Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Estrogen replacement therapy (ERT)¹ has been shown to be effective in both preventing postmenopausal osteoporosis and in reducing the risk of cardiovascular disease (Witt and Lousberg, 1997). However, without the concomitant administration of progesterone supplements, ERT has been associated with an increased stimulation of the endometrium, causing hyperplasia and risk of cancer. Thus, there is an interest in developing oral drugs that possess the beneficial effects of ERT, such as osteoporosis prevention, but do not have any detrimental effect on the uterus.

One such candidate, levormeloxifene (()-3,4-trans-7-methoxy-2,2-dimethyl-3-phenyl-4-{4-[2-(pyrrolidin-1-yl)ethoxy]phenyl}chromane, hydrogenfumarate), is a selective estrogen receptor modulator, with low intrinsic estrogenicity that has been shown to prevent osteopenia in the ovariectomized rat model of human osteoporosis (Bain et al., 1997), and to prevent aortic cholesterol accumulation in the ovariectomized rabbit model (Holm et al., 1997). In addition, levormeloxifene has an apparently unique estrogenic effect on the uterus of ovariectomized animals whereby uterine weight is increased with no evidence of epithelial proliferation or glandular stimulation (Bain et al., 1997; Korsgaard et al., 1997).

Levormeloxifene was selected as a development candidate for the prevention and treatment of postmenopausal osteoporosis and it has been postulated that it could provide an alternative to current ERTs because no epithelial or glandular proliferation in the uterus or associated tissue has been observed in animal species or postmenopausal human volunteers treated with levormeloxifene. The drug, whose structure is shown in Fig. 1, is the l-enantiomer of ormeloxifene, and the following preclinical studies were performed to characterize the disposition and excretion of this new selective estrogen receptor modulator, because there were very little preclinical metabolic data available for the new antiestrogens that are currently in clinical trials (Lindstrom et al., 1984; Tanaka et al., 1994; O'Donnell et al., 1998). However, the development of this compound has recently been stopped due to a number of adverse events being reported during phase III clinical trials, but new indications are currently being pursued as preclinical testing is near completion. It is anticipated that data generated within drug metabolism may contribute to the overall evaluation of new indications.

View larger version (11K): [in this window] [in a new window]   Fig. 1.   The structure of levormeloxifene. *, indicates position of radiolabel.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals. ¹⁴C-Radiolabeled levormeloxifene (3,4-trans-7-methoxy-2,2-dimethyl-3-phenyl-4{4-[2-pyrrolidin-1-yl)ethoxy]phenyl}chromane hydrogen-fumarate) (Fig. 1) was synthesized at Amersham (Amersham, UK) and purified in the Department of Isotope Chemistry, Novo Nordisk A/S. The radiochemical purity was >99%, as determined by HPLC analysis, and the specific activity was 53 mCi/mmol. Nonradiolabeled levormeloxifene and chromatographic reference compounds were synthesised by Dr. S. Treppendal (Department of Chemistry, Novo Nordisk A/S, Maaloev, Denmark). Reference compounds included (+, )-3, 4-trans-2, 2-dimethyl-3-phenyl-4-[4-{2-(pyrrolidin-1-yl)ethoxy}phenyl]-7-hydroxychromane hydrochloride; (+, )-3, 4-trans 3- phenyl-4-[4-{2-(pyrrolidin-1-yl)ethoxy}phenyl]-7-hydroxychromane hydrochloride; (+, )-3, 4-trans-2, 2-dimethyl-3-phenyl-4-(4-hydroxyphenyl)-7-hydroxychromane; and 3, 4-trans-2, 2-dimethyl-3-phenyl-4-(4-hydroxyphenyl)-7-methoxychromane, with respective codes NNC 46-0002, NNC 46-0003, NNC 46-0004, and NNC 46-0005.

Animals and Dosing. Sprague-Dawley rats weighing 200 to 230 g were obtained from Charles River UK Ltd. (Margate, UK) or from Moellegards Breeding Laboratories (Lille Skensved, Denmark). Rats were housed in groups of six or fewer in stainless steel cages or singly in glass metabolism cages (Jencons Ltd., Leighton Buzzard, UK) in air conditioned rooms maintained at 19-23°C, 40 to 60% relative humidity, and a 12-h light/dark cycle. Rats were acclimatized under these conditions for at least 2 days before dosing.

Blood Sampling and Drug Analysis for Pharmacokinetic Experiments. Blood samples were collected before, 4, 6, 24, and 48 h after the sixth administration (day 13), and before, 4, 6, 24, 48, 54, 72, and 96 h after the final administration (day 34). In general, three blood samples, one on days 13 to 15 and two on days 34 to 38 (1000 µl each) were collected per animal by removal of blood from the opthalmic venous plexus into heparinized Eppendorf tubes. Blood samples were centrifuged and the supernatant (plasma) aspirated and stored frozen (18°C). Plasma (400 µl) was applied to mixed-mode columns (SPEC C8/SCX, 30 mg, 3 ml (Ansys, Irvine, CA) conditioned with 0.5 ml of methanol followed by 0.5 ml of phosphate buffer (0.1 M, pH 2.0). The cartridges were then rinsed with 0.5 ml of 1 M acetic acid, 0.5 ml of acetonitrile, and 0.5 ml of buffer in that order. The analytes were finally eluted from the cartridges by 1 ml of methanol-triethylamine (98:2, v/v). The eluate was evaporated to dryness in a TurboVap LV evaporator and redissolved in 100 µl of acetonitrile/water (40:60), and 75 µl was applied to the HPLC system.

Data Handling for Pharmacokinetic Experiments. Mean concentration/time data were calculated and analyzed by noncompartmental methods using the software Topfit (version 2.0; Heinzel et al., 1993). The apparent maximal concentration (C_max) and the corresponding time (t_max) were determined visually from the concentration-time profile. Calculation of terminal half-life (t_1/2) was based on data obtained during days 34 to 38. The terminal half-life was calculated by means of log-linear regression using at least six data points. The total area under the plasma concentration versus time curve (AUC) was determined by the linear trapezoidal rule from time zero to last sampling point equal to or above the lower limit of quantitation, AUC_t, added as the residual area as estimated by log-linear extrapolation to infinity.

Radioanalysis. Radioactivity in liquid samples (urine, plasma, bile, metabolism cage washes, and expired air trap solutions) was quantified by mixing aliquots with scintillation system MI-31 or Pico Aqua (Packard Instruments Ltd., Pangbourne, UK and Downers Grove, IL, respectively) and conventional liquid scintillation counting. Rat whole body digestion was carried out at 50°C in a solution containing NaOH, water, methanol, and Triton X-405 and samples (1 g) were mixed with scintillation system MI-31.

WBA. Sprague-Dawley rats received single oral doses of [¹⁴C]levormeloxifene (1.4 mg/kg b.wt.) and WBA was performed at 2, 4, 24, 48, and 72 h after dose administration, essentially as described by Ullberg and Larrson (1981). Sections were prepared using a 9400 Cryostat Microtome (Bright Instruments Co., Huntingdon, UK). Sagittal sections (30 µm) were cut at six levels through the carcass, between the levels of the kidneys (males) or ovaries (females) and the spinal cord. Sections were mounted on Cellux tape (Aston Clinton, St. Albans, UK) and freeze-dried in a Lyolab B freeze-drier (Life Sciences Laboratories Ltd., Luton, UK) before placing them in contact with Kodak DEF5 film (Kodak Ltd., Hemel Hempstead, UK) and max X-ray film (Amersham International, Amersham, UK). The film was exposed for 41 days at 20°C before its development. Autoradiographs were evaluated by visual inspection.

Analysis of Metabolites. Tentative identification of metabolites in fecal extracts, urine, plasma, bile, and selected tissue extracts for the low and high doses was achieved by HPLC cochromatography with authentic standards. The elution times for levormeloxifene, NNC 46-0002, NNC 46-0003, NNC 46-0004, and NNC 46-0005 were typically 36, 22, 29, 32, and 55 min, respectively, although it was apparent on some of the HPLC runs that there was a shift in retention time, possibly due to the matrix in the injected samples. Additional identification was achieved by mass spectroscopy. Certain bile, urine, and plasma samples were deconjugated by mixing in equal proportions with -glucuronidase (Type H1, 2000 U/ml) provided by Sigma, and incubating overnight at 37°C in acetate buffer (pH 5). Feces and tissue extracts were made by homogenization of samples in diethylether, centrifugation to obtain supernatants, evaporation of solvent, and resuspension in the HPLC mobile phase. Plasma samples (native or enzyme-treated) were applied to Isolute-Confirm HCX mixed-mode solid-phase extraction columns (3 ml/300 mg size; Jones Chromatography, Hengoed, UK), which were preconditioned with methanol and phosphate buffer (0.1 M; pH 2.0). After sample loading, the column was rinsed with acetic acid (1 M) and phosphate buffer (0.1 M; pH 2.0) and the analytes were eluted with methanol/triethylamine (98:2, v/v).

HPLC conditions. Samples were analyzed using a LiChrospher 100 C₁₈ column (particle size 5 µm, 250 × 4 mm id; Merck, Darmstadt, Germany) with a µBondaPak C₁₈ Guard-Pak precolumn (Millipore, Waters, MA). A Thermo Separation Products HPLC system was used (Thermo Separation Products, Stone, UK) consisting of a pump, interface, and a UV 2000 variable wavelength UV detector and a Ramona-5 or -RAM on-line radioactivity detector (supplied by LabLogic Systems Ltd., Sheffield, UK) fitted with a solid scintillator flow cell. For the biliary excretion experiments a Hitachi-Merck HPLC system (supplied by Kebo Struers, Copenhagen, Denmark) consisting of a pump, interface, UV detector, and on-line radioactivity detector (Canberra Flo-One, Canberra, Copenhagen, Denmark) was used. The absorption (for detection of reference standards) was measured at 279 nm. A mobile phase gradient was used: from 0 to 55 min, a 3:7 ratio of 0.1 M ammonium formate, pH 3.3/acetonitrile; from 55.1 to 57.0 min, a 1:1 ratio was used; and from 57.1 to 59.0 min, a 7:3 ratio was used. A flow rate of 1.0 ml/min was used except from 27.6 to 55.1 min, when the flow rate was increased to 1.3 ml/min. The above conditions were modified slightly during the isolation of metabolites for mass spectrometry analysis. Individual radioactive peaks were collected manually, solvent was removed under nitrogen, and residues were redissolved in acetonitrile/water (1:1, v/v) before mass spectrometry. The radiochemical purity of each metabolite was assessed by both reinjection on HPLC and thin-layer chromatography analysis (not shown).

Mass spectrometry. Selected metabolites were analyzed by atmospheric pressure chemical ionization (APCI) mass spectrometry in both positive and negative ionization modes, where appropriate, using a TSQ7000 (Finnigan MAT, San Jose, CA) or an API 300 triple quadrapole liquid chromatography-tandem mass spectrometry (LC-MS-MS) mass spectrometer (Perkin-Elmer Sciex Instruments, Beaconsfield, UK) equipped with an Ionspray interface (Perkin-Elmer Sciex Instruments, Thornhill, Canada). Ionization conditions were optimized by varying the octapole offset (range 2.9 to 3.3 V in the positive ion mode, + 3.0 V in the negative ion mode), capillary (+20 to +91.8 in the positive ion mode, 20 to 96.8 V in the negative ion mode), and tube lens voltages (+72 to +181.5 V in positive ion mode, 72 to -187.1 V in negative ion mode). A portion of the sample (ca. 50 µl) was injected into the APCI interface at 1 ml/min, in acetonitrile/water (1:1, v/v). Mass spectra of the compounds of interest were recorded over an appropriate mass range for 2 min at a scan rate of 1 s/scan. An APCI mass spectrum of the compound of interest was obtained by averaging several scans across the region of the mass chromatogram where a response was observed, with appropriate background subtraction. This mass spectrum was examined to identify a candidate molecular ion, [M+H]⁺ or [M-H], for the compound of interest, and any other structurally significant fragment ions resulting from in-source collisionally induced dissociation. For metabolite identification experiments from bile, the mass spectrometer scanned in the range m/z 100 to 700 with a dwell time of 0.5 ms and a step size of 0.1 amu. The electrospray and orifice voltages were set to 5000 and 30 V, respectively. In the tandem mass spectrometry mode (product ion scan) the mass spectrometer scanned in the range m/z 50 to 500 (or higher, dependent on the mass of the precursor ion) with a dwell time of 0.5 ms and a step size of 0.1 amu. The electrospray and orifice voltages were set to 5000 and 30 V, respectively, and the fragmentation energy was +35 V.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Quantitative Tissue Distribution. Radiolabeled levormeloxifene was distributed throughout body tissues after oral administration at dose levels of 0.7 and 50 mg/kg b.wt., and the rate of absorption and the general distribution of radioactivity was similar at both dose levels. The concentrations of radioactivity in the tissues increased in a dose-proportional manner. Not surprisingly, the greatest amounts of radioactivity were found in the gastrointestinal tract contents and in those organs responsible for absorption and elimination soon after administration (Table 1).

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Mean concentrations (±S.D.) of radioactivity (tissue/plasma ratio) in selected tissues after oral administration of [14C]levormeloxifene (0.7 mg/kg or 50 mg/kg b.wt.) to male (m) and female (f) rats. Data shown for 4-h time point. SI, small intestine; LI, large intestine.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Proportions of dose present in combined selected tissues of male (m) and female (f) rats after oral administration of [14C]levormeloxifene (0.7 or 50 mg/kg) to rats. Selected tissues (including GI tract contents) as described in Materials and Methods. Data shown as mean ± S.D. , f (0.7 mg/kg); , m (0.7 mg/kg); , f (50 mg/kg); , m (50 mg/kg).

Qualitative Distribution. The general distribution of radioactivity from WBA agreed with the quantitative distribution data. The only tissues found to contain notable concentrations of radioactivity that were not sampled in the quantitative analysis were the pineal body, brown fat, and the preputial, meibomain, and Harderian glands.

View larger version (97K): [in this window] [in a new window]   Fig. 4.   WBA of a female rat 4 h after oral administration of [14C]-levormeloxifene. ad, adrenal; b, brain; bf, brown fat; bl, blood; bm, bone marrow; cac, caecum contents; elg, exorbital lacrimal gland; fa, fat; Hd, Harderian gland; ilg, intraorbital lacrimal gland; k, kidney; l, liver; lic, large intestine contents; lu, lung; mb, meibomian gland; my, myocardium; mu, muscle; ov, ovary; p, pancreas; pb, pineal body; pg, preputial gland; pit, pituitary; s.c., spinal cord; sg, salivary gland; sic, small intestine contents; sk, skin; sp, spleen; stc, stomach contents; th, thymus; ty, thyroid; ut, uterus.

View larger version (97K): [in this window] [in a new window]   Fig. 5.   Enlargement of female rat brain 2 h after oral administration of [14C]levormeloxifene. bm, bone marrow; cb, cerebellum; ce, cerebrum; ch, choroid plexus; cq, copora quadrigemina; ol, olfactory lobe; pit, pituitary.

¹⁴C-Excretion. After oral administration of radiolabeled levo-rmeloxifene (0.7- or 50-mg/kg dose), radioactivity was excreted predominantly in the feces (Table 2). Mean fecal excretion accounted for 68.5 to 64.1% of the administered dose during the first 48 h, and 96.8 to 96.1% of the dose at the end of 168 h for the low (0.7 mg/kg)- and high (50 mg/kg)-dose groups, respectively, with male and female data combined at each dose level. Approximately 1.0% of the administered dose was excreted in the urine in the low- and high-dose groups. In a comparison of male and female rats, at both dose levels radioactivity was excreted more rapidly by males than by females (Table 2). Interestingly, a lower proportion of the dose was excreted in the urine of males than females, 0.6 to 0.8% compared with 1.3 to 1.3% (low and high doses, respectively) over 168 h, respectively, and the amount retained by the carcasses was also lower in males at both dose levels.

Pharmacokinetics. After drug administration, C_max was generally observed 6 h after dosing. Pharmacokinetic parameter estimates are presented in Table 4. The half-life of elimination was long (24 h) and a doubling in dose resulted in an approximate doubling in exposure.

Metabolite Profiling and Identification. A number of different radioactive metabolites were tentatively identified by HPLC analysis, with subsequent structural confirmation by LC-MS-MS.

Bile. Proportions of radioactive components were generally similar at both the low- and high-dose groups in male and female animals (0-24 h post dose). One major metabolite (M1), in addition to two minor metabolites (M2 and M4) were isolated from bile. Typical HPLC chromatograms of bile from male (a) and female (b) rats, 0 to 6 h after oral administration of 0.7 mg/kg b.wt. [¹⁴C]levormeloxifene, are shown in Fig. 6. Mass fragmentation patterns for four metabolites are shown in Table 5 and the proportions of identified metabolites in the bile samples analyzed are shown in Table 6. The remaining radioactivity was excreted into feces during the experimental time period. Bile isolated from both male and female rats (both dose levels) consisted mainly of 7-desmethyllevormeloxifene glucuronide (M1) at all time points, with the additional quantitatively minor metabolites being identified as the glucuronides of hydroxylevormeloxifene (M2) and levormeloxifene (M4). Glucuronidase treatment of bile resulted in the identification of the aglucans for the isolated metabolites (M3 shown for reference purposes).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   HPLC of bile from male (A) and female (B) rats, 0 to 6 h after oral administration of 0.7 mg/kg b.wt. [14C]levormeloxifene. Radioactivity in each collected fraction was determined by off-line liquid scintillation counting. Letter indicates metabolite assignment. Minor metabolites not shown in figure.

Feces. Proportions of radioactive components and metabolite profiles were generally similar at both the low- and high-dose groups in male and female animals. At least 11 metabolites were detected in feces, based on fraction collection data and on-line radioactivity monitoring; however some of these metabolites remained unidentified during the study (Fig. 7, representative on-line HPLC chromatograms of feces from male (a) and female (b) rats, 0 to 24 h after oral administration of 0.7 mg/kg b.wt. [¹⁴C]levormeloxifene). There was some shift in retention time during HPLC analysis but this was attributed to a matrix effect and was compensated for by cochromatography with the authentic reference standards. Subsequent HPLC runs (minus reference standards) were then used for isolation of metabolites for LC-MS-MS analysis. Feces consisted mainly of 7-desmethyllevormeloxifene (M6, norlevormeloxifene) at all time points, but unchanged drug was the second most prevalent component (Fig. 7 and Table 6). Mass fragmentation patterns for six of the isolated metabolites are shown in Table 5. The proportions of identified metabolites (M5-M9, and levormeloxifene) in feces are shown in Table 6.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   HPLC, with on-line radioactivity detection, of feces from male (A) and female (B) rats, 0 to 24 h after oral administration of 0.7 mg/kg b.wt. [14C]levormeloxifene. Letters indicate metabolite assignment. Minor metabolites not shown in figure.

Urine. After both 0.7 and 50 mg/kg b.wt. doses of ¹⁴C-levormeloxifene, radioactivity excreted in urine, accounting for approximately 1% of the administered dose, was largely associated with chromatographically polar metabolites. There were marginal sex differences in the polar metabolites, although there were no differences between dosing groups. For example, polar components made up 91% of the total urinary radioactivity excreted from 0 to 24 h (Fig. 8A) in male rats after dosing at 0.7 mg/kg b.wt. In urine excreted by female rats 24- to 48-h postdose (0.7 mg/kg b.wt. dose), a slightly lower proportion of urinary radioactivity was associated with these polar components (about 73-75%) (Fig. 8B).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   HPLC, with on-line radioactivity detection, of urine from male (A) and female (B) rats, 0 to 24 and 24 to 48 h, respectively, after oral administration of 0.7 mg/kg b.wt. [14C]levormeloxifene. Letters indicate metabolite assignment.

Plasma. Unchanged drug was the major radioactive component present in the systemic circulation, at both dose levels, 4 and 24 h after dose administration. At 4 h, unchanged drug in the low- and high-dose groups, respectively, accounted for 71.9 to 37.7 and 61.5 to 49.4% of plasma extract radioactivity in male and female rats, respectively. Figure 9 shows HPLC profiles of plasma from male (a) and female (b) rats 4 h after oral administration of 0.7 mg/kg b.wt. [¹⁴C]levormeloxifene. There was again some shift in retention time during analysis of the different samples; however, this was compensated for by inclusion of reference standards. Other metabolites present in the systemic circulation included monohydroxylevormeloxifene (M5), and the minor metabolites (not shown in Fig. 9), 7-desmethyllevormeloxifene (M6), desmethylnorlevormeloxifene (M9), and 7-desmethyllevor-meloxifene glucuronide (M1, which increased over time). Additionally, an unknown metabolite was also evident at both levels at all time points. After 24 h the proportion of radioactivity associated with unchanged drug decreased, and proportions of metabolites correspondingly increased. Interestingly, greater concentrations of the parent drug remained in the circulation of female rats at increasing time reflecting a possible lower rate of metabolism.

View larger version (28K): [in this window] [in a new window]   Fig. 9.   HPLC, with on-line radioactivity detection, of plasma from male (A) and female (B) rats 4 h after oral administration of 0.7 mg/kg b.wt. [14C]levormeloxifene. Letters indicate metabolite assignment. Minor metabolites not shown in figure.

Tissues. Up to 72 h after administration of [¹⁴C]levormeloxifene, radioactivity was predominantly associated with the parent compound in nongastrointestinal tract tissues, in both sexes, at both dose levels. Figure 10 shows representative HPLC chromatograms of liver extracts from female rats 24 (a) and 72 h (b) after oral administration of 0.7 mg/kg b.wt. [¹⁴C]levormeloxifene. For example, [¹⁴C]levormeloxifene accounted for 85.8 to 93.9% of the total radioactivity in the liver of male and female rats after dosing at 0.7 mg/kg b.wt., declining to 76.2 to 88.4%, respectively, at 24 h. At later time points (24-h), there was an increase in the proportion of quantitatively minor metabolites, primarily 7-desmethyllevormeloxifene (M6) and monohydroxylevomeloxifene (M5 and M7). M6 was quantified by off-line radioactivity monitoring of collected fractions, because with the on-line radioactivity profile the response for M6 was only slightly above background. The metabolite profiles in lung and kidney were very similar to those in liver with unchanged drug accounting for the majority of the total radioactivity in the 2- to 4-h lung and 4-h kidney extracts.

View larger version (26K): [in this window] [in a new window]   Fig. 10.   HPLC, with on-line radioactivity detection, of liver extracts from female rats 24 (A) and 72 (B) h after oral administration of 0.7 mg/kg b.wt. [14C]levormeloxifene. Letters indicate metabolite assignment. Minor metabolites not shown in figure.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Paramount to the development of a suitable selective estrogen receptor modulator for the treatment of osteoporosis is an understanding of the distribution, metabolism, and excretion of the compound after oral administration to preclinical species, because the pharmacological profile of compounds that bind to estrogen receptors may be altered due to the formation of metabolites with higher estrogenic activity than the parent compound (Dodge et al., 1997).

The results from the current study indicate that within 2 h after oral dose administration, radioactivity was higher in all tissues than in the blood, indicating rapid distribution of levormeloxifene and/or metabolites. However, based on HPLC analysis of tissue extracts (liver, kidney, and lung), much of this radioactivity, up to 94% of the sample radioactivity in some tissue samples, was associated with parent compound (Fig. 10), as indeed was the radioactivity associated with the systemic circulation (Fig. 9). Peak radioactivity concentrations were generally achieved 4 h postdose in both male and female animals at dose levels of 0.7 and 50 mg/kg b.wt. Interestingly, peak radioactivity concentrations were generally circa 1.2 to 1.7 times greater in female rat tissue than in corresponding tissues from male animals (not shown), and this was also reflected in a slower elimination of radioactivity from female animals. The reason for these differences are unclear, although this may be related primarily to fundamental differences in the metabolism between sexes (Mugford and Kedderis, 1998). Alternatively differences in elimination rates may be related to the compound class, that of a selective estrogen receptor modulator and differences in tissue binding and discrete receptor interactions (Dodge et al., 1997) with respect to the distribution and concentration of estrogen receptors (Kuiper et al., 1997). However, after single oral doses any differences between the rate of elimination of drug is probably more likely attributable to sex differences in metabolism.

Radioactivity was slowly excreted into feces, presumably after conjugation of metabolites and excretion into bile (Tables 2, 3, and 6) with only approximately 1% of the administered dose being excreted via the renal route. The major metabolite isolated from feces was characterized by HPLC and LC-MS-MS as 7-desmethyllevormeloxifene (norlevormeloxifene), indicating a typical cytochrome P450 demethylation reaction on the methoxy group of levormeloxifene, and accounted for about 34 to 43% of fecal extract radioactivity and for about 25 to 33% of the dose. Unchanged drug was also excreted, mainly from 0 to 24 h, and accounted for about 6 to 12% of the dose. Together these two components accounted for approximately 50% of the radioactivity excreted in those feces samples analyzed. Additional metabolites isolated and identified by LC-MS-MS, and accounting for up to 5% of the excreted radioactivity in rat feces during the first 24 h included two separate monohydroxylevormeloxifene (hydroxylated on different benzene rings), and desmethylnorlevormeloxifene. The formation of this metabolite is highly unusual because the proposed structure would arise from C-demethylation and O-demethylation of levormeloxifene. C-demethylation is an unexpected metabolic reaction and in the absence of definitive evidence, the proposed structure should be regarded with caution. Additionally, a pyrrolidinone ring-opened metabolite of levormeloxifene was also isolated and identified. A proposed route of metabolism is shown in Fig. 11.

View larger version (22K): [in this window] [in a new window]   Fig. 11.   Proposed biotransformation pathway. Metabolite M3 not shown because this was identified as the aglucan of M4.

The pharmacokinetics of levormeloxifene were also determined in female rats, and drug measurements were performed on animals dosed with 1.0 and 0.5 mg/kg b.wt. levormeloxifene, reflecting the anticipated therapeutic dose range. Results indicated that C_max was generally observed 6 h after dosing, and the AUC values increased fairly proportionally to the dose. The half-life of elimination was long, being approximately 1 day. The plasma levels of the major metabolite 7-desmethyllevormeloxifene were in all pharmacokinetic samples below the lower limit of quantitation, confirming tissue distribution experiments, indicating that parent compound was the major circulating species and the major species in tissues even though the major metabolite excreted was 7-desmethyllevormeloxifene.

In conclusion, it would appear that levormeloxifene is an orally active compound and it can be predicted that the main site of metabolism is in the liver, with the major excretion pathway of parent compound and metabolites being via the fecal route in rodents. There appeared to be some minor gender differences in the distribution, metabolism, and excretion of radioactivity. However, there were no apparent changes in metabolism between dose levels. Similar studies in monkeys and human volunteers have indicted a similar excretion pathway and the formation of a number of comparable metabolites to those found in the rat species (Mountfield et al., 1999). What has not been established during these preclinical studies is the effect of the long half-life of this compound (219 h in volunteers; B.K., personal communication) on the overall consequences of repeated administration to patients. Interestingly, a similar compound being developed for osteoporosis, idoxifene, also has a long half-life, and clinical development was recently stopped due to adverse events in the clinic (SCRIP, 1999, 2431, p21).