Articles Pharmacokinetics and pharmacodynamics of recombinant human chorionic gonadotrophin in healthy male and female volunteers

Transcription

1 RBMOnline - Vol 4. No Reproductive BioMedicine Online; on web 8 January 2002 Articles Pharmacokinetics and pharmacodynamics of recombinant human chorionic gonadotrophin in healthy male and female volunteers Isabelle Trinchard-Lugan obtained her State Doctorate of Pharmacy in 1983 and in Pharmaceutical Sciences in A postgraduate fellowship in Endocrinology led her to integrate Clinical Research into the industry, first as a clinical research associate and then as a manager. She came to Clinical Pharmacology, where she has held various positions related to preparation, analysis, reporting of pharmacokinetic and pharmacodynamic clinical trials. Currently, she is heading the study management of the Clinical Pharmacology group at Serono International SA regarding exploratory and confirmatory pharmacokinetic and pharmacodynamic studies. She has participated as author or co-author in a number of papers. She also is the happy mother of two young boys. Dr Isabelle Trinchard-Lugan I Trinchard-Lugan 1,4, A Khan 2, HC Porchet 3, A Munafo 1 1 Serono International SA, 12, chemin des Aulx, 1228 Plan-les-Ouates, Geneva, Switzerland; 2 Medeval Ltd, Manchester University, Manchester, UK; 3 Debiopharm, Lausanne, Switzerland 4 Correspondence: Tel: ; Fax: ; Isabelle.Trinchard-Lugan@serono.com Abstract The pharmacokinetics and pharmacodynamics of recombinant human chorionic gonadotrophin (rhcg) were investigated in three studies of healthy volunteers. After single intravenous doses of 25, 250 and 1000 μg, rhcg and urinary HCG (uhcg) showed linear pharmacokinetics described by a bi-exponential model, although the area under the curve (AUC) for uhcg was ~29% lower than for rhcg. After intramuscular or subcutaneous administration (absolute bioavailability, 40 50% for both), rhcg pharmacokinetics could be described by a first-order absorption, one-compartment model. During multiple subcutaneous dosing, the amount of HCG increased by ~1.7-fold. A comparison of liquid and freeze-dried rhcg and freezedried uhcg showed pharmacokinetic bioequivalence. In down-regulated male subjects, single doses of 125 μg rhcg, given intravenously, intramuscularly or subcutaneously, produced comparable increases in serum testosterone, inhibin and 17βoestradiol, with little further increase during repeated subcutaneous administration (in female subjects, this produced a sustained comparable increase in serum androstenedione and testosterone concentrations). In conclusion, the pharmacokinetics and pharmacodynamics of rhcg are similar to those of uhcg and are not affected by the use of different formulations. In healthy subjects, rhcg produces pharmacodynamic responses consistent with HCG physiology and is suitable for use in the same clinical indications as uhcg. The secured source and high purity of rhcg may offer important advantages. Keywords: pharmacodynamics, pharmacokinetics, rhcg, uhcg 106 Introduction Human chorionic gonadotrophin (HCG) is a glycoprotein hormone that is secreted by the trophoblasts as early as 6 days after conception and that stimulates corpus luteum and early feto-placental function (Birken et al., 1990). It consists of a single α-subunit of 92 amino acids, which is common to all the pituitary glycoprotein hormones, and a specific β-subunit of 145 amino acids (Bahl, 1973; Pierce and Parsons, 1981). HCG shows substantial structural homology with human LH, and the two hormones bind to the same receptor. As a result of this similarity to LH, HCG is used pharmacologically in a number of clinical indications. In women, HCG is used to treat corpus luteum dysfunction, to mimic the pre-ovulatory LH surge in women undergoing ovulation induction or assisted reproduction techniques, or to prolong corpus luteum function in women undergoing assisted reproduction techniques (Hutchinson-Williams et al., 1990). In men, HCG can be used in support of FSH therapy to induce spermatogenesis, or to treat prepubertal cryptorchidism (Dickerman et al., 1983) that is not due to anatomical obstruction. To date, the only therapeutic formulations of HCG available have been derived from urinary HCG (uhcg). Recently, however, a recombinant HCG (rhcg) preparation has been developed (Ovitrelle, Serono International, Geneva, Switzerland). This agent is expressed in cultured Chinese hamster ovary (CHO) cells bearing the gene coding for human HCG (common α-subunit and specific β-subunit). The hormone is extracted from the culture medium by conventional

2 column chromatography techniques, yielding a preparation of high purity. rhcg offers all the recognized advantages of recombinant technology, e.g. remarkable batch-to-batch consistency, high purity and high specific activity (Fonjallaz and Loumaye, 2001; Hugues, 2001). It consists of the same protein moieties as in uhcg. However, slight differences in terms of glycosylation can occur (owing to different production techniques), because they are produced in different cellular environments. This paper describes studies of the pharmacokinetics and pharmacodynamics of rhcg in healthy volunteers. Materials and methods Three open, pharmacokinetic and pharmacodynamic studies were performed in healthy volunteers. Study 1 compared the pharmacokinetics of rhcg after intravenous administration of three doses with those of a standard dose of uhcg (5000 IU) in six males and six females. Study 2 assessed the pharmacokinetics and pharmacodynamics of rhcg after single subcutaneous, intramuscular and intravenous doses and repeated subcutaneous doses in six males and six females. Study 3 assessed the pharmacokinetic bioequivalence of liquid and freeze-dried formulations of rhcg after single subcutaneous doses in 12 males and 12 females. Also, the pharmacodynamic effects on testosterone, inhibin (A and B together), 17ß-oestradiol and androstenedione of both recombinant formulations and uhcg were evaluated. All studies were conducted according to the Declaration of Helsinki, and were approved by the Medieval Independent Ethics Committee, Manchester, UK. Experimental subjects Male and female subjects were aged between 18 and 40 years. They were within 15% of their ideal body weight, as determined by the Lorenz formula (Halme et al., 1986). Female subjects were required to have used an oestrogen progestogen monophasic oral contraceptive for at least two cycles before inclusion in the study. In Study 2, male subjects underwent down-regulation of their pituitary gonadal axis in order to suppress their testosterone secretion, as described below. Written informed consent was obtained from all subjects prior to inclusion in the studies. The volunteers good health was assessed by enquiry and by measurement of vital signs pre-dose and at 2, 6 and 12 h post-dose. Routine biochemistry, haematology and HCG antibody samples were collected pre-dose and post-study. Experimental designs Study 1 Six male subjects (age years, weight kg) and six female subjects (age years, weight kg) received single 25, 250 and 1000 μg (500, 5000 and 20,000 IU) intravenous doses of rhcg, and a single 5000 IU intravenous dose of uhcg, at intervals of 2 weeks. Blood samples for measurement of serum HCG concentrations were obtained before each dose and at 5, 15 and 30 min, 1, 2, 4, 6, 9, 12, 24, 36, 48, 72, 96, 120, 144, 168 and 192 h after dosing. Serum was separated and aliquoted for storage at 20 C until required for analysis. Urine was collected 0 2, 2 6, 6 12, 12 36, 36 60, 60 84, , , and h after administration of rhcg. The total volume and time of collection of each sample were noted, and samples were stored at 2 8 C prior to analysis. Study 2 Six male subjects (age years, weight kg) and six female subjects (age years, weight kg) took part in this study. Male subjects received four subcutaneous injections of goserelin, 3.6 mg (Zoladex implant, AstraZeneca, UK), at 28-day intervals prior to the first dose of rhcg; down-regulation of the pituitary gonadal axis was confirmed by measurement of serum testosterone ( 1.5 nmol/l) and LH ( 2.5 IU/l) at least 20 days after the first injection of goserelin. All subjects received single 125 μg doses of rhcg, by subcutaneous (in the anterior abdominal wall), intramuscular (in the buttock) and intravenous administration in random order at 2-week intervals. Blood samples were obtained before dosing, and at 0.5, 1, 2, 4, 6, 9, 12, 24, 48, 96, 120, 144 and 192 h after dosing; additional samples were obtained 5 and 15 min after intravenous administration. Subjects also received five 125 μg subcutaneous injections of rhcg every second day, starting 2 weeks after the last single dose. Blood samples were obtained before dosing and at 0.5, 1, 2, 4, 6, 9, 12, 24, 48, 96, 144, 192, 192.5, 193, 194, 196, 198, 201, 204, 216, 240, 264, 288, 312, 336 and 384 h after the first of these multiple doses. All blood samples were processed as described above. Study 3 Twelve male subjects (age years, weight kg) and 12 female subjects (age years, weight kg) received single 250 μg (5000 IU) subcutaneous doses of the rhcg liquid formulation, the freeze-dried formulation, and 5000 IU uhcg in randomized order at 3-week intervals. Blood samples were obtained before dosing and at 0.5, 1, 2, 4, 6, 9, 12, 24, 36, 48, 72, 96, 120, 144, 192, 240 and 288 h after each dose and were processed as described above. Measurement of serum HCG Serum and urine concentrations of HCG were measured by immunoradiometric assay (MAIAclone, Serono Biodata, Italy). Originally calibrated against an international standard (1st IRP/3rd IS 75/537), the immunoassay has subsequently been standardized against rhcg (Serono International) and internally validated according to good laboratory practice (GLP) guidelines and company standard operating procedures to set up assay specifications in a reproducible manner. In studies 1 and 3, the standard range was μg/l ( IU/l), the sensitivity was 0.05 μg/l (1 IU/l) and the total assay coefficient of variation (including intra- and inter-assay elements) was 5% or less. In study 2, the standard range was μg/l ( IU/l) and the sensitivity was 0.13 μg/l (2.5 IU/l). The total coefficient of variation (including intra- and inter-assay elements) for the immunoassay used in study 2 was 7.5%. In this study, HCG was also measured by an in-vitro bioassay (Dahl and Sarkissian, 1993), which consisted of mouse Leydig tumour cells (MA-10 clone strain) in which an increase in production of progesterone was measured in response to LH. Here the response was triggered not by LH but 107

3 108 by HCG contained in the serum sample. It was based on the non-specificity of the receptor as the HCG receptor is the same as the LH receptor. The standard range for this assay was μg/l ( IU/l), the sensitivity was 0.05 μg/l and the total coefficient of variation (including intra- and interassay elements) was 14.5%. Measurement of pharmacodynamic markers Serum testosterone was measured by a solid-phase coated-tube radioimmunoassay (Coat-A-Count, Diagnostic Products Corporation, Los Angeles, CA, USA). The standard range was nmol/l, the sensitivity was 0.5 nmol/l and the total assay coefficient of variation (including intra- and inter-assay elements) was <12%. Androstenedione, 17β-oestradiol and free T 4 were measured by radioimmunoassay (androstenedione, Diagnostic Laboratories, Webster, USA; 17β-oestradiol, Diagnostic Products Corporation, Los Angeles, CA, USA; free T 4, Kodak Clinical Diagnostics, Amersham, UK). The limits of quantification were 1.0 nmol/l, 100 pmol/l and 0.4 pmol/l, respectively, and the assay coefficient of variation was <11%. Inhibin was measured by an immunoenzymetric assay (Medgenix Diagnostics SA, Fleurus, Belgium), with a sensitivity of 0.4 U/ml. The assay coefficient of variation for inhibin was <17%. Measurement of safety Safety parameters were assessed by routine validated methods. Antibodies against rhcg were measured by an in-house validated radioimmunoprecipitation assay. Briefly radiolabelled HCG was incubated with the serum samples for a given time. In the presence of antibodies against HCG, labelled HCG binds to these antibodies to form a complex. This complex was precipitated by addition of polyethylene glycol and the radioactivity was measured in the precipitate: the higher the antibody titre, the higher the radioactivity. Pharmacokinetic analysis In all three studies, pharmacokinetic data were analysed by model-independent techniques; in addition, mathematical modelling was used in studies 1 and 2. Model-independent methods Peak serum concentrations (C max ) and time to peak concentrations (t max ) after intramuscular or subcutaneous administration were derived directly from the observed data. The area under the serum concentration time curve to the last quantifiable time point (AUC t ) was determined by the trapezoidal rule, and the terminal rate constant (λ z ) and terminal half-life (t 1/2 : calculated as ln (2)/λ z ) were estimated using Siphar version 4 software (Simed, Créteil, France). The total AUC was calculated as the sum of AUC t and C t /λ z, where C t is the last quantifiable concentration. Total serum clearance (CL) was calculated as dose/auc, and the terminal phase volume of distribution (V z ) as CL/λ z. The total area under the serum concentration time first moment curve (AUMC) was calculated as (C. t t/λ z + C t /λ 2 z ). The mean residence time (MRT) after intravenous administration was then calculated as AUMC/AUC, and the steady-state volume of distribution (V ss ) as MRT CL. Renal clearance (CL r ) was calculated as Ae/AUC, where Ae is the cumulative amount excreted unchanged in the urine. In study 2, the mean absorption time (MAT) for the single subcutaneous and intramuscular doses was calculated as (AUMC/AUC) MRT. The accumulation ratio during multiple subcutaneous dosing was calculated as (AUC /AUC 0 48 ) (first dose/mean dose), where AUC and AUC 0 48 are the AUC after the last and first doses, respectively, and the mean dose is the mean of the five administered doses. Modelling techniques In studies 1 and 2, weighted non-linear regression techniques were used to fit a bi-exponential model to the data obtained after intravenous administration of rhcg. The model was described by the formula C = C 1 e λ 1t + C z e λ zt, where C 1 and C z are the coefficients, and λ 1 and λ z are the two corresponding disposition rate constants. Terminal half-life, total serum clearance and steady-state volume of distribution were then calculated by the formulae described above. Modelderived AUC was calculated as C 1 /λ 1 + C z /λ z, and AUMC was calculated as C 1 /λ C z /λ z 2. The volume of distribution of the terminal phase was calculated as CL/λ z. The initial volume of distribution (V c ) was calculated as dose/(c 1 + C z ). In study 2, the data obtained after single intramuscular and subcutaneous doses were modelled using a Bateman function, C = C 0 (e λ zt e kat ), where C 0 is the pre-exponential coefficient, λ z the elimination rate constant and k a the absorption rate constant. Extravascular AUC was calculated as C 0 /λ z C 0 /k a. The half-lives associated with the exponents k a and λ z (t 1/2a and t 1/2 ) were calculated as ln (2)/k a and ln (2)/λ z, respectively. Pharmacokinetics after multiple subcutaneous doses were analysed according to the model C* = C 0 [1 e Nλ zτ /1 e λ zτ ] e λ z t* C0 [1 e Nkaτ /1 e kaτ ]. e kat*, where C * is the concentration during the N th dosing interval, and t * the time during that interval; the other parameters were derived as described above. A weighting scheme of 1/y 2 (calc) was used for all curve fitting. Model selection and selection of the weighting scheme were based upon visual inspection of the relationship between the fitted line and the experimental observations, visual inspection of the residual plots, and values for the information criteria and the statistics associated with the model parameters (Yamaoka et al., 1978; Everitt, 1998).

4 Statistical analysis of pharmacokinetic data Study 1 Dose-normalized AUC for the three rhcg treatments were logtransformed and compared by analysis of variance (ANOVA), with subject and treatment as factors. Model-independent values for AUC in the comparison of rhcg with uhcg were analysed in the same way, with factors for treatment, subject, period and treatment-by-period interaction. All analyses were performed using SAS (version 6.10) software. Study 2 Dose-normalized AUC, C 0, C max, t 1/2a, MAT and absolute bioavailability (F) after single subcutaneous and intramuscular administrations were compared by multivariate ANOVA on log-transformed data with factors for treatment, period, treatment-by-gender interactions, period-by-gender interactions, and treatment-by-period interactions (Hills and Armitage, 1979). In addition, 90% confidence limits were calculated for dose-normalized C max and AUC. T max values were analysed using Wilcoxon s matched-pairs test. The pharmacokinetics of rhcg after a single subcutaneous dose were also compared with those following the last dose of the multiple subcutaneous dosing regimen; since this regimen was always given during the last treatment period, the analysis included factors for treatment and gender only. All analyses were performed using SPSS version 5 software (SPSS, Chicago, USA). Study 3 Bioequivalence between the two formulations of rhcg was tested by the two one-sided tests procedure of the Food and Drug Administration (FDA) Bioequivalence Division (FDA, Code of Federal Regulations). C max and AUC were logtransformed and analysed by ANOVA, with factors for subject, sequence, period, treatment and gender. Least-squares means were calculated for each treatment, together with the adjusted differences between treatment means and the associated standard errors. The 90% confidence interval for the difference between the means was then calculated. A range of % for the ratio of treatment means was regarded as indicating equivalence between treatments. T max data were analysed using Wilcoxon s test. All analyses were performed using SAS (version 6.10) software. Pharmacodynamic analysis In studies 2 and 3, pharmacodynamic maximum effects and the time to maximum effect were derived directly from the observed data; no modelling was performed due to the known circadian rhythms of androstenedione and testosterone levels. Statistical analysis of the pharmacodynamic data Study 2 In males, the area under the concentration time curve calculated by trapezoidal summation, after subtraction of baseline values, represented the integrated response for each marker. The maximum effects and integrated responses after single doses were compared by analysis of variance (ANOVA) for a latin square design. Following multiple subcutaneous administration, the same parameters were compared with those obtained following a single subcutaneous dose by means of a two-tailed paired Student s t-test. P-values below 0.05 were regarded as significant. Results All coefficients of variation (CV) were below 12% for immunoassay except in one instance for inhibin (17%). The usual acceptance criterion is below 20%. The CV of the in-vitro bioassay was 14.5%, which is a good performance as far as a bioassay is concerned. Table 1. Modelling-derived pharmacokinetic parameters (means ± SD) of rhcg and uhcg after increasing intravenous administrations to healthy males and females (n = 12). Dose rhcg, 25 μg rhcg, 125 μg rhcg, 250 μg rhcg, 1000 μg uhcg, 5000 IU (study 1) (study 2) (study 1) (study 1) (study 1) C 1 (μg/l) 8.2 ± ± ± ± ± 11.2 t 1/2 1 (h) 3.9 ± ± ± ± ± 1.3 C z (μg/l) 1.7 ± ± ± ± ± 2.4 t 1/2 (h) 23 ± 4 29 ± ± 3 30 ± 2 31 ± 3 AUC (μg h/l) 100 ± ± ± 118 a 3657 ± ± 170 a CL (l/h) 0.28 ± ± ± ± ± 0.13 MRT (h) 20 ± 3 26 ± ± 3 23 ± 2 25 ± 2 V c (l) 2.9 ± ± ± ± ± 0.7 V ss (l) 5.7 ± ± ± ± ± 2.6 V z (l) 9.1 ± ± ± ± ± 5.7 CL r (l/h) ± 0.01 NA ± ± ± a Comparison by ANOVA gives a P-value of C 1, coefficient associated with the distribution constant λ 1 ; t 1/21, distribution half-life; C z, coefficient associated with the distribution exponent λ z ; t 1/2, elimination half-life; AUC, total area under the curve; CL, total serum clearance; MRT, mean residence time; V c, initial volume of distribution; V ss, steady-state volume of distribution; V z, volume of distribution estimated from terminal phase; Ae, amount excreted in urine; CL r, renal clearance. 109

5 Pharmacokinetics after intravenous administration Increasing intravenous doses of rhcg produced parallel serum profiles (Figure 1). Pharmacokinetic parameters of rhcg derived by modelling after intravenous administration in studies 1 and 2 are summarized in Table 1. The serum pharmacokinetics after intravenous administration of rhcg were well described by a bi-exponential model, except in one subject. The terminal half-life was approximately h but was slightly lower with the lowest dose (approximately 23 h). Total serum clearance was approximately 0.3 l/h, and renal clearance was approximately 0.03 l/h; thus, approximately 10% of a dose of rhcg is excreted unchanged in the urine. There were no significant differences in dose-normalized AUC between the different rhcg doses (P = 0.1), and no significant effect of gender. The mean serum profile after administration of 250 μg rhcg was similar to that seen after administration of 5000 IU uhcg (Figure 1). However, the variability in pharmacokinetics was low, and thus it was possible to detect a significant (P = 0.013) difference between the two drugs: uhcg tended to be distributed and eliminated slightly more slowly than did rhcg (Table 1). In study 2, HCG concentrations were measured both by immunoassay and in-vitro bioassay. The two techniques yielded similar serum profiles and hence similar pharmacokinetic parameters (results not shown). Pharmacokinetics after extravascular administration The pharmacokinetic (modelling-derived) parameters of rhcg after single intramuscular and subcutaneous doses, and multiple subcutaneous doses, are summarized in Table 2. In general, the pharmacokinetics were adequately described by a Bateman function. In a few subjects, however, it proved difficult to estimate the pre-exponential coefficient (C 0 ), and this was reflected by a large variability in this parameter. In male subjects, the absolute bioavailability tended to be higher after intramuscular administration of rhcg than after subcutaneous administration (0.52 ± 0.06 versus 0.38 ± 0.12, respectively), whereas in females the bioavailability was similar with the two routes (0.42 ± 0.13 versus 0.41 ± 0.09, respectively). However, the treatment-by-gender interaction did not reach statistical significance. During multiple subcutaneous administration, the pharmacokinetics of rhcg derived from immunoassay data did not differ significantly from those following a single dose (Figure 2). It can be seen that the bioassay data are fully superimposable on the immunoassay data, and as such associate an exposure to the drug with a corresponding in-vivo activity. When the bioassay data were used, however, the AUC during multiple dosing was approximately 30% lower than that after a single dose (P < 0.05). During repeated dosing, the accumulation ratio was 1.7 (± 0.6) based on immunoassay data, and 1.5 (± 0.4) based on bioassay data. Comparison of rhcg formulations Serum HCG profiles and model-independent pharmacokinetic parameters were comparable after administration of 250 μg rhcg, in freeze-dried and liquid formulations, and 5000 IU uhcg, in a freeze-dried formulation (Figure 3, Table 3). Table 2. Modelling-derived pharmacokinetic parameters (means ± SD) of rhcg after single intramuscular and subcutaneous doses, and multiple subcutaneous doses to healthy males and females (n = 12). Dose and route of administration 125 μg, im 125 μg, sc μg, sc C 0 (μg/l) 13.7 ± ± ± 3.7 t 1/2a (h) 8.4 ± ± ± 2.5 t 1/2 (h) 26.6 ± ± ± 4.2 AUC (h μg/l) ± ± ± MAT (h) 24.1 ± ± 10.4 NA F 0.47 ± ± 0.11 NA im, intramuscular; sc, subcutaneous; C 0, pre-exponential coefficient; MAT, mean absorption time; t 1/2a, half-life associated with absorption rate constant; F, absolute bioavailability; NA, not available. Remaining terms defined in Table Figure 1. Mean (± SD) concentration of HCG after intravenous administration of 25, 250 and 1000 μg (500, 5000 and 20,000 IU) rhcg (upper) and 5000 IU uhcg (lower).

6 Figure 3. Mean (±SD) serum HCG concentrations after administration of 250 μg rhcg, in liquid (A, ) and freezedried (B, ) formulations and 5000 IU uhcg in a freeze-dried formulation (C, ) to healthy male (12) and female (12) subjects. Figure 2. Serum hcg concentrations measured by immunoassay (upper panel) and bioassay (lower panel) after administration of rhcg, 125 μg, as single intravenous (A, ), intramuscular (B, ) and subcutaneous (C, ) doses, and after multiple subcutaneous doses of 125 μg (D, ). Table 3. Model-independent pharmacokinetic parameters (means ± SD) of rhcg after administration of 250 μg of rhcg, in freeze-dried and liquid formulations, and 5000 IU of uhcg, in a freeze-dried formulation. rhcg, liquid rhcg, uhcg, formulation freeze-dried freeze-dried formulation formulation C max (μg/l) 6.63 ± ± ± 1.89 t max (h) a 12 (9 36) 12 (9 24) 12 (6 36) AUC (μg h/l) 415 ± ± ± 82 t 1/2 (h) 29 ± 5 30 ± 5 35 ± 6 AUMC (μg h 2 /l) 21,488 ± ,959 ± ,913 ± 4662 MRT (h) 52 ± 8 53 ± 9 53 ± 6 a Medians and ranges. Statistical analysis of the mean ratios of C max and AUC for different treatment comparisons showed no significant difference between the treatments (Table 4), confirming that the three treatments were bioequivalent. There was, however, a statistically significant (P < 0.05) treatment-by-gender interaction for the liquid formulation only. C max tended to be higher with this formulation in males than in females (7.68 versus 5.58 μg/l, respectively), as did AUC (462 versus 368 μg h/l, respectively). There were no significant differences between genders in these parameters with freeze-dried rhcg and uhcg. Furthermore, a statistical comparison of t max Table 4. Mean ratios of C max and AUC, and associated 90% confidence intervals, for liquid and freeze-dried rhcg and freeze-dried uhcg after subcutaneous injection to healthy male (12) and female (12) subjects. Comparison C max AUC Liquid rhcg versus 1.01 ( ) 0.99 ( ) freeze-dried rhcg Liquid rhcg versus 0.95 ( ) 0.98 ( ) freeze-dried uhcg Freeze-dried rhcg 0.94 ( ) 0.99 ( ) versus freeze-dried uhcg AUC, area under the curve. values failed to reveal any significant differences among the three treatments. Effect of rhcg formulation on pharmacodynamic markers The effects of rhcg on testosterone, 17β-oestradiol and inhibin are summarized in Table 5. After a delay of approximately 12 h, concentrations of these hormones increased markedly and remained elevated as HCG concentrations declined (Figure 4). None of these hormones had returned to pretreatment levels within 8 days of dosing. There was, however, little accumulation of testosterone during multiple dosing (Figure 4); the mean maximum testosterone concentration during multiple dosing was 43 nmol/l, compared with 34 nmol/l after a single subcutaneous dose. Pre-dose concentrations of inhibin tended to decrease during multiple dosing, from approximately 4 U/ml before the second dose to 3 U/ml before the final dose. 17β-oestradiol showed the most sustained response, and modest accumulation was observed. Overall, the pharmacodynamic responses to rhcg were not 111

7 a. b. Figure 4. Serum concentrations of (a) testosterone, (b) 17βoestradiol and (c) inhibin in male subjects after single intravenous, intramuscular and subcutaneous injections of 125 μg (2500 IU) rhcg and five subcutaneous injections of the same dose. c. Figure 6. Serum concentrations of androstenedione (upper panel) and testosterone (lower panel) in 12 healthy female subjects following administration of 250 μg rhcg in liquid ( ) and freeze-dried ( ) formulations and 5000 IU uhcg in a freeze-dried formulation ( ). Serum androgen concentrations in control subjects were not measured in this phase I study primarily aimed at assessing bioequivalence between urinary and recombinant HCG. Therefore these data are not available and only a comparison with data from the literature is possible. 112

8 Figure 5. Serum concentrations of androstenedione in six healthy female subjects during repeated subcutaneous injections of 125 μg rhcg. Injections are represented by arrows. Table 5. Pharmacodynamic effects of rhcg on serum testosterone, 17β-oestradiol and inhibin in pituitary-desensitized healthy male subjects (n = 6). Dose and route of administration 125 μg, iv 125 μg, im 125 μg, sc μg, sc Testosterone Maximum concentration (nmol/l) 25 ± 8 33 ± 6 34 ± 6 43 ± 7 Time to maximum effect (h) a 72 (72 96) 84 (72 120) 84 (48 96) 240 (96 240) AUC (nmol h/l) 2257 ± ± ± ,648 ± 1671 AUC repeated /AUC single 3.0 ± 0.5 (P = ) 17β-oestradiol Maximum concentration (pmol/l) 127 ± ± ± ± 38 Time to maximum concentration (h) a 72 (48 96) 96 (72 120) 96 (24 96) 264 ( ) AUC (pmol h/l) 11,212 ± ,940 ± ,357 ± ,038 ± 15,723 AUC repeated /AUC single 3.9 ± 1.0 (P = 0.04) Inhibin Maximum concentration (U/ml) 3.6 ± ± ± ± 0.6 Time to maximum concentration (h) a 84 (12 144) 72 (48 96) 72 (48 96) 96 (48 192) AUC (U h/ml) 168 ± ± ± ± 133 AUC repeated /AUC single 2.7 ± 1.2 (P = 0.005) iv, intravenous; im, intramuscular; sc, subcutaneous. Results are presented as means and SD except where indicated. a Medians and ranges. AUC, area under the concentration time curve. directly proportional to the total exposure to the hormone, because the response elicited by multiple dosing was less than five times that elicited by a single dose (Table 1). The pharmacodynamic response to rhcg in females, assessed by measurement of serum androstenedione during repeated administration, was less pronounced; serum concentrations of androstenedione are shown in Figure 5. Figure 6 shows serum concentrations of testosterone and androstenedione in females on oestroprogestative pill after administration of 250 μg rhcg, in liquid and freeze-dried formulations, and 5000 IU uhcg, in a freeze-dried formulation. The three treatments produced consistent responses in both hormones. Androstenedione and testosterone concentrations started to decrease after HCG administration but were higher than baseline concentrations 24 h after dosing. This pattern persisted for 3 4 days, after which concentrations decreased towards baseline. Safety All treatments were well tolerated and there were no serious adverse events attributable to study drugs. Adverse events were mainly headaches for nearly 60% of the subjects. Through biochemistry and haematology monitoring, no trend could be detected on hepatic and renal function, electrolytes balance or coagulation. Drugs were not immunogenic, because no antibodies to study drugs were detected in this clinical setting. Chronic administration was done in infertile males, confirming the non-immunogenicity (Bouloux et al., 2002). Discussion After intravenous administration of rhcg, the pharmacokinetics were well described by a bi-exponential model. The terminal half-life was approximately 30 h with doses equal to or greater than 125 μg, although the 25-μg dose was associated with a shorter half-life of approximately 23 h. 113

9 114 This probably reflects the difficulty in measuring serum concentrations reliably with low doses, as these decline to near the limit of quantification at earlier time points than with higher doses, and hence data at later times are limited. When dose-normalized (data not shown), there was no statistical difference in drug exposure. The half-life and clearance of rhcg after intravenous administration were independent of dose. Therefore, the pharmacokinetics of rhcg were linear. After intravenous administration of uhcg, the pharmacokinetics of HCG were adequately described by the same bi-exponential model, and the serum concentration profiles of the two drugs were similar. However, the AUC of uhcg at a nominal dose of 5000 IU was approximately 29% lower than that of rhcg. This is probably due to slight differences in the pharmacokinetics of the two drugs, because uhcg tended to be distributed and eliminated more slowly than did rhcg. However, the nominal bioactivity of both products was determined by the Van Hell in-vivo bioassay (Van Hell et al., 1964), where the drug is already injected subcutaneously. Dosing on this basis corrects for some pharmacokinetic differences, resulting in concentration profiles that are similar despite possible differences in immunological content. Indeed, the apparent difference in AUC was not seen when the drugs were administered subcutaneously (Tables 3 and 4). After intramuscular or subcutaneous administration, the pharmacokinetics of rhcg could be described by a first-order absorption, onecompartment model, although difficulties were experienced in calculating the pre-exponential coefficient (C 0 ) in a few subjects. The absolute bioavailability with both extravascular routes was approximately 50%. Bioavailability in males appeared to be higher after intramuscular administration, compared with that after subcutaneous administration, although the treatment-by-gender interaction did not reach statistical significance. By contrast, in female subjects the bioavailability was similar with both routes and similar to the pharmacokinetics in males after subcutaneous administration. This apparent difference may be related to the larger amount of subcutaneous adipose tissue in females. In general, subcutaneous administration in females results in the drug being deposited in adipose tissue; similarly, intramuscular injections are usually given into the upper outer quadrant of the buttock, which has a higher adipose tissue content in females than in males. Thus, it seems likely that, in females, both routes of administration would have resulted in drug delivery to adipose tissue, resulting in similar bioavailabilities (Dobbs et al., 1994). During repeated subcutaneous administration, the accumulation ratio of HCG was approximately 1.7-fold, based on immunoassay data. Comparison of the pharmacokinetics derived from the immunoassay data following the first and last doses did not reveal any significant differences. Serum HCG concentration profiles and pharmacokinetic parameters were similar after administration of 250 μg rhcg in liquid and freeze-dried formulations, and 5000 IU uhcg in a freeze-dried formulation. Furthermore, the 90% confidence intervals for the ratios of C max and AUC with different treatments were within the bioequivalence range defined by the FDA as indicating equivalence ( ). Although a statistically significant treatment-by-gender interaction was seen with the liquid formulation, no such gender differences were seen with the other formulations. This demonstrates that from a clinical standpoint both formulations are interchangeable. Study 2 compared the pharmacokinetic parameters of rhcg when concentrations were measured by immunoassay and an in-vitro bioassay (data not shown). The two analytical techniques yielded essentially similar pharmacokinetic data, which indicates that pharmacokinetic analyses can be reliably based on immunoassay measurements. This is a valuable finding, because immunoassays are more reliable, reproducible and convenient to perform than are in-vitro bioassays. It is acknowledged that immunoassay may be less reliable if the target is the measurement of the activity of a molecule. In the case of pharmacokinetics, however, the aim is to assess reliably and reproducibly serum concentrations of a drug, and in this context the use of sensitive and specific immunoassay is more appropriate. Administration of rhcg produces consistent pharmacodynamic responses, irrespective of the route of administration or the formulation used. In study 2, all four regimens (single intravenous, intramuscular and subcutaneous doses, and repeated subcutaneous administration) produced responses that were of the same order of magnitude, the response after intravenous administration always being the lowest. The increase in serum testosterone concentration was comparable with that observed after administration of uhcg (Padrón et al., 1980; Ulloa-Aguirre et al., 1985); the observation that relatively little further increase occurred during repeated administration suggests that testosterone production rate is almost maximal after a single dose of rhcg. Desensitization may be an alternative explanation for the relatively modest increase in testosterone. However, a longer dosing regimen would be required to explore this possibility. Desensitization is also likely to result in a decrease in the response, which was not observed in this study. The lack of a biphasic response to HCG in these down-regulated subjects, and the latent period of approximately 12 h, resembles the pattern seen in men with hypogonadotrophic hypogonadism treated with uhcg (Padrón et al., 1980; Okuyama et al., 1981; Ulloa-Aguirre et al., 1985; Bouloux et al., 2002). Such a response would be consistent with the absence of a readily releasable steroid pool in the Leydig cells of such individuals, because previous exposure to LH or HCG is essential for the first peak of the biphasic response to HCG (Ulloa-Aguirre et al., 1985). This testosterone response has been modelled elsewhere (Gries et al., 1999). The inhibin response to rhcg was similar to the testosterone response, except that some degree of tolerance tended to develop during repeated administration. Inter-individual variability was more marked with inhibin than with testosterone, and the latent period less apparent. 17β-oestradiol showed the highest degree of accumulation during repeated administration of rhcg. The mean increase, however, was less than three times the limit of quantification of the assay, and the concentrations attained during repeated dosing were not markedly higher than the normal range in males ( pmol/l). Trough oestradiol concentrations tended to reach a plateau during repeated dosing, but it is not

10 possible to determine whether this reflects a maximal effect on oestradiol production. In adult men, approximately 70% of plasma oestradiol is formed by aromatization of testosterone and androstenedione (Ganong, 1991); hence the observed oestradiol concentrations may simply reflect a maximal effect of rhcg on testosterone production. Interpretation of the serum androstenedione concentrations observed in female subjects is complicated, because this steroid was measured only during repeated administration of rhcg, and the subjects were not down-regulated. In women taking oral contraceptives, androstenedione originates entirely from the adrenal glands and is subject to circadian variation (Yen and Jaffe, 1986). This circadian variation could be seen in these studies during the first 2 days of each rhcg treatment (Figure 6), although morning concentrations of androstenedione tended to increase during the first 3 days of repeated treatment. The profile over time was very similar for testosterone. However, the relative amplitude of the maximum rise was possibly more pronounced (proportionally) than for androstenedione, and although the peak effect also occurred at around the third day, its increase was somewhat delayed. This is possibly due to the different origin of circulating testosterone, as compared with that of androstenedione. Testosterone is produced, in a normal woman, roughly 50% by peripheral conversion of prehormones (principally androstenedione), 25% by the adrenal glands (and thus under adrenocorticotrophic hormone (ACTH) circadian control), and 25% by the ovary (Yen and Jaffe, 1986). Thus, in the present study in women on oral contraception, the rise observed after HCG administration was driven by both an ovarian stimulation and an increased substrate (androstenedione) amount available for conversion. Study 3 showed that the pharmacodynamic effects of rhcg were compatible with a single stimulus of the ovary. Effects were similar with freeze-dried and liquid formulations, and the effects of both formulations were comparable with those of uhcg. This is consistent with the results of pharmacokinetic analyses, which showed bioequivalence between the three preparations. In conclusion, these studies have characterized the pharmacokinetics of rhcg after intravenous, intramuscular and subcutaneous administration. The pharmacokinetics of rhcg are similar to those of uhcg, and rhcg has pharmacodynamic effects consistent with HCG physiology. Regarding testosterone response in males, owing to the apparent tachyphyllaxis and likely tolerance development, altering the frequency of administration could affect the time course of testosterone production and would require further clinical testing. The results of all studies show rhcg to be well tolerated, both systemically and locally, and there was no evidence that it was immunogenic. Overall, these findings support the use of rhcg in the same indications as uhcg. The secured source and high purity of rhcg (Ovitrelle ) might offer important advantages in these indications. Acknowledgements The authors wish to acknowledge the invaluable assistance of Drs K Dahl and M Walker, S Toon and their laboratory personnel, and Dr J-Y le Cotonnec in conducting the studies reported here. References Bahl OP 1973 Chemistry of human chorionic gonadotrophin. In Li CH (ed.) Hormonal proteins and peptides. Academic Press, New York, USA, pp Birken S, Krichevsky A, O Connor J et al Chemistry and immunochemistry of hcg, its subunits and its fragments. In Chin WW, Boime I (eds) Glycoprotein Hormones. Serono Symposia, Norwell, MA, USA, pp Bouloux P-M, Warne DW, Loumaye E 2002 Efficacy and safety of recombinant human follicle stimulating hormone in men with isolated hypogonadotropic hypogonadism. Fertility and Sterility 77 (2) (in press). Dahl KD, Sarkissian A 1993 Validation of an improved in vitro bioassay to measure LH in diverse species. Journal of Andrology 14, Dickerman Z, Bauman B, Sandovsky U et al Human chorionic gonadotrophin (hcg) treatment in cryptorchidism. Andrologia 15, Dobbs KE, Dumesic DA, Dumesic JA et al Differences in serum follicle-stimulating hormone uptake after intramuscular and subcutaneous human menopausal gonadotropin injection. Fertility and Sterility 62, Everitt BS 1998 The Cambridge Dictionary of Statistics. Cambridge University Press, Cambridge, UK. FDA, Code of Federal Regulations, Title 320, Chapter 1, Part 320: Bioavailability and bioequivalence requirements. Fonjallaz P, Loumaye E 2001 Recombinant hcg (OVIDREL) and recombinant interferon-beta1a (REBIF) (No. 13 in a series of articles to promote a better understanding of the use of genetic engineering). Journal of Biotechnology 87, Ganong WF 1991 Review of Medical Physiology. 15th edn Prentice Hall International, London, p Gries JM, Munafo A, Porchet HC et al Down-regulation models and modeling of testosterone production induced by recombinant human choriogonadotropin. Journal of Pharmacology and Experimental Therapeutics 289, Halme J, Hammond MG, Talbert LM et al Positive correlation between body weight, length of human menopausal gonadotropin stimulation, and oocyte fertilization rate. Fertility and Sterility 45, Hills M, Armitage P 1979 The two-period cross-over clinical trial. British Journal of Clinical Pharmacology 8, Hugues JN 2001 Recombinant human follicle-stimulating hormone: a scientific step to clinical improvement. Reproductive BioMedicine Online 2, Hutchinson-Williams KA, DeCherney AH, Lavy G et al Luteal rescue in in vitro fertilization embryo transfer. Fertility and Sterility 53, Okuyama A, Namiki M, Koide T et al A simple hcg stimulation test for normal and hypogonadal males. Archives of Andrology 6, Padrón RS, Wischusen J, Hudson B et al Prolonged biphasic response of plasma testosterone to single intramuscular injections of human chorionic gonadotropin. Journal of Clinical Endocrinology and Metabolism 50, Pierce JG, Parsons TF 1981 Glycoprotein hormones: structure and function. Annual Review of Biochemistry 50, Ulloa-Aguirre A, Mendez JP, Diaz-Sánchez V et al Selfpriming effect of luteinizing hormone-human chorionic gonadotropin (hcg) upon the biphasic testicular response to exogenous hcg. I. Serum testosterone profile. Journal of Clinical Endocrinology and Metabolism 61, Van Hell H, Matthijsen R, Overbeak GA 1964 Effects of human menopausal gonadotropin preparation in different bioassay methods. Acta Endocrinologica 47, Yamaoka K, Nakagawa T, Uno T 1978 Application of Akaike s information criterion (AIC) in the evaluation of linear pharmacokinetic equations. Journal of Pharmacokinetics and Biopharmaceutics 6, Yen SSC, Jaffe RB 1986 Reproductive Endocrinology Physiology, Pathophysiology and Clinical Management. 2nd edn WB Saunders, Philadelphia, USA, pp