Identifying Retinoic Acid Dependent Mechanisms of Neural Crest Epithelial-Mesenchymal Transition

Cancer metastasis is one of the distinguishing traits that make the disease difficult to treat. The steps of tumor invasion are similar to a process called epithelial to mesenchymal transition (EMT). Retinoic acid (RA) is shown to inhibit EMT and may cause an increased transcription of the gene, skiB. This gene may play a role in EMT regulation through RA synthesis activation. Thus, this study used in-situ hybridizations labeling RNA expression of the gene, skiB, and morpholinos to knock down expression. These experiments were performed in the presence of TP-0903, a multi-kinase inhibitor that increases retinoic acid biosynthesis. Results show increased expression, while knocking down the gene provides inconclusive results. By exploring further details on the molecular mechanisms of EMT, cancer treatment can be improved by creating better targeted therapies. Moreover, future research will involve investigating other genes that may be associated with EMT and increased synthesis of RA.

IDENTIFYING RETINOIC ACID DEPENDENT MECHANISMS OF NEURAL CREST EPITHELIAL-MESENCHYMAL TRANSITION by Scott Khuu A Senior Honors Thesis Submitted to the Faculty of The University of Utah In Partial Fulfillment of the Requirements for the Honors Degree in Bachelor of Science/Arts (depending on major) In Chemistry Department Approved: ______________________________ Rodney A. Stewart, PhD Thesis Faculty Supervisor _____________________________ Cynthia J. Burrows, PhD Chair, Department of Chemistry _______________________________ Thomas G. Richmond, PhD Honors Faculty Advisor _____________________________ Sylvia D. Torti, PhD Dean, Honors College May 2017 Copyright © 2017 All Rights Reserved ABSTRACT Cancer metastasis is one of the distinguishing traits that make the disease difficult to treat. The steps of tumor invasion are similar to a process called epithelial to mesenchymal transition (EMT). Retinoic acid (RA) is shown to inhibit EMT and may cause an increased transcription of the gene, skiB. This gene may play a role in EMT regulation through RA synthesis activation. Thus, this study used in-situ hybridizations labeling RNA expression of the gene, skiB, and morpholinos to knock down expression. These experiments were performed in the presence of TP-0903, a multi-kinase inhibitor that increases retinoic acid biosynthesis. Results show increased expression, while knocking down the gene provides inconclusive results. By exploring further details on the molecular mechanisms of EMT, cancer treatment can be improved by creating better targeted therapies. Moreover, future research will involve investigating other genes that may be associated with EMT and increased synthesis of RA. ii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 METHODS 8 RESULTS 12 DISCUSSION 18 ACKNOWLEDGEMENTS 21 REFERENCES 22 iii 1 INTRODUCTION The ability of cancer to invade other parts of the body is one of the distinguishing traits that make it a formidable disease. The steps of tumor invasion appear to be coordinated by a series of cellular and molecular processes characterized by changes in cell adhesion and migration, which imply an essential developmental process called epithelial-mesenchymal transition (EMT) (Polyak et al., 2009). Epithelial cells are characterized by the numerous structures that hold the cells together in an organized layer, such as desmosomes and tight junctions. Moreover, apical-basal polarization involving adhesion molecules like cadherins and integrins serve as strong determinates of epithelial cells. These cells are unable to remove themselves from the epithelial layer under normal conditions. Contrastingly, mesenchymal cells lack the membrane structures associated with epithelial cells and have different organization in their cytoskeleton. Furthermore, mesenchymal cells do not exist in an organized layer and are able to migrate in single cells or chains. (Friedl, 2004). EMT causes epithelial cells to lose intracellular junctions, apical-basal polarity and gain cytoskeleton reorganization like those of mesenchymal cells as a prerequisite for motility and invasion of surrounding tissue. (Lim et al., 2012). Likewise, cancer metastasis involves a multistep process where tumor cells spread from the primary tumor using circulation to form secondary tumors at a distant site (Khan et al., 2014). With this knowledge, EMT can grant cancer cells the significant properties of dissociating from the primary tumor site and promoting acquisition of stem cell like properties, therapeutic resistance, and increased survival, which contributes to a poor patient prognosis. 2 Several mechanisms are associated with EMT activation involving a series of extracellular signals and crosstalk signaling pathways. However, these signaling pathways have common endpoints such as the downregulation of E-cadherin expression and expression of EMT-associated regulators like snail (Theiry et al., 2006). Specifically, snail transcription factor genes repress E-cadherin and other cell-adhesion molecules to stimulate EMT in normal development and invasive tumors (Gupta et al., 2005). Regarding cancer, snail serves as an important EMT regulator of a variety of signaling molecules—Wnt, epidermal growth factor, fibroblast growth factor, and transforming growth factor beta (TGF-β) (Tania et al., 2014). Therefore, targeting signaling pathways in EMT may provide beneficial therapies for blocking EMT during cancer migration and/or colonization. The embryonic dorsal neural tube of vertebrates provides a useful model to study EMT as disruption of the neural crest development has many consequences, including cancer. Cells within this structure undergo predictable EMT to form neural crest (NC) cells sharing expression of transcription factors like snail1/2 and soxE that are able to migrate independently or in groups to become tissue derivatives, such as melanocytes, craniofacial skeleton, neurons and glia. (Green et al., 2015). Hence, signaling molecules including TGF-β regulate EMT drivers like snail that promote EMT to transform cells into expressing NC specifier genes like Sox in a proposed gene network. (Simoes-Costa et al., 2015). However, the details of where and how signaling molecules cause certain EMT transcription factors to be expressed in NC remains unclear. In order to target conserved signaling pathways in EMT, answers to these questions must be revealed to identify EMT inhibitors in vivo. 3 The Stewart Lab investigates these questions by developing a novel zebrafish EMT reporter. This model fluorescently labels neuroepithelial cells with snail:GFP before NC specifier genes are expressed with sox10:RFP to provide a clear observation of EMT independent of cell migration behaviors. The Tg(snail1b:GFP);Tg(sox10:RFP) reporter serves as an animal-based assay for EMT inhibitors. Using this assay, the Stewart Lab identified a multi-kinase inhibitor of EMT called TP-0903 that prevented the ability of NC cells to undergo EMT and delaminate out of the neural tube. Mollard et al. are the researchers who created TP-0903, also known as HCI-2084. The structure of the drug is shown here: Source: Mollard et al. “Design, Synthesis, and Biological Evaluation of a Series of Novel AXL Kinase Inhibitors.” ACS Med. Chem. Letters. Vol. 2, 907-912. 2011. 4 Further analysis by the lab shows TP-0903 activating retinoic acid (RA) dependent signaling and increasing the biosynthesis of RA where excess levels of RA causes NC EMT defects in drug treated embryos (Jimenez et al., 2016). The activation of RA signaling via TP-0903 provides a different perspective in stopping metastatic cancer through inhibiting EMT. Thus, repression of RA signaling is needed to achieve NC EMT in zebrafish embryos and suggests RA treatment could be a new therapy to block EMT in cancer patients. However, details remain unclear for how RA signaling regulates NC EMT and what transcriptional targets downstream of RA can inhibit NC EMT in zebrafish embryos. Understanding how RA signaling can inhibit cranial NC EMT is critical to understanding the therapeutic relevance of using RA to inhibit EMT in cancer patients. Retinoids are a group of vitamin A metabolites that cannot bind to retinoid nuclear receptors, but are characterized by their ability to convert into retinoic acid that binds to retinoic acid receptors (RARs) and retinoid X receptors (RXRs) exerting antagonistic or agonistic actions (Xu et al., 2004). This conversion from retinol-based structure to retinoic acid depends on the process of vitamin A conversion involving two oxidation steps by alcohol dehydrogenases (ADH), and subsequently retinaldehyde dehydrogenase (RALDH). The first step is reversible while the RALDH-driven step is not. The retinoic acid that is formed may then bind to RARs and RXRs, or may be catalyzed by cytochrome P450 enzymes CYP26 to reduce further activation of RAR/RXR pathway (Xu et al., 2012). The retinoic acid response elements (RAREs) associated with the bound receptors promote transcription. 5 Already, retinoids are used during cancer treatment thought to work by arresting cell proliferation and causing apoptosis. (Nasr et al., 2008; Reynolds, 2000). Cancers that are positively affected by retinoid include acute promyeloctic leukemia, neuroblastoma, breast cancer, and others. (Uray et al., 2015). Problems persist with retinoids like other cancer treatments—resistance and toxicity. However, TP-0903 acts through a new mechanism to activate RA-dependent transcription, thus bypassing the cell’s ability to transport and/or metabolize exogenous retinoids (Jimenez et al., 2016). As a result, retinoid toxicity and resistance may be alleviated by TP-0903. The specifics of the RAdependent transcription activated by the drug are unknown, but there may be a connection between TGF-β and the gene, skiB. The relation between RA and EMT is based on the signaling molecule TGF-β. As stated before, the substrate is known to be a major EMT-inducing signal that is able to initiate the activation of the three core families of EMT transcription—Twist, ZEB, and Snail (Xu, 2009). TGF-β binds to the cell membrane surface type II TGF- β receptor, which recruits type I TGF-β receptor and activates serine/threonine protein kinase (Yang et al., 2015). This leads to phosphorylation of SMADS, which are intracellular proteins that transduce extracellular signals from TGF-β ligand to nucleus for activation of downstream transcription. Furthermore, TGF-β and retinoids are involved in a complex crosstalk via SMADS. An example of these interactions involves a shared repression by TG-interacting factor (TGIF) acting as a transcriptional corepressor of SMAD signaling that prevents access to phosphorylated SMADS (Xu et al., 2012). TGF-β potentially has a duel function in cancer acting as a tumor suppressor and promoter. During the beginning stages of cancer, the tumor suppressor ability of TGF-β 6 is perpetuated by the loss or mutation of members of the TGF-β signaling pathway, specifically in pancreatic and colon cancers (Elliot et al., 2005). Thus, TGF-β serves as a tumor suppressor in terms of apoptosis, replication and proliferation, more specifically on inhibiting cell growth. Whereas, elevated levels of TGF-β are found during late stages of cancer in patients and are associated with increased invasiveness and poorer prognosis (Elliot et al., 2005). Taken together with previously stated knowledge, this observation provides insight to connecting cancer metastasis, EMT, and TGF-β. The gene, ski, is associated with TGF-β in terms of its role in negative regulation through SMADs. The c-ski protein by the Ski gene acts as a transcriptional co-repressor of TGF-β signaling (Akiyoshi et al., 1999). In terms of tumorigenesis, the role of ski becomes more complex possibly due to the dual role of TGF-β in cancer. Therefore, ski likely has a duel role but the molecular mechanisms about its effects on tumorigenesis remain unclear. On one hand, the details of ski’s involvement with causing cancer derives from its overexpression in multiple types of human tumor lines including melanoma and breast cancer (Luo, 2004). In further detail, ski repression of signaling activities for SMADs provides further support in promoting cancer by stopping the tumor-suppressing role of SMADs. On the other hand, ski may act as a tumor suppressor in some types of cells (Shinagawa et al., 2001). Therefore, there is a fine balance between having too much or too little expression of ski. Relating ski to retinoic acid, the gene is also known to interfere with RAR transcription leading to suppression of RA signaling (Dahl et al., 1998). This interference may be caused by ski co-localizing with HDAC3, a repressor protein associated with RAR/RXR dimer bound to RARE complex. Co-localization inhibits degradation of 7 HDAC3 via a proposed interaction between ski and Seven in Absentia Homologue 2 protein, thus promoting the suppression of RA signaling (Melling et al., 2013). Therefore, suggesting ski acts in a negative feedback loop with RA signaling. As a result, the increased biosynthesis of RA caused by TP-0903 may upregulate ski as part of a negative feedback response and contributes to the inhibition of EMT and twista by suppressing TGF-β signaling: This project serves to further determine the role of RA restraining EMT by treating embryos with TP-0903 and observing whether skiB, a paralogue of c-ski, becomes overexpressed allowing for the possibility to knock down the expression in hopes of returning normal EMT function. In addition, identifying transcripts affected by RA maintains an importance. 8 METHODS TP-0903/DMSO Treatment of Embryos Either wild type zebrafish embryos or Tg(snail1b:GFP);Tg(sox10:RFP) embryos were treated with DMSO or TP-0903. Treatment was done in 12 well plates for embryos staged at 10 hours post fertilization (hpf) with about 10 embryos/well and incubated at 28°C with 1.7 ml of egg water containing 7 μM TP-0903 or 1% DMSO as the control. Treatment was stopped for each batch of embryos at 11, 12, 14, and 18 hpf. For in situ’s, embryos were immediately dechorinated with pronase and fixed for hybridization overnight in 1.0 ml of 4% PFA at 4°C rocking in tin foil for approximately 100 embryos/microcentrifuge tube. For GFP/RFP imaging, embryos were further developed to 18 hpf in only egg water after the specific time of treatment was reached. Development of RNA in situ probes skiB RNA in situ probe was developed by cloning skiB DNA into pGEM-T Easy plasmid from Promega. Plasmid was linearized with NdeI for antisense strand and NcoL for sense strand. Once linearized, probe was made following Ambion probe making protocol using T7 polymerase for antisense and SP6 polymerase for sense. krox20 RNA in situ probe was made following same procedure with the following changes involving Pst1 for linearization and T3 polymerase to generate the antisense strand. 9 Whole-mount RNA double in situ hybridizations After embryos were fixed, they were washed 3x5 min. in PBST at room temperature (RT) then incubated in 100% methanol for 2 hours rocking. These embryos were then stored in -20°C for future use or proceeded to the next step of rehydrating embryos in PBST for 3x5 min. at RT. For each sample, approximately 20 embryos/microcentrifuge tube was used. Next, Hyb(-) and Hyb(+) were preheated to 68°C. 1 ml of Hyb(-) was added to each tube for 1 hour in 68°C. 500 ul of Hyb(+) was removed and 1 μl of FITC-labeled in situ probe for skiB + 1 μl of DIG-labeled in situ probe for krox20 was added. Incubate overnight at 68°C. The following day SSCT washes were preheated at 68°C. Hyb(+) and probe were replaced with 1 ml of 2X SSCT/50% Formamide for 2x30 min. at 68°C. All washes were at 1 ml unless indicated differently. The mixture was removed and replaced with 2X SSCT for 1x15 min. at 68°C. Mixture was removed and replaced with .2X SSCT for 2x30 min. Embryos were transferred to a 12-well plate and washed in 2 ml of MABT for 3x5 min. at RT. MABT was removed and replaced with block mixture for 1 hour (20 ml block included 14 ml of MABT, 4 ml 10% block, 2 ml FBS). Each well of embryos were incubated overnight at 4°C with 1:5000 anti-FITC antibody in block mixture. The next day, block mixture was removed and 2 ml of MABT were added to each well for 2x1 hour at RT. MABT was removed and 0.1 M Tris pH 9.5 was added for 3x5 min. at RT. 2 ml of vector BCIP/NBT staining was used in each well for either 1 hour at RT or until desired staining was achieved. Once stained, 2 ml of MABT wash was used for 2x20 min. During this time, 1X MAB + 10 μM EDTA was heated to 65°C. MABT was removed and the embryos were transferred to tubes. 10 minutes of incubation using 10 previous mixture was performed to inactive the anti-FITC antibody in 65°C heating block. Embryos were transferred back to inserts in 12 well plates and washed in 2 ml of MABT for 2x20 min. Block again like before, but using anti-DIG antibody instead. Block mixture removed and embryos were washed in 2 ml of MABT for 2x1 hour at RT. MABT was removed and 2 ml of .1 M Tris pH 8.2 was added for 3x5 min at RT. Using Fast Red staining Sigma tablets and following company’s procedure, embryos were stained and covered in foil for 1 hour or until desired staining was achieved. Once stained, solution was removed and 2 ml of PBST was added to each well for 3x5 min. Embryos were then stored in 80% glycerol in PBST and imaged under a confocal microscope. Morpholino Microinjections skiB MO antisense oligos were ordered from Gene Tools, LLC. Stock solution was made at 1 mM. Working solutions were made for concentrations of 6, 9, and 24 ng. Embryos were injected with MO between 0-2 hpf using microinjection tools and proceeded to the TP-0903/DMSO Treatment of Embryos procedure. GFP/RFP Fluorescent Imaging Once MO injected embryos were treated and grown to 18 hpf, 1% low melt agar was used for agar mounting of embryos to be imaged under fluorescent microscope. Procedure involved making lined squares of Vaseline on glass slides. One drop of agar was added then the embryo was added. The embryo was covered with more agar to fill 11 the box and quickly orientated to the desired position for imaging before the solution became solid. Once position was achieved, embryos were immediately imaged. Fluorescence Activated Cell Sorting Using embryos treated previously with either drug or DMSO, 200-300 embryos were transferred to a large petri dish for dechorination with pronase. Once dechorinated, embryos were transferred to microcentrifuge tube. Egg water was removed and the embryos were incubated in 500 ul of Ca2+-free Ringer’s at RT for 10 min. Tube was gently inverted for 5 min. during this time. Ringers were removed and replaced with another 500 μl of Ringers. Using a micropipette, yolks were disrupted by pipetting up and down until embryos were gone. Tubes were placed for 5 min. on Nutator at 150 rpm at 37°C. Trypsin was prewarmed to 28°C during this time. Embryos were centrifuged at 1000 rpm for 45 seconds and supernatant was removed. Cells were resuspended in 500 μl of Ringers. Centrifugation step was repeated with a final addition of 500 μl of Ringers. Prewarmed Trypsin was added to tubes and incubated at 28°C for 1 hour. During this time, cells were disrupted with micropipette every 5 minutes to avoid clumping. Reaction was stopped with the addition of .833 μl of 2 M CaCl2 and 166 μl of FBS. Mixture was filtered through 40 um nylon mesh filter in a falcon tube. Centrifugation occurred at 3000 rpm for 3 minutes and supernatant was removed. Cells were resuspended in 500 μl of .9X PBS/5% FBS and brought to FACS facility. Once the cells were collected in Zymo brand RNA lysing buffer, they were snapfreezed via liquid nitrogen and stored in -80°C freezer. 12 RESULTS In order to verify a connection between ski and its response to TP-0903, RNA in situ hybridizations were used to observe the difference in expression of the gene between the control and treatment. The land marker gene, krox20, provided the visual red lines showing the location of the dorsal hindbrain region of the developing zebrafish embryos. The black darkened areas of the embryos signify the skiB expression. Looking at Figure 1, the control embryos show an obvious distinction of skiB expression compared to the drug treated embryos, especially in the case of Figure 1D at 18h hpf. Interestingly, the control showed expression of skiB was the highest between 11 and 14 hpf according to Figure 1A, 1B, and 1C. Whereas, skiB expression becomes almost nonexistent at 18 hpf for the control embryos observed in part D. Taking these observations to account, the link between skiB expression and TP-0903 looked promising. The difficulty in the whole mount imaging was trying to distinguish the difference in expression between Figure 1A, 1B, and 1C. Therefore, the project resorted to a different method by excavating the yolk from the zebrafish embryos. Taking the whole mount in situ embryos and converting into flat mounts, this procedure provided an alternative, clearer observation of the skiB expression without the interference of the yolk. Looking at the flat mount images, the gene expression was highest in the dorsal hindbrain region of the zebrafish at 12 and 14 hpf according to Figure 2. This result provided the evidence to continue treating the embryos at 10 hpf for four hours in TP-0903 and imaged later for fluorescence since the scheduling was easier to perform experiment. With the evidence for skiB responding to the drug, the project moved forward to observe the relevance of the gene in regards to EMT. Therefore, the 13 project resorted to using the novel Tg(snail1b:GFP);Tg(sox10:RFP) reporter line of zebrafish and skiB MO to knock down the expression in hopes of recovering a partial normal phenotype. Yet, the results from the fluorescence images are inconclusive as seen in Figure 3. Lastly, FACS results reveal the population of double positive cells expressing GFP and RFP are in much smaller amounts compared to the rest of the cells. The parameters set in Figure 4 were suggested by the operators of the FACS facility and prove beneficial for setting future parameters for later experiments. This knowledge is important in obtaining more embryos from the transgenic line of fish to increase the desired cell population that will be sufficient for RNA sequencing. 14 Figure 1. In situ hybridization of zebrafish embryos at multiple stages of development All embryos were developed to 10 hpf then treated with 7 μM TP-0903 at varying times. For each group, the left column represents the control and the right column represents the experimental. skiB expression is in black and krox20 expression is the red land marker. A. 1 hour treatment. B. 2 hour treatment. C. 4 hour treatment. D. 8 hour treatment. 15 Figure 2. Flat Mounts of RNA in situ Hybridization The figure provides a more focused perspective of the dorsal region of the hybridized zebrafish embryos where skiB expression is the highest between 12 and 14 hpf. SkiB expression is in the black areas, while krox20 expression is shown in red. The developing head of the zebrafish is in the area of where krox20 is shown. 16 Figure 3. Tg(snail1b:GFP);Tg(sox10:RFP) zebrafish with morpholinos and TP-0903 All images were taken at 18 hpf. A. No MO + DMSO (control). B. 6 ng MO + No TP0903. C. No MO + 7μM TP-0903. D. 6 ng MO + 7μM TP-0903 E. 9 ng MO + 7μM TP0903 F. 24 ng MO + 7μM TP-090. Minute recovery of control phenotype seen with neural crest folds formed at E. and F with high doses of MOs. 17 Figure 4. Parameters for Fluorescence Activated Cell Sorting (FACS) for GFP and RFP Cells GFP/RFP double positive (DP) cells are the population stuck in the neural tube need for to generate transcripts affected by TP-0903 from RNA extractions. 18 DISCUSSION The objectives of the project revolve around understanding the underlying retinoic acid dependent mechanisms of NC EMT. By doing this, further knowledge can be derived towards determining how TP-0903 affects certain transcripts to initiate the increased biosynthesis of retinoic acid, which inhibits EMT as a result. Moreover, the project serves to identify the significance of ski due to its relation with EMT and TGF-β based on literature. The results bring forth details about a candidate gene, ski, that proves to be overexpressed by TP-0903. Knocking down this expression seems to produce inconclusive results at varying concentrations of MO. However, the embryos are able to continue development at the high concentration of 24 ng MO. Therefore, the project continues to increase the dosage of MO in hopes of producing some effect on returning normal EMT. Proving the significance of ski provides another piece of the puzzle towards understanding the relation of EMT and retinoic acid, while also uncovering further information on the specifics of how TP-0903 inhibits EMT based on the Stewart Lab’s novel EMT reporter line of zebrafish. Verifying that ski was affected by TP-0903 proves to be an important part of the project by initiating the following experiment involving MO. The overexpression of ski is shown to be in the area of the dorsal hindbrain region of the zebrafish embryos according to the landmark markers of krox20. Expression of skiB is observed to be relatively high in the tail region, but that is not the area of interest since the neural crest is located in the dorsal hindbrain region. The expression of skiB is the highest when the embryos have been incubated for four hours in the presence of the drug. An important observation can be added towards seeing whether the drug worked besides a higher expression—looking 19 at the fatty component of the yolk near the tail reveals an elongation. Flat mounts provide a better perspective without the interference of the color of the embryo in images. This knowledge of treatment is important for the MO experiment as the embryos could now be stopped at four hours of drug treatment, thus saving time. Next, returning the normal phenotype of EMT is attempted by using MOs. Figure 3A serves as the control showing the normal function of neural crest cells delaminating from the neural tube in waves. In the presence of the drug, the waves fuse together and the cells are stuck in the neural tube indicated by the higher presence of GFP in the dorsal hindbrain region viewed in the images. Upon adding the MO without the drug show normal function of the neural crest delamination. However, allowing the embryos to develop further show a ventralized phenotype as indicated by the paper by Dr. K and colleagues. This provides an important observation that skiB serves as a vital gene for normal development. The severity of the ventralization increases with the amount of MO added. Due to inexperience with imaging and agar mounting, the GFP/RFP images of the embryos can be improved and the colors are a bit off. However, the effects of the drug and the MO remain inconclusive since recovery of the normal phenotype is not observed. TP-0903 is a multikinase inhibitor and it’s known effect is increasing RA synthesis through some unknown method. Full recovery of the normal phenotype is highly unlikely since the project is targeting one gene, but the expected results should be some kind of effect on the function of EMT when using the MOs. Therefore, increasing the dosage of MO remains part of the future work in addition to verifying other genetic transcripts. 20 In order to generate more transcripts differentially regulated by TP-0903, FACs is used to sort out the population of cells expressing snai1B:GFP and sox10:RFP. The reason the project is interested in this population is because these cells are stuck between converting from epithelial cells to neural crest cells and may have the genes that are affected by the drug. By sorting out this specific population and performing RNA sequencing, the genes that are expressed the most in these cells can be uncovered. However, the results show that the amount of cells expressing GFP and RFP is small compared to the rest of the population. Thus, the project continues forward by identifying fish that are homozygous in both genes and generating more embryos by mating more fish. Overall, the research shows TP-0903 can induce ski in vivo starting at 11 hpf. The study shows the connection between RA and the induction of skiB where the gene may be important for delamination of neural crest cells. Although the fluorescent images are inconclusive, further research remains to be accomplished in verifying ski and other transcripts. 21 ACKNOWLEDGEMENTS Infinite gratitude goes towards Dr. Stewart and his lab for the years of laughs, knowledge, and overcoming obstacles. I would also like to specifically thank Dr. Morrison for providing the foundation of my research skills, Dr. Jimenez for passing on an exciting project, Dr. Modzelewska for her kind words of encouragement, and numerous undergraduates who kept me motivated. I am grateful for the opportunity to be surrounded by inspiring people. Thank you UROP and BioURP for funding. 22 REFERENCE 1. Polyak, K. and Weinberg, R. A. (2009). Transitions between epithelial and mesenchymal states: acquisition of malignant and stem cell traits. Nature Reviews Cancer 9, 265-273. 2. Friedl P. (2004). Prespecification and plasticity: shifting mechanisms of cell migration. Curr Opin Cell Biol 16, 14-23. 3. Lim, J. and Thiery, J. P. (2012). Epithelial-mesenchymal transitions: insights from development. Development 139, 3471-3486. 4. Khan, M. I., Adhami, V. M., Lall, R. K., Sechi, M., Joshi, D. C., Haidar, O. M., … Mukhtar, H. (2014). YB-1 expression promotes epithelial-to-mesenchymal transition in prostate cancer that is inhibited by a small molecule fisetin. Oncotarget, 5(9), 2462–2474. 5. Theiry, J. P. and Sleeman J. P. (2006). Complex networks orchestrate epithelial–mesenchymal transitions. Nature Reviews Molecular Cell Biology 7, 131-142. 6. Gupta, P. B., Kuperwasser, C., Brunet, J. P., Ramaswamy, S., Kuo, W. L., Gray, J. W., Naber, S. P. and Weinberg, R. A. (2005). The melanocyte differentiation program predisposes to metastasis after neoplastic transformation. Nat Genet 37, 1047-54. 7. Tania, M., Khan M. A. and Fu, J., (2014) Epithelial to mesenchymal transition inducing transcription factors and metastatic cancer. Tumor Biol 35(8), 7335-42. 23 8. Green, S. A., Simoes-Costa, M. and Bronner, M. E. (2015). Evolution of vertebrates as viewed from the crest. Nature 520, 474-82. 9. Simoes-Costa, M. and Bronner, M. E. (2015). Establishing neural crest identity: a gene regulatory recipe. Development 142, 242-57. 10. Mollard, A., Warner, S. L., Call, L. T., Wade, M. L., Bearss, J. J., Verma, A., ... Bearss, D. J. (2011). Design, synthesis, and biological evaluation of a series of novel AXL kinase inhibitors. ACS Medicinal Chemistry Letters 2(12), 907-912 11. Jimenez, L., Wang, J. Morrison, M. A., Whatcott, C., Soh K. K., Warner, S., Bearss D., Jette, C. A., Stewart R. A. (2016) Phenotypic chemical screening using zebrafish neural crest reporters identifies retinoid acid as an inhibitor of epithelial morphogenesis. Disease Models & Mechanisms 9(4), 389-400. 12. Xu, Q., Lucio-Cazana, J., Kitamura, M., Fine, L. G., Norman, J. T. (2004) Retinoids in nephrology: promises and pitfalls. Kidney Int 66(6), 2119-31 13. Xu, Q. and Kopp, J. (2012) Retinoid and TGF-β families: Crosstalk in development, neoplasia, immunity and tissue repair. Semin Nephrol 32(3), 287-294. 14. Nasr, R., Guillemin, M. C., Ferhi, O., Soilihi, H., Peres, L., Berthier, C., Rousselot, P., Robledo-Sarmiento, M., Lallemand-Breitenbach, V., Gourmel, B. et al. (2008). Eradication of acute promyelocytic leukemia-initiating cells through PML-RARA degradation. Nat Med 14, 1333-42. 15. Reynolds, C. P. (2000). Differentiating agents in pediatric malignancies: retinoids in neuroblastoma. Curr Oncol Rep 2, 511-8. 24 16. Uray, I. P., Dmitrovsky, E. and Brown. P. H. (2015) Retinoids and rexinoids in cancer prevention: from laboratory to clinic. Semin Oncol 43(1), 49-64. 17. Xu, J., Lamouille, S. and Derynck, R. (2009) TGF-beta-induced epithelial to mesenchymal transition. Cell Res 19(2), 156-72. 18. Yang, H., Zhan, L., Yang, T., Wang, L., Li, C., Zhao, J. ... Zhang, H. (2015). Ski prevents TGF-β-induced EMT and cell invasion by repressing SMAD- dependent signaling in non-small cell lung cancer. Oncology Reports, 34, 8794. 19. Elliot, R. L. and Blobe, G. C. (2005) Role of transforming growth factor Beta in human cancer. J Clin Oncol 23(9), 2078-93. 20. Akiyoshi, S., Inoue, H., Hanah, J., Kiyoshi, K., Nemoto, N. ... Kawabata, M. (1999) c-Ski Acts as a Transcriptional Co-repressor in Transforming Growth Factor-β Signaling through Interaction with Smads. J Biol Chem 4(49), 35269-77 21. Luo, K. (2004) Ski and SnoN: negative regulators of TGF-beta signaling. Curr Opin Genet Dev 14(1), 65-70 22. Shinagawa, T., Nomura, T., Colmenares, C., Ohira, M., Nakagawara, A. and Ishii, S. (2001) Increased susceptibility to tumorigenesis of ski-deficient heterozygous mice. Oncogene 20(56), 8100-8 23. Dahl, R., Kieslinger, M., Beug, H. and Hayman, M. J. (1998) Transformation of hematopoietic cells by the Ski oncoprotein involves repression of retinoic acid receptor signaling. Cell Biology 95(19). 11187-92 25 24. Melling, M. A., Friendship, C. R., Shepherd, T. G. and Drysdale, T. A. (2013) Expression of Ski can act as a negative feedback mechanism on retinoic acid signaling. Dev Dyn 242(6), 604-13 25. Christopher, K. (2000) Ectopic expression of c-ski disrupts gastrulation and neural patterning in zebrafish. Elsevier. 95, 147-162.