeric and chromatographic analysis to confirm the succinyl progesterone product (Fig. 1). An 11-estradiol hemisuccinate and 17-estradiol hemisuccinate were similarly produced but not reported here. For later studies, 21-succinyl 21-hydroxyprogesterone was purchased from Sigma-Aldrich. The GnRH analog [D-Lys6]GnRH and the two GnRH antagonists [Ac-D-Nal1, D-Cpa2, D-Pal3, Arg5, D-Lys6, D-Ala10]GnRH, designated peptide A, and [Ac-Pro1, D-Fpa2, D-Trp3, D-Lys6]GnRH, designated peptide B, (Table 1) were dissolved in 0.1 M phosphate buffer (pH 7.0) before addition of an equal volume of DMF. A 20-fold molar excess of 11-hydroxyprogesterone 11-hemisuccinate in the case of [D-Lys6]GnRH or (21-hydroxyprogesterone 21-hemisuccinate for GnRH antagonist conjugation) was dissolved in anhydrous DMF with equimolar 1-hydroxybenzotriazole and N,N-dicyclohexylcarbodiimide. The mixture was vortexed and left at room temperature for 1 h. After adjustment to a pH greater than 7 with tributylamine, the peptide-steroid mixture was left at 4 C for 20 h. The generic chemical synthesis used to conjugate progesterone to a Lys residue in GnRH analogs is shown in Fig. 2.
Purification and identification of products
An initial purification through a Sep-Pak C18 cartridge (Millipore UK Ltd., Harrow, Middlesex, UK) with ethylacetate followed by hexafluropropanol/DMF (70:30) was carried out before HPLC and mass spectrometry analysis of products. RP-HPLC purification was carried out on a Novapak C18 column (4 μm beads, 3.9 x 150 mm) connected to a Beckman Coulter System Gold LC125 pump and LC168 diode array detector (Beckman Coulter, Fullerton, CA). The buffer system was 0.1% trifluoroacetic acid in water as buffer A and 0.1% trifluoroacetic acid in acetonitrile as buffer B. The column was developed with a gradient of 10–100% buffer B over 30 min at a flow rate of 1 ml/min. Mass spectrometry was carried out on a Tofspec 2E matrix-assisted laser desorption ionization-time of flight mass spectrometer (Micromass UK Ltd., Manchester, UK) with a matrix of -cyano-4-hydroxycinnamic acid.
Additional GnRH antagonist progesterone conjugates
After successful proof of concept with peptide steroid conjugates A and B, larger batches were custom synthesized (Peninsula Labs, UK). Additional conjugates (C, D, and E) were designed and custom synthesized by Albachem Ltd. (Edinburgh, UK) using solid phase methodology and purified by HPLC to >98%. The structures of the peptides and sites of conjugation to 21-hydroxyprogesterone 21-hemisuccinate are shown in Table 1. Conjugates A and B explored the effects of conjugating via the -amino function of D-Lys at position 6, whereas C and D explored the effects of conjugating via position seven through the -amino function of L-Lys. Conjugate E examined the feasibility of conjugating to the NH2 terminus of peptide A lacking the first two amino acids, because we reasoned that the size and hydrophobicity of progesterone may substitute appropriately for these.
Determination of in vivo half-life in the male rabbit
To determine the biological half-life of GnRH agonist-progesterone conjugates, 0.5 ml saline containing 10,000,000 cpm of [125I-Tyr5, DAla6, MeLeu7, Pro9-N-ethylamide]GnRH or [125I-Tyr5, D-Lys6]GnRH-progesterone conjugate (iodinated as for [His5, D-Tyr6]GnRH) was injected as an iv bolus into the ear vein of eight male rabbits. Rabbits were sedated with 0.4 ml Aceprom 10 (Milborrow Animal Health, Rand, South Africa) injected im 3–4 min before the start of the experiment and again at 2 h. One-milliliter blood samples were collected in heparinized tubes from an indwelling cannula placed in a vein of the contralateral ear at 45 sec and 3, 5, 7, 9, 11, 13, 15, 17, 25, 30, 45, 60, 90, 120, 150, 180, and 210 min. Whole blood was counted directly to determine disappearance of the radio-iodinated analogs from the circulation. Experiments were approved by the University of Cape Town Animal Experimental Committee and carried out in accordance with the regulations of the Republic of South Africa.
Cell culture
Transient transfection.
Human GnRH receptor cDNA (17) for transient transfection was prepared using Maxi-prep columns (Qiagen, Chatsworth, CA). COS-7 cells were cultured in DMEM containing 10% fetal calf serum, glutamine, and penicillin/streptomycin. Cells were transfected with human GnRH receptor using Superfect (Qiagen) in Optimem media (Invitrogen Life Technologies, Paisley, UK) for 4 h (18). Transfected cells were cultured for a further 48 h in normal media before use in receptor binding and inositol phosphate assays.
Whole-cell receptor binding assays.
Whole-cell receptor binding assays (18, 19) used the [His 5, 125I-D-Tyr6]GnRH analog (20). Transiently transfected COS-7 cells in 12-well culture plates were incubated for 4 h on ice in HEPES/DMEM/0.1% BSA with 105 cpm per well [His5, 125I-D-Tyr6]GnRH and varying concentrations of unlabeled analogs. Cell monolayers were rapidly washed twice in ice-cold PBS and solubilized in 0.1 M NaOH, and radioactivity was counted to determine the amount of bound radioligand. Nonspecific binding (in the presence of consistently less than 10% of total binding) was determined using vector-transfected (pcDNA1/amp) COS-7 cells and was subtracted from total binding to give specific binding. Each data point was conducted in triplicate and experiments were repeated at least three times.
Total inositol phosphate assays
Determination of GnRH stimulation of total inositol phosphate production was as previously described (18). In brief, transiently transfected COS-7 cells were incubated with inositol-free DMEM containing 1% dialyzed heat-inactivated FCS and 1 μCi/well myo-[3H]inositol (Amersham Pharmacia Biotech, Piscataway, NJ) for 24–48 h. Medium was removed, and the cells were washed with 1 ml buffer (140 mM NaC1, 20 mM HEPES, 4 mM KC1, 8 mM glucose, 1 mM MgC12, 1 mM CaC12, and 1 mg/ml BSA) containing 10 mM LiCl and incubated for 1 h at 37 C in the same buffer containing 1 nM GnRH alone and with increasing concentrations of GnRH antagonist analogs. Reactions were terminated by the removal of medium and addition of 1 ml ice-cold 10 mM formic acid, and incubated for 30 min at 4 C. Total [3H]inositol phosphates were separated from the formic acid cell extracts on AG1-X8 anion exchange resin (Bio-Rad Laboratories, Hercules, CA) and eluted with a 1 M ammonium formate/0.1 M formic acid solution. Radioactivity was determined by liquid scintillation counting. Each data point was conducted in triplicate and experiments were repeated at least three times.
Plasma protein binding assay
Plasma protein binding was determined by the competitive binding of steroid conjugates in the presence of [1,2,6,7-3H]progesterone to pregnant guinea pig plasma (Charles River Laboratories, Wilmington, MA) (21). Progesterone binding globulin was purified by cation exchange chromatography (22). Twenty microliters of plasma equivalent was diluted with 2 ml dextran-coated charcoal solution and incubated at room temperature for 30 min to remove the endogenous steroids. The suspension was centrifuged at 3000 x g for 10 min and the supernatant was removed. One hundred μl of diluted plasma preparation was aliquotted into centrifuge tubes, followed by 1 pmol [1,2,6,7-3H]progesterone/100 μl PBS and varying concentrations of progesterone and the GnRH antagonist conjugates in 100 μl PBS. The tubes were vortexed and incubated at room temperature for 1 h, then for 15 min on ice. Seven hundred and fifty microliters of dextran-coated charcoal suspension (0.25 dextran T-70, charcoal decolorizing powder; Merck Ltd., Littleworth, UK) was added to the plasma solution and incubated for 10 min on ice, followed by centrifugation (3000 x g) at 4 C for 5 min. The supernatant (700 μl) was removed and radioactivity was quantified by liquid scintillation counting. Each data point was conducted in triplicate and experiments were repeated three times.
Progesterone receptor activation
Progestogenic activity was tested on the epithelial breast cancer cell line T47D, stably expressing the progesterone receptor and a chloramphenicol acetyltransferase (CAT) reporter gene (Dr. M. Beato, Institut fur Molekularbiologie und Tumorforschung, Marburg, Germany).
The T47D cells were maintained in RPMI 1640 medium supplemented with 2 mM glutamine, 10% FCS, penicillin, streptomycin, and insulin, transferrin, and sodium selenite supplement. Cells were plated in 100-mm dishes at a density of 2 million cells per dish, 24 h in advance of treatment with progesterone or hydroxyprogesterone-GnRH antagonist conjugates. Steroid or conjugate solution (100 μl) was added to 10 ml of medium and the cells incubated for a further 24 h.
Cells were washed in PBS (magnesium and calcium free) and 900 μl Reporter Lysis Buffer (Promega, Madison, WI) was applied. After 15 min of incubation at room temperature, the cells were scraped from the plate and transferred to a microcentrifuge tube on ice, vortexed for 10–15 sec, then heated at 60 C for 10 min to inactivate endogenous deacetylase activity. The cell lysates were then centrifuged for 2 min and the supernatant was transferred to microcentrifuge tubes containing [3H]chloramphenicol and n-butyryl CoA and made up to a total volume of 125 μl with water. Positive and negative controls were prepared containing a known amount of CAT or no cell extract respectively. The reaction mixtures were incubated at 37 C for 3 h, centrifuged briefly, and the reaction was terminated by the addition of 300 μl mixed xylenes. The xylene/reaction mixture was vortexed thoroughly and centrifuged for 3 min. Two hundred microliters of the xylene phase was removed and 100 μl 0.25 M Tris-HC1 (pH 8.0) was added before repeating the vortex and centrifuge procedure and removal of 150 μl of the xylene layer for liquid scintillation counting. The radioactivity measured in the negative control was subtracted from all values. Each data point was conducted in triplicate, and the experiment was repeated on three occasions.
In vivo studies in the male and female common marmoset
Marmosets were housed in the MRC Human Reproductive Sciences Unit Primate Centre. Experiments were carried out in accordance with the Animals (Scientific Procedures) Act, 1986, and were approved by the Local Ethical Review Process Committee. For collection of blood samples, marmosets were placed in a restraining device that allows routine sampling without stress (23). Samples were taken from the femoral vein into heparinized syringes and animals were returned to their cages between collections.
Female marmoset studies.
Preliminary dose-finding studies were undertaken in female marmosets to identify the effective dose required to inhibit corpus luteum function. Adult marmosets with regular ovulatory cycles as determined by measurement of plasma progesterone concentrations in blood samples collected three times per week were studied. A total of 0.25, 0.5, and 1.0 mg of GnRH antagonist-hydroxyprogesterone conjugate A or vehicle were administered by injection in 1 ml saline at two sc sites during the mid-luteal phase of the cycle. Blood samples were collected immediately before GnRH antagonist injection and at 4 and 8 h and d 1, 2, and 3 after treatment and thereafter at three times per week until the end of the posttreatment cycle. Plasma was assayed for progesterone as described previously (24). GnRH antagonist hydroxyprogesterone conjugate did not cross-react with the antibody in the RIA. As both doses inhibited progesterone, the 0.5 mg dose was selected for more detailed studies in male marmosets.
Male marmoset studies.
To determine the duration of action of the GnRH antagonist-progesterone conjugate at the GnRH receptor, 0.5 mg was administered as divided injections in 1 ml saline at two sc sites in six adult male marmosets. To compare the response with unconjugated antagonist, molar equivalents of unconjugated GnRH antagonist (0.41 mg) and progesterone (0.09 mg) were administered together to six adult male marmosets as for the conjugate. Unconjugated GnRH antagonist (0.5 mg) without progesterone was also administered to three marmosets. One 300-μl blood sample was collected immediately before and at 4 and 8 h after injection (d 0) and on d 1, 2, 3, 4, 6, 7, 8, and 10 after treatment. Plasma testosterone was assayed by RIA (25).
Statistical analysis
Effects of treatment on plasma testosterone concentrations were determined by ANOVA. Because normal testosterone concentrations in the male marmoset are extremely variable, data were log transformed before analysis and changes were compared with the mean of the two pretreatment values. The initial and secondary half-lives of GnRH analogs in rabbits were determined by PRISM 3.0 (GraphPad, San Deigo, CA). Differences were considered significant at a level of P < 0.05.
Results
Preliminary studies on receptor binding and half-life
Preliminary proof of concept studies established the successful chemical procedure for attachment of hydroxyprogesterones by C11 or C21 and estradiol by C11 or C17 to the -amino function of [D-Lys6]GnRH and the identification of products by mass spectrometry. These agonist conjugates were shown to retain high receptor binding affinity (kDa, 0.3 nM) and high efficacy in stimulating inositol phosphate production (ED50, 0.6 nM) by COS cells expressing the GnRH receptor (data not shown). The mean initial half-life of the [125I-Tyr5, D-Lys6]GnRH-hydroxyprogesterone 11-hemisuccinate conjugate was 9.1 ± 4.0 min (mean ± SEM, n = 5), which was longer but not significantly different from the half-life of the unconjugated [125I-Tyr5, D-Ala6, Me Leu7, Pro9-N-ethylamide]GnRH, which was 5.0 ± 1.8 min (mean ± SEM, n = 3) (Fig. 3). This lack of difference is anticipated as the initial rapid disappearance after bolus iv administration and is due predominantly to equilibration into extracellular spaces and first round renal clearance. The second phase longer half-life of the conjugate was 54.7 ± 13.2 min (mean ± SEM, n = 5), which was significantly greater (P = 0.0031) than that of the unconjugated peptide, which was 15.5 ± 7.0 min (mean ± SEM, n = 3). These findings thus established proof of concept and set the scene for the following more detailed studies with GnRH antagonist conjugates to hydroxyprogesterone.
Plasma protein binding
Plasma binding of GnRH antagonist-hydroxyprogesterone conjugates was studied with guinea pig progesterone binding globulin because of its high binding affinity and capacity compared with human plasma. Progesterone competed with [3H]progesterone with an IC50 of 96 ± 18 nM (n = 4). The conjugates retained progesterone binding globulin binding (IC50, 264-1020 nM), which was significantly higher (P < 0.05) than that of unconjugated progesterone (96 ± 18 nM) (Fig. 4, A and B, and Table 2). The unconjugated A and B peptides exhibited no competition for [3H]progesterone binding (Fig. 4C), and cortisol showed poor competition at 10 –5 M (data not shown). Robust binding data could not be obtained for [3H]progesterone with human transcortin.
Whole-cell GnRH receptor binding
GnRH antagonist 21-hydroxyprogesterone conjugates A and B had high binding affinities for the human GnRH receptor comparable to those of the unconjugated peptides (Fig. 5 and Table 2). Conjugates C, D, and E, in which 21-hydroxyprogesterone was conjugated to the -amino function of L-Lys in position seven or to the NH2 terminus of an antagonist lacking the first two amino acids, had low binding affinity (Fig. 5 and Table 2). Conjugate C, in which the natural carboxyl-terminal Arg-Pro sequence was substituted with Leu-Arg, retained affinity equivalent to that of D, although Pro in position 9 is conventionally regarded as essential for high binding affinity (26, 27, 28).
Inhibition of GnRH-stimulated inositol phosphate production
The antagonist activity of the conjugates was tested by measuring their ability to inhibit the stimulation of inositol phosphate by 1 nM GnRH in COS cells (Fig. 6 and Table 2). Conjugate A had an IC50 comparable to that of the unconjugated peptide (Fig. 6). The IC50 of conjugate B was improved relative to the unconjugated peptide (Fig. 6), suggesting it has greater antagonist efficacy as the binding affinities were similar. Conjugates C, D, and E competed poorly with GnRH in inhibiting inositol phosphate production (Fig. 6).
Progesterone receptor activation
All five conjugates were able to bind to and activate the progesterone receptor in T47D cells as measured by the stimulation of CAT enzyme activity (Fig. 7 and Table 2). Despite being tethered to the GnRH antagonists, their progestogenic activities were substantial, albeit less than that of progesterone. Only conjugates A and B were significantly (P < 0.03) less active than progesterone.
In vivo effects in the marmoset
Preliminary dose-finding studies were conducted in a female marmoset model. All three doses of conjugate A induced a rapid decline in plasma progesterone concentrations in female marmosets. The 0.25-mg dose induced a transitory suppression of progesterone (data not shown). Plasma progesterone concentrations were suppressed for 12 d (0.5 mg dose) and 17 d (1.0 mg dose) compared with 4–5 d of low follicular phase progesterone in the control cycles (Fig. 8).
In more detailed studies in male marmosets, administration of 0.5 mg of conjugate A, antagonist A, or antagonist A plus progesterone induced a rapid suppression in plasma testosterone concentrations (Fig. 9). Significant suppression of testosterone was maintained for 8–24 h after administration of peptide A or peptide A plus progesterone, but for at least 3 d after treatment with conjugate A. Plasma testosterone was significantly lower (P < 0.05) in the conjugate A-treated group on d 1, 2, and 3 after treatment (Fig. 9) when compared with treatments with peptide A alone or peptide A with progesterone, demonstrating the prolonged duration of action of the conjugate.
Discussion
GnRH peptide analogs are widely used in the treatment of infertility and hormone-dependent diseases (1, 2, 3, 4, 5). They also show promise as a new generation of contraceptives for men and women when combined with steroid hormone replacement (6, 7). Although current GnRH analogs have some excellent therapeutic properties, their peptide nature conveys rapid metabolic clearance and poor oral bioavailability. Consequently, they are administered by daily injection or by formulation in slow-release biodegradable polymers (11). These depot injections have several disadvantages: 1) in the discomfort and side-effects of a depot injection, 2) in the inability to terminate treatment when desired, and 3) in the lack of flexibility in dosage. The development of small molecule nonpeptide orally active GnRH antagonist analogs may overcome these problems (12). Nevertheless, because GnRH peptide analogs have many excellent properties, we considered that an alternative to nonpeptide analog development was to conjugate chemical moieties to GnRH antagonists, which would convey reduced metabolic clearance and oral bioavailability. Conjugation of GnRH analogs to steroid hormones may potentially be used to render plasma protein binding properties, reduce metabolic clearance and extend analog half-life in the circulation. Our current findings demonstrate that this is achievable and that such analogs can also be designed to be bifunctional in targeting both GnRH and steroid receptors.
The sites of conjugation of the GnRH analogs and steroid hormone are critical in maintaining GnRH receptor binding, plasma protein binding, and steroid receptor binding and activation. To this end, we explored the effects of conjugation via a D-amino acid in position six, and an L-amino acid in position seven of GnRH antagonists as well as the amino terminus in position 3 of a truncated GnRH antagonist, via a hydroxyl at C11 and C21 of steroids. These conjugates were successfully synthesized and biological characterization revealed that conjugation of [D-Lys6] of GnRH antagonists with C21 of 21-hydroxyprogesterone produced compounds with good plasma protein binding, high GnRH receptor binding affinity, potent inhibition of inositol phosphate production, and good potency in progesterone receptor activation. This suggests that the sites chosen and the long side chain attachment (eight carbon atoms plus one nitrogen) avoid disruption of crucial binding sites and steric hindrance.
All of the conjugates bound guinea pig progesterone binding globulin with relatively high affinity. The affinity of conjugate E was two to four times higher than that of the other conjugates, suggesting that there is less steric hindrance when conjugation is through the amino terminus of this truncated GnRH antagonist. Although binding to plasma proteins of man and other mammals was not studied, the extended half-life of a conjugate in rabbits together with the findings of binding to progesterone binding globulin and the extended duration of action in marmosets suggest that the conjugates would behave similarly in man. The preservation of the essential features for progesterone binding to the human cortisol binding globulin (20-oxo, 10 methyl, 3-oxo, and 4-ene groups) (14) in the conjugates reinforces this conclusion. Because a hydroxyl group in the 11 position impairs binding, whereas the 21-hydroxyl group has no effect (14), we chose to conjugate via a hydroxyl at C21.
Conjugates A and B bound to the human GnRH receptor with high affinity comparable with the unconjugated antagonists. Conjugation via the D-Lys6 in these molecules was expected to yield high affinity binding analogs as GnRH assumes a folded II/ conformation around Gly6 when bound to the GnRH receptor and substitution of Gly6 with a D-amino acid enhances this conformation and binding affinity (26, 27, 28). This property has been exploited previously with good effect in GnRH analog conjugation to toxins (29) and vitamin B12 (30). The long attachment chain used in our conjugate clearly avoids steric hindrance by the progesterone moiety. Position seven of GnRH analogs has not previously been explored as an attachment site. Because conjugates C and D had much lower binding affinities, conjugation via an L-Lys in position seven appears to hinder receptor binding. However, it does provide the potential for conjugation of two different moieties in positions six and seven (e.g. steroid and vitamin B12 to convey extended half-life, steroid bioactivity, and oral bioavailability), albeit with a loss in binding affinity. This feature was designed into conjugates C and D. Conjugate E, in which progesterone replaced the first two N-terminal residues of peptide A, had low binding affinity despite our attempts to maintain size and hydrophobicity (the molecular weight of the two N-terminal amino acids in the antagonist was 416.5, which is comparable with that of the hydroxyprogesterone hemisuccinate replacement at 413.5). This suggests that the progesterone moiety is incorrectly orientated, creates steric hindrance to binding, or makes inappropriate contacts with the GnRH receptor. Conjugate C is a completely novel GnRH analog. All of the several thousand active GnRH analogs retain Pro9 as in native GnRH because this residue was thought to be essential for bioactivity (27). Our data confirm this but demonstrate that reasonable binding affinity is retained when Pro9 is substituted with Arg (cf. C and D).
Conjugates A and B were equally or more effective than the unconjugated peptides in antagonizing GnRH stimulation of inositol phosphate production. These conjugates thus had the best potential for in vivo activity. Conjugate A was about 8-fold more efficacious in inhibiting inositol phosphate production and, therefore, was tested in marmosets. All of the conjugates were able to bind and activate the progesterone receptor although their activity was less than that of progesterone. This finding indicates that the conjugates are able to enter the cells intact or that the progesterone moiety is hydrolyzed from the antagonist to enter the cells. Although the latter possibility cannot be excluded, it seems less likely, because a monolayer of cells would have little capacity to hydrolyze a substantial amount of the conjugate in the comparatively large volume of medium. The conjugation of 21-hydroxyprogesterone to a long and flexible side chain does accommodate the requirement for binding to the progesterone receptor. The ligand-binding domain of the progesterone receptor has been crystallized, allowing the analysis of some aspects of ligand-receptor interactions (31). The most important feature for binding to the progesterone receptor is the C3 keto group, which forms a hydrogen bond with the conserved steroid receptor glutamine 725 (31, 32). Arginine 766 and phenylalanine 778 also make van der Waals contacts with the A ring (through intervening fixed water sites) to tightly couple the ligand to the receptor. The methyl-ketone substituent (projecting from C17) apparently interacts with cysteine 891 and threonine 894. Numerous other interactions are made with the A, B, C, and D ring. Interestingly, no contacts are made with the C21 methyl, which also influenced our decision to choose this site for conjugation and may account for the high bioactivity at the progesterone receptor. There is also a sizeable space in the progesterone receptor opposite C21 of the docked progesterone.
Preliminary studies in female marmosets revealed that 0.25 mg of conjugate A was sufficient to induce an immediate but transient inhibition of progesterone. Higher doses of 0.5 and 1.0 mg had a sustained inhibitory effect on plasma progesterone concentrations and delayed subsequent ovulation. On the basis of these findings, male marmosets were treated with 0.5 mg conjugate A or 0.5 mg peptide A. Conjugate A suppressed testosterone for at least 3 d (recovery on d 6), whereas the unconjugated peptide A suppressed testosterone for only 8 h.
The enhanced duration of action of conjugate A compared with that of the unconjugated antagonist A could be considered to be due to the combined biological activities of the antagonist and progesterone in inhibiting gonadotropin. However, it is unlikely that the progesterone activity would last for 3 d after a single im injection of about 0.09 mg (the progesterone component of 0.5 mg conjugate). To rule out this possibility, we compared the effects of 0.5 mg conjugate A with the equivalent dose of free antagonist A (0.41 mg) and free progesterone (0.09 mg) and showed that the conjugate is much more effective than the molar equivalent of free antagonist and progesterone. Because the testosterone inhibition profile by peptide A was the same as the combination of peptide A plus progesterone it is evident that the progesterone moiety of the conjugate extends the duration of action of the conjugate through prolongation of its presence in the circulation and not through its inhibitory progestogenic effects on gonadotropin secretion.
In conclusion, conjugation of selected hydroxyprogesterones to selected positions in GnRH antagonists via appropriate linkers produce new tools for research and potential therapeutic applications. These conjugates have plasma protein binding GnRH receptor antagonism, progestogenic activity, and extended duration of action in marmosets. The compounds also have the potential for conjugation to moieties that convey oral bioavailability. These compounds thus constitute a new class of GnRH analogs for research and potential therapeutic application.
Acknowledgments
Judit Erchegyi kindly synthesized a large batch of conjugate A and Cathy Lilly undertook the studies on half-lives of GnRH analogs.
Footnotes
This work was supported by the University of Cape Town, the UK Medical Research Council, Ardana Bioscience Ltd., and National Institutes of Health Grant RO1-HD 039899 (to J.R.).
Present address for K.E.R.: Inveresk, Elphinstone Research Centre, Tranent, East Lothian EH33 2NE, United Kingdom.
First Published Online October 13, 2005
Abbreviations: CAT, Chloramphenicol acetyltransferase; DMF, N-dimethyl formamide.
Accepted for publication October 3, 2005.
References
Emons G, Ortmann O, Becker M, Irmer G, Springer B, Laun R, Holzel F, Schulz KD, Schally AV 1993 High affinity binding and direct antiproliferative effects of LHRH analogues in human ovarian cancer cell lines. Cancer Res 53:5439–5446
Casper RF 1991 Clinical uses of gonadotropin-releasing hormone analogues. CMAJ 144:153–158
Conn PM, Crowley WF 1994 Gonadotropin-releasing hormone and its analogs. Annu Rev Med 45:391–405
Moghissi KS 1992 Clinical applications of gonadotropin-releasing hormones in reproductive disorders. Endocrinol Metab Clin North Am 21:125–140
Millar RP, Assefa D, Ott T, Pawson A, Troskie B, Wakefield I, Katz A 1998 GnRH and GnRH analogues: structure, actions and clinical applications. Horm Front Gynecol 5:77–83
Anderson RA, Baird DT 2002 Male contraception. Endocr Rev 23:735–762
Thau RB 1984 Luteinizing hormone-releasing hormone (LHRH) and its analogs for contraception in women: a review. Contraception 29:143–162
Hodgen GD 1990 Intermittent GnRH antagonist plus progestin contraception conserving tonic ovarian estrogen secretion and reducing progestin exposure. Contraception 41:623–631
Handelsman DJ, Swerdloff RS 1986 Pharmacokinetics of gonadotropin-releasing hormone and its analogs. Endocr Rev 7:95–105
Barron JL, Millar RP, Searle D 1982 Metabolic clearance and plasma half-disappearance time of D-Trp6 and exogenous luteinizing hormone-releasing hormone. J Clin Endocrinol Metab 54:1169–1173
Bouchard P, Fauser BCJM 2000 Gonadotropin-releasing hormone antagonist: new tools vs. old habits. Fertil Steril 73:18–20
Millar RP, Zhu Y-F, Chen C, Struthers RS 2000 Progress towards the development of non-peptide orally-active gonadotropin-releasing hormone (GnRH) antagonists: therapeutic implications. Br Med Bull 56:761–772
Westphal U 1983 Steroid-protein interaction: from past to present. J Steroid Biochem 19:1–15
Mickelson KE, Forsthoefel J, Westphal U 1981 Steroid-protein interactions. Human corticosteroid binding globulin: some physicochemical properties and binding specificity. Biochemistry 20:6211–6218
Mattox VR, Litwiller RD, Nelson AN 1979 A comparison of procedures for attaching steroidal glucosiduronic acids to bovine serum albumin. J Steroid Biochem 10:167–172
Rajkowski KM, Cittanova N 1981 The efficiency of different coupling procedures for the linkage of oestriol-16a-glucuronide and pregnanediol-3a-glucuronide to four different enzymes. J Steroid Biochem 14:861–866
Chi L, Zhou W, Prikhozhan A, Flanagan C, Davidson JS, Golembo M, Illing N, Millar RP, Sealfon SC 1993 Cloning and characterization of the human GnRH receptor. Mol Cell Endocrinol 91:R1–R6
Millar RP, Davidson J, Flanagan CIW 1995 Ligand binding and second-messenger assays for cloned Gq/G11-coupled neuropeptide receptors: the GnRH receptor. Methods Neurosci 25:145–162
Ott TR, Troskie BE, Roeske RW, Illing N, Flanagan CA, Millar RP 2002 Two mutations in extracellular loop 2 of the human GnRH receptor convert an antagonist to an agonist. Mol Endocrinol 16:1079–1088
Flanagan CA, Fromme BJ, Davidson JS, Millar RP 1998 A high affinity gonadotropin-releasing hormone (GnRH) tracer, radioiodinated at position 6, facilitates analysis of mutant GnRH receptors. Endocrinology 139:4115–4119
Hammond GL, Lahteenmaki PL 1983 A versatile method for the determination of serum cortisol binding globulin and sex hormone binding globulin binding capacities. Clin Chim Acta 132:101–110
Westphal U, Stroupe SD, Cheng SL 1977 Progesterone binding to serum proteins. Ann NY Acad Sci 286:10–28
Hearn JP 1977 A device for restraining small monkeys. Lab Anim 11:261–262
Smith KB, Lunn SF, Fraser HM 1990 Inhibin secretion during the ovulatory cycle and pregnancy in the common marmoset monkey. J Endocrinol 126:489–495
Lunn SF, Dixson AF, Sandow J, Fraser HM 1990 Pituitary-testicular function is suppressed by an LHRH antagonist but not by an LHRH agonist in the marmoset monkey. J Endocrinol 125:233–239
Karten MJ, Rivier JE 1986 Gonadotropin-releasing hormone analog design. Structure-function studies towards the development of agonists and antagonists: rationale and perspective. Endocr Rev 7:44–66
Sealfon SC, Weinstein H, Millar RP 1997 Molecular mechanisms of ligand interaction with the gonadotropin-releasing hormone receptor. Endocr Rev 18:180–205
Millar RP, Lu Z, Pawson AJ, Flanagan CA, Morgan K, Maudsley SR 2004 Gonadotropin-releasing hormone receptors. Endocr Rev 25:235–275
Qi L, Nett TM, Allen MC, Sha X, Harrison GS, Frederick BA, Carwford ED, Glode ML 2004 Binding and cytotoxicity of conjugated and recombinant fusion proteins targeted to the gonadotropin-releasing hormone receptor. Cancer Res 64:2090–2095
Russell-Jones GJ, Westwood SW, Farnworth PG, Findlay JK, Burger HG 1995 Synthesis of LHRH antagonists suitable for oral administration via the vitamin B12 uptake system. Bioconjug Chem 6:34–42
Williams SP, Sigler PB 1998 Atomic structure of progesterone complexed with its receptor. Nature 393:392–396
Bursi R, Groen MB 2000 Application of (quantitative) structure-activity relationships to progestagens: from serendipity to structure-based design. Eur J Med Chem 35:787–796(Karen E. Ratcliffe, Hamish M. Fraser, Ro)
Peptide Biology Laboratory (J.R.), Salk Institute, La Jolla, California 92037
Department of Medical Biochemistry (R.P.M.), University of Cape Town, Observatory 7925, South Africa
Abstract
GnRH peptide analogs are widely used to treat diverse clinical conditions. However, they have poor oral activity and exhibit rapid metabolic clearance, thus requiring injection and depot formulation. Because steroid hormones are bound to plasma proteins, we explored the possibility of conjugating hydroxylated progesterones to GnRH analogs to reduce metabolic clearance of the peptides. Conjugation of [D-Lys6]GnRH agonist to the 11-hydroxyl of 11-hydroxyl progesterone via a hemi-succinate bridge increased the plasma half-life after iv injection in rabbits by 3.6-fold while retaining high binding affinity, thus providing proof of concept. Five GnRH antagonists were then synthesized with 21-hydroxyprogesterone conjugated via C21-hydroxyl to positions six (conjugates A and B) and position seven (conjugates C and D) of GnRH antagonists. In the fifth compound the NH2 terminus of a GnRH antagonist lacking the first two amino acids was conjugated via the C21-hydroxyl to 21-hydroxyprogesterone (conjugate E). All five analogs bound to guinea pig progesterone binding globulin with relatively high affinities (264–1020 nM). Moreover, all five conjugates retained high progestogenic activity in stimulating a progesterone-response-element-driven chloramphenicol acetyltransferase reporter gene in the T47D breast cancer cell line. Conjugation via the -amino function of D-Lys6 (conjugates A and B) produced compounds with high binding affinity for the human GnRH receptor (15 and 7 nM) comparable to that of the unconjugated GnRH antagonists (4 and 26 nM). Conjugation via the -amino function of Lys7 (conjugates C and D) or the NH2 terminus of an N-terminally truncated antagonist (conjugate E) produced compounds of low binding affinity. Conjugates A and B also exhibited high functional antagonism of GnRH stimulation of inositol phosphate production in COS-7 cells expressing the human GnRH receptor (2.6 and 16 nM) compared with the unconjugated antagonists (1.3 and 122 nM). In accordance with their poor receptor binding affinity, conjugates C, D, and E had poor functional antagonism. Preliminary dose-finding studies in female marmosets showed transitory progesterone inhibition by 0.25 mg and prolonged suppression of 12 and 17 d by 0.5- and 1.0-mg doses. Injection of conjugate A in adult male marmosets (0.5 mg sc) rapidly suppressed plasma testosterone levels, which remained suppressed for at least 3 d. In contrast, the unconjugated parent antagonist alone or with progesterone suppressed testosterone for only 8 h to 1 d. The findings demonstrate that conjugation of progesterone to GnRH antagonists conveys plasma binding and progestogenic properties and increases their efficacy and duration of action in vivo. These new GnRH antagonists show promise as therapeutic agents for hormone-dependent diseases and as contraceptives.
Introduction
GnRH ANALOGS ARE extensively used in the treatment of sex hormone-dependent diseases such as prostate cancer, endometriosis, and leiomyoma, and for precocious puberty (1, 2, 3, 4, 5) and show promise as contraceptives for men and women (6, 7, 8). Although GnRH peptide agonists and antagonists are most efficacious in these applications, they have limitations in poor oral bioavailability and rapid metabolic clearance (9, 10). Consequently they are administered as injectable slow release formulations (11). Although this approach overcomes some of the problems, it limits the ability to vary dosages and does not allow withdrawal of treatment when desired. As a result, a number of pharmaceutical companies are developing nonpeptide orally active small molecule GnRH antagonists (12).
Alternative approaches are to conjugate moieties to GnRH peptide antagonists, to convey plasma binding capacity, to reduce metabolic clearance, and to convey active absorption from the gastro-intestinal tract. The majority of steroid hormones are bound to plasma binding proteins (>90%) and exhibit reduced metabolic clearance (13, 14). Therefore, we have conjugated GnRH antagonists to steroid hormones and demonstrated that this reduced their metabolic clearance and extended their duration of action. Moreover, steroid hormone bioactivity was maintained, thereby creating bifunctional molecules with potentially improved efficacy.
Materials and Methods
Synthesis of GnRH analog-steroid conjugates
The conjugation method was adapted from that of Mattox et al. (15) and Rajkowski and Cittanova (16). All chemicals were obtained from Sigma-Aldrich Co. Ltd. (Poole, Dorset, UK) with the exception of radiochemicals purchased from Amersham Pharmacia Biotech UK Limited (Little Chalfont, Buckinghamshire, UK) or unless otherwise stated. 11-Hydroxyprogesterone or 21 hydroxyprogesterone (1.5 M) in N-dimethyl formamide (DMF) was added slowly with stirring to 5 M succinic anhydride in DMF, and an alkaline pH was maintained with tributylamine. The mixture was placed on ice for 2 h and then stood for 12 h at room temperature. Forty-eight hours after the addition of an equal volume of water, the crystalline product was removed and subjected to spectrophotomeric and chromatographic analysis to confirm the succinyl progesterone product (Fig. 1). An 11-estradiol hemisuccinate and 17-estradiol hemisuccinate were similarly produced but not reported here. For later studies, 21-succinyl 21-hydroxyprogesterone was purchased from Sigma-Aldrich. The GnRH analog [D-Lys6]GnRH and the two GnRH antagonists [Ac-D-Nal1, D-Cpa2, D-Pal3, Arg5, D-Lys6, D-Ala10]GnRH, designated peptide A, and [Ac-Pro1, D-Fpa2, D-Trp3, D-Lys6]GnRH, designated peptide B, (Table 1) were dissolved in 0.1 M phosphate buffer (pH 7.0) before addition of an equal volume of DMF. A 20-fold molar excess of 11-hydroxyprogesterone 11-hemisuccinate in the case of [D-Lys6]GnRH or (21-hydroxyprogesterone 21-hemisuccinate for GnRH antagonist conjugation) was dissolved in anhydrous DMF with equimolar 1-hydroxybenzotriazole and N,N-dicyclohexylcarbodiimide. The mixture was vortexed and left at room temperature for 1 h. After adjustment to a pH greater than 7 with tributylamine, the peptide-steroid mixture was left at 4 C for 20 h. The generic chemical synthesis used to conjugate progesterone to a Lys residue in GnRH analogs is shown in Fig. 2.
Purification and identification of products
An initial purification through a Sep-Pak C18 cartridge (Millipore UK Ltd., Harrow, Middlesex, UK) with ethylacetate followed by hexafluropropanol/DMF (70:30) was carried out before HPLC and mass spectrometry analysis of products. RP-HPLC purification was carried out on a Novapak C18 column (4 μm beads, 3.9 x 150 mm) connected to a Beckman Coulter System Gold LC125 pump and LC168 diode array detector (Beckman Coulter, Fullerton, CA). The buffer system was 0.1% trifluoroacetic acid in water as buffer A and 0.1% trifluoroacetic acid in acetonitrile as buffer B. The column was developed with a gradient of 10–100% buffer B over 30 min at a flow rate of 1 ml/min. Mass spectrometry was carried out on a Tofspec 2E matrix-assisted laser desorption ionization-time of flight mass spectrometer (Micromass UK Ltd., Manchester, UK) with a matrix of -cyano-4-hydroxycinnamic acid.
Additional GnRH antagonist progesterone conjugates
After successful proof of concept with peptide steroid conjugates A and B, larger batches were custom synthesized (Peninsula Labs, UK). Additional conjugates (C, D, and E) were designed and custom synthesized by Albachem Ltd. (Edinburgh, UK) using solid phase methodology and purified by HPLC to >98%. The structures of the peptides and sites of conjugation to 21-hydroxyprogesterone 21-hemisuccinate are shown in Table 1. Conjugates A and B explored the effects of conjugating via the -amino function of D-Lys at position 6, whereas C and D explored the effects of conjugating via position seven through the -amino function of L-Lys. Conjugate E examined the feasibility of conjugating to the NH2 terminus of peptide A lacking the first two amino acids, because we reasoned that the size and hydrophobicity of progesterone may substitute appropriately for these.
Determination of in vivo half-life in the male rabbit
To determine the biological half-life of GnRH agonist-progesterone conjugates, 0.5 ml saline containing 10,000,000 cpm of [125I-Tyr5, DAla6, MeLeu7, Pro9-N-ethylamide]GnRH or [125I-Tyr5, D-Lys6]GnRH-progesterone conjugate (iodinated as for [His5, D-Tyr6]GnRH) was injected as an iv bolus into the ear vein of eight male rabbits. Rabbits were sedated with 0.4 ml Aceprom 10 (Milborrow Animal Health, Rand, South Africa) injected im 3–4 min before the start of the experiment and again at 2 h. One-milliliter blood samples were collected in heparinized tubes from an indwelling cannula placed in a vein of the contralateral ear at 45 sec and 3, 5, 7, 9, 11, 13, 15, 17, 25, 30, 45, 60, 90, 120, 150, 180, and 210 min. Whole blood was counted directly to determine disappearance of the radio-iodinated analogs from the circulation. Experiments were approved by the University of Cape Town Animal Experimental Committee and carried out in accordance with the regulations of the Republic of South Africa.
Cell culture
Transient transfection.
Human GnRH receptor cDNA (17) for transient transfection was prepared using Maxi-prep columns (Qiagen, Chatsworth, CA). COS-7 cells were cultured in DMEM containing 10% fetal calf serum, glutamine, and penicillin/streptomycin. Cells were transfected with human GnRH receptor using Superfect (Qiagen) in Optimem media (Invitrogen Life Technologies, Paisley, UK) for 4 h (18). Transfected cells were cultured for a further 48 h in normal media before use in receptor binding and inositol phosphate assays.
Whole-cell receptor binding assays.
Whole-cell receptor binding assays (18, 19) used the [His 5, 125I-D-Tyr6]GnRH analog (20). Transiently transfected COS-7 cells in 12-well culture plates were incubated for 4 h on ice in HEPES/DMEM/0.1% BSA with 105 cpm per well [His5, 125I-D-Tyr6]GnRH and varying concentrations of unlabeled analogs. Cell monolayers were rapidly washed twice in ice-cold PBS and solubilized in 0.1 M NaOH, and radioactivity was counted to determine the amount of bound radioligand. Nonspecific binding (in the presence of consistently less than 10% of total binding) was determined using vector-transfected (pcDNA1/amp) COS-7 cells and was subtracted from total binding to give specific binding. Each data point was conducted in triplicate and experiments were repeated at least three times.
Total inositol phosphate assays
Determination of GnRH stimulation of total inositol phosphate production was as previously described (18). In brief, transiently transfected COS-7 cells were incubated with inositol-free DMEM containing 1% dialyzed heat-inactivated FCS and 1 μCi/well myo-[3H]inositol (Amersham Pharmacia Biotech, Piscataway, NJ) for 24–48 h. Medium was removed, and the cells were washed with 1 ml buffer (140 mM NaC1, 20 mM HEPES, 4 mM KC1, 8 mM glucose, 1 mM MgC12, 1 mM CaC12, and 1 mg/ml BSA) containing 10 mM LiCl and incubated for 1 h at 37 C in the same buffer containing 1 nM GnRH alone and with increasing concentrations of GnRH antagonist analogs. Reactions were terminated by the removal of medium and addition of 1 ml ice-cold 10 mM formic acid, and incubated for 30 min at 4 C. Total [3H]inositol phosphates were separated from the formic acid cell extracts on AG1-X8 anion exchange resin (Bio-Rad Laboratories, Hercules, CA) and eluted with a 1 M ammonium formate/0.1 M formic acid solution. Radioactivity was determined by liquid scintillation counting. Each data point was conducted in triplicate and experiments were repeated at least three times.
Plasma protein binding assay
Plasma protein binding was determined by the competitive binding of steroid conjugates in the presence of [1,2,6,7-3H]progesterone to pregnant guinea pig plasma (Charles River Laboratories, Wilmington, MA) (21). Progesterone binding globulin was purified by cation exchange chromatography (22). Twenty microliters of plasma equivalent was diluted with 2 ml dextran-coated charcoal solution and incubated at room temperature for 30 min to remove the endogenous steroids. The suspension was centrifuged at 3000 x g for 10 min and the supernatant was removed. One hundred μl of diluted plasma preparation was aliquotted into centrifuge tubes, followed by 1 pmol [1,2,6,7-3H]progesterone/100 μl PBS and varying concentrations of progesterone and the GnRH antagonist conjugates in 100 μl PBS. The tubes were vortexed and incubated at room temperature for 1 h, then for 15 min on ice. Seven hundred and fifty microliters of dextran-coated charcoal suspension (0.25 dextran T-70, charcoal decolorizing powder; Merck Ltd., Littleworth, UK) was added to the plasma solution and incubated for 10 min on ice, followed by centrifugation (3000 x g) at 4 C for 5 min. The supernatant (700 μl) was removed and radioactivity was quantified by liquid scintillation counting. Each data point was conducted in triplicate and experiments were repeated three times.
Progesterone receptor activation
Progestogenic activity was tested on the epithelial breast cancer cell line T47D, stably expressing the progesterone receptor and a chloramphenicol acetyltransferase (CAT) reporter gene (Dr. M. Beato, Institut fur Molekularbiologie und Tumorforschung, Marburg, Germany).
The T47D cells were maintained in RPMI 1640 medium supplemented with 2 mM glutamine, 10% FCS, penicillin, streptomycin, and insulin, transferrin, and sodium selenite supplement. Cells were plated in 100-mm dishes at a density of 2 million cells per dish, 24 h in advance of treatment with progesterone or hydroxyprogesterone-GnRH antagonist conjugates. Steroid or conjugate solution (100 μl) was added to 10 ml of medium and the cells incubated for a further 24 h.
Cells were washed in PBS (magnesium and calcium free) and 900 μl Reporter Lysis Buffer (Promega, Madison, WI) was applied. After 15 min of incubation at room temperature, the cells were scraped from the plate and transferred to a microcentrifuge tube on ice, vortexed for 10–15 sec, then heated at 60 C for 10 min to inactivate endogenous deacetylase activity. The cell lysates were then centrifuged for 2 min and the supernatant was transferred to microcentrifuge tubes containing [3H]chloramphenicol and n-butyryl CoA and made up to a total volume of 125 μl with water. Positive and negative controls were prepared containing a known amount of CAT or no cell extract respectively. The reaction mixtures were incubated at 37 C for 3 h, centrifuged briefly, and the reaction was terminated by the addition of 300 μl mixed xylenes. The xylene/reaction mixture was vortexed thoroughly and centrifuged for 3 min. Two hundred microliters of the xylene phase was removed and 100 μl 0.25 M Tris-HC1 (pH 8.0) was added before repeating the vortex and centrifuge procedure and removal of 150 μl of the xylene layer for liquid scintillation counting. The radioactivity measured in the negative control was subtracted from all values. Each data point was conducted in triplicate, and the experiment was repeated on three occasions.
In vivo studies in the male and female common marmoset
Marmosets were housed in the MRC Human Reproductive Sciences Unit Primate Centre. Experiments were carried out in accordance with the Animals (Scientific Procedures) Act, 1986, and were approved by the Local Ethical Review Process Committee. For collection of blood samples, marmosets were placed in a restraining device that allows routine sampling without stress (23). Samples were taken from the femoral vein into heparinized syringes and animals were returned to their cages between collections.
Female marmoset studies.
Preliminary dose-finding studies were undertaken in female marmosets to identify the effective dose required to inhibit corpus luteum function. Adult marmosets with regular ovulatory cycles as determined by measurement of plasma progesterone concentrations in blood samples collected three times per week were studied. A total of 0.25, 0.5, and 1.0 mg of GnRH antagonist-hydroxyprogesterone conjugate A or vehicle were administered by injection in 1 ml saline at two sc sites during the mid-luteal phase of the cycle. Blood samples were collected immediately before GnRH antagonist injection and at 4 and 8 h and d 1, 2, and 3 after treatment and thereafter at three times per week until the end of the posttreatment cycle. Plasma was assayed for progesterone as described previously (24). GnRH antagonist hydroxyprogesterone conjugate did not cross-react with the antibody in the RIA. As both doses inhibited progesterone, the 0.5 mg dose was selected for more detailed studies in male marmosets.
Male marmoset studies.
To determine the duration of action of the GnRH antagonist-progesterone conjugate at the GnRH receptor, 0.5 mg was administered as divided injections in 1 ml saline at two sc sites in six adult male marmosets. To compare the response with unconjugated antagonist, molar equivalents of unconjugated GnRH antagonist (0.41 mg) and progesterone (0.09 mg) were administered together to six adult male marmosets as for the conjugate. Unconjugated GnRH antagonist (0.5 mg) without progesterone was also administered to three marmosets. One 300-μl blood sample was collected immediately before and at 4 and 8 h after injection (d 0) and on d 1, 2, 3, 4, 6, 7, 8, and 10 after treatment. Plasma testosterone was assayed by RIA (25).
Statistical analysis
Effects of treatment on plasma testosterone concentrations were determined by ANOVA. Because normal testosterone concentrations in the male marmoset are extremely variable, data were log transformed before analysis and changes were compared with the mean of the two pretreatment values. The initial and secondary half-lives of GnRH analogs in rabbits were determined by PRISM 3.0 (GraphPad, San Deigo, CA). Differences were considered significant at a level of P < 0.05.
Results
Preliminary studies on receptor binding and half-life
Preliminary proof of concept studies established the successful chemical procedure for attachment of hydroxyprogesterones by C11 or C21 and estradiol by C11 or C17 to the -amino function of [D-Lys6]GnRH and the identification of products by mass spectrometry. These agonist conjugates were shown to retain high receptor binding affinity (kDa, 0.3 nM) and high efficacy in stimulating inositol phosphate production (ED50, 0.6 nM) by COS cells expressing the GnRH receptor (data not shown). The mean initial half-life of the [125I-Tyr5, D-Lys6]GnRH-hydroxyprogesterone 11-hemisuccinate conjugate was 9.1 ± 4.0 min (mean ± SEM, n = 5), which was longer but not significantly different from the half-life of the unconjugated [125I-Tyr5, D-Ala6, Me Leu7, Pro9-N-ethylamide]GnRH, which was 5.0 ± 1.8 min (mean ± SEM, n = 3) (Fig. 3). This lack of difference is anticipated as the initial rapid disappearance after bolus iv administration and is due predominantly to equilibration into extracellular spaces and first round renal clearance. The second phase longer half-life of the conjugate was 54.7 ± 13.2 min (mean ± SEM, n = 5), which was significantly greater (P = 0.0031) than that of the unconjugated peptide, which was 15.5 ± 7.0 min (mean ± SEM, n = 3). These findings thus established proof of concept and set the scene for the following more detailed studies with GnRH antagonist conjugates to hydroxyprogesterone.
Plasma protein binding
Plasma binding of GnRH antagonist-hydroxyprogesterone conjugates was studied with guinea pig progesterone binding globulin because of its high binding affinity and capacity compared with human plasma. Progesterone competed with [3H]progesterone with an IC50 of 96 ± 18 nM (n = 4). The conjugates retained progesterone binding globulin binding (IC50, 264-1020 nM), which was significantly higher (P < 0.05) than that of unconjugated progesterone (96 ± 18 nM) (Fig. 4, A and B, and Table 2). The unconjugated A and B peptides exhibited no competition for [3H]progesterone binding (Fig. 4C), and cortisol showed poor competition at 10 –5 M (data not shown). Robust binding data could not be obtained for [3H]progesterone with human transcortin.
Whole-cell GnRH receptor binding
GnRH antagonist 21-hydroxyprogesterone conjugates A and B had high binding affinities for the human GnRH receptor comparable to those of the unconjugated peptides (Fig. 5 and Table 2). Conjugates C, D, and E, in which 21-hydroxyprogesterone was conjugated to the -amino function of L-Lys in position seven or to the NH2 terminus of an antagonist lacking the first two amino acids, had low binding affinity (Fig. 5 and Table 2). Conjugate C, in which the natural carboxyl-terminal Arg-Pro sequence was substituted with Leu-Arg, retained affinity equivalent to that of D, although Pro in position 9 is conventionally regarded as essential for high binding affinity (26, 27, 28).
Inhibition of GnRH-stimulated inositol phosphate production
The antagonist activity of the conjugates was tested by measuring their ability to inhibit the stimulation of inositol phosphate by 1 nM GnRH in COS cells (Fig. 6 and Table 2). Conjugate A had an IC50 comparable to that of the unconjugated peptide (Fig. 6). The IC50 of conjugate B was improved relative to the unconjugated peptide (Fig. 6), suggesting it has greater antagonist efficacy as the binding affinities were similar. Conjugates C, D, and E competed poorly with GnRH in inhibiting inositol phosphate production (Fig. 6).
Progesterone receptor activation
All five conjugates were able to bind to and activate the progesterone receptor in T47D cells as measured by the stimulation of CAT enzyme activity (Fig. 7 and Table 2). Despite being tethered to the GnRH antagonists, their progestogenic activities were substantial, albeit less than that of progesterone. Only conjugates A and B were significantly (P < 0.03) less active than progesterone.
In vivo effects in the marmoset
Preliminary dose-finding studies were conducted in a female marmoset model. All three doses of conjugate A induced a rapid decline in plasma progesterone concentrations in female marmosets. The 0.25-mg dose induced a transitory suppression of progesterone (data not shown). Plasma progesterone concentrations were suppressed for 12 d (0.5 mg dose) and 17 d (1.0 mg dose) compared with 4–5 d of low follicular phase progesterone in the control cycles (Fig. 8).
In more detailed studies in male marmosets, administration of 0.5 mg of conjugate A, antagonist A, or antagonist A plus progesterone induced a rapid suppression in plasma testosterone concentrations (Fig. 9). Significant suppression of testosterone was maintained for 8–24 h after administration of peptide A or peptide A plus progesterone, but for at least 3 d after treatment with conjugate A. Plasma testosterone was significantly lower (P < 0.05) in the conjugate A-treated group on d 1, 2, and 3 after treatment (Fig. 9) when compared with treatments with peptide A alone or peptide A with progesterone, demonstrating the prolonged duration of action of the conjugate.
Discussion
GnRH peptide analogs are widely used in the treatment of infertility and hormone-dependent diseases (1, 2, 3, 4, 5). They also show promise as a new generation of contraceptives for men and women when combined with steroid hormone replacement (6, 7). Although current GnRH analogs have some excellent therapeutic properties, their peptide nature conveys rapid metabolic clearance and poor oral bioavailability. Consequently, they are administered by daily injection or by formulation in slow-release biodegradable polymers (11). These depot injections have several disadvantages: 1) in the discomfort and side-effects of a depot injection, 2) in the inability to terminate treatment when desired, and 3) in the lack of flexibility in dosage. The development of small molecule nonpeptide orally active GnRH antagonist analogs may overcome these problems (12). Nevertheless, because GnRH peptide analogs have many excellent properties, we considered that an alternative to nonpeptide analog development was to conjugate chemical moieties to GnRH antagonists, which would convey reduced metabolic clearance and oral bioavailability. Conjugation of GnRH analogs to steroid hormones may potentially be used to render plasma protein binding properties, reduce metabolic clearance and extend analog half-life in the circulation. Our current findings demonstrate that this is achievable and that such analogs can also be designed to be bifunctional in targeting both GnRH and steroid receptors.
The sites of conjugation of the GnRH analogs and steroid hormone are critical in maintaining GnRH receptor binding, plasma protein binding, and steroid receptor binding and activation. To this end, we explored the effects of conjugation via a D-amino acid in position six, and an L-amino acid in position seven of GnRH antagonists as well as the amino terminus in position 3 of a truncated GnRH antagonist, via a hydroxyl at C11 and C21 of steroids. These conjugates were successfully synthesized and biological characterization revealed that conjugation of [D-Lys6] of GnRH antagonists with C21 of 21-hydroxyprogesterone produced compounds with good plasma protein binding, high GnRH receptor binding affinity, potent inhibition of inositol phosphate production, and good potency in progesterone receptor activation. This suggests that the sites chosen and the long side chain attachment (eight carbon atoms plus one nitrogen) avoid disruption of crucial binding sites and steric hindrance.
All of the conjugates bound guinea pig progesterone binding globulin with relatively high affinity. The affinity of conjugate E was two to four times higher than that of the other conjugates, suggesting that there is less steric hindrance when conjugation is through the amino terminus of this truncated GnRH antagonist. Although binding to plasma proteins of man and other mammals was not studied, the extended half-life of a conjugate in rabbits together with the findings of binding to progesterone binding globulin and the extended duration of action in marmosets suggest that the conjugates would behave similarly in man. The preservation of the essential features for progesterone binding to the human cortisol binding globulin (20-oxo, 10 methyl, 3-oxo, and 4-ene groups) (14) in the conjugates reinforces this conclusion. Because a hydroxyl group in the 11 position impairs binding, whereas the 21-hydroxyl group has no effect (14), we chose to conjugate via a hydroxyl at C21.
Conjugates A and B bound to the human GnRH receptor with high affinity comparable with the unconjugated antagonists. Conjugation via the D-Lys6 in these molecules was expected to yield high affinity binding analogs as GnRH assumes a folded II/ conformation around Gly6 when bound to the GnRH receptor and substitution of Gly6 with a D-amino acid enhances this conformation and binding affinity (26, 27, 28). This property has been exploited previously with good effect in GnRH analog conjugation to toxins (29) and vitamin B12 (30). The long attachment chain used in our conjugate clearly avoids steric hindrance by the progesterone moiety. Position seven of GnRH analogs has not previously been explored as an attachment site. Because conjugates C and D had much lower binding affinities, conjugation via an L-Lys in position seven appears to hinder receptor binding. However, it does provide the potential for conjugation of two different moieties in positions six and seven (e.g. steroid and vitamin B12 to convey extended half-life, steroid bioactivity, and oral bioavailability), albeit with a loss in binding affinity. This feature was designed into conjugates C and D. Conjugate E, in which progesterone replaced the first two N-terminal residues of peptide A, had low binding affinity despite our attempts to maintain size and hydrophobicity (the molecular weight of the two N-terminal amino acids in the antagonist was 416.5, which is comparable with that of the hydroxyprogesterone hemisuccinate replacement at 413.5). This suggests that the progesterone moiety is incorrectly orientated, creates steric hindrance to binding, or makes inappropriate contacts with the GnRH receptor. Conjugate C is a completely novel GnRH analog. All of the several thousand active GnRH analogs retain Pro9 as in native GnRH because this residue was thought to be essential for bioactivity (27). Our data confirm this but demonstrate that reasonable binding affinity is retained when Pro9 is substituted with Arg (cf. C and D).
Conjugates A and B were equally or more effective than the unconjugated peptides in antagonizing GnRH stimulation of inositol phosphate production. These conjugates thus had the best potential for in vivo activity. Conjugate A was about 8-fold more efficacious in inhibiting inositol phosphate production and, therefore, was tested in marmosets. All of the conjugates were able to bind and activate the progesterone receptor although their activity was less than that of progesterone. This finding indicates that the conjugates are able to enter the cells intact or that the progesterone moiety is hydrolyzed from the antagonist to enter the cells. Although the latter possibility cannot be excluded, it seems less likely, because a monolayer of cells would have little capacity to hydrolyze a substantial amount of the conjugate in the comparatively large volume of medium. The conjugation of 21-hydroxyprogesterone to a long and flexible side chain does accommodate the requirement for binding to the progesterone receptor. The ligand-binding domain of the progesterone receptor has been crystallized, allowing the analysis of some aspects of ligand-receptor interactions (31). The most important feature for binding to the progesterone receptor is the C3 keto group, which forms a hydrogen bond with the conserved steroid receptor glutamine 725 (31, 32). Arginine 766 and phenylalanine 778 also make van der Waals contacts with the A ring (through intervening fixed water sites) to tightly couple the ligand to the receptor. The methyl-ketone substituent (projecting from C17) apparently interacts with cysteine 891 and threonine 894. Numerous other interactions are made with the A, B, C, and D ring. Interestingly, no contacts are made with the C21 methyl, which also influenced our decision to choose this site for conjugation and may account for the high bioactivity at the progesterone receptor. There is also a sizeable space in the progesterone receptor opposite C21 of the docked progesterone.
Preliminary studies in female marmosets revealed that 0.25 mg of conjugate A was sufficient to induce an immediate but transient inhibition of progesterone. Higher doses of 0.5 and 1.0 mg had a sustained inhibitory effect on plasma progesterone concentrations and delayed subsequent ovulation. On the basis of these findings, male marmosets were treated with 0.5 mg conjugate A or 0.5 mg peptide A. Conjugate A suppressed testosterone for at least 3 d (recovery on d 6), whereas the unconjugated peptide A suppressed testosterone for only 8 h.
The enhanced duration of action of conjugate A compared with that of the unconjugated antagonist A could be considered to be due to the combined biological activities of the antagonist and progesterone in inhibiting gonadotropin. However, it is unlikely that the progesterone activity would last for 3 d after a single im injection of about 0.09 mg (the progesterone component of 0.5 mg conjugate). To rule out this possibility, we compared the effects of 0.5 mg conjugate A with the equivalent dose of free antagonist A (0.41 mg) and free progesterone (0.09 mg) and showed that the conjugate is much more effective than the molar equivalent of free antagonist and progesterone. Because the testosterone inhibition profile by peptide A was the same as the combination of peptide A plus progesterone it is evident that the progesterone moiety of the conjugate extends the duration of action of the conjugate through prolongation of its presence in the circulation and not through its inhibitory progestogenic effects on gonadotropin secretion.
In conclusion, conjugation of selected hydroxyprogesterones to selected positions in GnRH antagonists via appropriate linkers produce new tools for research and potential therapeutic applications. These conjugates have plasma protein binding GnRH receptor antagonism, progestogenic activity, and extended duration of action in marmosets. The compounds also have the potential for conjugation to moieties that convey oral bioavailability. These compounds thus constitute a new class of GnRH analogs for research and potential therapeutic application.
Acknowledgments
Judit Erchegyi kindly synthesized a large batch of conjugate A and Cathy Lilly undertook the studies on half-lives of GnRH analogs.
Footnotes
This work was supported by the University of Cape Town, the UK Medical Research Council, Ardana Bioscience Ltd., and National Institutes of Health Grant RO1-HD 039899 (to J.R.).
Present address for K.E.R.: Inveresk, Elphinstone Research Centre, Tranent, East Lothian EH33 2NE, United Kingdom.
First Published Online October 13, 2005
Abbreviations: CAT, Chloramphenicol acetyltransferase; DMF, N-dimethyl formamide.
Accepted for publication October 3, 2005.
References
Emons G, Ortmann O, Becker M, Irmer G, Springer B, Laun R, Holzel F, Schulz KD, Schally AV 1993 High affinity binding and direct antiproliferative effects of LHRH analogues in human ovarian cancer cell lines. Cancer Res 53:5439–5446
Casper RF 1991 Clinical uses of gonadotropin-releasing hormone analogues. CMAJ 144:153–158
Conn PM, Crowley WF 1994 Gonadotropin-releasing hormone and its analogs. Annu Rev Med 45:391–405
Moghissi KS 1992 Clinical applications of gonadotropin-releasing hormones in reproductive disorders. Endocrinol Metab Clin North Am 21:125–140
Millar RP, Assefa D, Ott T, Pawson A, Troskie B, Wakefield I, Katz A 1998 GnRH and GnRH analogues: structure, actions and clinical applications. Horm Front Gynecol 5:77–83
Anderson RA, Baird DT 2002 Male contraception. Endocr Rev 23:735–762
Thau RB 1984 Luteinizing hormone-releasing hormone (LHRH) and its analogs for contraception in women: a review. Contraception 29:143–162
Hodgen GD 1990 Intermittent GnRH antagonist plus progestin contraception conserving tonic ovarian estrogen secretion and reducing progestin exposure. Contraception 41:623–631
Handelsman DJ, Swerdloff RS 1986 Pharmacokinetics of gonadotropin-releasing hormone and its analogs. Endocr Rev 7:95–105
Barron JL, Millar RP, Searle D 1982 Metabolic clearance and plasma half-disappearance time of D-Trp6 and exogenous luteinizing hormone-releasing hormone. J Clin Endocrinol Metab 54:1169–1173
Bouchard P, Fauser BCJM 2000 Gonadotropin-releasing hormone antagonist: new tools vs. old habits. Fertil Steril 73:18–20
Millar RP, Zhu Y-F, Chen C, Struthers RS 2000 Progress towards the development of non-peptide orally-active gonadotropin-releasing hormone (GnRH) antagonists: therapeutic implications. Br Med Bull 56:761–772
Westphal U 1983 Steroid-protein interaction: from past to present. J Steroid Biochem 19:1–15
Mickelson KE, Forsthoefel J, Westphal U 1981 Steroid-protein interactions. Human corticosteroid binding globulin: some physicochemical properties and binding specificity. Biochemistry 20:6211–6218
Mattox VR, Litwiller RD, Nelson AN 1979 A comparison of procedures for attaching steroidal glucosiduronic acids to bovine serum albumin. J Steroid Biochem 10:167–172
Rajkowski KM, Cittanova N 1981 The efficiency of different coupling procedures for the linkage of oestriol-16a-glucuronide and pregnanediol-3a-glucuronide to four different enzymes. J Steroid Biochem 14:861–866
Chi L, Zhou W, Prikhozhan A, Flanagan C, Davidson JS, Golembo M, Illing N, Millar RP, Sealfon SC 1993 Cloning and characterization of the human GnRH receptor. Mol Cell Endocrinol 91:R1–R6
Millar RP, Davidson J, Flanagan CIW 1995 Ligand binding and second-messenger assays for cloned Gq/G11-coupled neuropeptide receptors: the GnRH receptor. Methods Neurosci 25:145–162
Ott TR, Troskie BE, Roeske RW, Illing N, Flanagan CA, Millar RP 2002 Two mutations in extracellular loop 2 of the human GnRH receptor convert an antagonist to an agonist. Mol Endocrinol 16:1079–1088
Flanagan CA, Fromme BJ, Davidson JS, Millar RP 1998 A high affinity gonadotropin-releasing hormone (GnRH) tracer, radioiodinated at position 6, facilitates analysis of mutant GnRH receptors. Endocrinology 139:4115–4119
Hammond GL, Lahteenmaki PL 1983 A versatile method for the determination of serum cortisol binding globulin and sex hormone binding globulin binding capacities. Clin Chim Acta 132:101–110
Westphal U, Stroupe SD, Cheng SL 1977 Progesterone binding to serum proteins. Ann NY Acad Sci 286:10–28
Hearn JP 1977 A device for restraining small monkeys. Lab Anim 11:261–262
Smith KB, Lunn SF, Fraser HM 1990 Inhibin secretion during the ovulatory cycle and pregnancy in the common marmoset monkey. J Endocrinol 126:489–495
Lunn SF, Dixson AF, Sandow J, Fraser HM 1990 Pituitary-testicular function is suppressed by an LHRH antagonist but not by an LHRH agonist in the marmoset monkey. J Endocrinol 125:233–239
Karten MJ, Rivier JE 1986 Gonadotropin-releasing hormone analog design. Structure-function studies towards the development of agonists and antagonists: rationale and perspective. Endocr Rev 7:44–66
Sealfon SC, Weinstein H, Millar RP 1997 Molecular mechanisms of ligand interaction with the gonadotropin-releasing hormone receptor. Endocr Rev 18:180–205
Millar RP, Lu Z, Pawson AJ, Flanagan CA, Morgan K, Maudsley SR 2004 Gonadotropin-releasing hormone receptors. Endocr Rev 25:235–275
Qi L, Nett TM, Allen MC, Sha X, Harrison GS, Frederick BA, Carwford ED, Glode ML 2004 Binding and cytotoxicity of conjugated and recombinant fusion proteins targeted to the gonadotropin-releasing hormone receptor. Cancer Res 64:2090–2095
Russell-Jones GJ, Westwood SW, Farnworth PG, Findlay JK, Burger HG 1995 Synthesis of LHRH antagonists suitable for oral administration via the vitamin B12 uptake system. Bioconjug Chem 6:34–42
Williams SP, Sigler PB 1998 Atomic structure of progesterone complexed with its receptor. Nature 393:392–396
Bursi R, Groen MB 2000 Application of (quantitative) structure-activity relationships to progestagens: from serendipity to structure-based design. Eur J Med Chem 35:787–796(Karen E. Ratcliffe, Hamish M. Fraser, Ro)