New drug discoveries and new approaches towards diagnosis and treatment have improved cancer therapeutics remarkably. One of the most influential and effective discoveries in the field of cancer therapeutics was antimetabolites, such as the antifolates. The interest in antifolates increased as some of the antifolates showed responses in cancers, such as mesothelioma, leukemia, and breast cancers. When pemetrexed (PTX) was discovered, our laboratory had established that the primary mechanism of action of pemetrexed is to inhibit thymidylate 22 synthase (TS) (E. Taylor et al., 1992). Preclinical studies have shown that PTX has a broad range of antitumor activity in human and murine models of cancer (Adjei, 2000; Adjei, 2004; S. Chattopadhyay, Moran, & Goldman, 2007; Miller et al., 2000). Accordingly, in February 2004, the FDA issued first-line treatment approval for pemetrexed in malignant pleural mesothelioma and in 2008 for first line treatment for locally advanced or metastatic NSCLC (reviewed in (Rollins & Lindley, 2005). As an antifolate this level of therapeutic activity of PTX against lung cancers was surprising and atypical (Hazarika, White, Johnson, & Pazdur, 2004). This led us to the question whether the effects of pemetrexed on other folate-dependent targets could explain the clinical activity of the drug. Our lab showed that, in addition to inhibiting thymidylate synthase, PTX also inhibits aminoimidazolecarboxamide ribonucleotide formyltransferase (AICART), the second folate-dependent enzyme of de novo purine synthesis. Inhibition of AICART leads to massive accumulation of its substrate 5-amino-4-imidazolecarboxamide ribonucleotide (ZMP), causing activation of AMP-dependent kinase (AMPK), which ultimately leads to suppression of mTORC1 signaling, a central regulator of cell growth and proliferation. This secondary mechanism could explain the unusual activity of PTX against mesothelioma and lung cancers. The large proportion of lung cancers are either null or mutant for p53 function. Therefore, this thesis focused on defining what the role of p53 is in the PTX-mediated AMPK activation and mTORC1 inhibition and how the loss of p53 affects mTORC1 signaling. These two questions proved to be interlinked. Chapter 2 investigates this relationship in detail. We found that, upon loss of p53, mTORC1 signaling is enhanced to a significant degree in colon carcinoma and lung cancer cell lines. Clearly, this observation required explanation. We found that the major factors responsible for these differences in mTORC1 activity upon loss of p53 23 were lower levels of two p53 target genes Tuberin (TSC2) and sestrin2. Immunoprecipitation studies of mTORC1 complexes from p53 wt and p53 null cells revealed quite interesting differences in the components of the mTORC1 complex. Immunoprecipitates from p53 null cells had higher levels of mTOR and lower levels of TSC2 and PRAS40 bound to raptor. This suggested that, in comparision to p53 competent cells, p53 null cells have more mTORC1 complex with enhanced activity due to decreased interaction of TSC2 and PRAS40, both of which are inhibitors of mTORC1. These observations explained the higher mTORC1 in p53 null cells and laid the foundation for determining the role of p53 in PTX-activated AMPK and mTORC1 inhibition. In the experiments described in Chapter 3, we found that PTX-mediated AMPK activation inhibited mTORC1 regardless of the p53 status in colon carcinoma cells. This suggested that mTORC1 inhibition by PTX was either independent of p53 mediated negative regulation of mTORC1 or was somewhere bypassing it. Therefore, we compared the effects of PTX with the classic AMPK activator aminoimidazolecarboxamide ribonucleoside (AICAR). In spite of a common mechanism of AMPK activation, namely, expansion of cellular ZMP levels, signaling from AMPK activated by PTX or AICAR were quite different. PTX-activated AMPK phosphorylated the mTORC1 component Raptor but not tuberin (TSC2), whereas AICARactivated AMPK phosphorylated both the targets. This differential behavior of two AMPK activators was due to differential behavior of p53 under these two treatments. Both, AICAR and PTX treatment led to increase in p53 levels but the p53 that accumulated after AICAR treatment was transcriptionally active while the p53 that accumulated after PTX treatment was not. Transcription of p53 targets, including TSC2 and sestrin2, was activated in AICAR- but not in PTX-treated cells. In the absence of p53 function, TSC2 was deficient and mTORC1 activity 24 enhanced, but Raptor phosphorylation by AMPK following PTX was robust and independent of both p53 and TSC2. Therefore we concluded that p53 deficiency suppresses TSC2 and upregulates mTORC1, but AMPK-phosphorylation of Raptor after pemetrexed treatment was sufficient to suppress mTORC1, even in TSC2 deficiency. This suggested pemetrexed as a drug for treatment of Tuberous Sclerosis, a genetic disease caused by functional inactivity of TSC1 or TSC2 due to point mutations in these genes. Mutation of p53 is one of the most common genetic alterations in human cancers and tumors. Cancers that express mutant p53 tend to be more aggressive, resistant to chemotherapy and show worse prognosis then p53-null tumors (Elledge et al., 1993; Olivier et al., 2006). This tumor-promoting activity of mutant p53 has been correlated with acquired and novel transcriptional activities of mutant p53. It has been shown that mutp53 can activate the transcription of cell growth promoting genes, such as, NFκB2, PCNA, MDR1, Axl, EGFR, hTERT, and HSP70, which are not usually transcriptional targets of wt p53. Interestingly, we found that whereas DNA damaging drugs enhance the acquired oncogenic transcriptional activities of mutp53, PTX interferes with this transcription activation. We also found in Chapter 4 that PTX can limit or block the DNA damaging drug-mediated increment of transcriptional activation of mutp53. This suggests that blockade of transcriptional activation of mutp53 by pemetrexed may provide an additional therapeutic benefit in mutp53 bearing cancers. As discussed in Chapter Three, although pemetrexed (with TdR) increases the levels of p53 and its binding to the promoter of its target gene, p21, this p53 is transcriptionally inactive. In order to understand the mechanism of the pemetrexed-mediated transcriptional defect of wt p53, we studied the PTX-mediated signaling towards ATM and ATR and their effects on their substrates Chk2 and Chk1, respectively. These studies suggested that the difference between 25 signaling under AICAR treatment and PTX treatment was that, unlike PTX, AICAR treatment was leading to DNA damage, followed by Chk2 phosphorylation at Thr68. We found there were three major differences between AICAR and pemetrexed (+ TdR) mediated signaling: AICAR caused DNA damage, followed by ATM mediated phosphorylation of Chk2 at Thr68 and phosphorylation of p53 at Ser15 all of which lead to activation of p53 transcriptional activity, events which do not take place under PTX treatment. Studies aimed at understanding the effects of PTX on wt and mutp53 transcriptional activities are discussed in detail in Chapters Three and Four of this dissertation. Overall, we concluded that PTX interferes with the transcription activity of wild type as well as gain-of-function mutant p53. The blockade of DNA damaging agent-mediated enhancement of mutp53 transcription activity by PTX, suggests the clinical relevance of PTX in carcinomas with mutp53. We suggest that this could be one of the contributing factors in the effects of PTX against human lung cancers.

La dihydrofolate réductase humaine (DHFRh) est une enzyme essentielle à la prolifération cellulaire, ce qui en fait une cible de choix pour le traitement de différents cancers. À cet effet, plusieurs inhibiteurs spécifiques de la DHFRh, les antifolates, ont été mis au point : le méthotrexate (MTX) et le pemetrexed (PMTX) en sont de bons exemples. Malgré l’efficacité clinique certaine de ces antifolates, le développement de nouveaux traitements s’avère nécessaire afin de réduire les effets secondaires liés à leur utilisation. Enfin, dans l’optique d’orienter la synthèse de nouveaux composés inhibiteurs des DHFRh, une meilleure connaissance des interactions entre les antifolates et leur enzyme cible est primordiale. À l’aide de l’évolution dirigée, il a été possible d’identifier des mutants de la DHFRh pour lesquels l’affinité envers des antifolates cliniquement actifs se voyait modifiée. La mutagenèse dite ¬¬de saturation a été utilisée afin de générer des banques de mutants présentant une diversité génétique au niveau des résidus du site actif de l’enzyme d’intérêt. De plus, une nouvelle méthode de criblage a été mise au point, laquelle s’est avérée efficace pour départager les mutations ayant entrainé une résistance aux antifolates et/ou un maintient de l’activité enzymatique envers son substrat natif, soient les phénotypes d’activité. La méthode de criblage consiste dans un premier temps en une sélection bactérienne à haut débit, puis dans un second temps en un criblage sur plaques permettant d’identifier les meilleurs candidats. Plusieurs mutants actifs de la DHFRh, résistants aux antifolates, ont ainsi pu être identifiés et caractérisés lors d’études de cinétique enzymatique (kcat et IC50). Sur la base de ces résultats cinétiques, de la modélisation moléculaire et des données structurales de la littérature, une étude structure-activité a été effectuée. En regardant quelles mutations ont les effets les plus significatif sur la liaison, nous avons commencé à construire un carte moléculaire des contacts impliqués dans la liaison des ligands. Enfin, des connaissances supplémentaires sur les propriétés spécifiques de liaison ont put être acquises en variant l’inhibiteur testé, permettant ainsi une meilleure compréhension du phénomène de discrimination du ligand. / Human dihydrofolate reductase (hDHFR) is an essential enzyme for cellular proliferation and it has long been the target of antifolate drugs for the treatment of various types of cancer. Despite the clinical effectiveness of current antifolate treatments, new drugs are required to reduce the side-effects associated with their use. An essential requirement for design of new antifolates is a better understanding of how these drugs interact with their targets. We applied directed evolution to identify mutant hDHFR variants with modified binding to some clinically relevant antifolates. A saturation mutagenesis approach was used to create genetic diversity at active-site residues of hDHFR and a new, efficient screening strategy was developed to identify the amino acids that preserved native activity and/or conferred antifolate resistance. The screening method consists in a high-throughput first-tier bacterial selection coupled with a second-tier in vitro assay that allows for rapid detection of the best variants among the leads, according to user-defined parameters. Many active, antifolate-resistant mutants of hDHFR were identified. Moreover, the approach has proven efficient in rapidly assessing kinetic (kcat) and inhibition parameters of the hDHFR variants (IC50). Structure-function relationship analysis based on kinetic investigation, available structural and functional data as well as modeling were performed. By monitoring which mutations have the greatest effect on binding, we have begun to build a molecular picture of the contacts involved in drug binding. By varying the drugs we test against, we gain a better understanding of the specific binding properties that determine ligand discrimination.

La dihydrofolate réductase humaine (DHFRh) est une enzyme essentielle à la prolifération cellulaire, ce qui en fait une cible de choix pour le traitement de différents cancers. À cet effet, plusieurs inhibiteurs spécifiques de la DHFRh, les antifolates, ont été mis au point : le méthotrexate (MTX) et le pemetrexed (PMTX) en sont de bons exemples. Malgré l’efficacité clinique certaine de ces antifolates, le développement de nouveaux traitements s’avère nécessaire afin de réduire les effets secondaires liés à leur utilisation. Enfin, dans l’optique d’orienter la synthèse de nouveaux composés inhibiteurs des DHFRh, une meilleure connaissance des interactions entre les antifolates et leur enzyme cible est primordiale. À l’aide de l’évolution dirigée, il a été possible d’identifier des mutants de la DHFRh pour lesquels l’affinité envers des antifolates cliniquement actifs se voyait modifiée. La mutagenèse dite ¬¬de saturation a été utilisée afin de générer des banques de mutants présentant une diversité génétique au niveau des résidus du site actif de l’enzyme d’intérêt. De plus, une nouvelle méthode de criblage a été mise au point, laquelle s’est avérée efficace pour départager les mutations ayant entrainé une résistance aux antifolates et/ou un maintient de l’activité enzymatique envers son substrat natif, soient les phénotypes d’activité. La méthode de criblage consiste dans un premier temps en une sélection bactérienne à haut débit, puis dans un second temps en un criblage sur plaques permettant d’identifier les meilleurs candidats. Plusieurs mutants actifs de la DHFRh, résistants aux antifolates, ont ainsi pu être identifiés et caractérisés lors d’études de cinétique enzymatique (kcat et IC50). Sur la base de ces résultats cinétiques, de la modélisation moléculaire et des données structurales de la littérature, une étude structure-activité a été effectuée. En regardant quelles mutations ont les effets les plus significatif sur la liaison, nous avons commencé à construire un carte moléculaire des contacts impliqués dans la liaison des ligands. Enfin, des connaissances supplémentaires sur les propriétés spécifiques de liaison ont put être acquises en variant l’inhibiteur testé, permettant ainsi une meilleure compréhension du phénomène de discrimination du ligand. / Human dihydrofolate reductase (hDHFR) is an essential enzyme for cellular proliferation and it has long been the target of antifolate drugs for the treatment of various types of cancer. Despite the clinical effectiveness of current antifolate treatments, new drugs are required to reduce the side-effects associated with their use. An essential requirement for design of new antifolates is a better understanding of how these drugs interact with their targets. We applied directed evolution to identify mutant hDHFR variants with modified binding to some clinically relevant antifolates. A saturation mutagenesis approach was used to create genetic diversity at active-site residues of hDHFR and a new, efficient screening strategy was developed to identify the amino acids that preserved native activity and/or conferred antifolate resistance. The screening method consists in a high-throughput first-tier bacterial selection coupled with a second-tier in vitro assay that allows for rapid detection of the best variants among the leads, according to user-defined parameters. Many active, antifolate-resistant mutants of hDHFR were identified. Moreover, the approach has proven efficient in rapidly assessing kinetic (kcat) and inhibition parameters of the hDHFR variants (IC50). Structure-function relationship analysis based on kinetic investigation, available structural and functional data as well as modeling were performed. By monitoring which mutations have the greatest effect on binding, we have begun to build a molecular picture of the contacts involved in drug binding. By varying the drugs we test against, we gain a better understanding of the specific binding properties that determine ligand discrimination.

Acquired cellular resistance to traditional chemotherapeutics is a common obstacle in the treatment of most cancer cell types. This resistance occurs as a result of changes in the underlying molecular mechanisms of disease progression. The development of novel chemotherapeutic approaches designed to enhance the efficacy of protypical anti-cancer drugs is important in order to overcome this issue. Such approaches will aid in understanding the biomolecular phenomena responsible for drug resistance and disease progression. Combining signaling pathway inhibitors has become an effective strategy for enhancing tumor cell death by targeting multiple pathways known to regulate cell survival. Pemetrexed, an FDA-approved anti-folate drug, targets thymidylate synthase (TS) and a secondary folate-dependent enzyme, 5’ aminoimidazole-carboximide ribonucleotide formyltransferase (AICART); both important for DNA synthesis. Studies performed by our collaborator demonstrated that TS inhibition causes intracellular accumulation of ZMP+ and activation of AMPK which is known to induce autophagy in mammalian cells. Previous studies from our lab and others showed that sorafenib, a multi-kinase inhibitor of Raf-1 and class III receptor tyrosine kinases, was able to induce a cytotoxic form of autophagy in a variety of tumor cell types. Combination treatment using pemetrexed and sorafenib in these cancer cells resulted in an enhancement of autophagy and cell lethality beyond that of individual drugs alone. Inhibition of autophagy suppressed the toxic interactions of these drugs in all cell types examined. Pemetrexed/sorafenib cotherapy also proved to be an effective treatment for triple negative breast cancer cells having advanced to a stage of estrogen independence. Fulvestrant-resistant MCF7 cells were more sensitive to the drug combination than parental, estrogen-dependent MCF7 cells. Breast cancer cells cotreated with pemetrexed and sorafenib exhibited enhanced MEK/ERK signaling, Src activation that was dependent on platelet-derived growth factor β (PDGFRβ) downregulation, elevated protein phosphatase 2A (PP2A) activity, and increased de novo ceramide synthesis. Studies using a mouse model of experimentally-induced breast cancer validated drug combination effectiveness through inhibition of tumor growth, while no deleterious effects on normal tissues were observed. The data presented demonstrates that pemetrexed/sorafenib cotreatment augments chemosensitivity in both in vitro and in vivo systems. Based upon these findings, a Phase I clinical trial involving pemetrexed and sorafenib in breast cancer patients with solid, recurrent tumors was begun in 2011. In conclusion, this work strongly supports a promising therapeutic utility for the pemetrexed/sorafenib combination in treatment of various cancer cell types.