Apixaban (BMS-562247; Fig. 1) is an oral anticoagulant in late-stage clinical development for the prevention and treatment of venous thromboembolism, stroke prevention in atrial fibrillation, and secondary prevention in acute coronary syndrome. It is a potent, oral, reversible, selective, and direct factor Xa inhibitor, which inhibits both free and prothrombinase-bound factor Xa activity, and shows considerable efficacy in the prevention of arterial and venous thrombosis at doses that preserved hemostasis in rabbits (Pinto et al., 2007; Wong et al., 2008). It is also effective and safe for the prevention and treatment of venous thrombosis in humans (Lassen et al., 2007; Büller et al., 2008). After oral administration of apixaban to human subjects, the parent compound was the major circulating component, and O-demethyl apixaban sulfate (Fig. 1) was a significant metabolite in human plasma; however, it was not as abundant relative to parent in plasma samples of mice, rats, female rabbits, and dogs (D. Zhang, N. Raghavan, L. Wang, K. He, and W. G. Humphreys, manuscript submitted for publication). Apixaban probably underwent O-demethylation catalyzed by cytochrome P450 enzymes to O-demethylated apixaban that was then conjugated by sulfotransferases (SULTs) to form a sulfate metabolite.

Sulfation is a major conjugation reaction of drug metabolism, and the SULTs are responsible for sulfation biotransformation of many xenobiotics and endogenous substrates such as steroids and neurotransmitters in humans (Rikke and Roy, 1996; Nagata and Yamazoe, 2000; Blanchard et al., 2004). The process of sulfation involves the transfer of a sulfonyl group of 3′-phosphoadenosine 5′-phosphosulfate (PAPS) to a hydroxyl or amino group of a molecule. This reaction generally results in a decrease in biological activity and an increase in hydrophilicity of xenobiotics (drugs) or endogenous compounds, so as to facilitate their excretion. SULTs play an important role in normal human homeostasis and drug metabolism (Glatt et al., 2000; Gamage et al., 2006). For example, SULT1E1 is responsible for the high affinity sulfation of β-estradiol (E2), estrone, and 17α-ethinylestradiol (Falany et al., 1995; Adjei et al., 2003; Schrag et al., 2004). SULT1A1 catalyzes the sulfation of many xenobiotics and is expressed in many tissues, including liver and small intestine (Chen et al., 2002; LeWitt, 2004).

SULTs are a superfamily of enzymes that catalyze the sulfate conjugation of various antibiotics and xenobiotics. Five gene families of SULTs, including SULT1, SULT2, SULT3, SULT4, and SULT5, have been identified in mammals. There is less than 45% amino acid sequence identity between families (Blanchard et al., 2004). In humans, 13 SULTs have been identified, all belonging to the SULT1, SULT2, or SULT4 families (Glatt et al., 2000, 2001; Gamage et al., 2006). Human SULT1A1 shows high homology with SULT1A2 (96%). Expression of SULT1A1, SULT1A3, SULT1B1, SULT1B2, SULT1E1, SULT1E4, SULT2A1, and SULT2A3 in adult human livers was documented and characterized (Richard et al., 2001; Honma et al., 2002; Tabrett and Coughtrie, 2003). SULT1A1 is the dominant SULT1A protein expressed in human liver (Gamage et al., 2006), and there are three important allelic variants, SULT1A1*1 (wild type), SULT1A1*2, and SULT1A1*3. SULT1A1*2 is defined by an R213H amino acid change (G to A change at nucleotide 638), and SULT1A1*3 variant is defined by a M223V amino acid change (A to G change at nucleotide 667) (Carlini et al., 2001). SULT1A1 allele frequencies have been reported in different populations (Carlini et al., 2001): 65.6% (white), 47.7% (African American), and 91.4% (Chinese) for SULT1A1*1; 33.2% (white), 29.4% (African American), and 8.0% (Chinese) for SULT1A1*2; and 1.2% (white), 22.9% (African American), and 0.6% (Chinese) for SULT1A1*3. These alloenzymes are associated with altered enzymatic activities (Raftogianis et al., 1997; Nagar et al., 2006). The general trend of V_max values estimated in previous study was *1>*3>*2, with trend of V_max/K_m values varied with substrates (Raftogianis et al., 1997). Six alloenzymes of SULT1A2 have been reported (Raftogianis et al., 1999). Human SULT1A3 showed 93% similarity to SULT1A1 (Blanchard et al., 2004; Gamage et al., 2006). No genetic polymorphisms have been reported for SULT1A3 and SULT1E1 at the DNA level. SULT2A1 is also a major enzyme located in human liver and involved in drug metabolism (Meloche et al., 2002). In this study, the sulfation activities of human cDNA-expressed SULTs, as well as liver S9 from mice, rats, rabbits, dogs, monkeys, and humans, on O-demethyl apixaban were investigated to determine the principal SULT enzymes involved in this sulfation reaction and to compare enzyme activities between species.

Materials and Methods

Materials. Apixaban (purity >99%), O-demethyl apixaban (purity >95%), and O-demethyl apixaban sulfate (purity >85%) were synthesized at Bristol-Myers Squibb (Princeton, NJ). The structures of apixaban and O-demethyl apixaban sulfate are shown in Fig. 1. PAPS, magnesium chloride, all the SULT inhibitors [quercetin dihydrate, estrone, and 2,6-dichloro-4-nitrophenol (DCNP)] were purchased from Sigma-Aldrich (St. Louis, MO). Pooled human liver S9 fraction (20 subjects) was purchased from BD Biosciences (Bedford, MA). Pooled liver S9 fractions of mice, rats, dogs, and monkeys were purchased from XenoTech, LLC (Lenexa, KS). Male and female rabbit liver S9 was purchased from In Vitro Technologies (Baltimore, MD). The highly active cytosolic extracts of Sf-9 insect cells infected with a baculovirus strain containing human cDNA encoding specific human SULTs (SULT1A1, SULT1A2, SULT1A3, SULT1E1, SULT2A1) were purchased from Invitrogen (Carlsbad, CA). Human factor Xa and α-thrombin were purchased from Hematologic Technologies (Essex Junction, VT). Human pancreatic trypsin was purchased from Calbiochem (Darmstadt, Germany). For factor Xa, thrombin, and trypsin, the respective peptide substrates were S2765 (N-α-benzyloxycarbonyl-d-arginyl-l-glycyl-l-arginine-para-nitroaniline-dihydrochloride), S2366 (l-pyroglutamyl-l-prolyl-l-arginine-para-nitroaniline hydrochloride), and S2222 (N-benzoyl-l-isoleucyl-l-glutamyl-glycyl-l-arginine-para-nitroaniline hydrochloride and its glutamyl methyl ester; 1:1 mixture), which were purchased from DiaPharma Group, Inc. (West Chester, OH). Acetonitrile was purchased from Honeywell Burdick & Jackson (Muskegon, MI). All the organic solvents and water were of high-performance liquid chromatography (HPLC) grade. O-Demethyl apixaban stock solutions (5 mM) were prepared in acetonitrile/water (2.5:1, v/v) for the concentration-dependent metabolite formation studies. The stock solutions of apixaban (0.1 mM) and O-demethyl apixaban sulfate (1 mg/ml) were prepared in acetonitrile/ water (1:1, v/v). The 250 μM stock solutions of quercetin, estrone, and DCNP were separately prepared in ethanol.

Metabolite Profiling in Rat and Human Plasma. Pooled plasma samples (at 24 h) of rats and humans after oral administration of [¹⁴C]apixaban were obtained from previous studies (Raghavan et al., 2009; D. Zhang, N. Raghavan, L. Wang, K. He, and W. G. Humphreys, manuscript submitted for publication). Plasma samples were extracted in duplicate by addition of 4 ml of acetonitrile/methanol (1:1, v/v) to 1 ml of plasma while the sample was mixed on a vortex mixer. After centrifugation at 2000g for 1 h, each supernatant fraction was removed and saved. The precipitate was resuspended in 2 ml of acetonitrile and 1 ml of methanol. After centrifugation of the mixture for 30 min at 2000g, the supernatant fraction was removed and combined with the first supernatant. The precipitate was resuspended in 2 ml of acetonitrile. After centrifugation of the mixture at 2000g for 30 min, the supernatant fraction was removed and combined with the first and second supernatants. The combined supernatant fraction was evaporated to dryness under nitrogen and reconstituted in 0.15 ml of acetonitrile and 0.05 ml of methanol. After centrifugation at 2000g for 5 min, a portion of 100 μl of supernatant was injected into the HPLC for metabolite profiling and identification.

Assessment of Pharmacological Activities ofO-Demethyl Apixaban andO-Demethyl Apixaban Sulfate. The enzyme activities of factor Xa, α-thrombin, and trypsin were determined as described previously (Pinto et al., 2007; Wong et al., 2008). In brief, the final concentration of human factor Xa, human α-thrombin, or human pancreatic trypsin in the assays was 0.1, 0.125, and 0.625 nM, respectively. The inhibition of factor Xa thrombin or trypsin was evaluated with substrate S2765 (100 μM), S2366 (200 μM), or S2222 (100 μM). Stock solutions of O-demethyl apixaban or O-demethyl apixaban sulfate were prepared in 20 mM phosphate buffer, pH 7.0, and kept frozen until the day the assays were performed. Assays were conducted at room temperature in 96-well microtiter plate spectrophotometers (Molecular Devices, Sunnyvale, CA) with simultaneous measurement of enzyme activities in control and O-demethyl apixaban or O-demethyl apixaban sulfate-containing solutions. Assays were initiated by adding enzyme to buffered solutions containing synthetic substrates in the presence or absence of O-demethyl apixaban or O-demethyl apixaban sulfate. Hydrolysis of the substrate resulted in the release of para-nitroaniline, which was monitored spectrophotometrically by measuring the increase in absorbance at 405 nm. The rate of absorbance change is proportional to the enzyme activity. A decrease in the rate of absorbance change in the presence of O-demethyl apixaban or O-demethyl apixaban sulfate is indicative of enzyme inhibition. Assays were conducted under conditions of excess substrate and O-demethyl apixaban or O-demethyl apixaban sulfate over enzyme. If negligible inhibition was observed at the highest inhibitor concentration tested, the inhibitory constant was conservatively estimated by assuming 20% inhibition at the highest dose. O-Demethyl apixaban sulfate was evaluated for inhibition of human factor Xa at concentrations up to 750 μM, and for inhibition of human thrombin and human trypsin at concentrations up to 30 μM. O-Demethyl apixaban was evaluated for inhibition of the factor Xa at a range of 3 to 10,000 nM and for inhibition of the trypsin at a range of 1 to 30,000 nM.

Sulfation ofO-Demethyl Apixaban by Human cDNA-Expressed SULTs. The sulfation activity of O-demethyl apixaban was determined using human cDNA-expressed SULTs (SULT1A1, SULT1A2, SULT1A3, SULT1E1, and SULT2A1). The incubation mixtures (0.5 ml, in triplicate) contained 50 mM potassium phosphate buffer, pH 7.0, 5 mM MgCl₂, SULT (15 μg of protein), 30 μM O-demethyl apixaban, and 2.5 mM PAPS. Acetonitrile in the incubation mixtures was 0.5%. Before addition of PAPS, the mixtures were preincubated at 37°C for 5 min. After 30-min incubation at 37°C with shaking (100 rpm), ice-cold acetonitrile (0.5 ml) was added to each incubation to stop the reaction, and the internal standard (IS, apixaban) was added to each sample to a final concentration of 100 nM. After centrifugation at 2000g for 15 min, an aliquot of supernatant (10 or 20 μl) was used for liquid chromatography/mass spectrometry (LC/MS) analysis.

Sulfation Assays in Human Liver S9 in Presence of SULT Inhibitors. The sulfation activity of O-demethyl apixaban was determined in human liver S9 in the presence of SULT inhibitors. The incubation mixtures (0.5 ml, in triplicate) contained 50 mM phosphate buffer, pH 7.0, 5 mM MgCl₂, human liver S9 (100 μg of protein), 30 μM O-demethyl apixaban, single SULT inhibitor, and 2.5 mM PAPS. The chemical inhibitors used were quercetin dihydrate (1 μM) for SULT1A1 and SULT1E1, estrone (1 μM) for SULT1E1 and SULT1A3, and DCNP (0.5 μM) for SULT1A1. Acetonitrile and ethanol in the incubation mixtures were 0.5% (v/v) and 0.4% (v/v), respectively. Before addition of PAPS, the mixtures were preincubated at 37°C for 5 min. After 30-min incubation at 37°C with shaking (100 rpm), ice-cold acetonitrile (1 ml) was added to each incubation to stop the reaction, and the IS (apixaban) was added to each sample to a final concentration of 100 nM. After centrifugation at 2000g for 15 min, an aliquot of supernatant (10 or 20 μl) was used for LC/MS analysis. Control incubations (without inhibitor or PAPS) were performed under similar conditions.

Sulfation Assays in Liver S9 of Different Species. Sulfation activities of O-demethyl apixaban were determined using liver S9 fractions from mice, rats, rabbits, dogs, monkeys, or humans, and O-demethyl apixaban as substrate. Incubation mixtures (0.5 ml, in triplicate) contained 50 mM potassium phosphate buffer, pH 7.5, 5 mM MgCl₂, liver S9 (100 μg of protein), 30 μM O-demethyl apixaban, and 2.5 mM PAPS. Before addition of PAPS, the mixtures were preincubated at 37°C for 5 min. After 30-min incubation at 37°C with shaking (100 rpm), ice-cold acetonitrile (0.5 ml) was added to each incubation to stop the reaction, and the IS (apixaban) was added to each sample to a final concentration of 100 nM. After centrifugation at 2000g for 15 min, an aliquot of supernatant (10 or 20 μl) was used for LC/MS analysis.

Substrate Concentration-Dependent Sulfate Formation. For enzyme kinetic studies, the incubation mixtures (0.5 ml) contained 50 mM phosphate buffer, pH 7.5, 5 mM MgCl₂, human liver S9 (150 μg of protein), or expressed SULTs (20 pmol), O-demethyl apixaban, and 2.5 mM PAPS. The final concentration of acetonitrile in these incubation mixtures was 0.5% (v/v). Before PAPS addition, the mixtures were preincubated at 37°C for 5 min. After PAPS addition, the samples were incubated at 37°C for 30 min with shaking (100 rpm); ice-cold acetonitrile (0.5 ml) then was added to each sample to stop reaction; and the IS (apixaban) was added to each sample to a final concentration of 100 nM. An aliquot of 20 μl of supernatant was used for LC/MS analysis. To determine linear conditions for protein concentration and incubation time, O-demethyl apixaban (10 μM) was incubated up to 50 min with 100, 200, 300, and 600 μg of human liver S9 protein/ml; O-demethyl apixaban (10 μM) was incubated up to 50 min with human cDNA-expressed SULTs at 8 to 20 μg of protein/ml: 8.8 μg of protein/ml for SULT1A1, 12 μg of protein/ml for SULT1A2, or 600 μg of protein/ml for human liver S9 and 50 min of incubation time was in a linear range; these conditions were used for enzyme kinetic studies. Thirteen substrate concentrations—0.01, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 30, 50, and 100 μM—were evaluated in triplicate for expressed SULT1A1 and SULT1A2; 11 substrate concentrations—0.5, 1, 2, 5, 10, 20, 30, 50, 100, 200, and 300 μM—were evaluated in triplicate for human liver S9.

Identification and Quantification of Sulfate Metabolite. Metabolites in human plasma samples of rats and human were analyzed as described previously (D. Zhang, N. Raghavan, L. Wang, K. He, and W. G. Humphreys, manuscript submitted for publication). Metabolites in samples of incubations were analyzed by LC/tandem MS method. For incubation samples, after termination of the incubation, an IS was added to the reaction mixture. LC/MS analysis was performed on a Finnigan LTQ mass spectrometer (Thermo Fisher Scientific, Waltham, MA) with an electrospray ionization probe in a positive ion mode. The HPLC system was an Agilent Technologies (Santa Clara, CA) 1100 series system equipped with two pumps, an autoinjector, a UV detector, and an ACE 3 C18 column (4.6 × 150 mm). The mobile phase consisted of two solvents: A) 0.4% formic acid in water, pH 3.2, and B) 100% acetonitrile. The gradient used was as follows: solvent B started at 5%, then linearly increased to 20% at 5 min, to 30% at 50 min, to 35% at 55 min, to 90% at 65 min, held at 90% for 2 min, and then decreased to 5% at 69 min. The mobile phase flow rate was 0.7 ml/min. The HPLC effluent was directed to the mass spectrometer through a valve set to divert the flow to waste from 0 to 5 min. The capillary temperature used for analysis was set at 250°C. The gas flow rate of nitrogen, spray current, and voltages were adjusted to give maximum sensitivity using apixaban as a standard. For metabolite quantification in incubation samples, the detection was performed at m/z 526 → 446 (O-demethyl apixaban sulfate), m/z 446 → 429 (O-demethyl apixaban), and m/z 460 → 443 (apixaban). The retention time of O-demethyl apixaban sulfate, O-demethyl apixaban, and apixaban was 10.50, 11.53, and 13.04 min, respectively. The quantification of O-demethyl apixaban sulfate was achieved with the same LC/MS system, HPLC column, and analysis gradient conditions. A standard curve of eight points of O-demethyl apixaban sulfate, ranging from 0.001 to 100 μg/ml, was generated by linear regression (r = 0.98). An aliquot (10–20 μl) of supernatant was injected for LC/MS analysis. The concentration of O-demethyl apixaban sulfate in each incubation sample was quantified from the peak area ratio (m/z 526 to 446) based on the calibration curve of O-demethyl apixaban sulfate.

Data Analysis. The formation rates of O-demethyl apixaban sulfate from incubations with a broad substrate concentration range were evaluated by fitting the data to the Michaelis-Menten equation, V = V_max · S/(K_m + S). Linear regression and nonlinear regression analyses were performed by SigmaPlot (version 8; SPSS Inc., Chicago, IL). The formation rate (nanomole per minute per milligram of protein of SULT or human liver S9) of O-demethyl apixaban sulfate in each incubation sample was calculated and plotted. Enzyme kinetic parameters were estimated using SigmaPlot nonlinear regression analysis to obtain K_m values and V_max values.

For substrate hydrolysis each protease was determined at room temperature by fitting the data to the Michaelis-Menten equation. For K_m determination initial hydrolysis rates from 0 to 5 min were used. For control, O-demethyl apixaban, and O-demethyl apixaban sulfate, the steady-state hydrolysis rates of 25 to 30 min were used. The IC₅₀ was calculated by fitting the percentage of activity versus control: IC₅₀ = 100[1 – (V_i/V_o)], where V_i is the observed velocity in the presence of the inhibitor, and V_o is the observed velocity in the absence of the inhibitor. The following relationship was used to calculate K_i values: K_i = IC₅₀/(1 + S/K_m) for a competitive inhibitor, where IC₅₀ is the concentration of inhibitor that produces 50% inhibition; K_i is the dissociation constant of the enzyme and inhibitor complex; S is the concentration of substrate; and K_m is the Michaelis-Menten constant for the substrate.

Results

Metabolite Profiling and Identification. The HPLC radiochromatographic profiles of plasma samples (at 24 h) from rats and humans are shown in Fig. 2. In rat plasma profile, apixaban was the dominant component (98%), and O-demethyl apixaban sulfate was minor (1.8%); in human plasma, major circulating components were apixaban (61%) and O-demethyl apixaban sulfate (34%). O-Demethyl apixaban was a minor metabolite detected by LC/MS analysis only in plasma samples at early time points (6 h) (Raghavan et al., 2009; D. Zhang, N. Raghavan, L. Wang, K. He, and W. G. Humphreys, manuscript submitted for publication).

The LC/MS analyses of plasma and incubation samples showed that a metabolic peak at 10.50 min had a molecular ion [M+H]⁺ at m/z 526 and a product ion at m/z 446 (526–80), indicating the loss of 80 Da. Based on this information, a sulfate conjugate of O-demethyl apixaban was proposed. Subsequent synthesis of an authentic standard (Raghavan et al., 2009) and comparison with LC/MS confirmed the proposed structure. This metabolite peak was detected in the incubations of liver S9 (mice, rats, rabbits, dogs, monkeys, and humans), SULT1A1, SULT1A2, SULT1A3, SULT1E1, and SULT2A1. A second peak at 11.53 min had a molecular ion [M+H]⁺ at m/z 446, and LC/tandem MS analysis showed a product ion at m/z 429. This second peak was O-demethyl apixaban. Apixaban (at 13.04 min, IS) showed a molecular ion [M+H]⁺ at m/z 460 and a major fragment ion at m/z 429.

Pharmacological Activities ofO-Demethyl Apixaban andO-Demethyl Apixaban Sulfate. The enzyme activities for factor Xa, thrombin, and trypsin were determined using para-nitroaniline–coupled peptides as substrates (S2765, S2366, and S2222). Under these conditions, K_m values were 23.6, 220, and 31 μM for factor Xa, thrombin, and trypsin, respectively. The significant inhibition for factor Xa was observed only at high concentrations of O-demethyl apixaban and O-demethyl apixaban sulfate (>20 μM). The value of IC₅₀ and K_i were approximately 300 and 58 μM, respectively, for O-demethyl apixaban sulfate. O-Demethyl apixaban sulfate at concentrations up to 30 μM did not inhibit thrombin or trypsin amidolytic activity toward their respective peptide substrates. For both enzymes, the K_i values of O-demethyl apixaban sulfate can be conservatively estimated to be greater than 10 μM. O-Demethyl apixaban had a K_i value of 20 nM for inhibition of factor Xa, 250-fold less potent than apixaban (0.08 nM) (Pinto et al., 2007), and a K_i value of >15,000 nM for inhibition of trypsin.

Sulfation Activity by Various Recombinant SULTs. Among five recombinant human SULTs screened for the sulfation activity, all generated O-demethyl apixaban sulfate. In human cDNA-expressed SULTs, the sulfation activities followed a decreasing order of SULT1A1 > SULT1A2 ≫ SULT1E1 > SULT1A3 ≈ SULT2A1 (Table 1; Fig. 3).

Inhibition Studies. The sulfation activity of O-demethyl apixaban was determined with human liver S9 in the presence of SULT inhibitors (Table 2; Fig. 4). Quercetin (1 μM) inhibited the formation of O-demethyl apixaban sulfate by 99%; DCNP (0.5 μM), a highly selective inhibitor for SULT1A1 (IC₅₀ < 0.1 μM), also inhibited the formation of O-demethyl apixaban sulfate by >90%; estrone, which has a high selectivity for SULT1E1, showed no effect at 1 μM concentration on the formation of O-demethyl apixaban sulfate. The inhibition results indicated that SULT1A1 is likely the major SULT catalyzing the sulfation of O-demethyl apixaban in humans.

Species Comparison ofO-Demethyl Apixaban Sulfation. The formation of O-demethyl apixaban sulfate was observed in all the liver S9 incubations, and the formation rates are shown in Table 3 and Fig. 5. Relatively low levels of sulfation activities (from ∼1–11 pmol/min/mg protein) were found in liver S9 of mice, rats, male rabbits, and female rabbits. Dog, monkey, and human liver S9 all showed significantly higher levels of sulfation activity for O-demethyl apixaban than mice, rat, and rabbit liver S9.

Substrate Concentration-Dependent Metabolite Formation. The formation rates of O-demethyl apixaban sulfate were measured over a range of substrate concentrations (0.5–300 μM) in human liver S9 and expressed SULT1A1 and SULT1A2. Formation of O-demethyl apixaban sulfate exhibited hyperbolic kinetics, and the data were fitted to the Michaelis-Menten equation. The K_m values for sulfation of O-demethyl apixaban were similar (37–41 μM) in human liver S9 and SULT1A1 (Table 4; Fig. 6); however, the V_max value in SULT1A1 was 53-fold higher than that in human liver S9 and 5-fold higher than that in SULT1A2. The higher catalytic efficiency (V_max/K_m) of SULT1A1 suggests that SULT1A1 is likely to play a major role for formation of O-demethyl apixaban sulfate in humans.

Discussion

The radioactivity profiles of plasma at 24 h postdose showed that both apixaban and O-demethyl apixaban sulfate were prominent circulation components in humans, whereas apixaban was the dominant component in rats. O-Demethyl apixaban was a minor metabolite detected by LC/MS analysis only in plasma samples of early time points (6 h) (Raghavan et al., 2009; D. Zhang, N. Raghavan, L. Wang, K. He, and W. G. Humphreys, manuscript submitted for publication). In addition, O-Demethyl apixaban and O-demethyl apixaban sulfate did not significantly inhibit purified human factor Xa. The affinity constants (K_i) were 20 nM and 58 μM for O-demethyl apixaban and O-demethyl apixaban sulfate, respectively, more than 250-fold less potent than apixaban (Pinto et al., 2007). In contrast, the K_i of apixaban for human factor Xa is 0.08 nM. Based on K_i values and human plasma profiles (apixaban accounting for more than 61% radioactivity), the antithrombotic effects of O-demethyl apixaban and O-demethyl apixaban sulfate are negligible in humans in clinical treatment. These results indicate that O-demethyl apixaban and O-demethyl apixaban sulfate did not possess any activity that could contribute to the pharmacological activity of apixaban. O-Demethyl apixaban sulfate did not produce significant inhibition of human thrombin or trypsin at concentrations up to 30 μM. The lack of affinity for factor Xa, thrombin, and trypsin suggests that O-demethyl apixaban sulfate is also unlikely to inhibit other serine proteases at low micromolar concentrations achieved after administration of apixaban in humans (Raghavan et al., 2009).

The identification of the SULTs involved in the sulfation of O-demethyl apixaban was carried out with initial screening of metabolic turnover by cDNA-expressed enzymes, followed by evaluation of the effects of chemical inhibitors on sulfation reaction in human liver S9. On initial screening with cDNA-expressed enzymes, multiple SULTs (SULT1A1, SULT1A2, SULT1A3, SULT1E, and SULT2A1) were capable of forming O-demethyl apixaban sulfate; however, SULT1A1 and SULT1A2 had a higher catalytic efficiency than other enzymes tested. Quercetin and DCNP were selective inhibitors of SULT1A1 with IC₅₀ values of approximately 100 nM (Walle et al., 1995; Schrag et al., 2004). In the present study, results showed that quercetin (1 μM) and DCNP (0.5 μM) selectively inhibited the formation of O-demethyl apixaban sulfate by >90% in human liver S9 incubations at 20 and 100 μM of O-demethyl apixaban (Fig. 4). In contrast, estrone, which is a high affinity inhibitor for SULT1E1 with a K_m value of approximate 6 nM (Schrag et al., 2004), had no significant effect on the formation of O-demethyl apixaban sulfate in human liver S9 incubations at 20 and 100 μM of O-demethyl apixaban (Fig. 3). These data suggested that SULT1A1 plays an important role in the formation of O-demethyl apixaban sulfate.

SULT1A1 had a K_m value of 36.8 μM for the formation of O-demethyl apixaban sulfate, similar to that in human liver S9 (41.4 μM). The estimated K_m value for SULT1A2 was 70.8 μM. The intrinsic clearance (Cl_int) values (V_max/K_m) of SULT1A1 and SULT1A2 for the formation of O-demethyl apixaban sulfate were 13 and 1.2 ml/min/mg protein, respectively, suggesting that the formation of O-demethyl apixaban sulfate in human liver was efficient. Because the Cl_int value of SULT1A1 was 10-fold higher than that of SULT1A2 and there is a high expressed level of SULT1A1 in the human liver, these findings suggest that SULT1A1 may play a significant role in apixaban metabolism and elimination.

The in vitro sulfation activities of O-demethyl apixaban were highly variable among species, with humans, monkeys, and dogs showing higher levels of the activity than rabbits, rats, and mice. These in vitro results are somewhat consistent with different amounts of the sulfate metabolite found in vivo, although the metabolite profile of dogs and monkeys would have been predicted to be more similar to human than that found. This metabolite represented approximate 3% of dose after a single 20-mg oral dose of [¹⁴C]apixaban in humans but was only detected as a very minor metabolite in mice, rats, or dogs after oral administration of [¹⁴C]apixaban in these animals (Zhang et al., 2009). It is possible that there was significant formation of O-demethyl apixaban sulfate in the liver and subsequently excreted in the bile and then hydrolyzed back to O-demethyl apixaban in intestines of animal species.

In summary, liver S9 fractions from human and animal species generated O-demethyl apixaban sulfate in incubations with O-demethyl apixaban. The studies with cDNA-expressed enzymes, SULT chemical inhibitors, and kinetic analysis showed that O-demethyl apixaban sulfate was mainly formed by SULT1A1.