Activation of the transcription factor Nrf2 regulates expression of phase II enzymes and other adaptive responses to electrophile and oxidant stress. Nrf2 concentrations are regulated by the thiol-rich sensor protein Keap1, which is an adaptor protein for Cul3-dependent ubiquitination and degradation of Nrf2. However, the links between site-specificity of Keap1 modification by electrophiles and mechanisms of Nrf2 activation are poorly understood. We studied the actions of the prototypical Nrf2 inducer tert-butylhydroquinone (tBHQ) and two biotin-tagged, thiol-reactive electrophiles N-iodoacetyl-N-biotinylhexylenediamine (IAB) and 1-biotinamido-4-(4'-[maleimidoethyl-cyclohexane]-carboxamido)butane (BMCC). Both IAB and tBHQ induce expression of ARE-directed GFP expression in ARE/TK-GFP HepG2 cells and both initiatednuclear Nrf2 accumulation and induction of heme oxygenase 1 in HEK293 cells. In contrast, BMCC produced none of these effects. Liquid chromatography tandem mass spectrometry (LC-MS-MS) analysis of human Keap1 modified by IAB or BMCC in vitro indicated that IAB adduction occurred primarily in the central linker domain, whereas BMCC modified other Keap1 domains. Treatment of FLAG-Keap1 transfected HEK293 with the Nrf2-activating compounds IAB and tBHQ generated high molecular weight Keap1 forms, which were identified as K-48-linked polyubiquitin-conjugates by immunoblotting and LC-MS-MS. Keap1 polyubiquitination coincided with Nrf2 stabilization and nuclear accumulation. In contrast, BMCC did not induce Keap1 polyubiquitination. Our results suggest that Nrf2 activation is regulated through the polyubiquitination of Keap1, which in turn is triggered by specific patterns of electrophile modification of the Keap1 central linker domain. These results suggest that Keap1 adduction triggers a switching of Cul3-dependent ubiquitination from Nrf2 to Keap1, leading to Nrf2 activation.The chemopreventive agent sulforaphane is an isothiocyanate, which was isolated from broccoli. Sulforaphane was demonstrated to induce ARE-regulated genes by stimulating the Keap1-Nrf2 system. This agent is a powerful electrophile that can react with thiols to form thionoacyl adducts. A specific sulforaphane modification pattern on Keap1 may trigger the activation of Nrf2. However, thionoacyl adducts are labile to hydrolysis and transacylation reactions, which prevent the identification of the sulforaphane modification patten on Keap1. In this study, we have developed a LC-MS-MS method to map sulforaphane modification sites formed on Keap1 in vitro. Our studies indicate that sulforaphane displays a different pattern of Keap1 modification than ARE/ERE inducers that modify Keap1 by alkylation. Moreover, the modification of Keap1 in vivo by sulforaphane does not trigger the ubiquitination of Keap1, which suggests a novel mechanism for Nrf2 stabilization by sulforaphane thionoacyl adduct formation.

Pancreatic cancer is the fourth leading cause of cancer-related mortality in the United Kingdom (UK) and the United States (US). Most of the cases are diagnosed in the metastatic stage (MPC) and the disease is associated with a significant economic and quality of life burden. Chemotherapy with gemcitabine (GEM) is the standard of care. Combination regimens such as cisplatin plus GEM (CIS+GEM), capecitabine plus GEM (CAP+GEM), nab-paclitaxel plus gemcitabine (NAB-P+GEM), FOLFIRINOX (FFX), and oxaliplatin plus GEM (OX+GEM) showed an improvement in the survival outcome when compared to GEM alone. The purpose of the pharmacoeconomic research program included in this dissertation-by-articles was to assess the benefits of these regimens as first-line treatment relative to their costs in the management of metastatic pancreatic cancer for the UK and the US-doing so by building upon a critical review of pharmaco-economic studies of treatments for MPC and using a newly developed algorithm for parametric model selection. The four objectives of this research program were: (1) to identify and critique previous economic evaluation studies that have been done in the pancreatic cancer setting; (2) to propose a transparent algorithm to justify parametric model selection in economic evaluation studies; (3) to conduct an economic evaluation of systemic chemotherapies as first-line treatment of MPC from the payer perspective for the UK; and (4) to conduct an economic evaluation of systemic chemotherapies as first-line treatment of MPC from the payer perspective for the US. This dissertation begins with an orientation to MPC, treatment options, innovation, and dissertation outline. From the critical review (Chapter 2) we concluded that no comprehensive economic evaluations for all treatment options for the UK and the US have been published; overall survival was the key driver for most of the economic evaluations; economic analyses were country-specific, not generalizable across countries, inconsistent in their use of model inputs and utility estimates; and their justification for selecting a specific parametric model. The latter led us to the development of a quantitative algorithm (Chapter 3) for objectively selecting parametric models to extrapolate the data beyond the time horizon reported in the clinical trials; showing also that the application of different parametric models impacted the economic evaluation estimates. Considering the likelihood of between-country differences, we then performed pharmacoeconomic evaluations of GEM alone, CAP+GEM, OX+GEM, CIS+GEM, NAB-P+GEM and FFX for the UK (Chapter 4), with its single-payer system, and for the US (Chapter 5), with its multiple-payer system. The UK analyses revealed that compared to CAP+GEM, OX+GEM and CIS+GEM were more costly but less effective. Treatment with either FFX or CAP+GEM was more costly but more effective than with GEM alone. In contrast, the US analyses showed that OX+GEM, CAP+GEM, NAB-P+GEM and FFX were found to be costlier and more effective than GEM alone. Compared to GEM alone, NAB-P+GEM was found to have the smallest incremental cost-effectiveness and utility ratio among all other regimens. We synthesize and examine the implications of these findings in Chapter 6.

Centrally released oxytocin (OT) is believed to attenuate stress-induced activation of the hypothalamic pituitary adrenal (HPA) axis as well as being anxiolytic. Therefore, it is expected that OT deficient (OT-/-) mice that do not synthesize or release OT centrally or peripherally will display enhanced HPA axis activation, as well as increased anxiety-related behavior compared to wildtype (OT+/+) mice. To test this hypothesis, OT-/- mice were exposed to shaker stress, (psychogenic stressor), cholecystokinin- (CCK) administration (physical stressor), or the elevated plus maze (EPM), a behavioral test of anxiety. Female OT-/- mice released more corticosterone than OT+/+ mice in response to shaker stress. Shaker stress exposure activated Fos in OT neurons of the paraventricular nucleus of the hypothalamus (PVN) of male and female OT+/+ mice and corticotropin-releasing hormone (CRH) within the PVN of male and female mice of both genotypes. In addition, shaker stress exposure revealed that Fos expression in the medial nucleus of the amygdala (MeA) was lower in female OT-/- than OT+/+ mice. Genotypic differences in corticosterone release and Fos activation of the MeA in response to shaker stress exposure were not observed in male mice. Furthermore, similar genotypic (and/or sex) differences were not revealed in response to CCK-administration. OT is also anxiolytic in female mice. Female OT-/- mice tested in the EPM displayed increased anxiety-related behavior compared to OT+/+ mice. In response to EPM exposure Fos expression in the MeA was greater in female OT-/- mice than OT+/+ mice. Surprisingly, male OT-/- mice tested in the EPM displayed decreased anxiety-related behavior compared to OT+/+ mice, but did not display genotypic differences in the Fos expression within the MeA. The results of this thesis suggest that OT is anxiolytic and attenuates HPA activation in female, but not male mice. Furthermore, it appears that OT plays a modulatory role in the processing of psychogenic stressors, but may not be involved in the processing of physical or systemic stressors. More specifically, it is possible that OT plays a role in behavioral and physiological responses that depend upon neuronal processing within the MeA.

The Cytochrome P450 (CYP) superfamily of drug metabolizing enzymes is responsible for the metabolism of a variety of drugs and endogenous compounds. The CYP2C enzyme subfamily (CYP2C8, CYP2C9, and CYP2C19) mediates the metabolism of approximately twenty percent of these compounds. While CYP2C9 and CYP2C19 have been well characterized in vivo, relatively little is known about the in vivo contribution of CYP2C8. However, as the number of substrates and interest in polymorphic expression has grown, so too has the importance of CYP2C8. The in vivo relevance of CYP2C8 can be estimated with a drug predominately metabolized by this enzyme as a probe substrate. Thus, the overall purpose of this research was to investigate the utility of rosiglitazone as an in vivo probe of CYP2C8 activity. To accomplish this goal, we characterized the pharmacokinetics of rosiglitazone in the presence of the CYP2C8 inhibitor, trimethoprim, the CYP inducer, St. Johns wort, and in subjects genotyped for variant CYP2C8 alleles. Novel liquid chromatographic methods were developed for the determination of rosiglitazone and trimethoprim plasma concentrations with fluorescence and ultraviolet wavelength detection, respectively. CYP2C8 genotyping was accomplished with a newly developed method based on Pyrosequencing technology, which facilitates high-throughput analysis in a cost-effective manner. Trimethoprim was an effective inhibitor of rosiglitazone metabolism in vitro and it increased rosiglitazone concentrations in vivo by 31%. In addition, there was a strong relationship (r2=0.97, p=0.0021) between trimethoprim plasma concentration and fold inhibition in subjects who did not carry the CYP2C8*3 allele, suggesting genotype influences the extent of CYP2C8 inhibition. Administration of St. Johns wort increased rosiglitazone clearance by 35%, but CYP2C8 genotype did not affect the magnitude of induction. Finally, genotype did not affect basal rosiglitazone metabolism. Since changes have been observed with other CYP2C8 metabolized drugs, polymorphic effects of CYP2C8 may be substrate dependent. In conclusion, these results support the use of rosiglitazone as an in vivo probe of CYP2C8 activity, as it is affected by CYP2C8 inhibitors and inducers. The clinical benefits of CYP2C8 substrates may be influenced by these and other CYP2C8 modulators and therefore rosiglitazone could serve as a probe to detect these interactions.

Leaf Huang Ph.D. INTRAVESICAL THERAPY OF INTERSTITIAL CYSTITIS Pradeep Tyagi, Ph.D. University of Pittsburgh, April 4th, 2005 Interstitial cystitis (IC) is an inflammatory disorder of bladder which affects middle-aged Caucasian women. Intravesical administration of drugs is a mainstay in its treatment either as adjunct to oral therapy or as second-line therapy. The vehicles currently used for this route of administration are ideally suited for hydrophobic drugs and typically maintain drug exposure in the bladder for very short duration of time. The present dissertation project was aimed at investigating the use of alternative vehicles for improving intravesical drug delivery of hydrophobic small molecular weight drugs such as capsaicin misoprostol in addition to the delivery of large molecular weight peptide nucleic acid (PNA) for antisense based therapy. The hydrophobic drugs selected for this study were delivered by intravesical route using liposomes, thermosensitive hydrogel and TAT peptide. The efficiency of drug delivery was assessed by measuring the physiological response of normal and diseased rat bladder by metabolic cages and the method of cystometrogram (CMG). Histology and immunohistochemistry of bladder and spinal cord sections was done to corroborate the response measured in the physiological measurement. Liposomes were demonstrated to be a superior vehicle for capsaicin and thermosensitive hydrogel was able to sustain the exposure of a hydrophobic drug in the bladder for prolonged time and increase the efficacy of misoprostol in rat model of cystitis induced by cyclophosphamide. An interesting observation made during the study was that liposomes in absence of drug were able to modulate physiological response of bladder and this observation was further investigated to define the charge on the lipid headgroup and the structural requirements of hydrophobic backbone in the lipids for reducing bladder hyperactivity induced by sequential infusion of protamine sulfate and high concentration of KCl. Overexpression of NGF in cyclophosphamide induced cystitis was downregulated in the urothelium of rat bladder using antisense based therapy with PNA, which was delivered with the aid of TAT peptide. Overall, the study concluded that liposomes cannot only be a treatment option but can also be used for delivery of hydrophobic drugs. The potential of hydrogels and cell penetrating peptides for intravesical drug delivery needs further investigation.

Hepatic regeneration is essential for the success of living donor liver transplantation (LDLT) in which the residual liver in donor and the segment that is transplanted into the recipient will grow in size to accommodate the requirements of the donor and recipient. We hypothesized that the clearance of drugs will be drastically reduced during the immediate post operative period in donor and recipient due to reduced liver mass and the decreased activity of the drug metabolizing enzymes in the liver. The objectives of this dissertation are to evaluate 1) the mRNA expression and in vitro and in vivo activity (pharmacokinetics of tacrolimus) of CYP3A; 2) the effect of hepatotropic drugs such as cyclosporine and tacrolimus on the activity of CYP3A; 3) the expression, and in vitro and in vivo activity (pharmacokinetics of mycophenolic acid) of various uridine diphosphate glucuronosyltransferases; and 4) the in vivo metabolism of acetaminophen in rats, during hepatic regeneration after partial hepatectomy. Our results indicate that 1) the activity of most of the drug metabolizing enzymes was decreased but recovered completely at different rates; 2) the mRNA expression mirrored the changes in in vitro activity of these enzymes; 3) the clearance of tacrolimus and mycophenolic acid was decreased but recovered completely at different rates; 4) the magnitude of reduction in in vivo clearance of tacrolimus and mycophenolic acid was much less than what was predicted from the loss of liver mass and reduction in enzyme activity; 5) cyclosporine and tacrolimus did not have any significant effect on the recovery of activity of CYP3A, and 6) there was increased production of toxic metabolites of acetaminophen during regeneration. The clinical implications of our study are as follows: 1) Drug dosing in LDLT patients should be carefully monitored. A less than proportionate decrease in dose relative to reduction in liver mass may be necessary for different drugs. 2) The drug elimination capacity of the liver recovers completely with time and normal hepatic function will be restored in subjects undergoing hepatic resection. 3) Recovery of hepatic function will proceed normally in presence of hepatotropic substrates such as cyclosporine A and tacrolimus.

The increase in the use of herbal products, particularly in patients taking conventional medicine, has increased the likelihood of drug-herb interactions. Herbal products sold to the public are often not a single chemical compound, but rather a complex mixture of hundreds of different constituents. Human microsomal systems have been employed as a cost and time efficient approach to prospectively evaluate individual constituents for the potential for interactions with drug metabolizing enzymes. In fact, it has been shown that certain herbal constituents are capable of direct inhibition of drug metabolizing enzymes in this system. However, extrapolation of the drug interaction potential to cellular systems or whole organisms is often difficult because the preparation of microsomes necessitates the destruction of the integrity of the living cell and the physiologically relevant processes within. The primary goal of this dissertation research was to investigate the effect of herbal products on human hepatic drug metabolizing enzymes and transporters using primary cultures of human hepatocytes. Cultured hepatocytes were exposed to the various herbal constituents acutely, to evaluate the direct effect on enzyme activity, or chronically, to evaluate the indirect effect on enzyme expression and subsequent activity. Additionally, in order to assess to scalability of our in vitro UGT1A results to humans, healthy human subjects were administered acetaminophen, a general UGT1A probe, before and after a 7-day course of milk thistle. These data demonstrate that herbal constituents can directly inhibit enzyme activity but also influence activity by indirectly modulating gene expression. In the case of St. Johns wort, human hepatocytes showed that while constituents were capable of enzyme induction, inhibition also occurred. However, in vivo, it is the former that predominates over the latter. Furthermore, our predictions of interactions in vivo for St. Johns wort have been validated through a number of clinical studies. The case of milk thistle, however, proved more complex. While our in vitro data showed the possibility of drug interactions with several drug metabolizing enzymes, little effect was found in vivo. The latter demonstrates the value of consideration of the entire pharmacologic profile of an herb before conclusions about clinical relevance are made.

The camptothecins are DNA topoisomerase I-interactive anticancer agents and have a wide range of antitumor activity. Currently approved camptothecin analogues (i.e., topotecan and irinotecan) are only available for IV administration. 9-Nitrocamptothecin (9NC) is administered orally and is partially metabolized to an active metabolite, 9-aminocamptothecin (9AC). As with other camptothecin analogues, 9NC and 9AC undergo a reversible, pH-dependent reaction between the active-lactone and inactive-hydroxy acid forms. In vitro and in vivo preclinical studies suggest that protracted administration of low doses of camptothecin analogues produces better antitumor activity than the less frequent administration of higher doses. Oral administration of 9NC could mimic the protracted schedule and maximize patient convenience. However, the optimal oral dose and schedule of 9NC and other camptothecin analogues are currently unclear. In addition, oral administration of camptothecin analogues has been characterized by extensive inter- and intra-patient variability in bioavailability. The primary goal of this dissertation research was to evaluate the pharmacokinetics and pharmacogenetics of 9NC and its 9AC metabolite in preclinical models and in patients as part of phase I and II trails. Daily administration of 9NC orally for 5 days per week for two consecutive weeks repeated every four weeks is tolerable and may be an active regimen in patients with gastric or pancreatic cancers. The responses seen in xenografts models evaluating the same regimen of 9NC as evaluated in the phase I study occurred at systemic exposures that are tolerable in patients. There was significant inter- and intra-patient variability in the pharmacokinetics of 9NC and 9AC when 9NC was administered under fasting conditions. Most of the drug remained in the 9NC form with an overall ratio of 9NC to 9AC of 4:1. Co-administration of 9NC with food reduces the oral absorption of 9NC; however there was no difference in the exposure of 9AC. The functional consequences of known single nucleotide polymorphisms in genes of known ATP-binding cassette (ABC) transporters were evaluated as potential sources of the pharmacokinetic variability of 9NC and 9AC. Our findings suggest that the inter-individual variability in the disposition of 9AC, but not 9NC, may be influenced, in part, by ABCG2 genotype. The factors associated with the high inter- and intra-patient variability remain unclear.

PHARMACOLOGICAL AND STRUCTURE-ACTIVITY RELATIONSHIP EVALUATION OF MICROTUBULE-STABILIZING AGENTS Charitha Madiraju, M.S. University of Pittsburgh, 2005 Microtubules are composed of alpha- and beta-tubulin subunits and are highly dynamic elements of the cytoskeleton that help maintain fidelity during cell division. This aspect of microtubules makes them a very useful target for the treatment of tumors. The potent microtubule-stabilizing agent (+)-discodermolide is a sponge-derived polyketide natural product and is a polyhydroxylated and polymethylated, C24:4 fatty acid lactone carbamate. The microtubule-stabilizing action of discodermolide was predicted by computational chemistry. Biochemical, pharmacological, and antiproliferative activity studies with discodermolide showed it to have potent hypernucleating, microtubule-stabilizing and microtubule bundling properties. Synthetic analogues of discodermolide were made to establish a structure-activity relationship (SAR) for discodermolide. The library consisted of analogues with modifications made on the lactone and carbamate/diene portions, with altered stereochemistry on C-11 and C-17, conformational analogues based on the solution structure of discodermolide, and macrocyclic analogues of discodermolide. Biological evaluation of the analogues suggested the diene, the carbamate on C-19 and the stereochemistry on C-11 and C-17 to be important for the activity of the parent molecule, and that the lactone moiety is amenable to modifications. Of interest were the conformational and macrocyclic analogues that showed promising microtubule-targeting activities. Furthermore, these studies paved way for the discovery of the related natural product dictyostatin as a potent microtubule stabilizer. Dictyostatin is a macrocycle with stereochemistry very much like discodermolide. Intensive biochemical and pharmacological studies showed dictyostatin to have comparable, or superior in some respects, properties to those of discodermolide and paclitaxel. Biological analyses of structural analogues of dictyostatin suggested that the macrolactone and the bottom half of dictyostatin have a critical role in the biological activity of the parent molecule. The top portion of the molecule was amenable for modifications without much loss of activity; however, this half of the molecule seemed crucial for the microtubule binding property of the parent molecule. These studies identified dictyostatin to be the most potent of the microtubule stabilizers studied. One interesting analogue, C16-normethyldictyostatin, was identified to have promising activity comparable to the parent molecule.

Herbal products have been shown to cause serious interactions when combined with conventional medications. A majority of these interactions are pharmacokinetic in nature and involve alteration in the activity of drug metabolizing enzymes. Valerian is a popular herbal product often used to treat insomnia and anxiety. Valerian extracts contain essential oils with sesquiterpenes such as valerenic acid and its derivatives. However, the drug interaction potential of valerian preparations is largely unknown. In human liver microsomes, valerenic acid forms a glucuronide conjugate suggesting that valerian extracts could interact with drugs that undergo glucuronidation. As glucuronidation is catalyzed by UDP- glucuronosyltransferase enzymes (UGT), the goal of this dissertation was to investigate the effect of valerian extracts on UGT activity. Acetaminophen was used as a probe substrate to measure UGT activity in these studies. A bioassay-guided fractionation approach was adopted to identify the major compounds in valerian extracts that are responsible for inhibition of UGT activity. The alcoholic extract of valerian was fractionated by liquid-liquid extractions followed by chromatographic methods. The organic extracts showed significant inhibitory activity compared to the aqueous extracts. Using various chromatographic and spectroscopic techniques, the major compounds present in the active fraction were identified as valerenic acid, acetoxyvalerenic acid and valerenal. The clinical implications of the inhibition of UGT enzymes by valerian extracts were investigated in a study in healthy human volunteers. Valerian administration resulted in an increased acetaminophen maximum plasma concentration (Cmax) and a decrease in time to reach the maximum plasma concentration (tmax), but did not affect the area under the plasma concentration-time curve (AUC) or half life. As these results were unexpected, human hepatocyte cultures were used to determine if enzyme induction potential of some components may offset the inhibition of UGT enzymes. We hypothesized that the inhibition observed in the microsomal study could be masked by an increase in enzyme activity due to induction of enzymes by chronic exposure to the extracts. In human hepatocyte cultures, valerian extracts inhibited UGT activity on acute exposure while chronic exposure increased UGT activity and mRNA levels. Our study indicates that there is no clinically significant interaction between acetaminophen and valereian. In vitro studies in human hepatocytes may better predict in vivo herb-drug interactions than studies in microsomes.