Abstract
Efflux transporters such as P-glycoprotein (P-gp) and multidrug resistance-associated proteins (Mrps) and their contributions to saquinavir (SQV) brain uptake were characterized. Cerebral flow rate was estimated from diazepam uptake and brain vascular volume was assessed using inulin. Mice brains were perfused with buffer containing SQV alone or coperfused with different concentrations of GF120918, a P-gp inhibitor or MK571, a specific Mrp family inhibitor. Inulin, a nonabsorbable marker, was also coperfused in all studies to assess whether the inhibitors altered the physical integrity of the blood-brain barrier (BBB). The estimated cerebral flow rate using diazepam was 250 ml·100g-1·min-1. The brain vascular volume, estimated using inulin, was almost constant (0.94 ± 0.03 ml·100 g-1, n = 12) during the perfusion study. SQV uptake kinetics was linear during the sampling period. Inclusion of 10 μM GF120918 in the perfusate resulted in a more than 7-fold increase in the brain distributional volume (i.e., uptake) of SQV. Inclusion of 100 μM MK571 in the perfusate increased SQV apparent brain uptake by more than 4.4-fold, suggesting, for the first time, that Mrp transporters may play an important role in the brain uptake and retention of SQV. Neither GF120918 nor MK571 altered the integrity of the BBB during the time course of the study. Although the current results reaffirm that SQV is a P-gp substrate, this is the first report implicating the Mrp transporter family in the limited brain uptake and retention of SQV in vivo in mice.
The human immunodeficiency virus (HIV)-1 protease inhibitor (PI) SQV was the first agent among HIV PIs approved and used for the treatment of AIDS and has very high selectivity and binding affinity to HIV-1 protease (Martin, 1992). Combination therapies using protease and reverse transcriptase inhibitors have proven successful in reducing the plasma viral load of HIV-infected patients (Collier et al., 1996). However, SQV has a number of biopharmaceutical limitations, related in many instances to poor and/or variable transport across important biological membranes. For example, oral absorption is often low and variable and penetration into the brain is very poor (Aungst, 1999). Several factors, including poor solubility, high protein binding, and extensive first pass metabolism contribute to poor transport across biological membranes (Barry et al., 1997). In addition, secretory membrane transporters, including P-gp and Mrps, have been implicated from in vitro studies (Kim et al., 1998; Williams et al., 2002). The expression of P-gp in capillary endothelial cells of the brain has been shown to be a critical factor in preventing the entry of SQV into the central nervous system (CNS) (Kim et al., 1998). CNS drug therapy is very challenging since few CNS-active drugs have appreciable permeability across the BBB (Bickel et al., 2001). Since the CNS represents a sanctuary site for HIV-1 infection, the incidence of CNS complications from secondary pathologies such as AIDS dementia is high (Gendelman et al., 1997). Hence, the poor permeation of SQV into the brain is likely to limit the effective treatment of AIDS.
The BBB, composed of a single layer of endothelial cells connected by tight junctions, restricts the transport of compounds from the circulating blood into the brain (Pardridge, 1995). The poor distribution of drugs into the CNS is also limited by specific membrane efflux transporters such as P-gp, the product of the multidrug resistance (MDR) gene (Kakee et al., 1996). Besides P-gp, which is expressed on the luminal side of endothelial cells, other ATP-binding cassette (ABC) transporters such as the family of Mrps (Borst et al., 2000) and breast cancer resistance protein (BCRP) (Eisenblatter and Galla, 2002) have been found in human and rodent brains. At this time, the functions and locations of the Mrps and BCRP in the brain have not been fully clarified.
Using an immunostaining method Miller et al. (2000, 2002) showed that P-gp and Mrp2 were expressed in the luminal side of fish, rat, and pig brain capillaries, whereas mRNA encoding Mrp2 in bovine brain capillary endothelial cells and rat brain capillaries has not been detected by other investigators (Zhang et al., 2000; Sugiyama et al., 2003). Mrp1 has been detected on bovine and rat brain capillary endothelial cells (Huai-Yun et al., 1998; Regina et al., 1998) and it has been also found on the basolateral side of epithelial cells in the choroid plexus (Rao et al., 1999). The mRNA encoding Mrp1, Mrp4, Mrp5, and Mrp6 in both bovine brain capillary endothelial cells and capillary homogenates were detected (Zhang et al., 2000). Nevertheless, the function of Mrp homologues in brain capillary endothelial cells in controlling the transport of drugs across the BBB remains to be investigated.
Transport studies using an in vitro model of the BBB suggested that the expression of P-gp (Glynn and Yazdanian, 1998) and Mrps was functional. However, since the expression of these efflux transporters may be up- or down-regulated during cell culture (Gutmann et al., 1999b), in vivo investigations into the role of the efflux transporters in facilitating SQV brain uptake and retention are required to fully understand the possible role of these proteins in the BBB.
The aim of this study was to elucidate the mechanism responsible for the low brain uptake of SQV using an in situ mice brain perfusion technique. The contribution of efflux transporters such as P-gp and Mrps to brain uptake of SQV was characterized by directly perfusing mice brains with a physiological buffer including SQV in the absence and presence of selective inhibitors of P-gp and Mrps.
Materials and Methods
Materials. [3H]Inulin (0.54 Ci/mmol) and [14C]diazepam (56 mCi/mmol) were purchased from Amersham Biosciences (Piscataway, NJ). [14C]Saquinavir (26.5 μCi/mg) was provided by Roche Discovery Welwyn (Welwyn Gardon City, UK). GF120198 and MK571 (base form) were provided by GlaxoSmithKline (Research Triangle Park, NC) and Merck (Whitehouse, NJ), respectively. d-Glucose was purchased from Sigma-Aldrich (St. Louis, MO). Solvable was purchased from PerkinElmer Life and Analytical Sciences (Boston, MA). All other chemicals and reagents were purchased from Fischer Scientific (Houston, TX) and were of analytical grade.
Animals. Male FVB mice (26-30 g) were purchased from Taconic Farms (Germantown, NY) and maintained under standard conditions of temperature and lighting with ad libitum access to food and water. All experimental procedures complied with the protocol (02-029) approved by the Institutional Review Board Use and Care of Animal Committee and housed in Association for Assessment and Accreditation of Laboratory Animal Care accredited facilities at Rutgers University.
In Situ Brain Perfusion. The in situ brain perfusion technique used in the experiments was similar to that described previously (Takasato et al., 1984; Dagenais et al., 2000) with modifications. Briefly, for surgical preparation, mice were anesthetized with ketamine/xylazine (140/8 mg/kg i.p.). After exposure of the left common carotid artery, the left external carotid artery was ligated at the bifurcation of the common carotid artery with the internal carotid artery. The left common carotid artery was ligated caudally. A polyethylene tube (0.30 mm i.d. × 0.70 mm o.d.) filled with heparin (20 units/ml) was catheterized into the left common carotid artery under the microscope. For perfusion experiments, the left hemisphere of the mouse brain was perfused with perfusion buffer containing test compounds at a selected flow rate through the catheter that was connected to the perfusion pump. The perfusion buffer consisted of bicarbonate-buffered physiological saline (128 mM NaCl, 4.2 mM KCl, 24 mM NaHCO3, 2.4 mM NaH2PO4, 1.5 mM CaCl2, and 0.9 mM MgSO4). d-Glucose (9 mM) was added before an experiment. The perfusate was bubbled with a mixture of 95% O2 and 5% CO2 for pH control (7.4) and maintained at 37°C. Immediately before initiation of the perfusion, the cardiac ventricles were severed to eliminate the contribution of contralateral blood flow. Multiple time-point experiments (15, 30, 60, and 90 s for diazepam and 30, 60, 90, and 120 s for SQV) were performed at a calibrated flow rate (Harvard pump PHD2000; Harvard Apparatus, Holliston, MA). At least 0.8 μCi/ml of [3H]inulin was added into the perfusates as a vascular space marker to obtain about 1000 dpm/tissue sample. The perfusates contained radiolabeled [14C]diazepam or [14C]SQV with or without GF120918 or MK571 to produce an appropriate drug concentration. The stock solutions of GF120918 (4 mM) and MK571 (13 mM) were prepared in DMSO. The stock solutions were diluted with the bicarbonate-buffered saline used for perfusion.
The perfusion was terminated by decapitation. The brain was removed from the skull, dissected and each hemisphere was weighed, and placed in a preweighed scintillation vial. Brain and perfusion fluid samples were digested in 0.7 ml of Solvable (PerkinElmer Life and Analytical Sciences) at 37°C for 24 h. Scintillation cocktail (5 ml) was added to each vial and radioactivities of 3H and 14C were determined simultaneously by dual liquid scintillation counting. All data are reported for the left hemisphere.
Calculations and Data Analysis. Data from the brain perfusion method was analyzed as described previously (Dagenais et al., 2000; Chen et al., 2002). Briefly, brain vascular volume (Vvasc, ml·100 g-1), defined as the ratio of the vascular marker concentration in brain to that in the perfusate, was determined using the following equation:
where X* is the amount of radiolabeled inulin in the brain (dpm·100 g-1), and C* is the perfusate concentration (dpm·ml-1). The unidirectional transfer coefficient Kin (ml·min-1·100 g-1) was calculated using the following relationship:
where Xbrain is the amount of radiotracer in the brain (dpm·100 g-1) corrected for vascular contamination (Xtotal - Vvasc·Cpf), Xtotal (dpm·100 g-1) is the total quantity of tracer measured in the tissue sample (vascular + extravascular), and Cpf is the tracer concentration in the perfusate (dpm·ml-1). In a single time point experiment, Xbrain/T replaced dXbrain/dt, where T was the perfusion time (minutes). Apparent brain distributional volumes (Vbrain, ml·100 g-1) were calculated from
Brain flux J (nmol/g brain/min) was calculated according to
where Cpf is the SQV concentration in the perfusate. Kinetic relationships between the brain uptake of SQV concentration were described by a nonlinear least-squares method. SQV brain flux J was divided into two components:
where Jd is the maximal brain flux of SQV when no efflux transporters are involved and it is transported by passive diffusion depending on the physicochemical properties of the drug and Jefflux is the transporter-mediated component (i.e., efflux transporters at the luminal membrane of the BBB endothelial cells). The best-fit line was obtained using a modified Michaelis-Menten equation as follows (GraphPad Prism 4; GraphPad Software Inc., San Diego, CA):
where Jmax is the maximum uptake rate for a saturable component, Km is the Michaelis constant, Kd is the first-order constant for the nonsaturable component, and Cpf is the concentration of SQV in the perfusates. SQV uptake data at 60 s in the inhibition studies were described using the following relationship:
where Kin, max is the maximal brain Kin (ml·min-1·100 g-1), C is the inhibitor concentration in the perfusate (μmol/l), and IC50 is the inhibitory concentration at half-maximal brain Kin (μmol/l). Estimates of SQV inhibition study parameter (IC50) were obtained by fitting a Michaelis-Menten equation to the unidirectional transfer coefficient Kin versus concentrations of inhibitors data by nonlinear least-squares regression using the same software. All the graphs were made using GraphPad Prism 4.
Data are presented as mean ± S.E.M. for three to five animals. When appropriate, analysis of variance or two sided Student t tests were used to determine the statistical significance of differences between experimental groups. Statistical significance was determined at the level of p = 0.05.
Results
Determination of the Perfusion Conditions. To study mechanisms restricting brain uptake of SQV, the perfusion conditions were first validated and optimized by investigating brain uptake of radiolabeled markers such as [14C]diazepam and [3H]inulin. [14C]Diazepam (0.3 μCi/ml) was used for selecting an appropriate perfusion flow rate and estimating regional cerebral perfusion rate since diazepam is known to be completely extracted during a single pass through the brain and has no significant back flow (Takasato et al., 1984). All perfusates contained [3H]inulin (1 μCi/ml) as a vascular marker because it does not significantly penetrate the BBB during short periods of perfusion. A linear relationship between initial brain transfer coefficient (Kin) of diazepam at 20 s and perfusion flow rate was observed (Fig. 1A). A perfusion flow rate of 2.1 ml·min-1, which results in Kin of diazepam similar to that in a previous report in mice (Dagenais et al., 2000) was chosen and used for the rest of experiments. The integrity of the BBB was checked by monitoring the distribution volume of [3H]inulin (Vvasc) for each mouse perfused. BBB integrity was well maintained at the higher flow rates. The regional cerebral fluid flow (Fpf; ml·g-1·s-1) was estimated in a separate perfusion experiment at 2.1 ml·min-1 (15-, 30-, 60, and 90-s time points). As shown in Fig. 1B, [14C]diazepam showed linear uptake kinetics through 90 s with a Kin of 250 ml·100 g-1·min-1 (Fig. 1B). This observation is also consistent with previously reported results (Dagenais et al., 2000). The brain vascular volume was estimated using [3H]inulin, and it was constant (0.94 ± 0.03 ml·100 g-1; mean ± S.E.M., n = 12) during perfusion for 15 to 90 s (Fig. 1B). A control experiment was performed to evaluate the effect of DMSO that was used for the preparation of stock solutions of inhibitors. The brain uptake of SQV was not affected by up to 4% of DMSO (data not shown).
Time Course of SQV Uptake. The brain uptake of [14C]SQV in mice was studied using the in situ mouse brain perfusion technique to choose a reliable perfusion time that allows for sufficient drug accumulation in the brain tissue (Fig. 2). Based on the results of the validation studies, a perfusion flow rate of 2.1 ml/min was selected for these experiments. The relationship between SQV brain uptake and perfusion time for up to 120 s was linear, whereas no BBB disruption was observed (Vvasc of [3H]inulin). The radioactivity was about 4000 dpm/tissue sample at 60 s of perfusion time when the perfusate contained 0.3 μCi/ml of [14C]SQV. Since sufficient accumulation of [14C]SQV was obtained after 60 s of perfusion, this perfusion time was chosen for all subsequent studies.
Concentration Dependence of SQV Uptake. The mice brains were perfused with various concentrations of SQV (1-36 μM) to determine whether the brain uptake of SQV was concentration-dependent. Figure 3 shows the rate of SQV uptake in the left hemisphere as a function of SQV concentration in the perfusate. As the concentration of SQV increased, the brain uptake rate of SQV increased steeply. The unidirectional transport coefficient (Kin) was significantly changed from 4.3 ± 0.2 ml·min-1·100 g-1 at 13 μM to 9.7 ± 0.8 ml·min-1·100 g-1 at 36 μM of SQV (Kin value; mean ± S.E.M., n = 4). The best-fit line (R2 = 0.97) was obtained using a modified Michaelis-Menten equation as described under “Calculations and Data Analysis”.
Effect of GF120918 on the Brain Uptake of SQV. HIV protease inhibitors such as SQV have shown very limited oral absorption and brain entry and the major efflux transporter present in the BBB has been reported to be P-gp (Kim et al., 1998). Therefore, this experiment was performed to affirm the role of P-gp on the brain permeation of SQV using the perfused brain model. To determine the effect of GF120918 on the brain uptake of SQV, mice brain perfusion studies of SQV were performed in the absence and presence of various concentrations of GF120918. The Vbrain value for SQV increased significantly as the concentration of GF120918 in the perfusate increased from 0 to 20 μM. The Vbrain values eventually plateaued at GF120918 concentrations greater than 10 μM (⋄, Fig. 4). Inclusion of GF120918 (10 μM) in the perfusates inhibited P-gp activity in the BBB and significantly increased the Vbrain values of SQV by more than 7-fold. When the Vbrain values of SQV in various concentrations of GF120919 were fitted to a modified Michaelis-Menten equation, an IC50 value of 2.69 ± 0.82 μM was obtained. The brain vascular volume (Vvasc of [3H]inulin, Fig. 4) during perfusion remained constant (1.24 ± 0.02 ml·100 g-1; mean ± S.E.M., n = 25). This result demonstrated that the presence of GF120918 did not alter the integrity of the tight junctions.
Effect of MK571 on the Brain Uptake of SQV. Mrps and P-gp have many substrates in common, including the HIV protease inhibitors. Previous reports from our laboratory and others suggest that Mrp2 efficiently transported SQV in overexpressed Mrp2 in vitro cell culture models (Huisman et al., 2002; Williams et al., 2002). We evaluated the brain uptake characteristics of SQV in the presence of various concentrations of MK571, a specific Mrp transporter family inhibitor using the brain perfusion technique to investigate the in vivo contribution of Mrp on brain uptake of SQV. Inclusion of MK571 (100 μM) in the perfusates increased the Vbrain values of SQV by more than 4.4- fold compared with SQV alone (⋄, Fig. 5). When the Vbrain values of SQV in the absence of presence of MK571 were fitted to a modified Michaelis-Menten equation, an IC50 value of 7.23 ± 4.58 μM was obtained. Data are mean ± S.E.M. (n = 3-4). As shown in the graph (Vvasc of [3H]inulin = 1.25 ± 0.03 ml·100 g-1; mean ± S.E.M., n = 20), addition of MK571 in the perfusates did not alter the integrity of the tight junctions.
Discussion
The mouse or rat brain perfusion technique is a very useful tool for studying the brain distribution of xenobiotics (e.g., drugs), since unlike in vitro models, drug transport across the BBB is measured in the intact brain with its full complement of enzymes, proteins, and transport systems. Along with the “leakiness” of cultured brain endothelial cells, any difference or change in expression of xenobiotic transporters compared with in vivo conditions [e.g., down-regulation of some nutrient transporters (Pardridge, 1995) or up-regulation of Mrp (Gutmann et al., 1999b)] make in vitro to in vivo extrapolation difficult. Compared with typical in vivo biodistribution studies that assess the amounts of brain-“associated” drug (i.e., adsorbed to tissues, in the brain vasculature), in situ perfusion methods measure actual drug uptake. The in situ brain perfusion technique has been used to investigate the mechanisms of transport of various drugs across the BBB (Smith, 1996). We validated the in situ perfused mice brain model before performing the studies to elucidate the mechanisms of brain uptake of SQV. A perfusion rate of 2.1 ml/min, at which a similar Kin value for diazepam was obtained in a previous report (Dagenais et al., 2000), was chosen since the physical integrity of the BBB was also maintained. [14C]Diazepam permeates the BBB by passive diffusion via the transcelluar route (Takasato et al., 1984), and the results of kinetic studies (Fig. 1B) confirmed linear diazepam brain uptake. Monitoring the brain vascular volume using vascular markers such as inulin assesses the integrity of the BBB during brain perfusions. The brain vascular volume of [3H] inulin did not change significantly during perfusion for 20 to 90 s and was comparable with the values [1.1 ml·100 g-1 in the mouse (Murakami et al., 2000) and 6-9 ml·100 g-1 in the rat (Takasato et al., 1984)] that were reported previously. Furthermore, in all experiments, the brain distribution volume of [3H]inulin remained low (1.3 ± 2 ml·100 g-1) in the perfusates containing SQV in the absence and presence of GF120918 or MK571, indicating that the integrity of the BBB was maintained during the perfusion.
The brain uptake of SQV was linear and unidirectional without any measurable disruption of the BBB for at least 120 s during perfusion. Since linearity was maintained, the possibility of underestimating brain SQV uptake was minimized. Potentially confounding systemic disposition effects (e.g., metabolism and protein binding) are avoided by directly perfusing the drug solution into the brain hemisphere. This was especially important for SQV since it shows high plasma protein binding (>98%; Kim et al., 1998) and extensive first pass metabolism. Therefore, the extent of brain uptake can be evaluated by directly assessing apparent brain distribution volume. The apparent brain distribution volume of SQV is low (4.3 ml·100 g-1 at 60 s), despite its favorable physicochemical properties (e.g., high lipophilicity with log P = 4.4 and molecular weight = 670.9). Under the same experimental conditions used in this study, the maximal possible value for Kin was estimated to be 250 ml·100 g-1·min-1.
Most of the published SQV brain uptake studies use traditional pharmacokinetic techniques in mice with a specific knocked out gene (e.g., mdr1). Brain concentrations were typically assessed by examining the total amount of SQV associated with brain tissue and then compared with a wild-type mice control. There are many shortcomings to this type of study design. For example, brain uptake can be overestimated since all tissue-associated SQV (i.e., bound to tissue, in the vascular space) is considered “absorbed”. In cases where assay sensitivity is not an issue, traditional pharmacokinetic studies can detect substantive differences in brain uptake but only in a semiquantitative manner. However, the issues of compensatory changes in the cytochromes P450 and the expression of other transporters in knockout animal models can obscure the interpretation of results (Schuetz et al., 2000; Cisternino et al., 2004) even further. In this report, the SQV brain transport mechanisms were elucidated in a quantitative manner, allowing for the observation of a novel transport mechanism. As shown in Fig. 3, transport rate of SQV in the left hemisphere did not increase proportionally to the concentration of SQV in the perfusate. The unidirectional transport coefficient increased by more than 2-fold at the highest concentration of SQV (36 μM) compared with no significant increase in the lower concentration range (1-13 μM). This greater than proportional increase in SQV brain uptake at higher concentrations is consistent with the saturation of a secretory transporter(s) or a metabolic enzyme since saturation leads to a greater rate of uptake. A similar relationship between relative oral absorption and dose was established after the administration of single doses of SQV, ritonavir, indinavir, nelfinavir, and amprenavir in humans. The relationship showed greater than dose proportional increases in oral absorption with SQV showing the highest nonlinearity among the HIV PIs (Williams and Sinko, 1999). This is in contrast to saturation of an absorptive transporter system where the rate of uptake is lower at higher concentrations (Sinko and Amidon, 1988).
Of the potential mechanisms responsible for this behavior, P-gp has been implicated by our group and others from in vitro studies (Polli et al., 1999; Williams et al., 2003) and from semiquantitative in vivo studies (Sparreboom et al., 1997). However, the importance of other efflux transporters on SQV brain uptake has not been conclusively demonstrated in vivo. Recent studies have demonstrated that the brain penetration of the HIV PIs such as indinavir, nelfinavir, and SQV is markedly affected by mdr1a P-gp since they are excellent substrates for this transporter (Kim et al., 1998; Lee et al., 1998). In addition, it was observed in rats that the distribution ratio of SQV in plasma/brain/cerebrospinal fluid was approximately 100:10:0.2 (Washington et al., 2000). Recently, BCRP (ABCG2), a member of the ABC family of drug transporters, was found to be highly expressed in the placenta and the luminal surface of the microvessel endothelium of the BBB where it may play a role in limiting the penetration of drugs. However, little is known about the function of BCRP in the BBB in vivo (Cisternino et al., 2004). For the purpose of this study, however, the role of BCRP in SQV brain uptake was considered minimal since Gupta et al. (2004) recently reported that HIV PIs, including SQV, were not substrates for BCRP.
Although the role of P-gp for limiting the brain entry of xenobiotics is well established (Schinkel, 1999), the role of the Mrps is much less well understood. The current results show that Mrps represent a significant permeation barrier for SQV across the BBB in vivo. Unlike P-gp, the role of the Mrps in BBB function should be considered with respect to differences in cellular localization of the various family members. Mrp2 is located on the apical cell membrane, whereas other Mrps, including Mrp1, Mrp3, and Mrp5, are located at the basolateral side (Borst et al., 2000). Although MK571is not specific for Mrp2, (i.e., it also inhibits Mrp1), it is reasonable to use as an inhibitor of Mrp2 in drug transport studies, especially in tissues such as brain capillary endothelial cells since Mrp1 is not predominantly expressed (Regina et al., 1998). In addition, SQV is known to be a substrate for Mrp2 in vitro in cell culture models overexpressing the transporter (Gutmann et al., 1999a; Huisman et al., 2002; Williams et al., 2002). Therefore, increased brain uptake of SQV in the presence of MK571 is likely due to inhibition of Mrp2.
In summary, results obtained in this study suggest that the in situ mouse brain perfusion model is a suitable technique in assessing the interaction of SQV with efflux transporters such as P-gp and Mrps in vivo. The finding that SQV is subject to transport by P-gp and Mrps in the brain suggests that P-gp and Mrps have an overlapping substrate spectrum, which is known from substrate recognition studies on P-gp and Mrps (Lee et al., 2001). Furthermore, our results suggest that increasing the low brain uptake of SQV may require parallel strategies for overcoming the effect of both P-gp and Mrp2.
Footnotes
-
doi:10.1124/jpet.104.076216.
-
ABBREVIATIONS: HIV, human immunodeficiency virus; PI, protease inhibitor; SQV, saquinavir; P-gp, P-glycoprotein; Mrp, multidrug resistance-associated protein; CNS, central nervous system; BBB, blood-brain barrier; MDR, multidrug resistance; ABC, ATP-binding cassette; BCRP, breast cancer resistance protein; GF120918, N-(4-[2-(1,2,3,4-tetrahydro-6,7-dimethoxy-2-isoquinolinyl)-ethyl]-phenyl)-9,10-dihydro-5-methoxy-9-oxo-4-acridine carboxamide; MK571, (E)-3-[[[3-[2-(7-chloro-2-quinolinyl)ethenyl]phenyl]-[[3-dimethylamino-3-oxopropyl]thio]methyl]thio]-propanoic acid; DMSO, dimethyl sulfoxide.
- Received August 13, 2004.
- Accepted November 2, 2004.
- The American Society for Pharmacology and Experimental Therapeutics