A Thesis submitted to The University of Arizona College of Medicine - Phoenix in partial fulfillment of the requirements for the Degree of Doctor of Medicine. / The Warburg effect is defined as the propensity for cancer cells to favor glycolysis over oxidative phosphorylation under aerobic conditions. Finding a way to reverse this effect would likely be very beneficial for cancer therapy. The PI3K/Akt pathway has been suggested to be responsible for the Warburg effect, and estrogen is a known regulator of this pathway. Estrogen, specifically 17 β-estradiol, has been shown to be protective at the level of the mitochondria. The purpose of this study was to try to use 17 β-estradiol to reverse aspects of the Warburg effect in two cancer lines. Various concentrations of 17 β-estradiol were added to the samples (0, 10nm, 100nm, 1μm) for various amounts of time (16-96h). Western blots probes for select subunits of the electron transport chain (ETC) showed no differences in cells with and without 17 β-estradiol across various times. Due to technical difficulties with cell lines, considerable troubleshooting was required, consuming the time available for further analysis. The available results do not suggest that 17 β-estradiol alone is able to reverse the Warburg effect.

Cancer cells exhibit an abnormal metabolic phenotype characterized by glycolysis and lactate fermentation in the presence of oxygen, a phenomenon known as the Warburg effect. This dysregulated metabolism plays an important role in every aspect of cancer progression, from tumorigenesis to invasion and metastasis. The Warburg effect is a common phenotype shared by most, if not all, cancer types. It is especially prominent in metastatic tumors, which are notoriously resistant to treatment and responsible for the majority of cancer-related deaths. Thus, metabolic therapies which target the Warburg effect could offer novel therapeutic options for most cancer patients, including those with aggressive or late-stage cancers. The ketogenic diet is a high fat, low carbohydrate diet that induces a physiological state of nutritional ketosis - decreased blood glucose and elevated blood ketones. It has been investigated as a cancer therapy for its potential to exploit the Warburg effect by restricting glucose availability to glycolysis-dependent tumors, and has been reported to slow cancer progression in some animal models as well as in anecdotal reports and small clinical studies in humans. Interestingly, there is some evidence that the elevation in blood ketones induced by the ketogenic diet contributes to its anti-cancer effects, suggesting that ketone supplementation could possibly inhibit cancer progression on its own. Rapid growth outstrips a tumor's ability to adequately perfuse its tissue, creating regions of tumor hypoxia which exacerbate the Warburg effect and promote a malignant phenotype. Hyperbaric oxygen therapy is the administration of 100% oxygen at elevated barometric pressure. It supersaturates the blood with oxygen, increasing its diffusion distance into the tissues, and can therefore be used to increase intratumoral pO2 and reverse tumor hypoxia. Here we present evidence that the ketogenic diet, ketone supplementation, and hyperbaric oxygen therapy work individually and in combination to slow progression and extend survival in the VM-M3 model of metastatic cancer. This study strongly suggests that these cost effective, non-toxic metabolic therapies should be further evaluated in animal and human studies to determine their potential clinical use.

Zusammenfassung 1) Fragestellung und zentrale Untersuchung Unter der Hypothese, dass die Transportrate des Glukosetransporters Typ 3 (GLUT3) abhängig von der Kopienanzahl (CNV) des für ihn kodierenden Gens SLC2A3 ist, wurden Zelllinien mit drei Kopien (Duplikation) mit Kontroll-Zelllinien mit nur zwei Kopien bezüglich ihrer Glukoseaufnahme miteinander verglichen (n=2; N=9). Hierzu wurde die zelluläre Glukoseaufnahme mittels radioaktiv markierter 2-Desoxyglukose in via Eppstein-Barr-Virus immortalisierten lymphoblastoiden Zelllinien (EBV-LCLs) gemessen. In den initialen Untersuchungen zeigt sich, dass das Protokoll an manchen Stellen zu viel Spielraum lässt. Die Methode wird daraufhin standardisiert und bezüglich einiger Parameter angepasst: g-Zentrifugeneinstellung, Mischen/Aliquotieren, Zellanzahl, Replikatanzahl, Inkubationszeit/-intervalle und Durchführungsdauer. 2) Wichtigste Ergebnisse Die funktionelle Untersuchung zur Duplikation des SLC2A3-Gens in Patienten mit Aufmerksamkeitsdefizit-/Hyperaktivitätsstörung (ADHS) zeigt schließlich im dynamischen Aushungerungsversuch der EBV-LCLs über vier Tage (Vergleich t2 zu t1) statistisch für die Gruppen eine deutliche Differenz mit mittlerer Effektstärke (Lineares Gemischtes Modell; p = 0,06; Cohens d = 0,37). Zum zweiten Messzeitpunkt (t2) zeigt sich statistisch zwischen den Gruppen eine sehr signifikante Differenz mit hoher Effektstärke (Lineares Gemischtes Modell; p < 0,006; Cohens d = 0,55). Damit konnte in dieser Arbeit nachgewiesen werden, dass die SLC2A3-Duplikation neben dem Gendosiseffekt auf mRNA-Ebene auch hypermorph funktionelle Veränderungen auf zellulärer Ebene nach sich zieht. Nachfolgende Untersuchungen sollten vor diesem Hintergrund mögliche Kofaktoren investigieren und auf Alterationen in nachgeschalteten Signalwegen abzielen. / 5.1 Research question and central investigation Under the hypothesis that the transport rate of the glucose transporter type 3 (GLUT3) is dependent on the copy number (CNV) of the gene encoding it, SLC2A3, cell lines with three copies (duplication) were compared with control cell lines with only two copies with respect to their glucose uptake (n=2; N=9). For this purpose, cellular glucose uptake was measured using radiolabeled 2-deoxyglucose in lymphoblastoid cell lines (EBV-LCLs) immortalized via Eppstein-Barr virus. The initial studies show that the protocol leaves too much leeway at some maneuvers. The method is then standardized and adapted with regard to the following parameters: g-centrifuge setting, mixing/aliquoting, cell number, replicate number, incubation time/intervals and execution time. 5.2 Main results The functional investigation for the duplication of the SLC2A3 gene in patients with attention-deficit/hyperactivity disorder (ADHD) finally shows in the dynamic starvation test of EBV-LCLs over four days (comparison t2 to t1) statistically for the groups a significant difference with a mean effect size (Linear Mixed model; p = 0.06; Cohen's d = 0.37). At the second measurement time point (t2), there is statistically a very significant difference with a high effect size (Linear Mixed Model; p < 0.006; Cohen's d = 0.55). Thus, this work demonstrated that the SLC2A3 duplication in addition to the gene dosage effect at the mRNA level, also induces hypermorphic functional changes at the cellular level. Subsequent studies should investigate possible cofactors and target alterations in downstream signaling pathways.

Ductal carcinoma in situ (DCIS) has been shown to be a precursor to invasive ductal cancer (IDC). Though the progression of DCIS to IDC is believed to be an important aspect of tumor aggressiveness, prognosis and molecular markers that predict progression are poorly understood. Therefore, determining the mechanisms by which some DCIS progress is critical for future breast cancer diagnostics and treatment. Singleminded-2s (SIM2s) is a member of the bHLH/PAS family of transcription factors and a key regulator of differentiation. SIM2s is highly expressed in mammary epithelial cells and lost in breast cancer. Loss of Sim2s causes aberrant mouse mammary development with features suggestive of malignant transformation, whereas over-expression of Sim2s promotes precocious alveolar differentiation, suggesting that Sim2s is required for establishing and enhancing mammary gland differentiation. We hypothesize that SIM2s expression must be lost in premalignant lesions for breast cancer to develop. We first analyzed Sim2s in the involuting mammary gland, which is a highly tumorpromoting environment. Sim2s is down-regulated during involution, and forced expression delays involution. We then analyzed SIM2s expression in human breast cancer samples and found that SIM2s is lost with progression from DCIS to IDC, and this loss correlates with metastasis. SIM2s expression in DCIS promoted a differentiated phenotype and suppressed genes associated with de-differentiation. Furthermore, loss of SIM2s expression in DCIS xenografts increased metastasis likely due to an increase in hedgehog signaling and matrix metalloproteinase expression. Interestingly, we found metabolic shifts with gain and loss of SIM2s in not only DCIS cells, but also MCF7 and SUM159 cells. SIM2s expression decreased aerobic glycolysis and promoted oxidative phosphorylation through direct upregulation of CDKN1a and senescence. Loss of SIM2s, conversely, promotes mitochondrial dysfunction and induction of the Warburg effect. This is the first time CDKN1a and cellular senescence have been indicated as causative to metabolic shifts within cancer cells. These studies show a new role for SIM2s in metabolic homeostasis, and this regulation is lost during tumorigenesis. These data indicate SIM2s is at the apex where aging, metabolism, and disease meet – regulating the delicate relationship between the three.

Seven genes involved in metabolism were tested as direct targets of the Notch signalling pathway. For each gene the occupancy of its enhancers by Su(H), its transcriptional response to Notch pathway and its biological functionality was verified in vitro and in vivo.

Under normal conditions, non-transformed cells rely on glycolysis followed by oxidative phosphorylation to generate ATPs. When oxygen is scarce or when cells are actively proliferating, cellular ATPs come mainly from glycolysis. Pyruvate is converted into lactate to allow glycolysis to continue. Interestingly, cancer cells have adapted to favor lactate production even at normal oxygen tensions, exhibiting a metabolic shift known as the Warburg effect. However, the metabolic state of other cellular constituents within the tumor remains mostly unknown. Cancer-associated fibroblasts (CAFs) are the most abundant stromal cells. They aid tumor growth and metastasis by providing growth factors, cytokine, ECM remodeling proteins and interacting with other tumor stromal cells. Here I show that the Warburg effect also operates in stromal fibroblasts of the tumor microenvironment. Using mass spectrometry, genetic mouse models, gene expression and methylation studies, I demonstrate that CAFs from human and mouse mammary tumors exhibit hyperactive glycolysis and a metabolic shift towards lactate production. Furthermore, this phenotype may be sustained through epigenetic modifications of endogenous hypoxia-inducible factor 1α, key regulatory enzymes fructose-bisphosphatase 1 and pyruvate kinase M2. Depletion of stromal fibroblasts or suppression of lactate production specifically in these cells alters the metabolic profile of not only the tumors but also the cancer cells and results in impeded tumor growth. These results collectively suggest that tumor growth is dependent on metabolic state and metabolic support of stromal fibroblasts, highlighting these cells as attractive therapeutic targets in controlling cancer progression.

Ovarian cancer is known as the "silent killer," due to its late diagnosis and frequent recurrence after initial treatment.  Finding a new way to diagnose and treat ovarian cancer in conjunction with current therapies is paramount.  By capitalizing on metabolic changes that occur during cancer progression, interventions can be developed.  The Nobel laureate Otto Warburg is credited with discovering an altered metabolic state within cancer cells known as the Warburg effect.  In the Warburg effect, cancer cells participate in an increased rate of aerobic glycolysis with an excess secretion of lactate, allowing for carbon flux into biosynthetic pathways.  Exactly which metabolic pathways are altered in ovarian cancer and at which stage in the progression of ovarian cancer they are occurring was unknown.  Therefore using the recently established mouse ovarian surface epithelial (MOSE) progression model, we were able to measure metabolic changes in varying states of disease and levels of aggressiveness.  As cells progressed from a benign early stage (MOSE-E), through a transitional intermediate stage (MOSE-I), to an aggressive late stage (MOSE-L), the MOSE cells became more glycolytic and lipogenic, establishing the MOSE model as a valuable model for studying ovarian cancer metabolism.  Treating the MOSE cells with the naturally occurring chemotherapeutic agent sphingosine decreased p-AKT  protein levels in the cell, decreased the glycolytic rate and decreased de novo cholesterol synthesis.  Cancer stem cells are known to be resistant to chemotherapy treatments and targeting their metabolism may be promising for combinatorial treatments.  Therefore, the metabolism of highly aggressive tumor-initiating cells (TIC), harvested from ascites of C57Bl/6 mice injected with MOSE-L cells were characterized.  Although the basal metabolism of the TICs was similar to the MOSE-L cells, TICs were more resistant to cell death as a consequence of external stresses and substrate depletion.  The TICs could also up-regulate oxygen consumption rate (OCR) when uncoupled and increase glycolysis when ATP Synthase was inhibited, highlighting their resiliency.  Taken together, we have identified targets for treatment strategies that could suppress the growth of primary tumors and may be effective against TICs, thereby suppressing tumor recurrence and possibly prolonging the life of women with ovarian cancer. / Ph. D.

Cancer cells are most clearly characterized by their abnormal and uncontrolled cell growth. One of the most notable theories that explains the vast proliferative capacity of tumorigenic cells is the Warburg effect, a significant shift in metabolism wherein cancer cells preferentially fuel cell division using aerobic glycolysis instead of aerobic respiration. This upregulation of glycolytic fermentation in aerobic environments is highly unusual - glycolysis is typically utilized in anaerobic conditions, but nonetheless dominates cancer metabolic activity in spite of the presence of oxygen. Since the discovery the Warburg effect in the 1920s, researchers have struggled to identify whether aerobic glycolysis is a cause or consequence of carcinogenesis. Interestingly, a new theory recently emerged that challenges this widely-accepted metabolic paradigm for cancer. Known as the reverse Warburg effect, this new mechanism shows that in carcinomas such as breast cancer, the Warburg effect occurs not in cancer cells, but rather in tumor-adjacent stromal fibroblasts. These cancer-associated fibroblasts (CAFs) in the greater tumor microenvironment produce lactate - a high-energy metabolite formed as a byproduct of aerobic glycolysis - to fuel aerobic respiration and rapid tumorigenesis in neighboring cancer cells. This emerging theory emphasizes the pivotal role of the tumor microenvironment in determining whether cancer cells undergo aerobic glycolysis or aerobic respiration. Central to this lactate-linked metabolic intersection are two critical enzymes that regulate a cell's metabolic commitment - lactate dehydrogenase (LDH) and pyruvate dehydrogenase complex (PDHc). In order to clarify the mechanisms through which CAFs induce tumorigenesis in breast cancer, we plan to carry out two specific aims: (1) evaluate the enzymatic activity of LDH and PDHc, and (2) compare LDH and PDHc enzyme content. Using co-culture techniques to study the breast cancer tumor microenvironment in vitro, we will compare the enzymatic activity and enzyme content of both MCF7 breast cancer cells and CAFs to identify whether the reverse Warburg effect occurs due to post-translational enzyme activation or increased enzyme synthesis.

<strong>Background:</strong> Malignant transformation of cells is typically characterised by aerobic glycolysis, resulting in supressed mitochondrial function, a state that helps resistance to apoptosis. This characteristic has been widely accepted as a hallmark of cancer and has been shown to be of critical importance in tumour development. The bioenergetic differences between normal and malignant cells are being exploited to identify potential cancer specific therapeutics. Improved in-vitro models are required to aid the identification and assessment of candidate drugs. In this project, we investigated the bioenergetic phenotypes of a panel of adult and paediatric cancer cell lines and evaluated the potential of 3D models as a platform for testing drugs that target cancer metabolism. We also investigated a novel method to assess mitochondrial function that enables the quantification of the level of oxygenation within the cell. <strong>Results:</strong> The results presented in this thesis show that not all cancers display this aerobic glycolytic phenotype. We found that while some cell lines displayed the Warburg phenotype others displayed high levels of oxidative metabolism. These bioenergetic profiles need to be considered when deciding which anti-cancer drugs to use in a chemotherapeutic regime. If a bioenergetic pattern can be identified it may one day form the basis of a screening strategy for tumours. Dichloroacetate (DCA) is a small molecule PDK inhibitor that was investigated in this study. It was found to be relatively non-toxic to cells cultured in 2D but had improved toxicity when the cells were cultured in a 3D environment. Lastly, we evaluated a new oxygen sensing nanoprobe, Mito-Xpress Intra, and the results demonstrate its potential as a non-invasive means of measuring oxygen concentrations within the cell in real time as well as highlighting some striking differences between applied ambient and measured intracellular oxygen concentrations. <strong>Conclusion:</strong> The findings suggest that not all cancers display the characteristic glycolytic phenotype. They also highlight the importance of controlling oxygen and glucose levels when evaluating metabolism and when drug testing.

The Warburg effect describes the survival advantage of cancer cells in that they can proliferate under low oxygen/hypoxic conditions via a less efficient pathway known as glycolysis. It has not yet been documented at which point, in an oxygen gradient, phenotypic changes occur. Measuring the intracellular redox potential (IRP) and its impact on cellular dynamics would provide greater insight into how disruption of redox homeostasis caused by changes in oxygen concentration leads to aberrant cell signalling and diseases such as cancer. Current techniques in measuring IRP include redox-sensitive fluorescent proteins such as roGFP which is glutathione-specific. Measuring the concentration of one redox couple is, however, not an accurate representation of IRP as it does not necessarily inform about the state of other redox couples. Furthermore, fluorescent biosensors can suffer from photobleaching and may interact with other oxidants. The IRP was measured, in this work, using our newly developed novel-class of surface enhanced Raman scattering nanoparticles which can quantitatively measure the redox potential of cells in vitro. A 'homemade' device was created to keep the cells under fixed pO2 whilst obtaining measurements. The IRP was correlated with the transcriptomic and downstream metabolic profiles of MCF7 breast cancer cells, under perturbed pO2, using 1H NMR spectroscopy (NMR), mass spectrometry (MS) and RNA-sequencing. Discriminatory metabolites were all associated with energy and glucose metabolism. Discriminatory microRNAs were all affiliated with the hallmarks of cancer; the regulation of some is controlled by transcription factors containing redox-sensitive motifs in their DNA binding domains. Multivariate analysis techniques were used to analyse the different data streams in a holistic way that allows the correlation of redox potential, metabolism and transcription.