Characterization of ALDH1A1 as a novel tumorigenic target mediating TAZ-induced lung tumorigenesis and stem cell phenotypes

Recent studies suggest that lung cancer stem cells (CSCs) may play major roles in lung cancer development, metastasis and drug resistance. Therefore, identification of lung CSC drivers may provide promising targets for lung cancer. TAZ (transcriptional co-activator with PDZ-binding motif) is a transcriptional co-activator and key downstream effector of the Hippo pathway, which plays critical roles in various biological processes. TAZ has been shown to be overexpressed in non-small cell lung cancer (NSCLC) and involved in tumorigenicity of lung epithelial cells. However, whether TAZ is a driver for lung CSCs and tumor formation in vivo is unknown. In addition, the molecular mechanism underlying TAZ-induced lung tumorigenesis remains to be determined. In this study, we provided evidence that constitutively active TAZ (TAZ-S89A) is a driver for lung tumorigenesis in vivo in mice and formation of lung CSC. Oncogenes upregulated in TAZ-overexpressing cells were identified with further validation. The most dramatically activated gene, Aldh1a1 (Aldehyde dehydrogenase 1 family member a1), a well-established CSC marker, showed that TAZ induces Aldh1a1 transcription by activating its promoter activity through interaction with the transcription factor TEA domain (TEAD) family member. Most significantly, inhibition of ALDH1A1 with its inhibitor A37 or CRISPR (Clustered Regularly Interspaced Short Palindromic Repeats) gene knockout in lung cancer cells suppressed lung tumorigenic and CSC phenotypes in vitro, and tumor formation in mice in vivo. In conclusion, this study identified TAZ as a novel inducer of lung CSCs and the first transcriptional activator of the stem cell marker ALDH1A1. Most significantly, we identified ALDH1A1 as a critical meditator of TAZ-induced tumorigenic and CSC phenotypes in lung cancer. Our studies provided preclinical data for targeting of TAZ-TEAD-ALDH1A1 signaling to inhibit CSC-induced lung tumorigenesis and drug resistance in the future.

Hormonal factors and insulin-like growth factor-1 in North American postmenopausal women

Background: Mechanisms underlying the effect of estrogen exposure on breast cancer risk remain unclear. Insulin-like growth factor-1 (IGF-1) levels have been positively associated with breast cancer and are a potential mechanism. Objectives: The objectives of this thesis are: 1) to explore whether the reproductive risk factors and the lifetime cumulative number of menstrual cycles (LCMC), as measures for long-term estrogen exposure, are associated with IGF-1 levels, and 2) to examine the effect of an aromatase inhibitor (AI) on IGF-1 levels, and the potential interaction with BMI. Methods: A cross sectional study and a randomized controlled trial nested with the MAP.3 chemoprevention trial were used to address objective 1 and 2, respectively. 567 postmenopausal women were selected. Anthropometric measurements, lifestyle factors, reproductive characteristics and serum IGF-1 concentrations were collected at baseline and one year. Objective 1. The LCMC was computed as a composite measure of the reproductive characteristics. Multivariable linear regression models were used to assess the association between IGF-1 levels and LCMC and the hormonal risk factors, while adjusting for potential covariates. Objective 2. Changes in IGF-1 were compared between the exemestane and placebo, and effect modification by BMI was tested with an interaction term. Results: Objective 1. Women aged 55 years or older at menopause had 16.26 ng/mL (95% CI: 1.76, 30.75) higher IGF-1 compared to women aged less than 50 years at menopause. Women in the highest category of menstrual cycles (≥500 cycles) had an average 19.00 ng/mL (95%CI: 5.86, 32.14) higher concentration of IGF-1 compared to women in the lowest category (<350). Exogenous hormones had no effect on postmenopausal IGF-1 levels. Objective 2. Exemestane significantly increased IGF-1 levels by 18% (95% CI: 14%-22%); while, placebo had no effect on IGF-1. The changes in IGF-1 were significantly different between the treatment arms (P<0.0001) and no significant interaction was observed between treatment and BMI on IGF-1 changes (P=0.1327). Conclusion: Objective 1. Larger number of menstrual cycles and a later age at menopause are positively associated with IGF-1. IGF-1 may be one mechanism by which prolonged estrogen exposure increases cancer risk. Objective 2. We conclude that the reduced cancer risk observed with AI therapy likely occurs in an IGF-1 independent mechanism. Further studies exploring the clinical consequences of increased IGF-1 on AI therapy are needed.

THE ROLE OF LIPOPROTEIN(a)/APOLIPOPROTEIN(a) IN ENDOTHELIAL DYSFUNCTION: MECHANISTIC STUDIES IN VASCULAR ENDOTHELIUM

Multiple lines of evidence suggest that elevated plasma lipoprotein(a) (Lp(a)) concentrations are a significant risk factor for the development of a number of vascular diseases including coronary heart disease and stroke. Lp(a) consists of a low-density lipoprotein (LDL)-like moiety and an unique glycoprotein, apolipoprotein(a) (apo(a)), that is covalently attached to the apolipoproteinB-100 (apoB-100) component of LDL by a single disulfide bond. Many studies have suggested a role for Lp(a) in the process of endothelial dysfunction. Indeed, Lp(a) has been shown to increase both the expression of adhesion molecules on endothelial cells (EC), as well as monocyte and leukocyte chemotactic activity in these cells. We have previously demonstrated that Lp(a), through its apo(a) moiety, increases actomyosin-driven EC contraction which, as a consequence, increases EC permeability. In this thesis, we have demonstrated a role for the strong lysine-binding site in the kringle IV type 10 domain of apo(a) in increasing EC permeability, which occurs through a Rho/Rho kinase-dependent pathway. We have further validated these findings using mouse mesenteric arteries in a pressure myograph system. We also have dissected another major signaling pathway initiated by apo(a) that involves in a disruption of adherens junctions in EC. In this pathway, apo(a)/Lp(a) activates the PI3K/Akt/GSK3β-dependent pathway to facilitate nuclear translocation of beta-catenin. In the nucleus beta-catenin induced the expression of cyclooxygenase-2 (COX-2) and the secretion of prostaglandin E2 (PGE2) from the EC. Finally, we have presented data to suggest a novel inflammatory role for apo(a) in which it induces the activation of nuclear factor-kappaB through promotion of the dissociation of IkappaB from the inactive cytoplasmic complex; this allows the nuclear translocation of NFkappaB with attendant effects on the transcription of pro-inflammatory genes. Taken together, our findings may facilitate the development of new drug targets for mitigating the harmful effects of Lp(a) on vascular EC which corresponds to an early step in the process of atherogenesis.

Functional characterization of a Kar3/Vik1-like Kinesin-14 heterodimer from the filamentous multinucleate fungus Ashbya gossypii

Kinesins are motor proteins that convert chemical energy from ATP hydrolysis into mechanical energy used to generate force along microtubules, transporting organelles, vesicles, and proteins within the cell. Kar3 kinesins are microtubule minus-end-directed motors with pleiotropic functions in mating and mitosis of budding and fission yeast. In Saccharomyces cerevisiae, Kar3 is multifunctionalized by two non-catalytic companion proteins, Vik1 and Cik1. A Kar3-like kinesin and a single Vik1/Cik1 ortholog are also expressed by the filamentous fungus Ashbya gossypii, which exhibits different nuclear movement challenges and unique microtubule dynamics from its yeast relatives. We hypothesized that these differences in A. gossypii physiology could translate into interesting and novel differences in its versions of Kar3 and Vik1/Cik1. Presented here is a structural and functional analysis of recombinantly expressed and purified forms of these motor proteins. Compared to the previously published S. cerevisiae Kar3 motor domain structure (ScKar3MD), AgKar3MD displays differences in the conformation of the ATPase pocket. Perhaps it is not surprising then that we observed the maximal microtubule-stimulated ATPase rate (kcat) of AgKar3MD to be approximately 3-fold slower than ScKar3MD, and that the affinity of AgKar3MD for microtubules (Kd,MT) was lower than ScKar3MD. This may suggest that elements that compose the ATPase pocket and that participate in conformational changes required for efficient ATP hydrolysis or products release work differently for AgKar3 and ScKar3. There are also subtle structural differences in the disposition of the secondary structural elements in the small lobe (B1a, B1b, and B1c) at the edge of the motor domain of AgKar3 that may reflect the enhanced microtubule-depolymerization activity that we observed for this motor, or they could relate to its interactions with a different regulatory companion protein than its budding yeast counterpart. Although we were unable to gain experimentally determined high-resolution information of AgVik1, the results of Phyre2-based bioinformatics analyses may provide a structural explanation for the limited microtubule-binding activity we observed. These and other fundamental differences in AgKar3/Vik1 could explain divergent functionalities from the ScKar3/Vik1 and ScKar3/Cik1 motor assemblies.

Higher brain neurons succumb to acute stroke-like insult while lower brain neurons strongly resist

Pyramidal neurons (PyNs) in ‘higher’ brain are highly susceptible to acute stroke injury yet ‘lower’ brain regions better survive global ischemia, presumably because of better residual blood flow. Here we show that projection neurons in ‘lower’ brain regions of hypothalamus and brainstem intrinsically resist acute stroke-like injury independent of blood flow in the brain slice. In contrast `higher` projection neurons in neocortex, hippocampus, striatum and thalamus are highly susceptible. In live brain slices from rat deprived of oxygen and glucose (OGD), we imaged anoxic depolarization (AD) as it propagates through these regions. AD, the initial electrophysiological event of stroke, is a depolarizing front that drains residual energy in compromised gray matter. The extent of AD reliably determines ensuing damage in higher brain, but using whole-cell recordings we found that all CNS neurons do not generate a robust AD. Higher neurons generate strong AD and show no functional recovery in contrast to neurons in hypothalamus and brainstem that generate a weak and gradual AD. Most dramatically, lower neurons recover their membrane potential, input resistance and spike amplitude when oxygen and glucose is restored, while higher neurons do not. Following OGD, new recordings could be acquired in all lower (but not higher) brain regions, with some neurons even withstanding multiple OGD exposure. Two-photon laser scanning microscopy confirmed neuroprotection in lower, but not higher gray matter. Specifically pyramidal neurons swell and lose their dendritic spines post-OGD, whereas neurons in hypothalamus and brainstem display no such injury. Exposure to the Na+/K+ ATPase inhibitor ouabain (100 μM), induces depolarization similar to OGD in all cell types tested. Moreover, elevated [K+]o evokes spreading depression (SD), a milder version of AD, in higher brain but not hypothalamus or brainstem so weak AD correlates with the inability to generate SD. In summary, overriding the Na+/K+ pump using OGD, ouabain or elevated [K+]o evokes steep and robust depolarization of higher gray matter. We show that this important regional difference can be largely accounted for by the intrinsic properties of the resident neurons and that Na+/K+ ATPase pump efficiency is a major determining factor generating strong or weak spreading depolarizations.

Elucidation of the interactions between the transcription factor E2A and the transcriptional co-activator CBP/p300

E2A is a transcription factor that plays a particularly critical role in lymphopoiesis. The chromosomal translocation 1;19, disrupts the E2A gene and results in the expression of the fusion oncoprotein E2A-PBX1, which is implicated in acute lymphoblastic leukemia. Both E2A and E2A-PBX1 contain two activation domains, AD1 and AD2, which comprise conserved ΦxxΦΦ motifs where Φ denotes a hydrophobic amino acid. These domains function to recruit transcriptional co-activators and repressors, including the histone acetyl transferase CREB binding protein (CBP) and its paralog p300. The PCET motif within E2A AD1 interacts with the KIX domain of CBP/p300, the disruption of which abrogates the transcriptional activation by E2A and the transformative properties of E2A-PBX1. The generation of a peptide-based inhibitor targeting the PCET:KIX interaction would serve useful in further assessing the role of E2A and E2A-PBX1 in lymphopoiesis and leukemogenesis. An interaction between E2A AD2 and the KIX domain has also been recently identified, and the TAZ domains of CBP/p300 have been shown to interact with several transcription factors that contain ΦxxΦΦ motifs. Thus the design of an inhibitor of the E2A:CBP/p300 interaction requires the full complement of interactions between E2A and the various domains of CBP/p300 to be elucidated. Here, we have used nuclear magnetic resonance (NMR) spectroscopy to determine that AD2 interacts with KIX at the same site as PCET, which indicates that the E2A:KIX interaction can be disrupted by targeting a single binding site. Using an iterative synthetic peptide microarray approach, a peptide with the sequence DKELQDLLDFSLQY was derived from PCET to interact with KIX with higher affinity than the wild type sequence. This peptide now serves as a lead molecule for further development as an inhibitor of the E2A:CBP/p300 interaction. Fluorescence anisotropy, peptide microarray technology, and isothermal titration calorimetry were employed to characterize interactions between both TAZ domains of CBP/p300 and the PCET motif and AD2 of E2A. Alanine substitution of residues within PCET demonstrated that the ΦxxΦΦ motif is a key mediator of these interactions, analogous to the PCET:KIX interaction. These findings now inform future work to establish possible physiological roles for the E2A:TAZ1 and E2A:TAZ2 interactions.

A FUNCTIONAL, COMPARATIVE AND CLINICAL ANALYSIS OF SPERM-BORNE OOCYTE ACTIVATING FACTOR, PAWP

Successful fertilization depends upon the activation of metaphase II arrested oocytes by sperm-borne oocyte activating factor (SOAF). Failure of oocyte activation is considered as the cause of treatment failure in a proportion of infertile couples. SOAF induces the release of intracellular calcium in oocyte which leads to meiotic resumption and pronuclear formation. Calcium release is either in the form of single calcium transient in echinoderm and amphibian oocytes or several calcium oscillations in ascidian and mammalian oocytes. Although the SOAF attributes are established, it is not clear which sperm protein(s) play such role. Sperm postacrosomal WW binding protein (PAWP) satisfies a developmental criteria set for a candidate SOAF. This study shows that recombinant human PAWP protein or its transcript acts upstream of calcium release and fully activates the amphibian and mammalian oocytes. Interference trials provided evidence for the first time that PAWP mediates sperm-induced intracellular calcium release through a PPXY/WWI domain module in Xenopus, mouse and human oocytes. Clinical applications of PAWP were further investigated by prospective study on the sperm samples from patients undergoing intracytoplasmic sperm injection (ICSI). PAWP expression level, analyzed by flow cytometry, was correlated to ICSI success rate and embryonic development. This study also explored the developmental expression of the other SOAF candidate, PLCζ in male reproductive system and its function during fertilization. Our findings showed for the first time that PLCζ most likely binds to the sperm head surface during epididymal passage and is expressed in epididymis. We demonstrated that PLCζ is also compartmentalized early in spermiogenesis and thus could play an important role during spermiogenesis. Detailed analysis of in vitro fertilization revealed that PLCζ disappears from sperm head during acrosome reaction and is not detectable during sperm incorporation into the oocyte cytoplasm. In conclusion, this dissertation provides evidence for the essential non-redundant role of sperm PAWP in amphibian and mammalian fertilization; recommends PAWP as a biomarker for prediction of ICSI outcomes in infertile couples; and proposes that sperm PLCζ may have functions other than inducing oocyte activation during fertilization.

Glycemic State Regulates Brain-Derived Neurotrophic Factor Responsiveness of Neurons in the Paraventricular Nucleus of the Hypothalamus

Brain derived neurotrophic factor (BDNF) is a member of the family of neurotrophins and binds to the tropomyosin-related kinase B (TrkB) receptor. Like other neurotrophic factors, BDNF is involved in the development and differentiation of neurons. Recently, studies have suggested important roles for BDNF in the regulation of energy homeostasis. The paraventricular nucleus (PVN) is critical for normal energy balance contains high levels of both BDNF and TrkB mRNA. Studies have shown that microinjections of BDNF into the PVN increase energy expenditure, suggesting BDNF plays a role in energy homeostasis through direct actions in this hypothalamic nucleus. We used male Sprague-Dawley rats to perform whole-cell current-clamp experiments from PVN neurons in slice preparation. BDNF was bath applied at a concentration of 2nM and caused depolarizations in 54% of neurons (n = 25; mean change in membrane potential: 8.9 ± 1.2 mV), hyperpolarizations in 23% (n = 11; mean change in membrane potential: -6.7 ± 1.4 mV), while the remaining cells tested were unaffected. Previous studies showing effects of BDNF on γ-aminobutyric acid type A (GABAA) mediated neurotransmission in PVN led us to examine if these BDNF-mediated changes in membrane potential were maintained in the presence of tetrodotoxin (TTX) sodium channel blocker (N = 9; 56% depolarized, 22% hyperpolarized, 22% non-responders) and bicuculline (GABAA antagonist) (N = 12; 42% depolarized, 17% hyperpolarized, 41% non-responders), supporting the conclusion that these effects on membrane potential were postsynaptic. We also evaluated the effects of BDNF on these neurons across varying physiologically relevant extracellular glucose concentrations. At 10 mM 23% (n = 11; mean: -6.7 ± 1.4 mV) of PVN neurons hyperpolarized in response to BDNF treatment, whereas at 0.2 mM glucose, 71% showed hyperpolarizing effects (n = 12; mean: -6.3 ± 2.8 mV). Our findings reveal that BDNF has direct impacts on PVN neurons and that these neurons are capable of integrating multiple sources of metabolically relevant input. Our analysis regarding glucose concentrations and their effects on these neurons’ response to other metabolic signals emphasizes the importance of using physiologically relevant conditions for study of central pathways involved in the regulation of energy homeostasis.