Discovery of a Small Molecule TUBB3/Βiii-Tubulin Modulator in Lung Cancer

Discovery of a small molecule TUBB3/βIII-tubulin modulator in lung cancer

Felicity Chao Lin Kao

A thesis submitted in fulfilment of the requirements for the degree of Doctor of Philosophy

School of Women’s and Children’s Health Faculty of Medicine The University of New South Wales

March 2016 TH E UNIVERSITY OF NEW SOUTH WALES

Thesis/Dissertation Sheet

Surname or Family name: Kao

First name: Felicity Other name/s: Chao Lin

Abbreviation for degree as g iven in the University calendar: PhD

School: School of Women's and Children's Health Faculty: Faculty of Medicine

Title: Discovery of a small molecule TU883/I311 1·tubulin modulator in lung cancer

Abstract Non-small Cell Lung Cancer (NSCLC) survival rates are dismal and chemotherapy resistance is a significant clinical problem. 13111-tubulin (encoded by TUBB3 gene) is aberranlly expressed and is associated with chemoresistance and tumour aggressiveness in NSCLC, where it has been identified as a bona fide target for chemosensitisation. Currently, there is no commercially available TUB83J13111-tubulin inhibitor. Regulation of 13111 -tubulin is poorly understood, making it difficult to target this protein. We sought to identify a chemical modulator of TU883J1311Hubulin expression, as it will be a valuable research tool to probe TU883J!3111-tubulin regulation . A novel small molecule TU883/13111-tubulin enhancer, WECC0017371 , was identified in our high throughput screen, based on its ability to modulate TUBB3 promoter activity. WECC001 7371 demonstrated the ability to significantly enhance TUBB3 mRNA and 13111-tubulin protein expression in a time· and dose-dependent manner. Additionally, WECC0017371 did not alter microtubule morphology but enhanced 13111-tubuli n immunostaining in two independent NSCLC cell lines, H460 and H1299, compared to control. Importantly, WECC0017371 enhanced 13111 -tubulin aberrant expression was functional and led to a significant decrease in in vitro sensitivity to DNA-damaging and tubulin-binding agents. Using chemical modifications, the efficacy, potency and selectivity of WECC0017371 was further improved for the study of TU883/13 111-tubulin regulation. The superior WECC0017371 analogue, ENTD014 was developed into an affinity chromatography probe for future studies to identify intracellular proteins that it binds to enhance TU883/1311Hubulin expression. Negative control probes were also developed to identify and eliminate proteins that bind non-specifically to the affinity probe. To understand how WECC001 7371 is exerting its effects on TUBB3 expression. NSCLC cells were treated, mRNA isolated and a microarray and bioinformatics approach was used. Coordinated gene expression changes in multiple members of the mitotic and p53 signalling pathways were detected in WECC0017371-treated cells. In conclusion, a novel chemical research tool has been developed in this thesis to better understand the molecular mechanisms underlying TUBB3/13111-tubulin regulation in NSCLC and 13111-tubulin-mediated pathobiology.

Declaration relating to disposition of project thesis/dissertation

I hereby grant to the University of New South Wales or its agents the right to archive and to make available my thesis or dissertation in whole or in part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all property rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation.

I also authorise U n iv~ i c rofilms to use the 350 word abstract of my thesis in Dissertation Abstracts International.------/ ../1 ~

The University recognises that there may be exceptional circumstances requ1nng restrictions on copying or conditions on use. Requests for restriction for a period of up to 2 years must be made in writing. Requests for a longer period of restriction may be considered in exceptional circumstances and require the approval of the Dean of Graduate Research.

FOR OFFICE USE ONLY Date of completion of requi rements for Award:

'I hereby grant the University of New South Wales or its agents the right to archive and to 111,ake available my thesis or dissertation in whole or part in the University libraries in all forms of media, now or here after known, subject to the provisions of the Copyright Act 1968. I retain all proprietary rights, such as patent rights. I also retain the right to use in future works (such as articles or books) all or part of this thesis or dissertation. I also authorise University Microfilms to use the 350 word abstract of my thesis in Dissertation Abstract International (this is applicable to doctoral theses only). I have either used no substantial portions of copyright material in my thesis--or I have obtained permission to use copyright material; where permission has not been granted I have applied/will apply for a partial restriction of the digital copy of my thesis or dissertation.'

Signed

Date ...... 4.. 1.. ~.':?. .(. ~ .~······················· ···· ··········· ·····

AUTHENTICITY STATEMENT

'I certify that the Library deposit digital copy is a direct equivalent of the final officially approved version of my thesis. No emendation of content has occurred and if there are any minor variations in formatting, they are the result of the conversion to digital format.'

Signed

Date ...... ~ .f. .l..~!.! .. l::? ...... :...... Table of Contents

Table of Contents ...... 1

Publications arising from this thesis ...... 6

Awards ...... 7

Acknowledgements ...... 8

List of Abbreviations ...... 10

Chapter 1 Introduction ...... 12

1.1 Thesis overview ...... 12

1.2 Lung cancer ...... 15

1.2.1 Staging of lung cancer ...... 16

1.2.2 Treatment strategies for NSCLC ...... 16

1.3 Tubulin/ microtubule system ...... 22

1.3.1 Structure and function of microtubules ...... 22

1.3.2 The β-tubulin family ...... 26

1.3.3 Microtubule interacting proteins ...... 30

1.3.4 Posttranslational modification ...... 31

1.3.5 Tubulin binding agents for NSCLC treatment ...... 34

1.3.6 Drug resistance in NSCLC and βIII-tubulin ...... 37

1.4 βIII-tubulin: expression and regulation ...... 42

1.4.1 Expression and function of βIII-tubulin ...... 42

1.4.2 Gene structure of human TUBB3 gene ...... 49 1

1.4.3 Regulation of TUBB3 expression in non-neoplastic cells ...... 52

1.4.4 Regulation of TUBB3 expression in cancer cells ...... 57

1.5 High throughput drug discovery ...... 70

1.5.1 βIII-tubulin as a drug target...... 70

1.5.2 High throughput screening ...... 71

1.5.3 Assay types ...... 72

1.5.4 Experimental design and planning ...... 74

1.5.5 Assay development and optimisation for HTS ...... 75

1.5.6 Execution of HTS ...... 77

1.5.7 Validation of drug candidate and lead optimisation ...... 78

1.6 Thesis perspective and significance ...... 79

1.7 Thesis aims ...... 80

Chapter 2 Materials and Methods ...... 81

2.1 Materials ...... 81

2.1.1 Cytotoxic drugs and other chemicals ...... 81

2.1.2 Tissue culture ...... 81

2.1.3 Plasmid sequencing and restriction digest ...... 82

2.1.4 Plasmid transfection and single clone selection ...... 82

2.1.5 Bioluminescence assay and high throughput drug screening ...... 83

2.1.6 Protein isolation and Western blotting...... 83

2.1.7 Electrophoresis ...... 84

2.1.8 RNA isolation, cDNA synthesis and real time-PCR ...... 84

2.2 Methods ...... 85

2.2.1 Maintenance of H460 NSCLC cells ...... 85

2.2.2 Validation of human TUBB3 promoter-Renilla Luciferase vector

construct ...... 85

2.2.3 Generation of H460 single clones stably expressing human TUBB3

promoter ...... 96

2.2.4 High throughput compound screening ...... 98

2.2.5 Cell viability and proliferation assays ...... 100

2.2.6 Quantitative real-time polymerase chain reaction (QRT-PCR) ...... 101

2.2.7 Western blotting ...... 104

2.2.8 Cell cycle analysis by flow cytometry ...... 107

2.2.9 Microscopy ...... 108

2.2.10 Drug-treated clonogenic assays ...... 109

2.2.11 Microarray ...... 110

2.2.12 Bioinformatics analysis of microarray data ...... 111

2.2.13 Statistical analysis...... 112

Chapter 3 Development of A Cell-Based Screen to Identify Small Molecule Inhibitors of

TUBB3 Promoter Activity ...... 113

3.1 Introduction ...... 113

3.2 Results ...... 114

3.2.1 Generation and clonal selection of H460/TUBB3p-luc and H460/GAPDHp-

luc cells 114

3.2.2 Preliminary screen- to validate H460/TUBB3p-luc cells for high throughput

screening ...... 118

3.2.3 Pilot screen- final validation of H460/TUBB3p-luc cells for high throughput

screening ...... 121

3.2.4 High throughput screening of TUBB3 promoter repressing agents...... 129

3.1 Discussion ...... 147

Chapter 4 Effect of Small Molecule and Bioactive Hits on TUBB3 Gene and βIII-Tubulin

Protein Expression ...... 149

4.1 Introduction ...... 149

4.2 Results ...... 150

4.2.1 Validation of bioactive hit RITA ...... 150

4.2.2 Validation of hit compound WECC0018639 ...... 164

4.2.3 Validation of hit compound WECC0017371 ...... 176

4.3 Discussion ...... 209

Chapter 5 Structure Activity Relationship Study of Imidazopyridines and Development of Affinity-Chromatography Probes for Chemical Proteomics ...... 213

5.1 Introduction ...... 213

5.2 Results ...... 214

5.2.1 WECC0017371 and a virtual library screen ...... 214

5.2.2 Structure-activity relationship studies of WECC0017371 analogues as

TUBB3/βIII-tubulin enhancers ...... 217

5.2.3 Development of a TUBB3/βIII-tubulin affinity chromatography probe and

control 253

5.3 Discussion ...... 274

Chapter 6 Gene expression changes following treatment with the TUBB3-enhancing

WECC0017371 small molecule ...... 278

6.1 Introduction ...... 278

6.2 Results ...... 279

6.2.1 Microarray analysis of H460 cells treated with WECC0017371 ...... 279

6.2.2 Identification of differentially expressed genes and pathway analyses of

naïve and WECC0017371-treated H460 cells at 48 h...... 279

6.2.3 Identification of differentially expressed genes and pathway analyses of

naïve and WECC0017371-treated H460 cells at 72 h...... 284

6.3 Discussion ...... 302

Chapter 7 Concluding Remarks and Future Directions ...... 309

References ...... 313

Appendices ...... 338

Appendix I ...... 338

Appendix II ...... 341

Appendix III ...... 344

Appendix IV ...... 345

Appendix V ...... 348

Publications arising from this thesis

Felicity CL Kao and Maria Kavallaris. How is TUBB3 regulated in cancer?

Manuscript under review

Abstracts

Felicity CL Kao, Tim Failes, Greg Arndt, Murray Norris, Maria Kavallaris. Discovery of small molecule inhibitors targeting microtubule proteins for the treatment of lung cancer. Lowy Drug Discovery Symposium 2013 University of New South Wales Poster

Poster presentation.

Felicity CL Kao, Tim Failes, Greg Arndt, Murray Norris, Maria Kavallaris. Discovery of small molecule TUBB3/βIII-tubulin modulator in lung cancer. Lowy Drug Discovery

Symposium 2015 University of New South Wales Poster presentation.

Felicity CL Kao, Tim Failes, Greg Arndt, Murray Norris, Maria Kavallaris. Discovery of small molecule TUBB3/βIII-tubulin modulator in lung cancer. American Association for Cancer Research annual meeting, Philadelphia, USA 2015 Poster presentation.

Awards

NHMRC Postgraduate Research Scholarship

Australian Postgraduate Award

Lung Foundation Australia Lung Cancer National Program Postgraduate Grant-in-Aid

The LH Ainsworth Cancer Research Scholarship

Faculty of Medicine’s 3 Minute Thesis Competition (Runner up)

School of Medical Science, Faculty of Medicine Scholarship

Acknowledgements

Immeasurable appreciation and deepest gratitude for the help and support are dedicated to the following people who in one way or another have contributed in making this study possible.

I would like to express my gratitude to my supervisor, Prof. Maria Kavallaris, for her brilliant mentorship and guidance, and specifically, for her warmth, trust and the opportunity to collaborate with professionals from multiple disciplines. I feel privileged and fortunate to have worked under her supervision and I thank her for encouraging me to think independently. I would also like to thank my cosupervisor, Prof. Murray Norris, for being there for me whenever I needed his advice and feedback. A huge thank you to past and present members of the Tumour Biology & Targeting programme. You are an amazing bunch and you have all helped shape the scientist I am today. Thank you for your support and for making my time here so enjoyable.

I have been incredibly lucky to collaborate with multiple talented scientists from diverse backgrounds. Without their help and expertise, this project would not have been possible.

I thank Dr. Tim Failes and Dr. Greg Arndt, for showing me the ropes with high- throughput drug screening. I thank A/Prof. Jonathan Morris for his guidance and for providing constructive feedback for my chemical probe work. A massive thank you to

Elysha Taylor, who performed all the chemical modifications in this thesis. Chelsea

Mayoh, I thank her for performing bioinformatics analyses for my microarray data. Dr.

Sharon Sagnella, I thank her for helping me with cell imaging and confocal microscopy and for being there for me every step of the way. A special thank you to Frieda, Walter,

Elysha, Jonathan and Wee for proofreading my thesis, their constructive comments and feedback made the write-up easier and smoother.

Outside research, I would like to thank my fabulous friends, Ann-Louise Nguyen, Chris

Jaeger, Helga Prochesca and Peggy Liu. Thank you for keeping me in touch with the human society. I would not have enjoyed this journey as much, if it were not for your support, laughter and friendships. An enormous thank you goes to Sara Chesterman, Ben

Glagovs and Jacob Glagovs, for keeping me sane with laughter, homemade breakfast and many great conversations! I would also like to thank my roommate, Bella Garson, for her warmth and kindness during my final year of study. Walter Muskovic, thank you for your loving support, encouragement and for believing in me! Thank you for keeping me grounded and for being there for me through ups and downs. Likewise, I hope I have and will continue to provide the same for you. A big thank you to Walt’s family, especially

Clara, Frances, Lucia and nonno Lino, for their warmth and for cheering me along the final stretch of my PhD. Last but definitely not least, I would like to express my deepest gratitude and appreciation to my parents, Richard and Sabrina, and brother, Philip, for their continual support, unconditional love and encouragement during my PhD. I love them all dearly.

List of Abbreviations

AR Androgen receptor ARE Androgen response elements ALK Anaplastic lymphoma kinase β2M β2 microglobulin bHLH Basic helix-loop- helix BSA Bovine serum albumin ChIP Chromatin immunoprecipitation DMSO Dimethyl sulfoxide ECM Extracellular matrix EGFR Epidermal growth factor receptor EML4 Echinoderm microtubule-associated protein like 4

ER Oestrogen receptor FCS Foetal calf serum FDA Food and Drug Administration GAPDH Glyceraldehyde-3-Phosphate Dehydrogenase GBP1 Guanylate-binding protein 1 GTP Guanosine-5'-triphosphate

GDP Guanosine diphosphate

H460/TUBB3p-luc cells H460 cells stably expressing TUBB3 promoter- Renilla luciferase expression constructs H460/GAPDHp-luc cells H460 cells stably expressing GAPDH promoter- Renilla luciferase expression constructs HIF Hypoxia-inducing factor HRE Hypoxia response element HTS High throughput screening

MAP Microtubule associated protein mRNA Messenger RNA miRNA MicroRNA

NSCLC Non-small cell lung cancer

PBS Phosphate buffered saline PgP P-glycogprotein pRB Retinoblastoma tumour suppressor protein PTM Posttranslational modification RE-1 Repressor element 1 REST Repressor element 1 silencing transcription factor

RNA ribonucleic acid RPMI Rosewell Park Memorial Institute RT-PCR real-time polymerase chain reaction

SAR Structure activity relationships SCLC Small cell lung cancer SEM Standard error of the mean SEMA6A Semaphorin-6A

SM Skim milk SOX9 Sex determining region Y-box 9 TBA Tubulin-binding agent TBST Tris-buffered saline- 0.01% Tween-20 TKI Tyrosine kinase inhibitor TNM Tumour, node, metastasis

UTR Untranslated region

Chapter 1

Introduction

1.1 Thesis overview

Lung cancer is the leading cause of cancer death worldwide. Recent cancer data estimated

1.83 million new cases of lung cancer and 1.59 million deaths from lung cancer globally

(Stewart, 2014). Lung cancer accounts for more deaths than prostate, breast and colon cancer combined, the next three leading causes of cancer death (Siegel et al., 2011).

Advanced stage non-small cell lung cancer (NSCLC) accounts for 80% of cases and more than half of these patients have developed metastases at the time of diagnosis, and so chemotherapy remains the most common treatment option. First line chemotherapy for

NSCLC commonly involves a combination of a tubulin-binding agent (TBA) and a DNA damaging agent. Despite this treatment, lung cancer remains one of the most lethal cancers. The relative 5-year survival rate has remained at <15% for the past three decades with a median survival of 8 to 10 months (Howlader et al., 2014). Resistance, both intrinsic and acquired, to current chemotherapeutics is one of the major barriers to improving long term-outcome for NSCLC patients. Clearly, there is a need for new treatment strategies to combat drug resistance and improve efficacy of existing therapies in the treatment of lung cancer.

Alterations in the expression of microtubule proteins in tumours are important contributors to chemotherapy resistance (reviewed in Kavallaris, 2010; Parker et al.,

2014). Microtubules are cytoskeletal proteins that comprise α- and β-tubulin heterodimers

(αβ-heterodimers). These highly dynamic polymers are constantly growing and shortening. Microtubule dynamics are involved in many important cellular processes, in

12 particular cell division, making it an attractive target in anticancer therapies (reviewed in

Jordan and Wilson, 2004). Clinically important chemotherapy agents, such as taxanes and vinca alkaloids all exert their toxic effect by binding the β-subunit of the αβ-heterodimers, disrupting microtubule dynamics and inducing mitotic arrest and apoptosis (reviewed in

Kavallaris, 2010).

There are seven different isotypes of β-tubulin, displaying tissue-specific expression. For example, βI-tubulin is constitutively expressed in many tissues, whereas βIV-tubulin expression is restricted to hematopoietic tissues. βII- and βIII-tubulin are highly expressed in neuronal tissues and in the case of βIII-tubulin, the Sertoli cells of the testes (reviewed in Kavallaris, 2010). In cancer, βIII-tubulin is expressed in both neuronal and non- neuronal tumours. In tumours of neuronal origin, βIII-tubulin expression is constitutive and linked to morphological changes of neuronal differentiation and reduced tumour proliferation (Caracciolo et al., 2010; Katsetos et al., 2003). In contrast, in non-neuronal tumours, βIII-tubulin expression is aberrant and represents dedifferentiation associated with malignant transformation. Aberrant βIII-tubulin expression is now recognised as a key determinant of drug resistance and tumour aggressiveness in several cancers, including NSCLC (Hayashi et al., 2009), breast (Shalli et al., 2005), prostate (Ploussard et al., 2010) and ovarian (Mozzetti et al., 2005; Roque et al., 2014) cancers. There is strong clinical evidence in lung, ovarian and breast cancer, that patients with aberrant

βIII-tubulin expression exhibit poorly differentiated tumour tissue, high grade malignancy, shorter disease progression, unfavourable prognosis and worse overall survival (reviewed in Karki et al., 2013; Mariani et al., 2011). Laboratory data have established the importance of βIII-tubulin in regulating sensitivity to chemotherapy in multiple epithelial cancers. In NSCLC, overexpression of βIII-tubulin has been demonstrated to mediate chemosensitivity to both TBAs and DNA-damaging agents in 13 vitro and in vivo (Gan et al., 2007; McCarroll et al., 2010). Knockdown of βIII-tubulin increased drug sensitivity to chemotherapeutic agents via an increase in apoptosis in vitro and in vivo in NSCLC (Gan et al., 2011; Gan et al., 2010; Gan et al., 2007; McCarroll et al., 2010). Despite increasing evidence linking aberrant βIII-tubulin expression to chemosensitivity in NSCLC, the biology of how this protein is regulated in NSCLC is poorly understood.

To date, studies of βIII-tubulin expression have primarily focused on neuronal cells, in which it is constitutively expressed, and commonly used as a neuronal differentiation marker. Little is currently known about the regulatory mechanisms responsible for its ectopic expression in NSCLC. The purpose of this project is two-fold: firstly, to identify small molecule specific βIII-tubulin modulators and secondly, to improve our understanding of the molecular mechanism of βIII-tubulin’s regulation in NSCLC. These studies will assist in identifying promising new drug targets and improving current therapies to increase the long-term survival of NSCLC patients.

1.2 Lung cancer

Lung cancer arises from the cells of the respiratory epithelium and is divided into two broad categories: small cell lung cancer (SCLC) and non-small-cell lung cancer

(NSCLC). SCLC is a highly malignant tumour derived from cells exhibiting neuroendocrine characteristics and accounts for 15% of all lung cancer cases. NSCLC accounts for the remaining 85% of lung cancer cases and is further divided into three major histological subtypes: adenocarcinoma, squamous cell carcinoma and large cell carcinoma (Travis et al., 2013). The incidence of adenocarcinoma has increased greatly in the past few decades, making it the most prevalent type of NSCLC, followed by squamous cell carcinoma and large cell carcinoma (Howlader et al., 2014). Recent analysis of Canadian and European cases indicated that adenocarcinoma is the most prevalent lung cancer subtype in never-smokers in women and squamous cell carcinoma is predominately found in male smokers (Pesch et al., 2012). Cigarette smoking, including second-hand smoke, is the dominant risk factor for lung cancer worldwide, accounting for 90% of lung cancer cases in high-income countries. A wide range of other agents are also recognised risk factors for lung cancer, such as indoor and industrial air pollutants, diesel engine exhaust and asbestos. The World Health Organisation estimates that lung cancer deaths worldwide will continue to rise as a result of increasing global tobacco use in developing nations.

Despite some improvement in frontline treatments over the past few decades, lung cancer survival rates remain dismal. In SCLC, the five-year survival rates for localised and distant disease are 27% and 2.8%, respectively (Howlader et al., 2014). Only 4% of SCLC cases are diagnosed at the localised stage, while the remaining 96% of SCLC patients present with advanced-stage disease at diagnosis. In NSCLC, the five year survival rate

15 for localised disease is 58.7%, but only 19% of NSCLC cases are diagnosed at this early stage (Howlader et al., 2014). The remaining 81% of NSCLC patients present with advanced-stage disease at diagnosis, with more than half of these with distant metastases.

This significantly larger patient population has a dismal five-year survival rate of only

4.7%.

1.2.1 Staging of lung cancer

Clinical staging of NSCLC patients is critical to determining the operability, treatment options and prognosis of the disease. NSCLC is staged using the tumour, node, metastasis

(TNM) classification and it is considered the most accurate prognosis assessment tool.

The T component describes the size of the primary tumour and the extent of local invasion. The N component describes the degree of involvement of regional lymph nodes and the M component describes the presence or absence of metastases (Sobin and

Wittekind, 2002). A number is assigned to each letter to indicate the size and extent of the primary tumour and the degree of metastasis (Rami-Porta et al., 2014). Different combinations of the T, N, and M are put together to stage the tumour into one of four stages of disease: I, II, III and IV with stage I reflecting the best prognosis and stage IV the worst, as summarised in Table 1.1.

1.2.2 Treatment strategies for NSCLC

Current treatment for NSCLC typically comprises surgery, radiotherapy, chemotherapy and/or targeted therapy. Therapeutic decisions for lung cancer are dependent on disease stage, histological type and molecular characteristics of tumours. Surgical resection is only available for functionally fit patients with early stage localised disease 16

Table 1.1 Staging of lung cancer

Five-year Primary tumour TNM stage survival Description size rate Stage IA (T1N0M0) ≤ 2cm 61% Tumours are confined to the lung, with no evidence of lymph node Stage IB (T2N0M0) > 2cm but ≤ 3cm 38% involvement or metastases. T1 primary tumour with intrapulmonary and ipsilateral hilar lymph Stage IIA (T1N1M0) > 3cm and ≤ 5cm 34% node metastasis. T2N1M0: > 5cm Stage IIB (T2N1M0 or Tumours are confined to the lung with or without localised hilar and ≤7cm; 22% T3N0M0) lymph node involvement. T3N0M0: >7cm Stage IIIA (T3N1M0, T1N2M0, T2N2M0 and 13% Tumour with no evidence of metastasis outside thorax. Two or more T3N2M0) >7 cm individual tumour nodules are present in the same lobe of a lung. Stage IIIB (T4 any N M0 Tumour invades a main bronchus. 1-8% and any T N3M0) All tumours that present with metastasis. M1 is further classified into Stage IV (any T any N M1a and M1b subclasses. The M1a subclass consists of tumours with >7 cm 1% M1) pleural nodules or malignant pleural effusion. M1b involves distant metastasis.

(D'Addario and Felip, 2009). A large number of randomised trials and meta-analyses have shown that supplementing surgery with postoperative/adjuvant chemotherapy improves overall and recurrence-free survival in patients (Pignon et al., 2008; Vansteenkiste et al.,

2013). Based on this strong body of evidence, adjuvant chemotherapy is considered the standard treatment for patients with localised early-stage NSCLC. One recent meta- analysis has suggested that preoperative/neoadjuvant chemotherapy may also benefit overall survival in patients with resectable NSCLC (mainly stage IB-IIIA) (Burdett et al.,

2014). However, a limited number of trials to date have compared the effect of neoadjuvant chemotherapy and surgery, surgery alone or adjuvant chemotherapy with surgery. Therefore, it is unclear whether the timing of administration of chemotherapy, neoadjuvant or adjuvant, affects survival. Recent studies observed no difference in overall survival between the adjuvant and neoadjuvant arm (Felip et al., 2010; Lim et al., 2009).

At present, adjuvant chemotherapy remains the treatment of choice in resectable NSCLC.

For advanced stage NSCLC patients (stage IIIB and IV), chemotherapy and radiotherapy may be the only treatment option. A large meta-analysis of randomised trials demonstrated a survival advantage in favour of cisplatin, a platinum-based therapy compared with best supportive care alone (Spiro et al., 2004). Furthermore, randomised trials have established that response rates, time to progression and survival rates are improved with combination therapy over single-agent therapy (Gatzemeier et al., 2000;

Lilenbaum et al., 2005; Sandler et al., 2000; Sederholm, 2002; Wozniak et al., 1998). The modest response rates and survival benefits achieved with cisplatin-based therapy have sparked the search for new agents with improved activities and more effective drug combinations. For example, Rudd et al. found that the combination of carboplatin plus gemcitabine conferred a significantly longer survival and better quality of life in patients with advanced NSCLC, than cisplatin plus mitomycin and ifosfamide (Rudd et al., 2005). 18

Today, platinum-based therapy (carboplatin, cisplatin, oxaliplatin) is the mainstay of chemotherapy for advanced and metastatic NSCLC, and is usually given in combination with a TBA (paclitaxel, docetaxel, vinorelbine), a camptothecin analogue (irinotecan, topotecan), gemcitabine or pemetrexed (Chang, 2011; Pfister et al., 2004; Spira and

Ettinger, 2004). Duration of chemotherapy has also been addressed, where 3-4 cycles of chemotherapy proved to be equally effective, with decreased toxicity, compared with 6 cycles (Rossi et al., 2014; Smith et al., 2001; Socinski et al., 2002). Docetaxel, pemetrexed and a combination of gemcitabine and vinorelbine are used as second-line chemotherapy for patients with recurrent or refractory disease following initial platinum- containing chemotherapy regimens (Barlesi et al., 2006; Fossella et al., 2000).

Historically, NSCLC was considered to be a single disease entity and the same treatment has been used for all patients for the past few decades. The median overall survival was persistently less than 12 months and a therapeutic plateau had been reached. NSCLC was only recognised as a collection of diseases that can be classified according to their genetic abnormalities in the last few years (Travis et al., 2013). Multiple mutations occur in

NSCLC and the spectrum of genetic mutations is very different between smokers and never-smokers, as well as between squamous cell carcinoma and adenocarcinoma (Ding et al., 2008; Sun et al., 2007). For example, epidermal growth factor receptor (EGFR) mutations are almost exclusively seen in never-smokers with lung adenocarcinomas (Sun et al., 2007). This identification of distinctive genetic abnormalities between different types of NSCLC has enabled personalised medicine approaches for lung cancer, due to major differences in treatment responses. The first breakthrough towards personalised treatment took place in 2004 when two landmark studies reported the presence of activating mutations in the tyrosine kinase domain of EGFR in patients who had a dramatic response to the EGFR tyrosine kinase inhibitor (TKI), gefitinib (Lynch et al., 19

2004; Paez et al., 2004). Since then, multiple randomised studies have shown that, compared to standard chemotherapy, EGFR TKIs (gefitinib, erlotinib and afatinib) significantly improve the response rates to treatment and progression-free survival in patients with activating mutations in EGFR (Maemondo et al., 2010; Mitsudomi et al.,

2010; Mok et al., 2009; Rosell et al., 2012; Zhang et al., 2012b; Zhou et al., 2011). This phenomenon was not observed in patients without mutations in EGFR. Gefitinib is now approved in 75 countries, as first-line targeted therapy for advanced adenocarcinoma patients with activating mutations in EGFR.

In lung cancer, the echinoderm microtubule-associated protein like 4 (EML4)-anaplastic lymphoma kinase (ALK) gene rearrangements are exclusively seen in lung adenocarcinomas. These gene rearrangements lead to constitutive activation of ALK, uncontrolled cell proliferation and inhibited apoptosis (Soda et al., 2007). Crizotinib, a tyrosine kinase inhibitor, was recently approved by the US Food and Drug Administration

(FDA) for advanced NSCLC with EML4-ALK rearrangements (Kwak et al., 2010; Sasaki and Janne, 2011; Shaw and Solomon, 2011). Two recent phase III clinical trials have demonstrated crizotinib as significantly superior to chemotherapy in both pretreated

(Shaw et al., 2013) and treatment-naïve patients (Solomon et al., 2014) with advanced stage EML4-ALK rearranged NSCLC. Both EML4-ALK gene rearrangements and activating mutations in EGFR are uncommon, based on ethnicity, occurring in about 4-

7% and 10-20% of all NSCLC patients, respectively (Sequist et al., 2007). Hence, there is no clear first-line targeted chemotherapy for the majority of lung cancer patients. Most patients will require two or three lines of therapy as their disease acquires resistance to chemotherapies.

Drug resistance to TBAs, both intrinsic and acquired, represents a significant clinical problem and is the primary cause of treatment failure in NSCLC patients. Mechanisms mediating TBA resistance can occur at multiple stages (reviewed in Katsetos and Draber,

2012; Kavallaris, 2010). Altered expression of a microtubule protein, βIII-tubulin, is a poor prognostic indicator that is also associated with resistance to TBAs in NSCLC. The mechanism regulating aberrant βIII-tubulin expression is currently poorly understood.

This thesis will identify small molecule βIII-tubulin modulators to study the regulation of

βIII-tubulin in NSCLC.

1.3 Tubulin/ microtubule system

Microtubules, together with microfilaments and intermediate filaments make up the cytoskeletal network in all eukaryotic cells. The microtubule network plays important roles in vital and fundamental cellular processes such as regulating cell growth and movement as well as key signalling events. Their crucial role in cell division makes microtubules ideal targets for anticancer drugs. Therefore, chemotherapies targeting microtubules remain an important component in the treatment of many paediatric and adult malignancies, including NSCLC.

1.3.1 Structure and function of microtubules

The basic structural building blocks of microtubules are α- and β-tubulins. Tubulins exist either as free tubulin heterodimers (αβ-heterodimers) in the cell cytoplasm, or as polymerised structures, constituting microtubules. To polymerise, αβ-heterodimers bind in a polar head-to-tail fashion, forming linear protofilaments (Figure 1.1). The cylindrical and left-handed helical wall of a microtubule is composed of 13 parallel protofilaments

25 nm in diameter (Risinger et al., 2009). Microtubules possess a very imperative feature of polarity, which is mandatory for their biological function. In a protofilament, one end has the α-tubulin exposed, while the other end has the β-tubulin exposed, these ends are called the minus and plus ends, respectively (Figure 1.1). Microtubules are continuously undergoing the process of growing and shrinking throughout the cell cycle. This property, referred to as dynamic instability, is critical to many microtubule functions. During interphase, microtubules are nucleated at the centrosome (minus end) and radiate toward the cell periphery (plus end), forming an extensive network throughout the cell (Robinson et al., 1995; Zhou et al., 2002).

Figure 1.1 Microtubule structure.

Microtubules are composed of repeating globular α- and β-tubulin heterodimers that polymerise into protofilaments. The cylindrical tube is usually formed by 13 protofilaments. The α- and β-tubulin-exposing ends are called the minus and plus ends, respectively. Microtubules are continuously undergoing the process of growing and shrinking throughout the cell cycle. This property, referred to as dynamic instability, is critical to many microtubule functions. Image modified from (Akhmanova and Steinmetz,

2008).

Plus end

Protofilament β-tubulin α-tubulin

αβ-heterodimer

8 nm

4 nm

25 nm Minus end

αβ-heterodimer

Microtubule

Interphase microtubules are involved in the maintenance of cell shape and in the trafficking of proteins and organelles (reviewed in Nogales, 2001). Motor proteins translocate cell components along microtubules, and protein-protein interactions with other adaptor proteins regulate this microtubule behaviour.

In microtubule polymerisation, both α- and β-tubulin subunits of the αβ-heterodimer are bound to a Guanosine-5'-triphosphate (GTP) molecule. The GTP bound to α-tubulin is stable and does not get hydrolysed. The GTP bound to β-tubulin is hydrolysed to

Guanosine diphosphate (GDP) shortly after assembling, which results in the addition of

αβ-heterodimers (Sontag et al., 2005). GDP-bound tubulins are prone to depolymerisation and ones that are situated at the end of a microtubule will fall off. A GDP-bound tubulin in the middle of a microtubule cannot spontaneously pop out. Since tubulin adds onto the end of the microtubule only in the GTP-bound state, a cap is formed at the end of the microtubule to prevent it from depolymerisation. When hydrolysis catches up to the end of the microtubule, it starts a rapid depolymerisation and shrinkage. This change from growth to shrinkage is called “catastrophe”. GTP-bound tubulin can be added to the end of the microtubule again, providing a new cap to protect the microtubule from shrinking.

This is called “rescue”.

Microtubule dynamics vary greatly throughout the cell cycle. During interphase, microtubule dynamics are slow. Microtubule dynamics then increase up to 100 fold during mitosis (Rusan et al., 2001). At the onset of mitosis, newly disassembled microtubules form networks of microtubule spindles, ensuring proper attachment and segregation of chromosomes during cell division (Jordan and Wilson, 2004; Zhai et al.,

1996). Microtubule dynamics and roles are regulated and influenced by endogenous protein interactions (described in section 1.3.3) and tubulin-binding agents (described in

25 section 1.3.5). Posttranslational modifications on tubulins influence these interactions

(described in section 1.3.4).

1.3.2 The β-tubulin family

Human microtubules are composed of combinations of eight α-tubulin isotypes and seven

β-tubulin isotypes. The different isotypes are encoded by different genes, located on different chromosomes possessing unique cellular and tissue distributions (Ludueña,

1998). This study focuses on β-tubulins, as β-tubulins are the targets of TBAs and tubulin isotype alterations have been implicated in drug resistance and aggressive disease

(Kavallaris, 2010; Parker et al., 2014). Functional human β-tubulin mRNA is approximately 1.7 kb in length (Cowan et al., 1981) with a protein weight close to 50 kDa

(Cleveland et al., 1980). β-tubulin polypeptides share high amino acid homology

(reviewed in Ludueña, 2013), however, they are distinguished from one another by highly divergent sequences in the carboxy-terminal tails (the last 20-24 amino acids) (Sullivan and Cleveland, 1986). When αβ-heterodimers polymerise into microtubules, the carboxy- terminal tails of both α- and β-tubulins protrude out from the microtubule wall (Nogales,

2001) (Figure 1.2), and are thought to impart functional differences to microtubules.

Table 1.2 gives an overview of β-tubulin isotypes and their differential tissue distribution.

Many of these β-tubulin isotypes are enriched in certain cell or tissue types, for example,

βII- tubulin is enriched in brain and epithelial cells (Banerjee et al., 1988), βIV-tubulin in ciliary and flagellar structures (Raff et al., 1997), and βVI-tubulin in platelets and haematopoietic cells (Leandro-García et al., 2012). In brain, the isotype composition of

β-tubulin was reported as: βI 3%, βII 58%, βIII 25% and βIV-tubulin

Figure 1.2 Illustration of the crystal structure of an αβ-heterodimer in a microtubule

This figure illustrates the fitting of an αβ-heterodimer onto the three-dimensional structure of a microtubule obtained from cryoelectron microscopy reconstruction. When

αβ-heterodimers polymerise into microtubules, the carboxy-terminal tails of both α- and

β-tubulins protrude out from the microtubule wall and are exposed at the outer surface of the microtubule. These moieties contain key interaction sites of many microtubule- associated proteins, motor proteins, severing enzymes, microtubule plus end-tracking proteins (+TIPs) and posttranslational modifications. Hence, carboxy-terminal tails of both α- and β-tubulins are thought to impart functional differences to microtubules.

Adapted with permission from Macmillan Publishers Ltd: Nature Reviews Molecular

Table 1.2 Human β-tubulin isotypes

Gene Accession Chromosome Protein name Tissue distribution name number loci

Constitutive expression;

TUBB βI- tubulin NM_178014 6q21.33 predominant isotype in many

cell lines

TUBB2A βII- tubulin NM_001069 6q25.2 Major isotype of neurons and

TUBB2B βII- tubulin NM_178012 6q25.2 epithelial cells

Neurons and testicular Sertoli TUBB3 βIII- tubulin NM_006086 16q24.3 cells

TUBB4 βIVa- tubulin NM_006087 19q13.3 Brain-specific

Constitutive expression; testes, TUBB2C βIVb- tubulin NM_006088 9q34.3 lung, heart, brain, skin

TUBB6 βV- tubulin NM_032525 18q11.21 Uterus

Haematopoiesis-specific cell

TUBB1 βVI- tubulin NM_030773 20q13.32 types; megakaryocytes and

platelets

Adapted with permission from Macmillan Publishers Ltd: Nature Reviews Cancer

13% (Banerjee et al., 1988). The physiological role of each tubulin isotype and the regulatory mechanisms controlling their distinct expression are poorly understood. While it appears that β-tubulin isotypes can function in a similar manner, inherent differences among the isotypes, such as assembly properties and microtubule dynamics, in combination with complex interactions with cellular proteins, may dictate the overall functional significance of a given isotype in a cell.

1.3.3 Microtubule interacting proteins

A wide variety of proteins are known to interact with the tubulin/microtubule system to regulate microtubule dynamics, growth, shrinkage and stabilisation (reviewed in Alfaro-

Aco and Petry, 2015). These include motor proteins, microtubule associated proteins

(MAPs) and a large heterogeneous group of proteins, such as chaperonins, histones, glycolytic enzymes and kinases. The two major families of microtubule motors are the kinesins and dyneins (reviewed in Verhey and Hammond, 2009). Motor proteins generate force upon interaction with microtubules and these forces are used for diverse cellular functions, such as intracellular transport (reviewed in Sheetz et al., 1989), ciliary beating

(Lindemann and Lesich, 2010) and the self-organisation of mitotic spindles (Surrey et al.,

2001). MAPs are a group of specialised proteins that bind to the wall of microtubules and influence their stability and dynamic behaviour, as well as affecting chemotherapy sensitivity and tumour growth in cancer (Bhat and Setaluri, 2007). The heterogeneous group of MAPs comprises both microtubule stabilising proteins, for example tau,

MAP1A-1C, MAP2a-2c, MAP4 and MAP7, as well as microtubule destabilising/severing proteins, such as katanin, stathmin, spastin and Kin I. Microtubule stabilising proteins, MAP2 and tau, are predominantly expressed in neuronal cells and in

30 their phosphorylated form, can bind to the microtubule wall and stabilise microtubules

(Wang and Liu, 2008). Tau is predominantly associated with axonal microtubules and dendrites, where it is involved in signalling functions. Tau overexpression has also been correlated with poor outcome in breast cancer, where it may influence taxane sensitivity by decreasing the drug’s affinity for β-tubulin (Rouzier et al., 2005).

One example of a microtubule destabilising protein is the cytosolic phosphoprotein, stathmin. Stathmin regulates the mitotic spindle by binding to αβ-heterodimers, increasing microtubule catastrophes and inducing microtubule depolymerisation (Curmi et al., 1999; Rubin and Atweh, 2004). Recent studies have highlighted stathmin’s contribution to tumour progression and in some instances, chemotherapy resistance, particularly to tubulin-binding agents in a number of malignancies (Alli et al., 2007;

McGrogan et al., 2008). Together, microtubule stabilising and destabilising proteins work in a coordinated fashion to construct cell-cycle specific and functional microtubule structures (Kinoshita et al., 2001; Niethammer et al., 2007).

1.3.4 Posttranslational modification

In order for the microtubule cytoskeleton to maintain a variety of cellular functions, the diversity and heterogeneity of α- and β-tubulin isotypes is increased by posttranslational modifications (PTMs). PTMs regulate protein-protein interactions with microtubules, thereby modulating microtubule properties and conferring context-specific cellular functions (reviewed in Janke, 2014). Some of the common tubulin PTMs includes; acetylation, polyamination, phosphorylation, glycylation, polyglutamylation and tyrosination/detyrosination. Although known for several decades, the regulation and biological function of these tubulin PTMs is still largely unchartered. This thesis focuses

31 on βIII-tubulin, hence only PTMs found on β-tubulins will be covered here (Table 1.3).

With the exception of polyamination on glutamine 15 (Song et al., 2013), acetylation on lysine 252 (Chu et al., 2011) and phosphorylation on serine 172 (Caudron et al., 2010), all of the PTMs discussed here take place on the carboxy-terminal tails of β-tubulin

(Figure 1.2) (reviewed in Garnham and Roll-Mecak, 2012 and ; Janke, 2014).

Polyglutamylation is the addition of secondary glutamate side chains catalysed by tubulin tyrosine ligase-like family enzymes (Audebert et al., 1993; van Dijk et al., 2007). Tubulin polyglutamylation is specifically enriched in neurons, centrioles and axonemal structures

(van Dijk et al., 2007) and is thought to regulate microtubule functions in mitosis, cell cycle, neurogenesis and ciliary and flagellar movement. Polyglutamylation has been suggested to affect kinesin-1-mediated transport in cultured neurons (Maas et al., 2009;

Sirajuddin et al., 2014), and enhance the turnover of microtubules via the activation of microtubule-severing enzymes, such as spastin and katanin (Lacroix et al., 2010). In cycling cells, detyrosination, acetylation and polyglutamylation are enriched in mitotic spindles and in the cytokinetic midbody (Bobinnec et al., 1998; reviewed in Janke and

Bulinski, 2011). These PTMs are thought to regulate the interactions between microtubules and diverse MAPs, contributing to precisely-controlled cell cycling.

Different PTMs and tubulin isotypes can influence one another in a combinatorial manner

(Sirajuddin et al., 2014). A recent in vitro study reported that the unique positively charged lysine residue on βIII-tubulin carboxy-terminal tails can markedly suppress kinesin-1 motility (Sirajuddin et al., 2014), while polyglutamylation reverses this interaction. The combination of negative (βIII-tubulin) and positive regulators, in

Table 1.3 Reported post-translational modifications (PTMs) of β-tubulin

PTMs PTM location and enzymes Potential effects and functions References  Lysine 252 (K252) Acetylation Regulate microtubule assemble efficiency (Chu et al., 2011)  Enzyme: acetyltransferase San  Glutamine 15 (Q15) Polyamination Contribute to microtubule stabilisation (Song et al., 2013)  Enzyme: transglutaminases  Serine 172 (S172), threonine 409 (T409) and serine 420 (S420) on β-tubulin (Alexander et al., Regulate microtubule dynamics during  Serine 444 (S444) and tyrosine 437 1991; Caudron et al., Phosphorylation cell division (S172); neurite extension (Y437) on βIII-tubulin 2010; Yoshida et al., (S444)  Enzymes: cyclin-dependent kinase 1 and 2003) G-protein-coupled receptor kinase 2 Function remains unknown. (Bosch Grau et al.,  Glutamine 437 (Q437) on β-tubulin Mutagenesis studies suggests potential 2013; Hammond et Polyglycylation  Enzyme: tubulin tyrosin ligase-like function in stabilising axonemal al., 2008; Redeker et family microtubules al., 1994) Regulates beating behaviour in cilia via  Glutamine 435 and 438 (Q435 and Q438) (Ikegami et al., 2010; the regulation of flagellar dynein motors; Polyglutamylation  Enzymes: tubulin tyrosin ligase-like Sirajuddin et al., regulate turnover of microtubules via family 2014) microtubule severing enzymes

33 this case polyglutamylation, may enable regulation of motor protein function. Overall, recent work on tubulin PTMs has shed light on how these modifications could regulate, fine-tune and coordinate the remarkably complex functions of microtubules in cells.

1.3.5 Tubulin binding agents for NSCLC treatment

Microtubules play critical roles in many important cellular processes including mitosis, making them attractive drug targets for anticancer therapies. Drugs targeting the tubulin/microtubule system, known as tubulin-binding agents (TBAs), are important chemotherapeutic drugs that suppress microtubule dynamics, perturb kinetochore- microtubule attachment and disrupt chromosome segregation. This activates a checkpoint pathway that delays cell cycle progression and induces programmed cell death in rapidly dividing cells (reviewed in Dumontet and Jordan, 2010; Katsetos and Draber, 2012).

TBAs have a long history of effective use in the treatment of numerous cancers including

NSCLC. This section focuses on only those TBAs that are used for NSCLC treatment in the clinic. Readers are referred to two reviews for detailed summaries of the class of TBAs as a whole (Altmann and Gertsch, 2007; Amos, 2011).

Tubulin binding agents are broadly classified into microtubule-stabilising and microtubule-destabilising agents, based on their effects on microtubule polymer mass at high concentrations. Stabilising TBAs include taxanes and epothilones. They bind to the microtubule wall, enhance microtubule polymerisation and suppress microtubule dynamics. Taxanes (paclitaxel, cabazitaxel and docetaxel) are well established in the treatment of NSCLC, breast and prostate cancer, both as single agents and as a partner in platinum-based combination therapy. Taxanes stabilise microtubules by disrupting the dynamic equilibrium between soluble αβ-heterodimers and their polymerised form

(Horwitz et al., 1993), blocking cells in the late G2 or mitotic phase of the cell cycle. The effect of taxanes on microtubules, is concentration-dependent, whereby high concentrations stabilise the polymerised microtubule, maintain the polymerised state and increase polymer mass (Horwitz, 1994). At low concentrations (nM), taxanes inhibit microtubule dynamics without affecting the microtubule polymer mass and it is this inhibition that leads to mitotic arrest by interfering with the function of the mitotic spindle

(Jordan et al., 1993).

A novel class of non-taxane microtubule-stabilising agents, epothilones, are natural compounds derived from myxobacterium Sorangium cellulosm. The epothilones have a mechanism of action similar to that of taxanes. Several epothilones, ixabepilone, patupilones and sagopilone, have advanced to clinical trials for multiple solid tumours, including NSCLC (reviewed in Edelman and Shvartsbeyn, 2012). In phase II trials, ixabepilone has demonstrated encouraging clinical activity in advanced NSCLC (Spigel et al., 2012), even in patients with taxane-treated and platinum-refractory tumours

(Vansteenkiste et al., 2007), however, this agent did not progress to phase III trials.

Clinically important destabilising TBAs include vinca alkaloids (vincristine, vinblastine, vinorelbine and vinflunine) and the most recently approved TBA, eribulin. Vinca alkaloids were originally isolated from the plant Catharanthus roseus. They bind to β- tubulin on the αβ-heterodimers enhancing microtubule depolymerisation and suppressing microtubule dynamics (reviewed in Jordan and Wilson, 2004). Similar to microtubule- stabilising TBAs, vinca alkaloids affect microtubules in a concentration-dependent manner. Low (sub-stoichiometric) concentrations suppress microtubule dynamics without affecting the microtubule polymer mass, while high concentrations depolymerise microtubules and decrease polymer mass (Jordan and Wilson, 2004). Vinflunine is a

35 novel fluorinated vinca alkaloid that has seen promising clinical activity in several chemotherapy-refractory cancers, including advanced NSCLC (Bennouna et al., 2006), breast (Blasinska-Morawiec et al., 2013) and bladder cancer (Hegele et al., 2014;

Moriceau et al., 2015; Retz et al., 2015).

Another microtubule-destabilising agent distinct from the vinca alkaloids, eribulin, has recently been approved by the US FDA as a third-line therapy for metastatic breast cancer patients (Donoghue et al., 2012). It has also advanced to phase II clinical trials for numerous malignancies, including NSCLC (reviewed in Swami et al., 2012). Unlike conventional TBAs (paclitaxel, vinca alkaloid and epothilones), eribulin potently inhibits microtubule dynamics by suppressing microtubule polymerisation without affecting depolymerisation and sequestering tubulin into non-functional aggregates (Dabydeen et al., 2006; Jordan et al., 2005; Okouneva et al., 2008).

Binding sites for TBAs have been revealed on β-tubulin (reviewed in Downing, 2000;

Field et al., 2013) and include the taxane-binding site, the vinca domain and the colchicine site. The taxane binding site resides within the lumen of the microtubule (Nogales, 2001).

It is situated in a deep hydrophobic pocket at the lateral interface between adjacent protofilaments (Nogales et al., 1998). Taxane binding strengthens the interaction between adjacent protofilaments, which promotes microtubule assembly and stability. Epothilones have a similar binding site to taxanes but not identical (Nettles et al., 2004). The vinca domain is located on the plus end interface on the GTP-binding site in β-tubulin (Gigant et al., 2005; Jordan et al., 1986). The colchicine-binding site is located between the α- tubulin and β-tubulin interface (Uppuluri et al., 1993).

The precise mechanism of cell death as a consequence of TBA-induced mitotic arrest remains a subject of debate (Gascoigne and Taylor, 2009). As a general rule, tumour cells

36 are more susceptible to TBA-induced cell death compared to normal cells. A broad variation in cell death response to TBAs has been described in different tumour cell lines and within the same cell line (Gascoigne and Taylor, 2008). Factors such as the type and concentration of drug used, the type of analysis carried out, time-point in cell cycle and the duration of mitotic arrest may all influence the fate of a cell after TBA-induced mitotic arrest. In some instances, cell fate is dictated by two competing signalling networks: the activation of cell death pathways and those that regulate degradation of cyclin B1 and exit from mitosis (Gascoigne and Taylor, 2008). Moreover, recent data suggests that tumour cells in in vitro culture systems could respond differently to TBA than their in vivo counterparts, utilising different mode of apoptosis induction following TBA treatment

(Chittajallu et al., 2015; Janssen et al., 2013). Such difference may be attributed by the lack of complex tumour microenvironment in in vitro systems, compared to in vivo.

1.3.6 Drug resistance in NSCLC and βIII-tubulin

Despite incremental improvements in clinical outcome achieved by optimising deployment, further clinical utility of current chemotherapeutic agents is hindered by the persistent emergence of drug resistance. Drug resistance, both intrinsic and acquired, is the primary cause of treatment failure in NSCLC patients. A range of resistance mechanisms have been established in NSCLC, including active efflux of chemotherapeutic agents, modification of drug targets, changes in mitotic checkpoint signals, drug sequestration, detoxification of cytotoxic agents via enzymes (Hida et al.,

1993; Matsumoto et al., 1997) and enhanced DNA repair mechanisms (Joerger et al.,

2011; Okuda et al., 2008; Sullivan et al., 2014; Vilmar et al., 2010). This thesis introduction will primarily focus on resistance associated with alterations in

37 tubulin/microtubule proteins in NSCLC. Other mechanisms have been reviewed by

Chang et al. (2011) and Seve et al. (2005).

Mechanisms mediating TBA resistance can occur at multiple levels (reviewed in Katsetos and Draber, 2012; Kavallaris, 2010), including active drug efflux, apoptosis-induction deficiency and altered expression of tubulin isotypes. Active efflux of chemotherapeutic agents is associated with resistance to TBAs in many different types of cancer and it can occur via ATP-binding cassette transporters, such as P-glycogprotein (PgP) and multidrug-resistance proteins (reviewed in Gottesman et al., 2002; Jordan and Wilson,

2004). Both taxanes and vinca alkaloids are substrates for PgP-mediated efflux (Fojo and

Menefee, 2005; Harker and Sikic, 1985; Ueda et al., 1987). However, clinical evidence suggests that PgP-mediated mechanisms play only a minor role in resistance to TBAs in

NSCLC (reviewed in Seve and Dumontet, 2005). Another potential mechanism of drug resistance to TBAs occurs through alterations in the target of these agents, the tubulin/microtubule system. Evidence has associated β-tubulin mutations, namely βI- tubulin, with resistance to TBAs in vitro (Kavallaris et al., 2001; Martello et al., 2003;

Mozzetti et al., 2005; Verdier-Pinard et al., 2003). Nevertheless, to date, β-tubulin mutations have not been identified in NSCLC clinical samples (reviewed in Berrieman et al., 2004; Kelley et al., 2001; Kohonen-Corish et al., 2002; Šale et al., 2002; Tsurutani et al., 2002) and are unlikely to play the dominant role in acquired drug resistance in the clinic.

Altered expression of tubulin isotypes is now considered a major form of resistance to

TBAs in multiple cancers, including NSCLC (reviewed in Parker et al., 2014). Compared to α-tubulin isotypes, β-tubulin isotypes are more extensively studied in resistance to

TBAs, largely due to the fact that TBAs bind to the β-tubulin subunit to exert their toxic

38 effect. Altered levels of βII-, βIVb- and βV-tubulins have been associated with altered sensitivity to TBAs in NSCLC cell lines and tumour samples (Cucchiarelli et al., 2008;

Gan and Kavallaris, 2008; Gan et al., 2011). However, our understanding of the clinical relevance of these β-tubulin isotypes in drug resistance is still limited and further investigation is required. In contrast, βIII-tubulin expression has been shown to be the most relevant marker in terms of prognostic and predictive value and for treatment decision making in advanced NSCLC (Karki et al., 2013; Koh et al., 2010; Reiman et al.,

2012; Seve et al., 2005b; Seve et al., 2010; Vilmar et al., 2011). A prognostic factor is defined as that which impacts disease outcome regardless of treatment strategy, while a predictive factor is one which predicts the activity of a specific agent. Clinical data have validated and established that aberrant βIII-tubulin expression is associated with drug resistance to TBAs, including taxanes and vinorelbine, and reduced survival in advanced stage NSCLC patients (Dumontet et al., 2005; Okuda et al., 2008; Seve et al., 2005a; Seve et al., 2005b; Zhang et al., 2012a).

Several laboratory studies have explored potential functional explanations for this clinical observation. Initially, resistance mediated by βIII-tubulin was thought to be restricted to taxanes, as βIII-tubulin was linked to dynamic instability of microtubules. In vitro studies observed that microtubules enriched in αβIII-tubulin heterodimers exhibit decreased assembly of microtubules, weakening the microtubule stabilising effect of taxanes (Hari et al., 2003; Kamath et al., 2005). Multiple studies have reported that NSCLC cell lines selected for resistance to taxanes had significantly increased expression of βIII-tubulin, compared to parental drug-sensitive cells (Burkhart et al., 2001; Kavallaris et al., 1997;

Wehbe et al., 2005). Recent laboratory data have also established the importance of βIII- tubulin in regulating sensitivity to chemotherapy in multiple epithelial cancers (discussed in section 1.4.1). Importantly, suppression of βIII-tubulin using antisense 39 oligonucleotides partially restored drug sensitivity in paclitaxel-resistant NSCLC cells

(Kavallaris et al., 1999). A landmark study, by Gan et al. (2007) demonstrated a direct functional role for βIII-tubulin in mediating sensitivity to chemotherapy in NSCLC. In the study, βIII-tubulin expression was stably and potently suppressed using siRNA in two independent NSCLC cell lines, H460 and Calu-6, which aberrantly express βIII-tubulin.

Knockdown of βIII-tubulin significantly sensitised these cells to broad classes of TBAs, including taxanes, vinca alkaloids, epothilone B (Gan et al., 2011; Gan et al., 2007).

Additionally, knockdown of βIII-tubulin using siRNA not only significantly increased

NSCLC cell sensitivity to broad classes of TBAs, but also to DNA-damaging agents

(cisplatin, doxorubicin and etoposide) that are structurally and mechanistically distinct from TBAs (Gan et al., 2007; McCarroll et al., 2010). This strongly suggests that βIII- tubulin plays a role in mediating sensitivity to chemotherapy. McCarroll et al. (2010) further showed that stable knockdown of βIII-tubulin increased sensitivity to both paclitaxel and cisplatin in vivo (McCarroll et al., 2010). This βIII-tubulin knockdown mediated sensitivity to chemotherapy was associated with an increase in susceptibility of cells to undergo drug-induced apoptosis (Gan et al., 2007; McCarroll et al., 2010) and is specific to βIII-tubulin, as silencing of other β-tubulin isotypes did not yield the same general protective effect (Gan and Kavallaris, 2008; Gan et al., 2011; Gan et al., 2007).

Strikingly, when βIII-tubulin expression was rescued, the βIII-tubulin knockdown- mediated sensitivity to cisplatin was abolished, confirming βIII-tubulin is directly involved in drug sensitivity (McCarroll et al., 2010). One possibility for the changes in drug sensitivity when βIII-tubulin levels are altered was that βIII-tubulin influences microtubule dynamics. Interestingly, suppression of βIII-tubulin did not affect microtubule dynamics in H460 cells in the absence of drug treatment, but significantly suppressed microtubule dynamics and increased apoptosis when cells were treated with 40 low dose paclitaxel, vincristine and epothilone B (Gan et al., 2010). This increase in apoptotic cells appeared to be independent of mitotic arrest, and low drug doses did not alter microtubule dynamics in control cells (Gan et al., 2010). At higher doses of paclitaxel, vincristine or epothilone B, βIII-tubulin suppression markedly increased apoptotic cell death (Gan et al., 2011; Gan et al., 2010). The authors concluded that knockdown of βIII-tubulin enhances the effectiveness of TBAs via two mechanisms: suppression of microtubule dynamics at low drug doses and a mitosis-independent mechanism of cell death at higher drug concentrations. βIII-tubulin mediated sensitivity to epothilones was also reported in cervical (Risinger et al., 2008) and breast cancer cells

(Lopus et al., 2015), however, the same effect was not observed in other cancer cells. For example, epothilones were found to be extremely active in ovarian cancer models

(Mozzetti et al., 2008) and other gynaecologic malignancies (Carrara et al., 2012;

Dumontet et al., 2009) that overexpress βIII-tubulin. Drug resistance is multifactorial, hence difference in cell lines and other drug resistance mechanisms are thought to account for this divergence.

βIII-tubulin accounts for only 7.8% of all β-tubulin isotypes present in H460 cells

(Nicoletti et al., 2001). The fact that manipulation of the expression of this minor isotype yielded significant changes in NSCLC response to chemotherapy further highlights its role in mediating drug sensitivity. These data suggest that βIII-tubulin plays a cellular survival factor role, exerting a broad protective effect against chemotherapy, and when overexpressed confers resistance to chemotherapy in NSCLC cells. Collectively, these data have validated βIII-tubulin as a bona fide target for chemosensitisation in NSCLC.

1.4 βIII-tubulin: expression and regulation

βIII-tubulin (encoded by the TUBB3 gene) is by far the most comprehensively studied β- tubulin isotype across a range of human malignancies. As reviewed in the preceeding sections, it is a multifunctional protein. This section covers the expression of βIII-tubulin and its roles in malignant and non-malignant cells, followed by a review of what is currently known about the regulation of TUBB3/βIII-tubulin, with a particular focus on cancer.

1.4.1 Expression and function of βIII-tubulin

During the evolution of the β-tubulin isotypes, βIII-tubulin is postulated to have appeared between 530 and 460 million years ago as an adaptive mechanism, protecting microtubules against a significant increase in O2 levels in the Earth’s atmosphere

(reviewed in Ludueña, 2013). βIII-tubulin is expressed in fish, amphibians, birds and mammals (Tuszynski et al., 2006). Its unique protein structure (Ser239, Thr429 and

Cys124) is highly conserved in all vertebrates and is thought to be functionally significant

(reviewed in Ludueña, 2013). βIII-tubulin is abundant in the brain but only found in neurons, where it accounts for 25% of the total β-tubulin pool (Ludueña et al., 1982). It is also found in neurons of the peripheral nervous system (Katsetos et al., 2003), Sertoli cells of the testes (Sullivan and Cleveland, 1986), epidermal and hair follicle melanocytes

(Akasaka et al., 2009; Locher et al., 2013) and olfactory ensheathing cells (Omar et al.,

2013). βIII-tubulin expression in both melanocytes and the nervous system is differentiation-dependent. Recent reports have highlighted the pivotal role of TUBB3 in neural development and mutations of this gene were shown to be causative of brain malformations, neurological disabilities and ocular motility disorder (Chew et al., 2013;

Niwa et al., 2013; Poirier et al., 2010; Tischfield and Engle, 2010). As further evidence of its critical and specific role, TUBB3 knockdown impairs neural progenitor proliferation, which cannot be rescued or restored by other β-tubulin isotypes (Saillour et al., 2014).

In cancer, βIII-tubulin is expressed in both neuronal and non-neuronal tumours. In tumours of neuronal origin, such as medulloblastomas, neuroblastomas and retinoblastomas, βIII-tubulin expression is constitutive and linked to morphological changes of neuronal differentiation and reduced tumour proliferation (Caracciolo et al.,

2010; Katsetos et al., 2003). In contrast, in non-neuronal tumours, where cells of origin do not express this β-tubulin isotype, βIII-tubulin expression is aberrant and represents dedifferentiation associated with malignant transformation and acquisition of a stem cell- like phenotype. In the clinic, βIII-tubulin is now recognised as a key determinant of drug resistance, tumour aggressiveness and tumour progression (reviewed in Hayashi et al.,

2009; Karki et al., 2013; Kavallaris, 2010; Mariani et al., 2011). Non-neuronal tumours that aberrantly express βIII-tubulin include epithelial tumours of the lung, including

NSCLC (Cucchiarelli et al., 2008; Seve and Dumontet, 2010), ovary (De Donato et al.,

2012; Du et al., 2015; Raspaglio et al., 2014), breast (Tommasi et al., 2007; Wang et al.,

2013), uterine cervix (Ferrandina et al., 2007), prostate (Ploussard et al., 2010; Tsourlakis et al., 2014), pancreas (Lee et al., 2007; McCarroll et al., 2015b), alimentary tract (Jirasek et al., 2009; Zheng et al., 2012) and gliomas of the brain (Katsetos et al., 2007; Yan et al.,

2011). There is strong clinical evidence in lung, ovarian and breast cancer, that patients with aberrant βIII-tubulin expression exhibit poorly differentiated tumour tissue, high grade malignancy, shorter disease progression, unfavourable prognosis and worse overall survival (reviewed in Kavallaris, 2010; Mariani et al., 2011; Seve and Dumontet, 2008).

1.4.1.1 βIII-tubulin and chemoresistance

Overexpression of βIII-tubulin in tumour specimens often correlates with decreased progression-free or overall survival and is associated with resistance to chemotherapeutic agents (Karki et al., 2013; Mozzetti et al., 2005; Reiman et al., 2012; Roque et al., 2014;

Rosell et al., 2003; Vilmar et al., 2011). Laboratory data have established the importance of βIII-tubulin in regulating sensitivity to chemotherapy in multiple epithelial cancers, such as ovarian cancer (Mozzetti et al., 2008), breast cancer (Shalli et al., 2005; Stengel et al., 2010), prostate cancer (Ploussard et al., 2010; Ranganathan et al., 1998a;

Ranganathan et al., 1996; Ranganathan et al., 1998b), pancreatic cancer (Lee et al., 2007;

Liu et al., 2001; McCarroll et al., 2015b), and lung cancer (Gan et al., 2011; Gan et al.,

2010; Gan et al., 2007; Kavallaris et al., 1999; McCarroll et al., 2010).

Our understanding of the role of βIII-tubulin and chemoresistance in cancer cells has evolved significantly in the last three decades. Originally, βIII-tubulin has been thought to mediate taxane resistance in tumours by constitutively enhancing microtubule dynamics. This view was based on in vitro studies, where removal of βIII-tubulin from unfractionated tubulin resulted in increased microtubule assembly both in the absence and presence of paclitaxel (Banerjee et al., 1990). Furthermore, in vitro studies on isolated microtubules have shown that microtubules enriched in αβIII-tubulin heterodimers are unstable polymers (Derry et al., 1997; Panda et al., 1994), capable of evading the microtubule-stabilising effect of Taxol. However, studies on Chinese hamster ovary cells and microtubule dynamics showed that βIII-tubulin overexpression does not change the inherent properties on microtubule dynamic instability, but reduced the ability of paclitaxel to suppress microtubule dynamics and induced drug resistance (Kamath et al.,

2005). The difference in the effect of βIII-tubulin on microtubule dynamics between cell- free microtubule assembly studies and cell studies may be due to the presence of 44 additional cellular factors. Kavallaris and co-workers (2010) showed that βIII-tubulin silencing does not change the intrinsic microtubule dynamic instability, but enhanced cell apoptosis and TBA-induced suppression of microtubule dynamics, as well as mediating a mitosis-independent mechanism of cell death. Furthermore, knockdown of βIII-tubulin significantly increases NSCLC cell sensitivity to broad classes of TBA and DNA- damaging agents that are mechanistically distinct from TBA (Gan et al., 2007; McCarroll et al., 2010). Together, Kavallaris and co-workers propose βIII-tubulin as a survival factor that rescues tumour cells from death signals triggered by broad classes of chemotherapeutic agents.

Ferlini and co-workers propose a different view that βIII-tubulin is a prognostic marker in certain cancer types but not a predictive marker for taxane resistance (De Donato et al.,

2012; Karki et al., 2013; Mariani et al., 2012). βIII-tubulin can be triggered and conditionally expressed under stressing microenvironments (described later in section

1.4.1.3) and act as a gateway for prosurvival signals into the cytoskeleton. De Donato et al. (2012) demonstrated that βIII-tubulin can interact and form complexes with prosurvival factor guanylate-binding protein 1 (GBP1), allowing the incorporation of

GBP1 into microtubules. Once in the cytoskeleton, GBP1 can bind to prosurvival kinases such as PIM1 and initiate signalling pathways, which ultimately lead to drug resistance.

In contrast to the mechanism proposed by Kavallaris and co-workers, Ferlini and co- workers assert that βIII-tubulin does not work alone but requires partners to play its role.

In spite of their differences, both views agree that βIII-tubulin constitutes an important prosurvival factor in cancer, whose aberrant expression confers drug resistance and a survival advantage.

1.4.1.2 βIII-tubulin and metastasis

In addition to its role in drug resistance, emerging evidence suggests that βIII-tubulin is involved in tumourigenic and metastatic potential. Levels of βIII-tubulin in clinical samples correlate with aggressive tumour growth, metastasis and poor overall survival in lung, ovarian, prostate and breast cancer, regardless of response to chemotherapy (Egevad et al., 2010; Kanojia et al., 2015; Katsetos et al., 2011; Riihimaki et al., 2014; Tsourlakis et al., 2014). McCarroll et al. first reported in 2010 that stable knockdown of βIII-tubulin significantly delayed tumour growth and reduced tumour incidence of subcutaneously xenografted tumours in the absence of chemotherapy (McCarroll et al., 2010). This effect is not a result of reduced cell proliferation, as stable βIII-tubulin knockdown did not affect cell proliferation in vitro (McCarroll et al., 2010).

Functional and differential proteomic studies have revealed that βIII-tubulin is involved in regulating the expression of key proteins associated with tumour growth and metastatic behaviour including the adhesion-associated tumour suppressor, maspin (McCarroll et al., 2015a) as well as adhesion proteins, integrin and L1 cell adhesion molecule (Kanojia et al., 2015), raising the possibility that βIII-tubulin may be functionally significant in these processes. To investigate the functional role of βIII-tubulin in the metastatic cascade, three separate studies used siRNA to silence βIII-tubulin expression in NSCLC cells (McCarroll et al., 2015a), pancreatic cancer cells (McCarroll et al., 2015b) and brain metastatic breast cancer cells (Kanojia et al., 2015). Detachment and adhesion of cancer cells to extracellular matrix (ECM) components are critical steps in the metastatic cascade. McCarroll et al. (2015a) and Kanojia et al. (2015) showed that suppression of

βIII-tubulin altered adhesion ability and markedly reduced migratory capacity. In the context of breast cancer, βIII-tubulin influences adhesion ability through the integrin-Src signalling axis (Kanojia et al., 2015). Typically, the detachment of normal epithelial cells 46 from the ECM results in anoikis, a form of programmed cell death that is induced by anchorage-dependent cells detaching from the surrounding ECM. Tumour cells with high metastatic potential are resistant to anoikis, allowing them to survive under non-adherent conditions in blood and lymph, to seed at distant sites and form metastasis. Interestingly, suppression of βIII-tubulin resulted in significantly increased sensitivity to anoikis in tumour cells (Kanojia et al., 2015; McCarroll et al., 2015a; McCarroll et al., 2015b). In

NSCLC cells, anoikis was reversed by re-expression of βIII-tubulin, indicating a direct correlation between this protein and anoikis sensitivity. This correlation is further supported by a recent in vivo study, where βIII-tubulin promotes tumour growth in an anchorage-independent manner (McCarroll et al., 2015a). These results suggest that βIII- tubulin mediates resistance to anoikis in tumour cells. Mechanistically, βIII-tubulin was demonstrated to influence anoikis sensitivity by regulating the PTEN/AKT signalling axis

(McCarroll et al., 2015a). Most importantly, tumour cells with stable suppression of βIII- tubulin exhibited compromised metastatic capabilities in xenograft mouse models of

NSCLC (McCarroll et al., 2015a), pancreatic (McCarroll et al., 2015b) and brain metastatic breast cancer (Kanojia et al., 2015). Moreover, the decreased metastatic load was correlated with improved survival in the brain metastatic breast cancer model.

Together, these preliminary results provide evidence that βIII-tubulin overexpression confers metastatic potential to cancer cells at multiple levels, and this protein is a possible marker for metastasis.

1.4.1.3 βIII-tubulin and cellular stress

Evidence is accumulating that βIII-tubulin can be conditionally induced and preferentially translated in some cancer cells as an adaptive mechanism against stressors such as

47 hypoxia (Bordji et al., 2014; Katsetos et al., 2009; Raspaglio et al., 2008; Raspaglio et al.,

2014), hypoglycaemia (Raspaglio et al., 2010), serum starvation (Mariani et al., 2012) and genotoxic stress induced by chemotherapy (Ranganathan et al., 1996; Ranganathan et al., 1998b; Saussede-Aim et al., 2009b). Increased expression of βIII-tubulin has been reported in glioblastoma cells that are adjacent to areas of ischemic necrosis (Katsetos et al., 2009). These glioblastoma cells are thought to be inherently resistant to irradiation and chemotherapy. Moreover, a proteomic study demonstrated that βIII-tubulin was able to form protein-protein interactions with a number of proteins known to play a role in the adaptation to cellular stress in paclitaxel-resistant cancer cells (Cicchillitti et al., 2008).

Although unconfirmed, these studies suggest a functional link between βIII-tubulin and tumour survival, where βIII-tubulin may be part of a survival pathway, protecting tumour cells against deleterious stressors and providing a survival advantage. In one study, βIII- tubulin has been described to provide a survival advantage to tumour cells even in the absence of cellular stress (McCarroll et al., 2015b). Further investigation is required to reveal mechanisms underlying the functional relationship between βIII-tubulin expression and tumour cell survival. The distribution of βIII-tubulin can vary within different types of tumours. Some tumours exhibit increased βIII-tubulin expression throughout the whole section, while in other tumour types, βIII-tubulin expression is observed predominately at the invasive tumour edge (Portyanko et al., 2009). Coupled with the fact that silencing of βIII-tubulin has been shown to reduce tumour invasion in vitro (Kanojia et al., 2015), it is possible that βIII-tubulin plays a role in tumour invasion, although more studies are required to address this phenotype.

Collectively, these data suggest that βIII-tubulin has a broad role in tumour pathobiology and specific targeting of this protein may hold the potential for subduing malignant tumour behaviour, enhancing sensitivity to chemotherapy and hence improving overall 48 clinical outcome in cancer patients. Understanding the mechanisms by which

TUBB3/βIII-tubulin expression is regulated in cancer will help to unravel its function and may highlight potential therapeutic approaches to target this protein. Molecular mechanisms behind its regulation are starting to be revealed at multiple levels and these insights will be discussed next.

1.4.2 Gene structure of human TUBB3 gene

There are seven different isotypes of β-tubulin, each of which is encoded by a different gene, located on a different chromosome and contains a unique promoter sequence

(Ludueña, 1998). Like other β-tubulin family genes, the human TUBB3 gene consists of four exons, three intervening introns, a 3’- and a 5’ untranslated region (UTR) and a promoter region (Lewis et al., 1985) (Figure 1.3). The distinguishing feature of each isotype is their unique promoter region. The human TUBB3 gene (encoding for βIII- tubulin) contains 14,089 nucleotides located on chromosome loci 16q24.3 (sequence between 89988417 to 90002505, NCBI ID: NC_000016.9). The TUBB3 gene sequence encodes for 450 amino acids and a 50.4 kDa protein. A shorter isoform of 378 amino acids derived from the alternative splicing of exon 1 is devoid of a part of the N-terminus and may represent mitochondrial expression (Cicchillitti et al., 2008), the functional significance of which remains unknown.

Figure 1.3 Primary structure of β-tubulin genes.

Structurally, β-tubulin genes consist of four exons (blue boxes), three intervening introns

(thin lines), a promoter (thick line with inverted arrows), a 5’- and a 3’ untranslated region

(green hatched boxes).

Exon 1 Exon 2 Exon 3 Exon 4 5’ UTR 3’ UTR Promoter Intron 1 Intron 2 Intron 3

1.4.3 Regulation of TUBB3 expression in non-neoplastic cells

This section covers the regulation of TUBB3 expression in non-neoplastic cells. Studies are divided into neuronal and non-neuronal cells (summarised in Table 1.4). Insights into potential mechanisms of TUBB3 regulation have been gained through cloning and characterisation of the rat and human gene promoter and coding region (Dennis et al.,

2002; Ranganathan et al., 1998b). The differential expression of βIII-tubulin in neurons and Sertoli cells of the testes suggests unique transcriptional regulatory mechanisms exist in different tissues.

1.4.3.1 Transcriptional regulation of TUBB3 in neuronal cells

TUBB3 is predominately expressed in neurons, and its expression is differentially regulated during central and peripheral neuron development. In neurons of the central nervous system, TUBB3 levels decline with increasing maturity. In neurons of the peripheral nervous system, a reverse pattern is found (Jiang and Oblinger, 1992). βIII- tubulin is important in neuronal lineage commitment, where neuronal progenitor cells undergo terminal mitosis to adopt a neuronal phenotype (Haendel et al., 1996). In nearly all developing neurons, TUBB3 is expressed either immediately before or during terminal mitotic division (Easter et al., 1993; Moody et al., 1989). This suggests that TUBB3 expression may be positively regulated by transcription factors necessary for committing a neural progenitor cell to a mature neuron and for initiating morphologic differentiation.

One potential regulatory candidate is Scratch, a Snail family zinc finger transcription factor that is specifically expressed in postmitotic and newly differentiating neurons

(Nakakura et al., 2001a). Coexpression of Scratch and βIII-tubulin protein has been reported in neuronal differentiation of P19 embryonal

Table 1.4 Potential TUBB3 regulatory mechanisms in normal cells

TUBB3 regulatory Regulatory protein/gene Downstream effects Cell type References regions Neuronal Enhance TUBB3 Promoter (unconfirmed) Scratch Neuronal (Nakakura et al., 2001b) expression Enhance TUBB3 (Uittenbogaard and Promoter (unconfirmed) MATH-2 Neuronal expression Chiaramello, 2002)

Retinoblastoma tumour Induction of TUBB3 Promoter (unconfirmed) Neuronal (Toma et al., 2000) suppressor protein expression

Downregulate TUBB3 Promoter (unconfirmed) Id2 Neuronal (Toma et al., 2000) expression Activate TUBB3 promoter Promoter (unconfirmed) SP1 and AP2 Neuronal (Dennis et al., 2002) activity Non-neuronal Androgen response Androgen and androgen Induce TUBB3 expression Testicular Sertoli cells (De Gendt et al., 2011) elements in intron 1 receptors

Repressor element 1 Immortalised human Repressor element in Repress TUBB3 silencing transcription embryonic kidney 293 (Shibazaki et al., 2012) intron 1 expression factor cells

First 13 translated (Theodorakis and Exon 1 Autoregulation CHO cells nucleotides of TUBB3 Cleveland, 1992)

53 carcinoma cells following retinoid acid treatment (Nakakura et al., 2001b). Moreover, overexpression of Scratch, by itself, is sufficient to induce βIII-tubulin expression in a subset of P19 cells in a differentiation-dependent manner (Nakakura et al., 2001b). The basic helix-loop- helix (bHLH) differentiation transcription factor, MATH-2 has also been shown to stimulate βIII-tubulin expression during neuronal differentiation

(Uittenbogaard and Chiaramello, 2002). One study demonstrated that the minimal TUBB3 promoter region required for neuronal differentiation contains an E box motif (CACTTG)

(Dennis et al., 2002). While both MATH-2 and Scratch regulate their target gene expression via E-box enhancer motifs (CANNTG) (Nakakura et al., 2001a), direct evidence demonstrating that MATH-2 and Scratch positively modulate TUBB3 via this

E-box motif is still lacking.

Another potential TUBB3 transcriptional regulator is the retinoblastoma tumour suppressor protein (pRB) (Table 1.4). During neurogenesis, pRB acts as a key cell cycle regulator that ensures neuronal progenitors exit the cell cycle and undergo terminal mitosis. Collaboration of pRB proteins with a bHLH transcription factor, such as MATH-

2, has been shown to be essential for the induction of terminal mitosis and TUBB3 expression (Toma et al., 2000). Additionally, the endogenous inhibitory HLH protein,

Id2, has been demonstrated to bind to and inhibit the bHLH- and pRB-coordinated induction of neuronal gene expression (Toma et al., 2000), suggesting a potential role for

Id2 as a TUBB3 negative regulator.

Cloning and characterisation of the 5’ flanking region of rat TUBB3 has revealed its minimal promoter region and several potential regulatory motifs (Dennis et al., 2002).

These include putative binding sites for transcription factors SP-1, AP-2, central nervous system enhancer, a TATA box, as well as a Pit-1 element, PEA3 sites, and an E box

54 element (Dennis et al., 2002). Identification of multiple SP1 putative binding motifs in both rat and mouse TUBB3 promoter regions suggest a potential regulatory role for these transcription factors (Dennis et al., 2002; Liu et al., 2007). SP1 is known to function by binding GC-rich sequences (GGGCGG) and recruiting essential machineries to TATAA boxes to initiate transcription of its target genes (Näär et al., 1998). One functional study reported that binding of three of the five GC-rich sequences located near the proximal

TATAA box are sufficient to confer TUBB3 promoter activity during neuronal differentiation (Dennis et al., 2002). Further, mithramycin, an SP1-selective inhibitor was shown to downregulate multiple genes in cortical neurons, including TUBB3 (Sleiman et al., 2011). Hence, it is possible that SP1 is involved in TUBB3 transcription activation.

However, whether this is a direct or indirect effect remains to be determined.

1.4.3.2 Transcriptional regulation of TUBB3 in non-neuronal cells

To date, TUBB3 regulation is best studied in neuronal cells and little is known in regards to its regulation in non-neuronal cells. In adult testicular Sertoli cells, TUBB3 expression is directly regulated by an androgen-dependent and androgen receptor (AR)-dependent pathway (De Gendt et al., 2011). In silico and mutagenesis studies revealed four androgen response elements (ARE) in TUBB3 intron 1, one of which (AGAAGGCTGTGTTCT, position +5439 to +5453) is functional and facilitates direct binding of ARs in vivo (De

Gendt et al., 2011). Androgen and AR-regulated TUBB3 expression plays a critical role in spermatogenesis (De Gendt et al., 2011).

In immortalised embryonic kidney 293 (HEK293) cells, constitutive TUBB3 expression appears to be regulated in a cell-cycle dependent manner, with maximal expression in

G2/M phase of the cell cycle (Shibazaki et al., 2012). This cell-cycle dependent TUBB3

55 expression is regulated at the chromatin level and is directly controlled by repressor element 1 silencing transcription factor (REST) via a repressor element 1 (RE-1) in

TUBB3 intron 1 (Shibazaki et al., 2012). REST is a global transcriptional silencer that represses neuron-specific gene expression in non-neuronal cells. Typically, REST forms complexes with chromatin-modifying enzymes, such as HDACs, coREST, mSin3a,

MeCP2 and suppresses neuronal gene expression by epigenetic mechanisms (Ballas et al., 2005). In the context of HEK293 cells, REST was shown to dissociate from the

TUBB3 RE-1 element during G2/M phase and recover in G1 phase, achieving timed

TUBB3 expression (Shibazaki et al., 2012). Mechanisms facilitating REST association and dissociation from the TUBB3 RE-1 element were not addressed in this study.

Knockdown studies indicated that cell-cycle dependent TUBB3 expression is required for mitosis and normal cell growth (Shibazaki et al., 2012). It is unclear whether REST- regulated TUBB3 expression is a cell-specific effect or can be applied to all non-neuronal cells. Future studies are needed to better understand role of REST in TUBB3 regulation.

1.4.3.3 Translational regulation of βIII-tubulin in normal non-neuronal cells

β-tubulin polypeptides are regulated by a general translational mechanism without isotype distinction. Briefly, the total level of intracellular tubulin is tightly controlled by an autoregulatory mechanism (Caron et al., 1985; Gay et al., 1987). In eukaryotic cells, an increase in the cytoplasmic concentration of tubulin subunits leads to a rapid and specific degradation of ribosome-bound tubulin mRNA polypeptides (Theodorakis and

Cleveland, 1992). This autoregulation is dependent on the first 13 translated nucleotide residues (encoding amino acids Met-Arg-Glu-Ile) within exon 1 of all nascent β-tubulin peptides (Pittenger and Cleveland, 1985; Theodorakis and Cleveland, 1992; Yen et al.,

1988). While the key cellular factor involved in this autoregulatory pathway remains unknown (Theodorakis and Cleveland, 1992), it is clear that neither introns, 5’UTR, nor the internal coding sequence of the β–tubulins, are required for this regulation (Pittenger and Cleveland, 1985). βIII-tubulin protein levels may also be partially controlled by the ubiquitin-proteasome system (Shibazaki et al., 2012).

1.4.4 Regulation of TUBB3 expression in cancer cells

Despite the well-established link between βIII-tubulin overexpression, drug resistance and poor clinical outcomes in patients, the regulation of TUBB3 expression in cancer cells remains poorly understood. It is becoming apparent that mechanisms driving aberrant

TUBB3 expression in tumours are complex and may vary depending on cell type and gender. This section will discuss TUBB3 regulation in cancer. Studies are stratified by

TUBB3 gene structure from the 5’ to 3’ direction (summarised in Table 1.5).

1.4.4.1 The TUBB3 promoter and 5’ untranslated region

Induction of TUBB3 expression has been widely reported in numerous tumour cell lines by both short term (Ranganathan et al., 1998b) and long term (Kavallaris et al., 1997;

Ranganathan et al., 1998a; Ranganathan et al., 1996; Shalli et al., 2005) exposure to

TBAs. Mechanisms underlying this observation were not addressed in these studies, however, factors responsible for this response may not be unique to βIII-tubulin as the levels of several other β-tubulin isotypes were also significantly increased

Table 1.5 Potential TUBB3 regulatory mechanisms in cancer

TUBB3 regulatory regions Downstream effects Cell type References TUBB3 promoter  Hypomethylated SP1 and AP2 Aberrant TUBB3 expression Ovarian cancer cells (Izutsu et al., 2008) binding sites TUBB3 intron 1  Demethylation of CpG islands in Ovarian cancer (Akasaka et al., 2009; Izutsu et al., Aberrant TUBB3 expression intron 1 and chromatin acetylation cells 2008)  Histone deacetylation of REST Melanoma, ovarian cancer, (Akasaka et al., 2009; Gao et al., Overexpression of TUBB3 putative binding sites & loss of REST colonic crypts 2012; Hatano et al., 2011)

 Oestradiol Aberrant TUBB3 expression Breast cancer cells (Saussede-Aim et al., 2009a)

 Androgen and androgen receptors Aberrant TUBB3 expression Colorectal cancer (Mariani et al., 2012) TUBB3 3’UTR  Hypoxia induced HIF-1α/ HIF- (Bordji et al., 2014; Forde et al., Glioblastoma cells, ovarian 2α/SOX9 binding to HIF response Aberrant TUBB3 expression 2010; Raspaglio et al., 2008; and prostate cancer cells elements Raspaglio et al., 2014) Preferential translation of  Hypoglycaemia and HuR Ovarian cancer cells (Raspaglio et al., 2010) TUBB3 Breast, ovarian and (Cochrane et al., 2010; Cochrane et  miRNA-200c Targeted silencing of TUBB3 endometrial cancer cells al., 2009) Others  SEMA6A TUBB3 inducer Ovarian cancer cells (Mozzetti et al., 2008)  K-Ras mutant protein Upregulation of TUBB3 protein Immortalised bronchial cells (Levallet et al., 2012)

(Ranganathan et al., 1998a; Ranganathan et al., 1996; Shalli et al., 2005). Results should be interpreted with caution, as very high doses of TBAs were used in some studies. For example, in MCF7 cells, TUBB3 gene expression has been shown to be inducible following acute exposure to 1 μM of vinorelbine, vinblastine or colchicine or 1 nM of paclitaxel (Lobert et al., 2011; Saussede-Aim et al., 2009b). The doses of vinblastine and vinorelbine used in these studies are extremely high concentrations. At 1μM, vinblastine is known to completely depolymerise microtubules and increase microtubule polymer mass in vitro (Jordan et al., 1991; Toso et al., 1993). The concentration used is not clinically relevant and the vinca alkaloid-induced TUBB3 expression is likely to be a compensatory response to microtubule depolymerisation, or an “off- target” effect on the transcriptional machinery or signalling pathways. Using the same dose of vinorelbine

(1μM) in a mutagenesis study, Saussede-Aim et al. (2009) further reported that vinca alkaloid treatments can enhance TUBB3 promoter activity via an AP-1 binding site located within the fragment 639 to 810 bp of the promoter (Saussede-Aim et al., 2009b).

Future investigation using chromatin immunoprecipitation (ChIP) is required to identify transcription factors responsible for this vinca alkaloid-induced TUBB3 expression at clinically relevant doses.

Like vinca alkaloids, Taxol has also been reported to alter the level of TUBB3 expression in tumours. For example, Kavallaris et al. (1997) reported that, while the level of individual β-tubulin isotypes remained the same in normal ovary and primary untreated ovarian tumours, analysis of ovarian carcinoma specimens from the same patient before and after chemotherapy revealed that TUBB3 and TUBB2c gene expression increased significantly in Taxol-resistant tumours post-treatment. As patients develop Taxol resistance after several cycles of Taxol/platinum combination therapy, it is difficult to differentiate whether the increased TUBB3 expression observed was a direct consequence 59 of chemotherapy-mediated regulation, a result of the resistant phenotype, or due to other off-target effects. Kavallaris et al. went on to show that partial suppression of TUBB3 in

NSCLC selected for resistance to Taxol and overexpressing this gene, were sensitised to

Taxol (Kavallaris et al., 1999), linking TUBB3/βIII-tubulin expression with Taxol sensitivity.

Methylation status of putative binding sites within the TUBB3 gene may also play a key role in its regulation. Izutsu et al. (2008) discovered that multiple putative SP1 and AP2 binding sites located within the TUBB3 promoter region are hypomethylated in several ovarian cancer cell lines, but not in non-cancerous ovarian tissues. Given that SP1 and its

DNA-binding activities are inducible under oxidative stresses and DNA-damage (Ryu et al., 2003), it is possible that under chemotherapeutic insults, hypomethylated TUBB3 promoter regions with enhanced SP1 signalling may contribute to aberrant TUBB3 expression in ovarian cancer. Future studies are required to clarify whether SP1 and AP2 can directly bind to those hypomethylated regions and drive aberrant TUBB3 expression.

Moreover, it will be important to determine if hypomethylation of TUBB3 occurs in patient samples with upregulated TUBB3 expression.

1.4.4.2 TUBB3 intron 1

Repressor element 1 silencing transcription factor (REST) is a global transcriptional silencer that represses neuron-specific gene expression in non-neuronal cells. Typically,

REST forms complexes with chromatin-modifying enzymes, such as HDACs, coREST, mSin3a, MeCP2 and suppresses neuronal gene expression in epithelial cells by epigenetic mechanisms (Ballas et al., 2005). REST-mediated mechanisms and chromatin remodelling have been demonstrated to play an important role in TUBB3 regulation in

60 several epithelial cancer cells (Akasaka et al., 2009; Gao et al., 2012; Izutsu et al., 2008;

Shibazaki et al., 2012). In ovarian cancer cells, DNA demethylation at the CpG island within TUBB3 intron 1 has been shown to result in βIII-tubulin overexpression, with chromatin acetylation accelerating the process (Akasaka et al., 2009; Izutsu et al., 2008).

In silico analysis revealed a putative binding site for REST within the CpG island of

TUBB3 intron 1, called repressor element 1 (RE-1) (Izutsu et al., 2008). Histone deacetylation of this RE-1 motif has been shown to partially contribute to TUBB3 overexpression in melanoma (Akasaka et al., 2009). This poses the question- is dysregulated REST-TUBB3 the primary cause of aberrant TUBB3 expression in tumours of non-neuronal origin? The loss of REST in cancers has been linked to the aberrant expression of neuronal genes in the clinic. A negative correlation between REST and

TUBB3 expression has been reported in skin, ovarian, and small cell lung cancer biopsy samples (Akasaka et al., 2009; Gao et al., 2012; Hatano et al., 2011; Kreisler et al., 2010), while in normal non-neoplastic tissues TUBB3 is barely detectable. Additionally, REST gene deletion and frame-shift mutations are frequently observed in colon and small cell lung cancers (Coulson et al., 2000; Westbrook et al., 2005). In mouse colonic crypts, targeted REST genetic ablation has resulted in upregulation of TUBB3 expression (Gao et al., 2012). Furthermore, TUBB3 expression can be independently induced upon REST siRNA treatment in cancer cells (Akasaka et al., 2009; Gao et al., 2012). Together, these findings suggest REST as a transcriptional silencer of TUBB3 and that dysfunctional

REST, in conjunction with epigenetic modifications in TUBB3 intron 1, may be important mechanisms underlying aberrant TUBB3 expression in tumours of non-neuronal origin.

In cervical cancer HeLa cells, TUBB3 expression appears to be regulated in a cell-cycle dependent manner, with maximal expression in G2/M phase of the cell cycle (Shibazaki et al., 2012). Shibazaki et al. (2012) reported that this cell-cycle dependent TUBB3 61 expression is regulated at the chromatin level and is directly controlled by REST via a

RE-1 motif in TUBB3 intron 1 (Shibazaki et al., 2012). In the context of HeLa cells, REST was shown to dissociate from the TUBB3 RE-1 element during G2/M phase and recover in G1 phase, achieving timed TUBB3 expression (Shibazaki et al., 2012). Future studies are needed to better understand the mechanisms facilitating REST association and dissociation from the TUBB3 RE-1 element. Further, knockdown studies indicated that cell-cycle dependent TUBB3 expression is required for mitosis and normal cell growth

(Shibazaki et al., 2012). It is unclear whether REST-regulated TUBB3 expression is a cell-specific effect or can be applied to all non-neuronal cells.

Gonadal steroids and their corresponding receptors have also emerged as potential drivers of TUBB3 expression in neoplasia. In breast cancer cells, Saussede-Aim et al. (2009a) described an oestrogen-dependent TUBB3 regulatory pathway, where TUBB3 mRNA and protein expression are inducible upon oestradiol exposure. In silico analysis revealed that consensus oestrogen responsive elements are absent in the 3’- and 5’-UTR of TUBB3.

However, several binding sites for transcription factors known to be implicated in indirect oestrogen-regulations such as AP-1, NF-κB and SP1 were identified in the first intron of

TUBB3 (Saussede-Aim et al., 2009a). In the same study, oestradiol-induced TUBB3 expression could not be reproduced in oestrogen receptor (ER) negative breast cancer cell lines, and was abrogated after exposure to the ER antagonists tamoxifen and fulvestrant in several ER-expressing breast cancer cell lines. These findings suggest that oestradiol- induced TUBB3 expression is ER-dependent. The authors proposed that ERs may regulate

TUBB3 in an indirect manner, facilitating transcription factor binding to nearby corresponding sites in intron 1 and subsequent TUBB3 transcription activation.

Conflicting results were reported in invasive breast cancer specimens, where high TUBB3 expression was identified in both ER positive and ER negative breast tumour specimens 62

(Wang et al., 2013), raising the question whether ER is relevant to TUBB3 regulation in the clinic. This disparity could be explained by the different biology in cell models and clinical specimens. In the study by Wang et al. (2013), specimens were collected from patients with different pathological stages, with or without neoadjuvant chemotherapy, all of which could potentially contribute to high TUBB3 expression. In addition, patients in the study by Wang et al. (2013) were not treated with oestrogen and therefore further studies are required to assess the clinical value of ER in TUBB3 regulation in breast cancer.

In colorectal cancer, elevated TUBB3 expression is associated with invasive phenotypes in both genders (Cleveland, 1987), while metastatic disease and poor outcomes are associated with females exclusively (Mariani et al., 2012). In vitro analysis of 23 colorectal cancer cell lines suggested that TUBB3 is activated after exposure to androgens in males (Mariani et al., 2012), as with oestrogens in breast cancer cells (Saussede-Aim et al., 2009a). In both male and female colorectal cancer cell lines, stable silencing of androgen receptors (AR) yielded significant downregulation of TUBB3, raising the possibility that ARs play a significant role in driving TUBB3 expression. Importantly, in male colorectal cancer cells, the AR-dependent TUBB3 regulatory pathway is constitutively activated via testicular androgen, while in female colorectal cancer cell lines TUBB3 is only inducible upon serum starvation. This finding suggests that different kinds of AR regulatory regions might exist and are able to induce TUBB3 expression in response to external stimuli. Future mutagenesis and ChIP studies are required to identify

AR binding regions within the TUBB3 gene.

1.4.4.3 3’ untranslated region of TUBB3

As described in section 1.4.1, growing evidence suggests βIII-tubulin expression is a key adaptive response that is activated on cellular exposure to a toxic microenvironment. In solid tumours, cells often grow in a hypoxic microenvironment and cells with a highly efficient hypoxia-inducing factor (HIF)-orchestrated survival program possess an advantage to offset its selective pressure. HIF-1α and HIF-2α have recently been implicated in the transcriptional regulation of TUBB3 via the 3’UTR of TUBB3 and are thought to protect tumours against hypoxic injury (Bordji et al., 2014; Danza et al., 2012;

Forde et al., 2010; Raspaglio et al., 2008; Raspaglio et al., 2014). In A2780 ovarian cancer cells (A2780 cells), hypoxia has been shown to strongly induce βIII-tubulin gene and protein expression and this phenotype was directly linked to cisplatin and paclitaxel resistance (Raspaglio et al., 2008; Raspaglio et al., 2014). This process was shown to be transcriptionally regulated through the binding of HIF-1α to a hypoxia response element

(HRE) within the 3’ UTR of TUBB3 (TAGGCCACGTGTGAG, position +168 after the stop codon) (Raspaglio et al., 2008). Epigenetic regulation could also account for the event, as methylation at the 3’ enhancer region was shown to abolish hypoxia-induced and HIF-1α-mediated TUBB3 expression in ovarian cancer cells, prostate cancer cells and prostate tumour biopsies (Forde et al., 2010; Raspaglio et al., 2008). An alternative transcriptional mechanism regulating TUBB3, involving HIF-2α and sex determining region Y-box9 (SOX9) transcription factors, has also been recently described (Raspaglio et al., 2014). In ovarian cancer specimens, high levels of TUBB3 mRNA and protein were significantly associated with increasing levels of SOX9 and HIF-2α (Raspaglio et al.,

2014). In ovarian cancer cells, both SOX9 and HIF-2α silencing abrogated hypoxia- activated TUBB3 expression, suggesting a role as TUBB3 positive regulators. In silico analysis revealed a SOX9 specific binding site within the TUBB3 promoter region

(TTAGAACAAGGGCCTTCTGATACTT, position -980 to -968 upstream of the transcriptional start site) (Raspaglio et al., 2014). Although direct engagement of SOX9 to the TUBB3 promoter is yet to be confirmed, gene-reporter and site-directed mutagenesis studies all support its involvement in TUBB3 regulation in hypoxia

(Raspaglio et al., 2014). In glioblastoma clinical specimens, a direct link between hypoxia and elevated βIII-tubulin expression was demonstrated (Katsetos et al., 2009). One recent report implicated HIF-2α, but not HIF-1α in hypoxia-induced βIII-tubulin expression in glioblastoma cells (Bordji et al., 2014). It was suggested that HIF-2α achieved this transcriptional control of TUBB3 through binding to the two overlapping HREs located in the 3’UTR of the gene. It should be noted that both HIF-1α and HIF-2α/SOX9 mediated

TUBB3 regulation could be a cell-specific response, as it is not inducible upon hypoxia in some cell lines expressing high basal levels of βIII-tubulin (Bordji et al., 2014; Danza et al., 2012; Levallet et al., 2012; Raspaglio et al., 2008; Raspaglio et al., 2014; Shen and

Yu, 2008).

During hypoglycaemic stress, TUBB3 mRNA has been shown to be regulated via a posttranscriptional mechanism involving the RNA-binding protein HuR in A2780 cells

(Raspaglio et al., 2010). As an adaptive response to glucose deprivation, HuR selectively induces TUBB3 expression by binding to a motif within the TUBB3 3’ UTR. This not only allows HuR to protect TUBB3 mRNA from degradation, but also to directly escort

TUBB3 mRNA from the nucleus into cytoplasmic polysomes for preferential translation at the expense of other mRNA (Raspaglio et al., 2010). This hypoglycaemia-induced

TUBB3 expression is HuR-dependent and has been shown to directly contribute to cisplatin, paclitaxel and thiocolchicine resistance in several ovarian cancer cell lines.

Importantly, this phenomenon is reversible upon silencing of the TUBB3 3’ HuR putative binding site (Raspaglio et al., 2010), suggesting its significance in hypoglycaemia. The 65 same study also reported a positive correlation between cytoplasmic HuR staining and high TUBB3 expression with the worst outcomes for ovarian cancer patients. Together, these findings suggest that the HuR-dependent regulatory pathway contributes to aberrant

TUBB3 expression in ovarian cancer under hypoglycaemic conditions, encouraging selection of drug resistant tumour cells and contributing to poor clinical outcomes.

Whether this finding translates to other types of tumours remains to be determined.

1.4.4.3.1 TUBB3 3’UTR and microRNA

MicroRNAs (miRNAs) have emerged as potential post-transcriptional regulators of

TUBB3 mRNA (Cochrane et al., 2010; Cochrane et al., 2009; Köhler et al., 1996; Lobert et al., 2011). miRNAs are single-stranded RNA molecules that function via base-pairing with complementary sequences within target mRNA molecules, which results in gene silencing via translational repression or targeted degradation (Pillai et al., 2007; Rana,

2007). Cochrane et al. (2009 and 2010) demonstrated that miRNA-200c has a complementary binding site in the 3’UTR of TUBB3 (5’-CCUGCAGUAUU-3’), the binding of which facilitates targeted silencing of TUBB3. Notably, restoration of miRNA-

200c in breast, ovarian and endometrial cancer cells and in two independent xenograft models resulted in a marked reduction of TUBB3 mRNA and protein levels, restored anoikis sensitivity (Cittelly et al., 2012; Howe et al., 2011), re-established “epithelial” identity and restored sensitivity to TBAs (Cochrane et al., 2010; Cochrane et al., 2009).

Changes in miRNA-200c were reported in a number of cancer cell lines and clinical specimens. Specifically, several studies reported that low miRNA-200c expression is significantly associated with high βIII-tubulin protein levels, resistance to TBAs, high incidence of recurrence and poor survival in ovarian cancer patients (Cittelly et al., 2012;

Leskelä et al., 2011; Prislei et al., 2013). These findings suggest miRNA-200c negatively regulates TUBB3 expression and loss of miRNA-200c may result in βIII-tubulin overexpression in ovarian, breast and endometrial cancer.

Different miRNA have been shown to regulate different β-tubulin isotypes. For example, in MCF7 breast cancer cells, miRNA-100 can potently repress the mRNA expression of

TUBB, TUBB2a, TUBB2b and TUBB2c (Lobert et al., 2011). As mentioned in the preceding paragraph, miRNA-200c can negatively regulate TUBB3 mRNA expression, but whether the silencing is exclusive to TUBB3 remains to be elucidated. In silico analysis indicated that the miRNA-200c putative binding sites were also predicted in the

3’UTR of TUBB and TUBB2a, however experimental validation showed no association between miRNA-200 family members and TUBB and TUBB2a protein levels (Leskelä et al., 2011). Either the predicted binding site is non-functional or additional factors are required for this regulation.

Using immunohistochemistry, a recent study reported that in ovarian cancer patients with nuclear HuR expression, high levels of cytoplasmic miRNA-200c and low TUBB3 levels exhibit a better than average outcome (Prislei et al., 2013). The authors proposed a model for combined regulatory activity by miRNA-200c and HuR on TUBB3 expression in ovarian cancer. Under normal glycaemic conditions, HuR proteins are localised in the nucleus and miRNA-200c mediates TUBB3 repression. In hypoglycaemia, HuR proteins are predominantly localised in the cell cytoplasm. This may antagonise miRNA-200c- mediated TUBB3 repression, promote TUBB3 mRNA translocation into ribosomes and enhance their stability. This may in turn yield an upregulation of TUBB3 expression.

Future studies are required to test this hypothesis.

1.4.4.4 Other potential regulators

Several other factors have also been proposed to play a role in TUBB3 regulation. For example, overexpression of Semaphorin-6A (SEMA6A) is correlated with TUBB3 upregulation in ovarian cancer cells, while the reverse is observed in SEMA6A-silenced cells (Mozzetti et al., 2008). This suggests SEMA6A may positively regulate TUBB3

(Mozzetti et al., 2008). In NSCLC cells, Slug overexpression was shown to markedly suppress TUBB3 gene and protein expression (Tamura et al., 2013). Additional studies are required to assess whether Slug can directly bind to the only E-box motif in the

TUBB3 promoter region. The K-Ras signalling pathway has also been proposed to play a role in TUBB3 regulation in NSCLC (Levallet et al., 2012). In clinical samples, K-Ras mutations are strongly and frequently associated with positive TUBB3 expression

(Levallet et al., 2012). In immortalised human bronchial cells, expression of a K-Ras mutant protein was shown to significantly increase TUBB3 protein level, while TUBB3 mRNA remained unchanged (Levallet et al., 2012). This observation raises the possibility that TUBB3 protein translation or turnover may be controlled by K-Ras-induced signalling cascades. In further support of this notion, siRNA knockdown of K-Ras and pharmacologic inhibition of K-Ras downstream effectors resulted in βIII-tubulin protein downregulation (Levallet et al., 2012). Studies on these candidates are intriguing and further investigations are required to validate their role in TUBB3 regulation in NSCLC and potentially other cancer.

The regulation of TUBB3/βIII-tubulin expression is complex in both normal and cancer cells, as its expression can be stimulated or repressed by multiple factors. Altered epigenetic modifications, altered transcriptional controls, in conjunction with altered signalling pathways may all contribute to disrupted TUBB3 expression in cancers and subsequent survival advantages. Current knowledge in this field still remains puzzling 68 and different cell lines used could account for differences. Improved understanding of the mechanisms whereby βIII-tubulin is regulated in cancer is critical to the identification of novel drug targets and new therapeutic strategies for the treatment of aggressive and drug refractory cancers.

1.5 High throughput drug discovery

1.5.1 βIII-tubulin as a drug target

Aberrant βIII-tubulin expression is now recognised as a key determinant for drug resistance, aggressive disease and poor overall survival in NSCLC patients (Dumontet et al., 2005; Okuda et al., 2008; Seve et al., 2005a; Seve et al., 2005b; Zhang et al., 2012a).

Laboratory evidence has validated βIII-tubulin as a bona fide target in the treatment of

NSCLC. Preferential targeting of this tubulin isotype has been extremely challenging as:

(1) little is known about how βIII-tubulin expression is regulated in cancer; (2) the protein structure of tubulin isotypes share a high degree of amino acid homology; and (3) the crystal structure of the isotype-defining domain at the carboxyl terminus remains unresolved, thus, it is difficult to rationally design drugs that target this moiety. A number of approaches, including high throughput screening and in silico modelling have been explored to identify agents that maintain cytotoxic activity in drug-resistant tumours that overexpress βIII-tubulin. This had led to the discovery of anti-tubulin agents, IDN5390 and other seco-taxanes (taxane derivatives) (Ferlini et al., 2005; Pratesi et al., 2003).

IDN5390 have demonstrated anti-cancer activities against paclitaxel-resistant cell lines overexpressing βIII-tubulin (Ferlini et al., 2005; Pepe et al., 2009; Pratesi et al., 2003).

Molecular modelling studies predicted that βIII-tubulin would be targeted more specifically by IDN5390 than by paclitaxel (Ferlini et al., 2005; Pepe et al., 2009). To date, in vitro data on IDN5390 binding to βIII-tubulin are limited and in vivo data are lacking. Future studies are required to provide direct evidence on IDN5390 specific targeting of βIII-tubulin in vitro and in vivo.

A recent study exploited the knowledge that βIII-tubulin is part of a prosurvival pathway, facilitating the incorporation of the GBP-1 GTPase into microtubules and subsequent

70 binding and activation of downstream prosurvival kinases such as PIM1 (Andreoli et al.,

2014; De Donato et al., 2012) (described in section 1.4.1). This approach has led to the identification of a potential inhibitor of the GBP-1:PIM1 interaction, aza-podophyllotoxin

(De Donato et al., 2012). To the best of our knowledge, there is currently no chemical molecule that can specifically modify βIII-tubulin expression. High throughput screening is a suitable approach in identifying molecules that can modify a validated target, such as

βIII-tubulin, even in the absence of structural information or in depth knowledge of the target. This thesis utilised a high throughput cell-based screening approach and the rest of this section will give an overview of our current knowledge in regards to the design, optimisation and conduct of a high throughput screen.

1.5.2 High throughput screening

High throughput screening (HTS) refers to an efficient automated screening process in which a large number of drug-like compounds may be tested against a biological target of interest. HTS allows the identification of biologically active molecules as candidates for further investigation using chemical, biological or pharmacological experiments. The results of HTS provide useful starting points for drug design. HTS normally involves chemical compound libraries, industrial scale liquid-handling robots and sophisticated automation. HTS emerged in the late 1980s, and has matured into the principal hit identification method and a crucial source of drug discovery leads within pharmaceutical and biotechnology companies (Pereira and Williams, 2007). In the late 1990s, the maturation of liquid handling automation and large screening libraries allowed HTS to be adopted by researchers in academic institutions. The goal of HTS is to screen and select

“lead-like” chemical structures that will feed into drug discovery and development

71 pipelines in a therapeutic setting. In contrast to traditional methods of drug discovery, which rely on trial-and-error testing of chemical compounds in vitro or in vivo, high throughput drug screening begins with a validated biological target and a screening assay.

A lack of “druggability” and structural information for many novel molecular targets led to HTS becoming the method of choice for identification of small molecule modulators of these targets. Consequently, HTS attracted large sums of capital investment in advanced HTS technologies, including automation, miniaturisation and assay methodology. Today, most instruments in tissue culture and liquid-handling robotics allow the execution of high quality HTS. The standard format of HTS is 384-well plates, though the use of 1536- and 3456-well plate formats is not uncommon. Miniaturised plate-based assays are an efficient and cost-effective way to lead identification. Returns on investment in HTS are evident in the increasing number of leads, clinical candidates and marketed drugs arising from HTS. For example, Viramune, an anti-HIV drug, was a direct result of early HTS efforts (Grozinger et al., 2006). In addition, identified chemical molecules can also be used as molecular tools or probes to address biological questions.

1.5.3 Assay types

Assays developed for HTS can be divided into two broad categories; biochemical assays and cell-based assays. Biochemical assays are target-based in vitro assays and historically have been the predominant assay used in HTS. Such assays have been used to examine enzymatic activity, protein-protein interactions, nuclear receptors, ion channels and receptor-ligand binding. Biochemical assays are direct and specific to the target of interest. Due to the homogenous nature of the reactions, biochemical assays can be miniaturised easily with little variability. However, the activity of a small molecule in

72 reconstituted in vitro assays does not always reflect its activity in a cellular context, as multiple factors are overlooked. For example, this includes the requirement for cellular co-factors, membrane permeability, off-target effects, and cytotoxicity. In recent years, cell-based assays have emerged as a more physiologically-relevant assay and represent more than half of all HTS currently performed. In contrast to traditional biochemical assays, cell-based assays aim to identify modulators of a pathway or target-of interest in an environment with closer resemblance to physiological conditions, complete with intact regulatory networks and feedback mechanisms. For example, when performing a functional assay that involves the activation of EGFR, the entire pathway of interest can be examined, as opposed to a single-step biochemical assay. The benefit of such assays is two-fold, not only do these increase the identification of targets, but also provides novel chemical structures for lead identification. Moreover, cell-based assays allow for the selection of compounds that can cross cellular membranes and provide indications of cytotoxicity.

Cell-based assays can measure a variety of cellular features, such as cell proliferation, production of markers and activation of specific signalling pathways. Many cell-based readout assays rely on reporter gene technology. Reporter gene assays are based on the cloning of transcriptional control elements to a reporter gene. They are widely used to study gene expression at the transcriptional level (Wolff et al., 2008). These assays allow the study of regulation of a target gene by placing the cDNA of a reporter protein under the transcriptional regulation of the target gene’s regulatory elements in a vector that is stably or transiently transfected into mammalian cells. Most of these assays rely on the use of bioluminescent proteins, such as Firefly luciferase and Renilla luciferase, derived from Photinus pyralis and the sea pansy Renilla reniformis, respectively (reviewed in Fan and Wood, 2007). These enzymes catalyse the oxidation of their substrates and produce 73 a light reaction that can be measured directly and quantitatively. Renilla luciferase is a 36 kDa enzyme that catalyses the oxidation of coelenterazine to yield coelenteramide, which emits blue light with a spectral maximum of 480 nm (Lorenz et al., 1991). Due to signal amplification of cell-signalling cascades, reporter gene assays are highly sensitive and thus ideal for assay miniaturisation (Wu et al., 2010). Other key advantages of reporter gene assays include enhanced reliability, convenience, dynamic range and adaptability to

HTS. Despite these advantages, reporter gene assays are based on signal-transduction events that occur downstream of receptor activation and require gene expression. This causes long response times, spanning from hours to days, during which there is a possibility that a response could be interfered with by other intracellular pathways.

Another weakness of reporter gene technology is the high variability of cell response. To improve robustness, an internal reference signal should be included to correct the analytical response and separate specific response from non-specific interference. The introduction of a constitutively expressed second reporter gene can resolve this issue. A successful example of using this technology is the identification of a small molecule inhibitor, Disulfiram, for Metalloproteinase Matrix-9 promoter expression in a panel of cancer cells (Nair et al., 2008).

1.5.4 Experimental design and planning

Several important parameters, such as screening strategies and detection methods must be taken into consideration when designing a small molecule screen. Firstly, it is essential to identify a screening strategy that maintains the appropriate biological context while balancing feasibility of reagent availability and adaptation to automation. For cell-based assays, these include the choice of biological system (primary human cells, immortalised cells, transformed cells or model organisms), assay type (functional, reporter gene or

74 morphologic), assay readout system (uniform well measurements or high content imaging), follow-up experiments (secondary screen, target identification, dose-response experiments and validation tests) and determination of data analysis strategies for data interpretation.

A variety of detection methods can be employed in cell-based assays. These can be broadly divided into uniform well- and high content measurements, where one or multiple measurements are made per well, respectively. Uniform well measurement can be conducted by fluorescence, bioluminescence or spectrophotometric methods (reviewed in Fan and Wood, 2007). High content screening refers to techniques where multiple measurements are obtained from a single well, often representing sub-populations of cells or sub-cellular features (reviewed in Brodin et al., 2015). In comparison to uniform well measurements, high content screening is slower but it provides a greater wealth of contextual information, that allows a more profound understanding of the behaviour and mechanism of action of small molecules. High content screening involves image-based measurements using automated high resolution microscopy and robotic handling.

1.5.5 Assay development and optimisation for HTS

Assay development refers to the process where prospective assay designs for a biological target are tested and evaluated. The selected design is then further optimised in terms of throughput, time, cost, sensitivity, signal dynamic range and replicability, then adapted to the screening instruments. Assay development is critical to the success of HTS, where hundreds of thousands of data points are collected and analysed at once. Given the expensive nature of HTS (estimate for $1-2 AUD per compound, i.e. $30,000 to $60,000 for a 30,000 compound screen, excluding labour and chemical reagent costs), vigilant

75 assay development and a small scale pilot screen (5K compound library) is usually implemented before performing a primary screen to ensure data is of high-quality and biological relevance.

A variety of experimental parameters need to be assessed to achieve optimal signal to background levels. These include cell seeding density, determination of the incubation time for drug treatment and determination of time points at which measurements are obtained. The inclusion of positive and negative control assays and vehicle assays is imperative to provide a gauge to scale and normalise data from run-to-run, plate-to-plate and well-to-well. Additionally, the stability or half-life of assay reagents and readout signals is evaluated over the time course of the screen. The time lapse between the first and last well on a plate and the first and the last plate of a batch are examined and planned up-front to ensure both accuracy and consistency between plates and batches.

Furthermore, the effect of compound solvent, such as DMSO, is also assessed. A titration of DMSO from 0 to 2% is recommended for cell-based assays.

Once optimal experimental conditions have been identified, miniaturisation of assays ensues. Miniaturisation provides key benefits, such as reducing the investment of labour and time thereby lowering the cost. It is essential that the level and robustness of reaction signal and signal-to-background ratio remain constant despite substantial reduction in reagent volume and cell numbers. Compatibility of liquid-handling equipment and the miniaturised assay should be assessed to ensure the accuracy of dispensing small volumes of compounds (5nl - 0.5μl). Adjustment of instrument settings to the optimal height, speed, volume and position of tips and/or pin tool dispensers is required. This allows even distribution of compounds without splashing of reagents or damaging cells.

Miniaturisation to a 1536-well format is achievable but less commonly utilised, as the accuracy of these assays is likely to be compromised.

Assay quality and reproducibility is another critical factor to achieve successful HTS. The most widely used and accepted statistical analysis of assay robustness and readiness is the Z’ factor (Zhang et al., 1999), and the variation is tested using positive- and negative control assays (compound solvent). Briefly, Z’ is a statistical analysis used to validate assay reliability for HTS. This encompasses several different factors, such as reproducibility, robustness and sensitivity. This assessment of multiple aspects regarding assay quality means that Z’ is more widely used than other validation assays (Zhang et al., 1999). The Z’ value is determined by the separation of a positive activity and a background control in the absence of test compounds, and is calculated using the formula:

Z’=1 − ( ) , where SD stands for standard | | deviation. An assay with Z’=1 indicates an ideal assay for HTS, 0.5 ≤ Z’< 1 indicates an excellent assay, while 0 < Z’ < 0.5 is considered marginal and may be suitable for HTS but further optimisation and additional tests are required. Assays with Z’<0 are essentially unsuitable for HTS. After the assay has been optimised with an acceptable Z’ value, a small scale pilot screen containing 2,000 to 10,000 compounds should be performed to validate the assay in high throughput.

1.5.6 Execution of HTS

The execution of HTS takes thorough planning. The screener is required to prepare sufficient supply of consumables, reagents, and cells for both primary and secondary screens as well as retesting and ten-point dose response experiments. In the primary screen, each compound is usually tested in single wells, while the secondary screen is 77 obtained using triplicates. In addition to the primary cell-based assay, a housekeeping cell-based assay can also be included in the secondary screen to rule out compounds possessing off-target effects. This additional housekeeping cell-based assay needs to be tested and validated the same way as the primary cell-based assay for HTS. Appropriate data analysis and interpretation should be carefully picked well in advance for all stages of the screen.

1.5.7 Validation of drug candidate and lead optimisation

After leads have been obtained from HTS, follow up validation studies are designed and decided based on the nature of the question asked, whether it is to identify a new biological target, a novel therapeutic agent or an innovative research tool. For a cell-based assay that screens for a drug that modulates the level of a protein, it is important to clarify whether the effect of a hit observed was the result of an off-target event. The identified hit compound can interfere/interact with protein candidates other than the target protein while producing the same phenotype. Techniques such as chemical proteomics (pulldown studies and mass spectrometry), RT-PCR and Western blotting are commonly used to elucidate the identity of the protein that is being targeted. This information allows a better understanding of the specificity of the hit compound and assists in downstream drug discovery process.

1.6 Thesis perspective and significance

Lung cancer is the leading cause of cancer-related death worldwide. Advanced NSCLC accounts for more than 80% of lung cancer cases. Despite the advances in frontline treatments, overall survival rate for patients with NSCLC remains dismal. Resistance to chemotherapy is the major cause of treatment failure in lung cancer. In order to improve the outcomes for patients, there is an urgent need to discover new effective treatment strategies. The key objective of my PhD project is to identify novel small chemical molecules and bioactives that can specifically and potently modulate TUBB3 promoter activity, in a NSCLC cell line, H460, which aberrantly expresses endogenous βIII- tubulin.

The potential outcome of this project is two-fold. The identification of a βIII-tubulin inhibitor will lead to the development of a new therapeutic that modifies the expression of βIII-tubulin. The identification of a βIII-tubulin enhancer will lead to the development of a chemical probe to elucidate TUBB3/βIII-tubulin regulation in NSCLC cells. These studies may lead to the identification of novel therapeutic targets and improve targeting of drug resistant tumours that overexpress this tubulin isotype. The significance of this study is highlighted by the lack of effective treatments for advance stage lung cancer and the subsequent social economic burden that this causes to society. Ultimately, newly identified cellular factors will present new therapeutic targets for drug resistant lung cancers, with enormous potential to benefit patient outcomes by sensitising tumours to conventional chemotherapeutics.

1.7 Thesis aims

The central goal of this thesis is to ultimately target a tubulin-mediated drug-resistance mechanism in NSCLC that is refractory to conventional chemotherapeutics. The key objective is to identify small chemical molecules and bioactives that can specifically and potently modify βIII-tubulin promoter activity. This thesis addressed the key object with the following aims:

1. To develop a cell-based high throughput screen, using TUBB3 promoter activity as a

readout.

2. To identify small molecule chemicals that can modulate βIII-tubulin expression using

cell-based screen.

3. To characterise "hit" compounds and establish structure-activity relationships.

4. To identify cell-based factors that regulate βIII-tubulin expression.

Chapter 2

Materials and Methods

2.1 Materials

2.1.1 Cytotoxic drugs and other chemicals

Cytotoxic drugs were obtained as follows: Paclitaxel and Epothilone B (Calbiochem, San

Diego, CA); Vinblastine (David Bull laboratories, VIC, Australia); Cisplatin (Pfizer,

West Ryde, NSW, Australia); Dimethylsulphoxide (DMSO) and Actinomycin D (Sigma-

Aldrich, Castle Hill, NSW, Australia).

2.1.2 Tissue culture

Reagents for tissue culture were purchased as follows: Rosewell Park Memorial Institute

(RPMI), foetal calf serum (FCS), and Trypan blue (Invitrogen Life Technology, Carlsbad,

CA); phosphate buffered saline (PBS) and puromycin dihydrochloride (Sigma-Aldrich,

Castle Hill, NSW, Australia); sterile TC-treated 96-well, 24-well, 12-well, 6-well plates,

T25, T75 and T125 tissue culture flasks for transfection and single cells cloning (Corning

Life Sciences, Acton, MA, USA); aspirator, 10 mL and 25 mL pipettes (Corning Life

Sciences, Acton, MA, USA); 384-well white flat bottom polystyrene TC-treated microplates, 96-well flat clear bottom black polystyrene TC-treated microplates and T75 tissue culture flasks for high throughput drug screening (Corning Life Sciences, Acton,

MA, USA);

2.1.3 Plasmid sequencing and restriction digest

Reagents for plasmid sequencing were purchased as follows: human βIII-tubulin promoter- Renilla luciferase- puromycin resistant vector (SwitchGear Genomics, Menlo

Park, CA, USA); BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems,

Mulgrave, VIC, Australia); nuclease-free water (Promega, Madison, WI, USA); forward primer 5’- GATCCGACGCTTTGTTTCTTCT-3’ and reverse primer 5’-

GTCGAGCACGTTCATCTGCTT-3’ for βIII tubulin promoter sequence amplification

(GeneWorks, Hindmarsh, SA, Australia); ethanol and ethylenediaminetetraacetic acid

(EDTA) (Sigma-Aldrich, Castle Hill, NSW, Australia).

Reagents for restriction digest were purchased as follows: nuclease-free water (Promega,

Madison, WI, USA); BGL-II restriction endonuclease (New England BioLabs, Hitchin,

Herts, United Kingdom); MLU-I restriction endonuclease (Invitrogen, Mulgrave, VIC,

Australia); buffer E (Roche, Dee Why, NSW, Australia); agarose powder (Thermo

Scientific, Rockford, IL, USA); Tris-acetate-EDTA (TAE) buffer (40mM Tris acetate,

1mM EDTA)

2.1.4 Plasmid transfection and single clone selection

Reagents for plasmid transfection into H460 NSCLC cells were purchased as follows:

Lipofectamine 2000 and Opti-MEM I reduced serum medium (Invitrogen, Mulgrave,

VIC, Australia); human TUBB3 promoter-Renilla luciferase-puromycin vector, human

GAPDH promoter-Renilla luciferase plasmid-puromycin vector, human random promoter (scrambled sequence)-Renilla luciferase-puromycin vector, and empty promoter-Renilla luciferase-puromycin vector (SwitchGear Genomics, Menlo Park, CA,

USA); spectrophotometer (BioRad Laboratories, Gladesville, NSW).

2.1.5 Bioluminescence assay and high throughput drug screening

Reagents for bioluminescence assay and high throughput drug screening were purchased as follows: Promega Renilla-GLO Luciferase Assay System (Promega, Madison, WI,

USA); DMSO and Actinomycin D (Sigma-Aldrich, Castle Hill, NSW, Australia); Victor3 fluorescent plate reader (PerkinElmer, Waltham, Massachusetts, USA).

2.1.6 Protein isolation and Western blotting

Reagents for protein isolation, western blot analysis were purchased as follows: 4-15%

Tris-HCl gradient polyacrylamide gels and dual colour precision plus protein standards

(Bio-Rad, Hercules, CA, USA); Ponceau S sodium salt, ethylene glycol-bis (β- aminoethyl ether)- N,N,N’,N’-tetraacetic acid (EGTA), protease inhibitor cocktail, phenylmethylsulfonyl fluride (PMSF, Tween-20 and Trizma base (Sigma-Aldrich, Castle

Hill, NSW, Australia); sodium dodecyl sulphate (SDS) (ICN Biomedicals, Aurora, OH,

USA); noniodet P-40 (Fluka, Buchs, Switzerland); glycerol (Fronine, Riverstone, NSW,

Australia); hybond C extra nitrocellulose membranes and Pierce ECL Plus Western

Blotting Substrate Kit (Thermo Scientific, Rockford, IL, USA); 3MM blotting paper

(Whatman, Madistone, United Kingdom); 6-well and 96-well clear flat bottom polystyrene TC-treated microplates (Corning Life Sciences, Acton, MA, USA).

Primary antibodies were obtained as follows: anti-α-tubulin (clone DM1A) and anti-class

IV β-tubulin (βIV-tubulin; clone ONS.1A6) (Sigma Aldrich, Castle Hill, NSW, Australia); anti-class II β-tubulin (βII-tubulin; clone 7B9) and anti-class III β-tubulin (βIII-tubulin; clone TUJ1) (Covance, Richmond, CA, USA); anti-class I β-tubulin (βI-tubulin; clone

SAP4G5) (Abcam, Cambridge, MA, USA). Secondary antibody, mouse IgG, HRP-linked whole antibody from goat (GE Healthcare, Bioscience, Buckinghamshire, UK).

2.1.7 Electrophoresis

Reagents for electrophoresis were purchased as follows: boric acid and bromophenol blue, ammonium persulfate, ethidium bromide, and N,N,N’,N’- tetramethylethylenediamine (TMED) (Sigma-Aldrich, Castle Hill, NSW, Australia); 1kb

DNA ladder (Invitrogen, Mulgrave, VIC, Australia).

2.1.8 RNA isolation, cDNA synthesis and real time-PCR

Reagents for RNA isolation, cDNA synthesis and PCR were purchased as follows:

RNeasy Plus Mini Kit (Qiagen, Hilden, Germany); nuclease-free water (Promega,

Madison, WI, USA); High Capacity cDNA Reverse Transcription Kit (Applied

Biosystems, Foster City, CA, USA); Power SYBR Green RT-PCR Master Mix (Applied

Biosystems, Warrington, UK).

2.2 Methods

2.2.1 Maintenance of H460 NSCLC cells

A H460 TUBB3PROM-LUC readout system, or human NSCLC H460 cells stably transfected with TUBB3 promoter-Renilla luciferase constructs were maintained in RPMI media, supplemented with 10% FCS. The cells were grown at 37°C in humidified atmosphere with 5% CO2. H460/TUBB3p-luc cells was cultured under selective pressure, 1μg/μL of puromycin at all times, except stated. All cultures were routinely screened for

Mycoplasma spp. and found to be free of contamination.

2.2.2 Validation of human TUBB3 promoter-Renilla Luciferase vector construct

High throughput screening of chemical small molecule libraries is an exciting strategy towards the discovery of potential new pharmaceutical agents capable of specifically and potently inhibiting aberrant expression of βIII-tubulin. Prior to developing a small molecule functional readout system, the human TUBB3 promoter-Renilla luciferase expression construct was first validated via sequencing and restriction digest.

2.2.2.1 Restriction digest and 2% agarose gel

Restriction digest was performed to validate the size of human TUBB3 promoter

(Appendix I), GAPDH promoter (Appendix I), random promoter (scramble sequence that does not encode any functional protein) (Appendix I) and empty promoter cloned into the

BGL-II/MLU-I cloning site (Appendix I) of the pLightSwitch_Prom_puro_RenSP vector

(Figure 2.1).

Figure 2.1 pLightSwitch_Prom_puro_RenSP vector construct map.

Promoter sequences were cloned into the BGL-II/MLU-I cloning site of the 4.7kb pLightSwitch_Prom_puro_RenSP vector, upstream of the synthetic Renilla luciferase reporter gene (RenSP) and the puromycin resistance gene. Upon introduction of the plasmid intro mammalian cells, the promoter together with cellular machineries drive the downstream expression of luciferase reporter gene, resulting in Renilla luciferase protein expression. Upon substrate (coelenterazine and oxygen) addition, the promoter activity is measured as bioluminescence readout.

The 20 μL restriction digest reaction mix consisted of: 2 μL 10x buffer, 0.5 μL MLU-1 and 0.5 μL BGL-II restriction endonuclease, 500 ng of Renilla luciferase-puromycin plasmids expressing either human TUBB3 promoter, GAPDH promoter, random promoter or empty promoter. Reaction mix was incubated in 37°C water bath overnight.

An 0.8% agarose gel was prepared by dissolving agarose powder in 1x TAE buffer using microwave. The gel was briefly cooled, gel red added and swirled to combine. Running agarose gel was poured into a gel tray carefully and allowed to set at room temperature for 1 h. 2.5 μL of 6x gel loading dye was added to restriction digest product and 1kb

DNA ladder separately before being loaded in 0.8% agarose gel. Electrophoresis was performed at 80 V for 1.5 h. Restriction digest products were then visualised by Gel Doc

XR System (BioRad Laboratories, Gladesville, NSW).

Restriction digest was performed to validate the size of TUBB3-, GAPDH-, random- and empty promoter insertions cloned into the BGL-II/MLU-I cloning site of the pLightSwitch_Prom_puro_RenSP vector. The insert sizes for TUBB3-, GAPDH-, random- and empty promoter inserts were 975, 1063, 966 and 0 bp, respectively. The pLightSwitch plasmid is 4762 bp in size. Restriction digest products were electrophoresed on a 0.8% agarose gel and visualised by Gel Doc XR System (Figure 2.3). The GAPDH promoter- pLightSwitch_Prom_puro_RenSP plasmid was partially digested using endonuclease BGL-II (Figure 2.3, lane 2) and MLU-I (Figure 2.3, lane 3), respectively.

The partially digested product consists 4762 bp pLightSwitch plasmid_Prom_puro_RenSP and 1063 bp of GAPDH promoter insert. Together, the molecular size of partial digest product is 5825 bp. Indeed, this was reflected in both

Figure 2.3 lane 2 and 3 of the agarose gel, where one single band was observed above the

5000 bp marker. Lane 1 illustrates complete restriction digest products of GAPDH promoter inserts (1063 bp) and pLightSwitch_Prom_puro_RenSP plasmids (4762 bp). 88

The two single bands located just above 1000 bp and 5000 bp suggests complete restriction digests with their corresponding molecular sizes. The single band in Figure 2.3 lane 4 represents intact GAPDH promoter-pLightSwitch_Prom_puro_RenSP plasmids.

The uncut supercoiled plasmid had run slightly farther/ faster than the linearised version seen in lane 2 and 3.

The TUBB3 promoter-pLightSwitch_Prom_puro_RenSP plasmid was partially digested using endonuclease BGL-II (Figure 2.3, lane 6) and MLU-I (Figure 2.3, lane 7), respectively. The partially digested product consist 4762 bp pLightSwitch_Prom_puro_RenSP plasmid and 975 bp of TUBB3 promoter inserts.

Together, the molecular size of the partial digest product is 5737 bp. This was demonstrated in the agarose gel where one single band was observed above the 5000 bp.

The single band at roughly below 1000 bp mark in lane 5 represents the TUBB3 promoter insert (975 bp) and the single band located around 5000 bp represents the pLightSwitch_Prom_puro_RenSP plasmids (4762 bp). This indicates that the complete restriction digests using both endonucleases BGL-II and MUL-I was effective. In Figure

2.3 lane 8, no endonucleases were used, and therefore the single band at above 5000 bp mark represents intact TUBB3 promoter-pLightSwitch_Prom_puro_RenSP plasmids.

Similar results were observed for random promoter-pLightSwitch_Prom_puro_RenSP plasmid digests (Figure 2.3, lane 9 to 12). A clear single band above 5000 bp was detected in both BGL-II (Figure 2.3, lane 10) and MLU-I (Figure 2.3, lane 11) endonucleases treated plasmids, indicating partially digested plasmid product. Using complete restriction digest (Figure 2.3, lane 9), the two clear bands in lane 9 correspond to pLightSwitch_Prom_puro_RenSP plasmid (4762 bp) and random promoter insert (966

89 bp). The single band in lane 12 represents intact random promoter- pLightSwitch_Prom_puro_RenSP plasmids (4762 bp).

No promoter inserts were cloned into the empty promoter- pLightSwitch_Prom_puro_RenSP plasmids and a clear single band was visualised in partial digest products (Figure 2.3, lane 14 and 15), complete digest products (Figure 2.3, lane 13) and intact empty promoter-pLightSwitch_Prom_puro_RenSP plasmids (Figure

2.3, lane 16). In summary, restriction digest coupled with electrophoresis analysis have demonstrated the correct size for each promoter insertion. Together with DNA plasmid sequencing, these data validates TUBB3-, GAPDH-, random and empty promoter- pLightSwitch_Prom_puro_RenSP vector constructs for generation of cell-based readout systems.

2.2.2.2 Amplification of human TUBB3 promoter expressing plasmid DNA

followed by plasmid sequencing

Forward and reverse primers were manually designed around the human TUBB3 promoter-Renilla luciferase- puromycin vector, as shown in Table 2.1. The 20 μL reaction mixture consisted of 1 μL DNA plasmid (386 ng/ µL), 3.2 μL each forward and reverse primer, 1 μL BigDye® Terminator v3.1 Cycle Sequencing Kit, 3.5 μL 5x buffer and 11.3

μL RNase-free water. The human TUBB3 promoter region was amplified using polymerase chain reaction (PCR): amplification over 25 cycles at 96°C for 10 s, 50°C for

5 s and 60°C for 4 min. To purify, the PCR product was washed and precipitated in 60

μL of 100% ethanol and 5 μL of 125 mM EDTA for 15 min at room temperature. The precipitated PCR product was then collected from 20 min of centrifugation at 14, 000 g, follow by removal of supernatant. DNA pellet was further washed with 160 μL

Figure 2.2 BGL-II and MLU-I endonuclease restriction sites.

(A) BGL-II endonuclease recognises the palindromic sequence AGTCT and cuts between the A and G on both the top and bottom DNA strands (black arrows, restriction sites).

This leaves a sticky end on each strand, GATCT, allowing the ligation with a complementary overhang of the pLightSwitch_Prom_puro_RenSP vector. (B) MLU-I endonuclease recognises the palindromic sequence ACGCGT and cleaves between A and

C of both strands.

A B 5’…AGATCT…3’ 5’…ACGCGT…3’ 3’…TCTAGA…5’ 3’…TGCGCA…5’

Figure 2.3 Restriction digest of TUBB3-, GAPDH-, random- and empty promoter- pLightSwitch plasmids.

GAPDH promoter-pLightSwitch plasmid (4) was subjected to partial digest using endonucleases BGL-II (2) and MLU-I (3), respectively and complete digest using both

BGL-II and MLU-I (1). TUBB3 promoter-pLightSwitch plasmid (8) was subjected to partial digest using endonucleases BGL-II (6) and MLU-I (7), respectively and complete digest using both BGL-II and MLU-I (5). Random promoter-pLightSwitch plasmid (12) was subjected to partial digest using endonucleases BGL-II (10) and MLU-I (11), respectively and complete digest using both BGL-II and MLU-I (9). Empty promoter- pLightSwitch plasmid (16) was subjected to partial digest using endonucleases BGL-II

(14) and MLU-I (15), respectively and complete digest using both BGL-II and MLU-I

(13). DNA ladder (M): 5000, 3000, 2000, 1000 and 500 bp. The pLightSwitch plasmid itself is 4762 bp in size.

Table 2.1 Primer sequences for plasmid sequencing of the human TUBB3 promoter-

Renilla luciferase- puromycin vector construct

Primer name Primer sequence

Forward primer- TUBB3 5’- GATCCGACGCTTTGTTTCTTCT- 3’

promoter

Reverse primer- Renilla 5’- GTCGAGCACGTTCATCTGCTT- 3’

luciferase

95 freshly made 70% ethanol three times, collected by 10 min of centrifugation at 14, 000 g.

The ethanol was carefull aspirated off, the DNA pellet air-dried and stored at 4°C away from light. The DNA sample was then sequenced at Ramaciotti Centre in the University of New South Wales to confirm the presence of human TUBB3 promoter sequences.

The TUBB3 promoter-Renilla luciferase vector construct was amplified and sequenced at

Ramaciotti Centre in the University of New South Wales to confirm the presence of human TUBB3 promoter sequences. The result of this procedure revealed that the TUBB3 promoter region cloned in the Renilla luciferase-puromycin expression vector is identical to the human βIII-tubulin promoter sequence.

2.2.3 Generation of H460 single clones stably expressing human TUBB3 promoter

2.2.3.1 Determination of puromycin selection dose

To select H460 cells that have stably taken up the TUBB3 promoter-Renilla luciferase construct, puromycin treatment was used as a selective pressure. The minimum concentration of puromycin required to eliminate untransfected H460 cells was determined by the following method. H460 cells (6 × 105 cells/ well) were seeded in 6- well plates and exposed to 0, 0.5, 1, 1.5, 2, 2.5 µg/ mL of puromycin. The cells were incubated for 14 days, with the selective medium being replaced every 3 days. Using

Trypan blue assays, % cell viability for each puromycin treatment was examined on a daily basis using the equation: % cell viability= (total cell number – dead cell number)/ total cell number. The lowest concentration of puromycin that gave significant cell death within 3-5 days and killed all the cells within the 14 days was determined to be optimal for selecting stable transfectants. The optimal puromycin concentration for selection of stable transfectant for H460 cells was 1 µg/ mL. 96

2.2.3.2 DNA plasmid transfection, stable clone selection

Next, a cell-based readout system was developed, where TUBB3 promoter-Renilla luciferase expression construct was stably transfected into H460 NSCLC cells. 6 x 105

H460 cells were plated in a single well of a 6 well plate one day before transfection. 24 h post-plating, cells were transfected with 5 μL Lipofectamine 2000, 2 μg of plasmid and

428 μL of Opti-MEM. 24 h post-transfection, the old medium containing Opti-MEM and

Lipofectamine 2000 was removed. Cells were cultured under 37°C, 5% CO2. When cells were 80% confluent in the 6-well plate, cells were harvested and expanded in T25 then

T75 flasks. Cells were further expanded in two T75 flasks, one for cryogenic storage and the other for single cell cloning.

During stable clone selection, cells were selected under puromycin pressure (1μg/μL) at all times. For single cell cloning, stably transfected H460 cells were harvested and diluted to 1 cell/ 100 μL RPMI/ well, then plated in five 96 well plates. Single clones were selected from a single well into a well of a 24-well plate for expansion. One hundred single-cell clones expressing the TUBB3 promoter-luciferase construct were screened and isolated based on their luciferase activity and growth rate. The growth rates of single-cell clones were examined using Alamar Blue assays. Briefly, 2000 stably transfected H460 single-cell clones were plated in 100 µL RPMI in each well of a 96 well plate. Twenty

μL of Alamar Blue assay was added to each well at time points: 1, 24, 48, 72, 96 h post- plating. After incubation for 5 h before, absorbance with a Spectrophotometer was measured at 590- 570 nM. Clonal doubling time was plotted as mean ± SEM and statistically analysed by unpaired t-test at a significant level of p < 0.05, using GraphPad

Prism. Clones that exhibited a similar growth rate to wild-type H460 cells were selected.

Further, bioluminescence assays were performed on single-cell clones to assess their

TUBB3 promoter activities. Briefly, 2000 stably transfected H460 single-cell clones were plated in 100 µL RPMI in separate wells of a 96-well plate. Twenty-four h post-plating,

100 µL of Promega Renilla Glo Luciferase assay was added to each well for 10 min at room temperature before reading using the Victor3 plate reader. Assay reagent and cell culture were returned to room temperature before performing the assay. Clones expressing desirable level of Renilla luciferase and possessing a similar growth rate to the wild-type H460 parental cells were picked, expanded and cryopreserved for additional experiments. Selected clones were grown continuously in the presence of puromycin, however, prior to experimental assays; the cells were grown in the puromycin-free medium for 7 days. Together, a high luciferase-expressing clone possessing a similar growth rate to the wild-type H460 cells was picked and used in the small molecule screen.

Additionally, endogenous βIII-tubulin protein levels were compared between H460 parental cells and H460 cells being single and double transfected with the TUBB3 promoter-luciferase expression construct, using Western blotting (refer to section 3.2.1).

Cell lysates were collected at 72 h post-transfection and 10 μg of whole cell lysates were loaded for each sample. As controls, H460 clones stably expressing GAPDH- , random- and empty promoter-Renilla luciferase expression constructs have also been generated.

These controls were included in the validation and secondary drug screen to minimise non-specific inhibitors targeting luciferase and GAPDH gene expression.

2.2.4 High throughput compound screening

2.2.4.1 Determination of DMSO and Actinomycin D dose for control assays

Negative (DMSO) and positive control (Actinomycin D, a general transcription repressor) assays were established for the small molecule screen. Increasing

98 concentrations of DMSO (0%, 0.1%, 0.2%, 0.4%, 0.8%, 1.0%) and Actinomycin D (0,

0.1, 0.5, 1.0, 2.0, 10.0 µM) were examined over a series of time points (6, 12, 18, 24 h).

Several assay parameters were examined, including temperature, seeding density, time points and signal-to-background ratio. The difference between the mean of positive and negative controls demonstrated the largest signal dynamic range over 24 h of treatment, i.e. low signal-to-background ratio. The final negative and positive assay conditions include: 0.1% DMSO and 0.1µM Actinomycin D treatment over 24 h, at 37°C and 5%

CO2.

2.2.4.2 Primary screening of bioactive and chemical libraries

A moderate Renilla luciferase-expressing clone possessing a similar growth rate to the wild-type H460 cells was picked and used in the primary small molecule screen. The primary screen on 30,224 chemical small molecules and additional 2480 bioactive compounds (LOPAC Library and TOCRIS) were subsequently performed, using liquid handling robotics from the ACRF Drug Discovery Centre. The cut-off of “hits” was defined as compounds that are able to yield 75% or greater luminescence reduction relative to DMSO control (0.1% DMSO).

2.2.4.3 Secondary counter screen

In order to rule out false positive hits in those 406 primary “hit” compounds, a secondary screen was performed. In addition to the TUBB3 readout system, GAPDH readout system was also included. In the secondary screen, compounds that potently repressed TUBB3 promoter activity and had minimal effects on GAPDH promoter activity were sought after. Leads were selected based on the difference between the average fold inhibition of

GAPDH and TUBB3-promoter activities, calculated using, Δ fold inhibition= GAPDH average fold inhibition- TUBB3 average fold inhibition. A cut-off of Δ fold inhibition

>0.28 was used. This value was arbitraily selected with consultation with the Drug

Discovery Centre. Twenty eight small molecule hits and ten bioactive hits were obtained in the secondary screen.

2.2.4.4 Dose response experiments and hit selection

The primary and secondary screens were performed at a single dose of 10 μM. In order to validate hit compounds, ten-point dose response experiments were performed, using both TUBB3 and GAPDH readout systems. The experiment was repeated in triplicates with Promega Renilla Glo Luciferase assay and bioluminescence was measured using

Victor3 fluorescent plate reader. Data acquired from the ten-point dose response experiments were normalised to DMSO control (0.1% DMSO) and the half maximal inhibitory concentrations (IC50) for TUBB3 and GAPDH readout systems were calculated.

The IC50 ratio was then calculated using, IC50 ratio= (IC50 of GAPDH readout)/ (IC50 of

TUBB3 readout). The 28 IC50 ratios displayed distinct disparity, either greater than 4 fold inhibition or smaller than 1 fold inhibition, therefore the IC50 ratio cut-off was set at 4 fold or greater. Two small molecule inhibitors were prioritised for further analysis, eliciting potent suppression of TUBB3 promoter activity with minimal inhibitory effects on GAPDH controls.

2.2.5 Cell viability and proliferation assays

Cell viability and cell proliferation were assessed by Trypan blue assays using a haemocytometer. All experiments were repeated at least three times. Briefly, cells 100 resuspended in media were stained with equal volume of Trypan blue assay and loaded onto a haemocytometer covered in a cover slip. Viable and non-viable cells were counted and calculated using equations: % cell viability= (total cell number – non-viable cell number) / total cell number; % cell proliferation= viable cell number/ viable cell number in non-treated controls.

2.2.6 Quantitative real-time polymerase chain reaction (QRT-PCR)

2.2.6.1 RNA isolation and cDNA synthesis

To examine the relative levels of βIII-tubulin gene expression in H460 cells after drug treatment, RNA extraction was performed. Total RNA was extracted using the Qiagen

RNeasy Plus Mini Kit, according to the manufacturer’s instructions. Briefly, cells were lysed and genomic DNA was removed by filtering through a column. RNA was precipitated with 70% ethanol. The eluted RNA was further purified with the provided buffers and eluted into the collection column. The purified RNA was then resuspended in nuclease-free water and concentration was determined using spectrophotometric absorbance measured at 260 nm, using NanodropTM (Model ND-1000, Wilmington, DE,

USA). A A260/A280 ratio of 1.8-2.1 is desirable for pure RNA. Total RNA was reverse transcribed to cDNA using the High Capacity cDNA Reverse Transcription Kit. cDNA reaction was prepared from 0.5 µg of total RNA in a 20 µL reaction mixture, containing

2.0 μL of 10X reverse transcriptase buffer, 0.8 μL of 25X dNTP mix, 2.0 μL of 10X RT

Random Primers, 1.0 μL of MultiScribeTM Reverse Transcriptase and 4.2 μL of water.

To verify complete removal of genomic DNA, each RNA sample was subjected to a mock reverse transcription by omitting the reverse transcriptase from the reaction mix.

Additionally, total RNA harvested from non-treated samples served as internal control

101 for qRT-PCR later on. All samples were then subjected to a PCR amplification, using a thermocycler, Gene Amp® PCR system 9700 version 3.01, Applied Biosystems). The reaction conditions were 25°C for 10 min, 37°C for 120 min, 85°C for 5 min. cDNA samples were stored at -20°C for future qRT-PCR experiments.

2.2.6.2 Quantitative Real Tim –Polymerase Chain Reaction (QRT-PCR)

To examine the relative levels of TUBB3 expression in H460 cells after drug treatment, quantitative real time- PCR (qRT-PCR) was performed. The reaction was performed with

2.0 μL of cDNA from each sample and specific primers for TUBB, TUBB3, TUBB2a,

TUBB2b and TUBB2c. β2-microglobulin (β2M) was used as the housekeeping control.

β2M has been validated to be an appropriate housekeeping gene for use in NSCLC cell lines within our lab (McCarroll et al., 2010). There was no significant change in the expression of β2M upon drug treatment, making them appropriate housekeeping genes.

The qRT-PCR reaction master mix contained 2.5 μL of forward and reverse primer mix

(Table 2.2), 12.5 μL of SYBR Green RT-PCR Master Mix and 8.0 μL of nuclease-free water. Two μL of cDNA template was pipetted into designated wells of a 96-well microplate, followed by 23.0 μL of the qRT-PCR

102

Table 2.2 Oligonucleotide primers used for PCR amplification

Gene QantiTect Primer Set from Qiagen Accession number Product size

TUBB Hs_TUBB_1_SG (QT00089775) NM_178014 120 bp

TUBB2a Hs_TUBB2A_1_SG (QT00081088) NM_001069 121 bp

TUBB2b Hs_TUBB2B_1_SG (QT01034803) NM_178012 89 bp

TUBB2c Hs_TUBB2C_1_SG (QT00224385) NM_006088 100 bp

TUBB3 Hs_TUBB3_1_SG (QT00083713) NM_006086 78 bp

β2-microglobulin (β2M) Hs_B2M_1_SG (QT00088935) NM_004048 98 bp

GAPDH Hs_GAPDH_2_SG (QT01192646) NM_002046 119 bp

103 reaction master mix. A negative control assay was included in the assay by replacing cDNA template with nuclease-free water. All reaction was carried out as duplicates and a standard curve was generated from two-fold serial dilutions of non-treated cDNA. The qRT-PCR reaction was performed using the ABI PRISM 7500 Sequence Detector

(Applied Biosystems Inc, Foster City, CA, USA). The samples were heated for 2 min at

50°C, 10 min at 95 °C, followed by 40 cycles of two-step PCR consisting of amplification and detection at 95 °C for 15 s and 60 °C for 1 min. A dissociation stage was included to detect any double stranded DNA including primer dimers, contaminating genomic DNA, and PCR product from mis-annealed primers. Dissociation stage included 15 s at 95 °C,

15 s at 60 °C, 15 s at 95 °C. The relative mRNA expression levels were calculated using the absolute quantitation method, where unknown samples were extrapolated from a standard curve generated from two-fold serial dilutions of untreated control.

2.2.7 Western blotting

Whole cellular protein was extracted by trypsinising and resuspending cell pellets at a density of 5 x 106 H460 cells/ mL ice-cold lysis buffer (0.1 M PIPE, pH 8.0; 10% NP-40;

0.85 M KCl; Milli-Q® water) containing PMSF and protease inhibitor and incubated for

20 min on ice. Genomic DNA in the protein lysate was sheered by 10 short pulses, using a MicrosonTM Ultrasonic Homogeniser on ice. Whole cell lysate was transferred into a sterile Eppendorf tube, centrifuged at 13, 300 rpm at 4 °C for 20 min to remove cell debris.

Protein containing supernatant was transferred to a separate sterile Eppendorf tube and quantified using BCA protein assay kit, according to manufacturer’s instructions.

Absorbance values were determined using a Benchmark Microplate reader equipped with

Microplate Manager III (version 1.57) program. Absorbance values were measured at 570

104 nm. Aliquots of 10 µg of protein lysate were prepared and stored at -20°C for further experiments.

Prior to electrophoresis, 10 µg of protein lysate was defrosted and mixed with 2x sample buffer (0.125 M Tris, pH 6.8, 4% SDS, 20% glycerol, 0.1% bromophenol blue, 10% 2- mecaptoethanol). Protein sample was denatured on a heating block at 100°C for 5 min, then briefly vortexed to collect. Sample was kept on ice. A total of 10 µg of protein and

8 μL of Benchmark marker were separated by electrophoresis using a 12% SDS- polyacrylamide gel (30% acrylamide mix, 1.5 M Tris pH 8.8, 10% SDS, 10% ammonium persulfate, TEMED) for 2 h at 100 V in a pre-cast gel running apparatus (Model Mini-

PROTEAN II Cell, BioRad) using running buffer (0.025 M Tris, 0.192 M glycine, 0.1%

SDS).

The separated protein was electrotransferred onto a nitrocellulose membrane using the

Hoefer Mini Transphor unit (Model TE22; Hoefer Scientific Instruments) containing transfer buffer (0.025 M Tris/ glycine, pH 8.2; 20% methanol) at 70mA for overnight at

4°C. Following transfer, the nitrocellulose membrane was stained with 0.5% (w/v)

Ponceau S in 1% acetic acid for 5 min to confirm efficient transfer and equal protein loading. The membrane was then blocked in 5% BSA/PBS for 1 h at room temperature on a shaker. Membranes were then probed with primary antibodies (Table 2.3) in 1 mL of 0.5% BSA/PBS for 1 h at room temperature.

The blot membrane was washed three times for 10 min in TBST and probed with secondary antibodies (Table 2.3) in 1 mL of 0.5% BSA/PBS in a sealed plastic pouch for

1 h at room temperature. The blot was further washed three times for 10 min in TBST.

Overlay membrane with ECL Plus solution for 5 min at room temperature. The housekeeping protein, GAPDH was used as a loading control, and blots were scanned

105

Table 2.3 Antibodies used for western blotting

Antibody (Ab) Animal Primary Ab Primary Ab Secondary Ab Secondary Ab

source dilution diluent dilution diluent

Anti-SP1 rabbit 1:2,500 0.5%BSA/ PBS 1:2500 0.5%BAS/ PBS

Anti-GAPDH (clone 6C5) mouse 1:20,000 0.5%BSA/ PBS 1:20,000 0.5%BAS/ PBS

Anti-total β-tubulin (clone TUB 2.1) mouse 1:1,000 0.05% TBST 1:5,000 0.05% TBST

Anti-class II β-tubulin (clone 7B9) mouse 1:1,000 0.5% SM/TBST 1:7,500 0.5% SM/TBST

Anti-class III β-tubulin (clone TUJ1) mouse 1:1,000 0.05% TBST 1:7,500 0.05% TBST

Anti-class IVb β-tubulin (clone mouse 1:500 0.5% SM/TBST 1:2,000 0.5% SM/TBST

ONS.1A6)

Anti-GAPDH (clone 6C5) mouse 1:20,000 0.05% TBST 1:20,000 0.05% TBST

FBS- Fetal bovine serum; PBS- Phosphate buffer saline; BSA- Bovine serum albumin; SM- Skim milk; TBST- Tris-buffered saline- 0.01%

Tween-20.

106 using the typhoon scanner (Model 9410, GE Healthcare) and quantified using Image

Quant software version 5.2 (Molecular Dynamics Inc., Sunnyvale, CA). The protein expression was determined by dividing the densitometric value of the target protein by that of the control protein (GAPDH). Whenever applicable, the normalised ratios were expressed against the mock-transfected cells in which the protein level was arbitrarily set as 1.0. The experiments were repeated in triplicate with protein isolated from three independent extractions.

2.2.8 Cell cycle analysis by flow cytometry

For analysis of DNA content by propidium iodide staining, H460 cells were seeded in 6- well plates containing 4x104 cells per well for 24 h. Twenty-four hours post-plating, H460 cells were treated with hit compounds for 24, 48 and 72 h. On the day of analysis, both adherent and floating cells were harvested, washed with warm PBS and fixed with 1 mL of ice-cold 70% methanol and stored in -30°C for 20 min. Cells were suspended in 10 mL of PBS, then centrifuged at 1500 RPM for 5 min at room temperature to discard supernatant and isolate cell pellets. Cell pellets were resuspended and stained with a solution containing 0.4% Triton X-100, 50 μg/mL propidium iodide, and 2 μg/mL

DNAse-free RNase for 15 min at room temperature in dark. After, 200 uL of PBS was added to each sample and pipetted to mix. DNA content was measured by a FACSCalibur flow cytometer. The flow rate was <200 nuclei and 20,000 cells for each sample were analysed. The Cell Quest program was used to quantitate the distribution of cells in sub-

G1 (dead cells), G0/G1, S and G2/M.

107

2.2.9 Microscopy

2.2.9.1 Phase contrast

Cells were grown in 6 well plates and then imaged using a 10x objective on the Axiovert

S100 inverted phase-contrast microscope (Zeiss) equipped with a SPOT RT Slider CCD scientific digital camera system driven by SPOT Basic software (Diagnostic instruments,

Inc. Sterling Heights, MI).

2.2.9.2 Immunofluorescence

Cells were seeded at a density of 2,000 H460 cells/ well/ 500 μL RPMI into sterile 4-well

Lab Tek II™ chamber slides (Thermo Fisher Scientific Inc.). 48 h post-seeding, cells were treated with 100 μM of small molecule WECC001371 for 48 h at 37 °C and 5% CO2. On the day of staining, media was removed and cells were briefly rinsed with 500 μL of PBS.

Cells were then fixed with 100% ice-cold methanol for 12 min at -30°C. Methanol-fixed cells were then blocked with 10% FCS/PBS at room temperature for 10 min. Cells were then incubated with primary monoclonal antibodies (diluted in 5% FCS/PBS) against

βIII-tubulin (1:1,000) and βII-tubulin (1:1,000) for 30 min in a humidified chamber.

Humidified chamber contains Rediwipes soaked with water, which were lined in a closed container. Slides were washed with three times 10 min with 0.1% Tween-20/PBS at room temperature on an Orbit P4 shaker. Primary antibodies of βIII-tubulin and βII-tubulin were detected using Alexa Fluor 568 (orange/red) goat anti-mouse IgG (Invitrogen) diluted 1:1,000 in 5% FCS/PBS and incubated for 30 min at room temperature in dark.

Slides were washed three times 10 min in 0.1% Tween-20/PBS at room temperature with constant shaking. For dual staining, cells were then probed with primary monoclonal antibody (diluted in 5% FCS/PBS) against α-tubulin (1:400) for 30 min in dark humidified chamber. Cells were then washed three times 10 min in 0.1% Tween-20/PBS

108 at room temperature on an Orbit P4 chamber. Primary antibody of α-tubulin was detected using Alexa Fluor 488 (green) goat anti-mouse IgG (Invitrogen) diluted 1:500 in 5%

FCS/PBS for 40 min in the dark. Slides were washed further three times 10 min in 0.1%

Tween-20/PBS at room temperature and rinsed briefly in water to remove any salt residue. A Kimwipe was used to carefully blot excess water from the slide, avoiding drying. Slides were mounted using VECTASHIELD® mounting medium with 4’,6- diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA). Cytoskeletal structures were then visualised and imaged using an Olympus FluoView™ FV1000 confocal microscope with a 63x 1.35 NA oil objective.

2.2.10 Drug-treated clonogenic assays

In vitro chemosensitivity assay of the transfected cells to various drugs was determined by the clonogenic assay. This is a more accurate assay when assessing replicative ability compared to the conventional short-term MTT [3-4,5-dimethylthiazol-2-yl]-2,5- diphenyltetrazolium bromide] assay. 4 x 104 H460 cells were plated into each well of a

6-well plate. 24 h post-plating cells were treated with small chemical molecule

WECC0017371 for 72 h. 72 h post-treatment, 150 H460 cells were replated into each well of a 6-well plate in drug-free RPMI. Cells were allowed to attach for 6 h prior to combination drug treatment. This ensured that equal number of cells was plated for all drug treatments and the controls before drug exposure. Serial dilutions of various drugs were prepared fresh and added to cells. Following an 8-day incubation at 37°C, the drug- containing medium was removed and surviving colonies were simultaneously fixed and stained with 0.5% crystal violet in methanol for 20 min. Crystal violet was aspirated and cells were washed with water to remove excess stain and air-dried overnight. Individual

109 stained colonies in each well were manually counted. Colonies of ≥ 50 cells were counted in the surviving fraction. To evaluate the effect of WECC0017371 in combination with chemotherapies on H460 clonogenic potential, the surviving fraction was calculated as follows: colony number/ (number of cells seeded x plating efficiency), where plating efficiency is equivalent to the colony number divided by the number of cells seeded in the drug free-medium.

2.2.11 Microarray

H460 cells were seeded in 6-well plates at a density of 4 x 104 cells/well/2 mL of RPMI

(10% FCS). Twenty-four hours post-plating, H460 cells were treated with 100 μM of

WECC0017371. Non-treated H460 cells and WECC0017371 treated H460 cells were harvested at 48 and 72 h post-treatment. Media was removed and cells were washed with

2 x 1 mL ice-cold PBS to remove any residual RPMI. Cells were lysed in Rapid Lysis

Buffer® with 10% β-mercaptoethanol and total RNA was extracted using the RNeasy Plus

Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer’s instructions. The purified RNA was resuspended in nuclease-free water and RNA concentration, 260:280 and 260:230 ratios were determined using NanodropTM (Model ND-1000, Wilmington,

DE, USA). For microarray analysis, 260:280 and 260:230 ratios must be great than 1.8.

Routine RT-PCR was performed on all samples to assess the effect of WECC0017371 on

TUBB3 mRNA level prior to microarray analysis. In addition, Total RNA samples were run on the Agilent Bioanalyzer RNA Nano 6000 chip to assess the integrity of the total

RNA. RNA with RNA integrity numbers >8 (out of 10) is required for DNA microarray analysis. Total RNA of WECC0017371-treated and untreated H460 cells was reverse transcribed to cDNA and analysed using Illumina HT-12 genome arrays. Labeling,

110 hybridization, and scanning were performed by the Ramaciotti Centre for Gene Function

Analysis (University of New South Wales, Sydney, NSW, Australia), with each sample hybridized to an individual array.

2.2.12 Bioinformatics analysis of microarray data

Liasing with a Children’s Cancer Istitute bioinformatician, Chelsea Mayoh, microarray results were analysed. The Bioconductor implementation of Limma

(http://www.bioconductor.org) (Ritchie et al., 2015; Smyth, 2004) was used to identify the differentially expressed (DE) genes between treatment with WECC0017371and without at 48 and 72 hours of non-small cell lung cancer samples in triplicate. Microarray analysis was performed on Illumina Bead Chip Array so these data were processed using neqc (Shi et al., 2010) method for background correction using negative probes then quantile normalization and log2 transformations. This method removes control probes only leaving regular probes used to measure differential expression of genes as profiled from Illumina HumanHT 12 V4.0.R2.15002873.B. To identify DE genes and estimate fold change (FC) between no treatment and treatment at 48 hours and 72 hours, a linear model was fit to each probe in the array, and empirical Bayes was then applied to the standard errors, a moderated t-statistic was applied with Benjamini Hochberg used for multiple test correction. These data were further filtered on a FC of |1.3| (as determined by TUBB3 FC between treated and non-treated at 48 hours). Analysis was done using R

(https://www.r-project.org/).

DAVID (Huang da et al., 2009a; Huang da et al., 2009b) the online bioinformatic tool to aid researchers with functional annotations of microarrays was used for KEGG pathway analysis and GO annotation (Biological Process, Cellular Component and Molecular

111

Function). A Gene Set Enrichment Analysisc (GSEA) (http://www.broad.mit.edu/gsea/) was then performed using the gene sets C2 (curated gene sets) , C4 (computational gene sets), C5 (GO gene sets), and C6 (oncogenic signatures). Cytoscape (Shannon et al., 2003) an open-source bioinformatics software was used to visualize the molecular interactions and networks of the results outputted from the GSEA.

2.2.13 Statistical analysis

Statistical analysis were performed using the GraphPad Prism program version 5.0 (San

Diego, USA). For in vitro experiments, results are expressed as means of at least three independent experiments ± standard error of the mean (SEM). Unpaired, two-tailed

Student’s t tests were used to determine the statistical differences between various experimental and control groups, with <0.05 considered statistically significant.

112

Chapter 3

Development of A Cell-Based Screen to Identify Small

Molecule Inhibitors of TUBB3 Promoter Activity

3.1 Introduction

Lung cancer is the leading cause of cancer death worldwide. Treatment for the most common form, non-small cell lung cancer (NSCLC) includes a combination of surgery, chemotherapy, and radiotherapy. Tubulin binding agents (TBA) such as paclitaxel, epothilones, eribulin and vinorelbine are often used in combination with platinum compounds. Despite the clinical success of TBAs and advances in chemotherapies, the prognosis for the majority of patients diagnosed with advanced stage disease remains dismal. The emergence of drug-resistant tumours is the primary cause of treatment failure.

As discussed in the Introduction (section 1.3), aberrant βIII-tubulin (encoded by TUBB3 gene) expression is now recognised as a key determinant of drug resistance and tumour aggressiveness in a number of epithelial cancers, including NSCLC (reviewed in Karki et al., 2013). There is strong clinical evidence in lung, ovarian and breast cancer that patients with aberrant βIII-tubulin expression exhibit poorly differentiated tumour tissue, high grade malignancy, shorter disease progression and a poor overall survival rate

(Reiman et al., 2012; Vilmar et al., 2011). These clinical findings have been supported with laboratory data (Gan et al., 2007; Kavallaris et al., 1999; McCarroll et al., 2010;

McCarroll et al., 2015b; Ploussard et al., 2010), where βIII-tubulin has been validated as a bona fide target for chemosensitivity and cell survival in cancers that aberrantly express this protein (section 1.3).

113

Despite the relevance of βIII-tubulin in drug resistance, preferential targeting of this tubulin isotype has been extremely challenging. There is currently no chemical molecule that can specifically modulate βIII-tubulin expression. Investigation of mechanisms of

βIII-tubulin regulation could lead to identification of key cellular factors responsible for its aberrant expression in cancer, and therefore exposing new therapeutic targets. Here, we have taken a novel approach to investigate mechanisms of TUBB3/βIII-tubulin regulation by exploiting the fact that various tubulin isotypes are encoded by different genes, each with its own unique promoter. High throughput screening and chemical proteomics were used in this project to elucidate cellular regulators of βIII-tubulin. This chapter describes the development and conduct of a cell-based screen to identify novel small chemical molecules and bioactives that can specifically and potently modulate

TUBB3 promoter activity, in a NSCLC cell line, H460, which aberrantly expresses endogenous βIII-tubulin.

3.2 Results

3.2.1 Generation and clonal selection of H460/TUBB3p-luc and H460/GAPDHp- luc cells

The following results section 3.2.1 will describe the design and development of the

TUBB3 promoter cell-based screen, while section 3.2.2 and 3.2.3 will describe the optimisation and validation of assays for the high throughput screening. Section 3.2.4 will outline the results of the main screen which progressed in three stages, the primary screen

(section 3.2.4.1), the secondary screen on a refined pool of candidates obtained from the primary screen (section 3.2.4.1) and ten-point dose response experiments and cell viability studies for final hit selection (section 3.2.4.2, 3.2.4.3 and 3.2.4.4).

114

In order to identify small chemical molecules and bioactives that can specifically and potently modulate TUBB3 promoter activity, cells were developed for high-throughput screening. The strategy is outlined in Figure 3.1A. Briefly, a vector construct containing the minimal TUBB3 promoter region (Figure 3.1B) (Dennis et al., 2002) cloned upstream of a Renilla luciferase reporter gene was transfected into H460 NSCLC cells and stable clones were generated. During stable clone selection, cells were selected under puromycin pressure at all times. One hundred single-cell clones expressing the TUBB3 promoter-Renilla luciferase expression construct were screened and isolated based on their luciferase activity and growth rate. TUBB3 promoter activity is measured via bioluminescence as outlined in Figure 3.1C. A high Renilla luciferase-expressing clone possessing a similar growth rate to the wild-type H460 parental cells was selected. This clone of cells is henceforth referred to as H460/TUBB3p-luc cells (p, promoter; luc, luciferase) and was designated for use in the small molecule screen of TUBB3 modulators.

To minimise false positive small molecule modulators, a control cell model was also generated, using the promoter region of the housekeeping GAPDH gene (refer to materials and methods section 2.2.3.2). One hundred single-cell H460 clones stably expressing GAPDH promoter-Renilla luciferase expression constructs were generated as described for TUBB3. Again, a high Renilla luciferase-expressing clone possessing a similar growth rate to wild-type H460 cells was selected and referred to as

H460/GAPDHp-luc cells (p, promoter; luc, luciferase). This control cell model was included in the secondary drug screen to identify and remove non-specific inhibitors targeting luciferase and GAPDH gene expression.

115

Figure 3.1 Strategy to develop cell-based readout assays for high throughput screening.

(A) To develop a small molecule functional readout assay, a TUBB3 promoter-luciferase expression construct was stably transfected into H460 NSCLC cells and stable clones were selected under puromycin pressure. Bioluminescence assays were used to validate

TUBB3 promoter-luciferase reporter activity in stable clones. A high luciferase- expressing clone possessing a similar growth rate to the wild-type H460 cells was picked and used in the small molecule screen. This clone is referred to as H460/TUBB3p-luc cells. (B) The TUBB3 promoter region (blue arrowed line) cloned into pLightSwitch_Prom_puro_RenSP vectors contains the first 838 bp of human TUBB3 promoter from the transcriptional start site, a translational start site and 117 bp of 5’ untranslated region. This sequence was selected as it includes the minimal promoter region (black arrowed line). (C) In H460/TUBB3p-luc cells, activation of TUBB3 promoter leads to the transcription of downstream luciferase reporter gene, producing luminescence after substrate addition. H460/TUBB3p-luc cells are used in the high throughput screen to identify small molecules that can modulate TUBB3 promoter activity. Schematic diagrams indicate hypothetical outcomes of the small molecule screen. The change in bioluminescence intensity between treated and non-treated

H460/TUBB3p-luc cells is representative of changes in TUBB3 promoter activity.

116

117

Further, to assess the effect of stable expression of TUBB3 promoter- and GAPDH promoter- Renilla luciferase expression constructs, endogenous levels of β-tubulin isotypes were examined using western blotting. Figure 3.2A-E demonstrated that the endogenous level of βIII-tubulin, βII-tubulin, βIV-tubulin, βI-tubulin and total β-tubulin remained unchanged in both H460/TUBB3p-luc and H460/GAPDHp-luc cells compared to wild-type H460.

3.2.2 Preliminary screen- to validate H460/TUBB3p-luc cells for high throughput screening

Prior to carrying out a high throughput small molecule screen thorough optimisation, stringent assessment and statistical analysis of H460/TUBB3p-luc cells was required.

H460/TUBB3p-luc cells were treated with 0.1% DMSO (solvent control) and 0.1µM

Actinomycin D (general transcription inhibitor, positive control) for 24 h before measuring change in bioluminescence. Experiments were performed in triplicate and repeated using both manual (n=3) and automated equipment (n=3). Assay variability within and between plates was calculated using the formula: Z’= 1 −

( ) . Briefly, Z’ is a statistical analysis used | | to validate assay reliability for high throughput screening (HTS), including several different factors, including reproducibility, robustness and sensitivity (Zhang et al.,

1999). The Z’ factor has become the industry standard means of measuring assay quality on a plate bases. The Z’ factor has a range of 0 to 1; an assay with Z’=1 indicates an ideal assay for HTS; 0.5 ≤ Z’< 1 indicates an excellent assay; while 0 < Z’< 0.5 is considered marginal and may be suitable for HTS but further optimisation and additional tests are required. Assay with Z’<0 is unsuitable to be used for HTS. In

118

Figure 3.2 The effect of stable expression of TUBB3 promoter- and GAPDH promoter- luciferase expression constructs on endogenous levels of β-tubulin isotypes.

To investigate any potential effect of stable expression of TUBB3 promoter- and GAPDH promoter-luciferase expression constructs on other β-tubulin isotypes, western blotting was performed to measure the endogenous level of β-tubulin isotypes in wild-type H460 cells (WT), H460/TUBB3p-luc cells (TUBB3-luc) and H460/GAPDHp-luc cells

(GAPDH-luc). Representative western blots show the endogenous level of (A) βIII- tubulin, (B) total β-tubulin, (C) βI-tubulin, (D) βII-tubulin and (E) βIV-tubulin in

H460/TUBB3p-luc and H460/GAPDHp-luc cells remained unchanged compared to wild- type H460 cells. GAPDH was used as a loading control.

119

A TUBB3- GAPDH- WT luc luc βIII-tubulin GAPDH

B TUBB3- GAPDH- WT luc luc Total β-tubulin GAPDH

TUBB3- GAPDH- C WT luc luc βI-tubulin GAPDH

E TUBB3- GAPDH- WT luc luc βII-tubulin

GAPDH

D TUBB3- GAPDH- WT luc luc βIV-tubulin

GAPDH

120 summary, using H460/TUBB3p-luc cells, well-to-well, plate-to-plate, day-to-day and batch-to-batch variations were negligible. The Z’ factor equalled 0.72, indicating that the cell-based assay is valid for high throughput screening of TUBB3 promoter activity modulating agent.

3.2.3 Pilot screen- final validation of H460/TUBB3p-luc cells for high throughput screening

Prior to the large 30K compound library screen, a small pilot screen of 1200 FDA- approved bioactives was performed as a pilot study to provide final validation of

H460/TUBB3p-luc cells under screening conditions. The purpose of this screen is to gather information on relevant technical parameters that could be used to guide the larger and more expensive primary screen. The Prestwick library was selected for the pilot study as it consists 100% FDA-approved bioactives, representing the greatest possible degree of drug-likeliness. The active compounds were selected for their high chemical and pharmacological diversity as well as for their known bioavailability and safety in humans.

The Prestwick library was designed to reduce the risk of “low quality” hits, reduce the cost of the initial screening, and accelerate lead discovery. In the small pilot screen, the

1200 FDA-approved drugs were screened under the experimental conditions similar to those set for the primary screen. The average values for the positive and negative controls were consistent across all assay plates and the average Z’ value was 0.74, indicating acceptable data quality. Seventy six compounds elicited a 75% or greater reduction in luminescence signal and may be suitable for follow up studies (Figure 3.3). Additionally, no compound achieved a 100% or greater enhancement in luminescence activity.

121

Accordingly, these results provided confirmation that H460/TUBB3p-luc cells were suitably robust for use in HTS of the 30K compound library.

Out of the 76 hits from the Prestwick library (Figure 3.4), many are known cytotoxics and thus need to be distinguished from a true hit. To address this, H460/GAPDHp-luc cells were included in the secondary screen to minimise non-specific TUBB3 promoter repressing agents targeting luciferase and/or GAPDH gene expression. Hypothetically, a genuine “hit” compound is one that shows maximum repression or enhancement on

TUBB3 promoter activity and minimal effect on GAPDH promoter activity (Figure 3.5).

Prior to the secondary screen, H460/GAPDHp-luc cells were also validated using the

Prestwick library and the Z' value equalled 0.65, indicating the cell-based assay was suitably robust for use in HTS.

122

Figure 3.3 Dot plot of pilot screen of FDA-approved bioactive library. An initial pilot screen using the FDA-approved bioactive library was conducted using

H460/TUBB3p-luc cells. In the dot plot, each blue data point represents a single bioactive compound. Data are plotted against Z score, which is a log2 scaling of the relative luminescence. The relative luminescence value for each compound data point is normalised to the average of non-treated control wells (DMSO control). The cut-off (red dotted line) for a hit that can repress TUBB3 promoter activity was set at Z score -2, which corresponds to a 75% reduction in bioluminescence relative to DMSO control. The cut- off (green dotted line) for a hit that can enhance TUBB3 promoter activity was set at Z score -1, which corresponds to a 100% increase in bioluminescence relative to DMSO control. Hence, blue data points below the red dotted line and above the green dotted line may be suitable for follow up studies. Data collected for the pilot screen was done without replicates

123

124

Figure 3.4 A pie graph shows distribution of the type of bioactive hits that repress bioluminescence readout in H460/TUBB3p-luc cells from Prestwick library.

The pie graph illustrates distribution of hit compound target classes using H460/TUBB3p- luc cells. Out of the 76 hits from Prestwick library pilot screen, many are known cytotoxics, such as apoptotic agents, microtubule stabilisers and destabilisers, as well as antibiotics, antifungals and anti-parasitic drugs.

125

126

Figure 3.5 Hypothetical results illustrating “hit” selection. During the drug screen, small molecule inhibitors that potently repress TUBB3 promoter activity, but minimally affect GAPDH promoter activity will be selected. Graphs indicate hypothetical outcomes to illustrate what would be considered a true hit versus a false positive hit. (A) Using H460/TUBB3p-luc cells and H460/GAPDHp-luc cells, TUBB3 and GAPDH promoter activities are measured via bioluminescence. (B) While a false- positive hit affects both GAPDH and TUBB3 promoter activities, (C) a genuine “hit” shows inhibition or enhancement on TUBB3 promoter activity with minimal effect on

GAPDH promoter activity.

127

C enhancer

inhibitor

128

3.2.4 High throughput screening of TUBB3 promoter repressing agents

3.2.4.1 Primary screen

Having optimised and validated H460/TUBB3p-luc cells and H460/GAPDHp-luc cells for high throughput screening, a primary screen on a diverse library of 30,224 chemical small molecules and 3680 bioactives was subsequently performed. Using liquid handling robotics from the ACRF Drug Discovery Centre, cells were screened for luciferase expression as outlined in Figure 3.1C to identify hits that modify TUBB3 promoter activity. The cut-off for “hits” that repress TUBB3 promoter activity was defined as compounds that are able to elicit a 75% or greater reduction in bioluminescence signal relative to DMSO control (Figure 3.6). Due to cost consideration, H460/GAPDHp-luc cells were not included in the primary screen but were included in the secondary screen, which focuses on a narrower group of compounds. The result of the secondary screen will be discussed in the following section (section 3.2.4.1). From the primary screen, 406 and

308 TUBB3 promoter activity repressing hits were successfully obtained in the diverse chemical libraries and bioactive libraries (Prestwick, LOPAC and TOCRIS library), respectively. The cut-off of “hits” that enhance TUBB3 promoter activity was defined as compounds that are able to induce a 100% or greater increase in bioluminescence signal relative to DMSO control (Figure 3.6). Briefly, 226 TUBB3 promoter activity enhancing hits were obtained in the primary screen of small molecule chemicals. The Z’ measure for the primary screen was 0.8, indicating excellent screen quality. Depending on the nature of the assay and the stringency of the criteria used, the predicted hit rate for a random chemical library using a cell-based assay typically ranges from 0.05 to 2%

(reviewed in Michelini et al., 2010). The expected hit rate for a bioactive library is generally higher, in the order of 1-10%, or even higher. In summary, the hit rates achieved in this primary 129

Figure 3.6 Dot plot of primary screen of 30,224 small molecule compounds

An initial primary screen of 30,224 small molecule compounds was conducted using

H460/TUBB3p-luc cells to identify modifiers of TUBB3 promoter activity. In the dot plot, each red data point represents one small molecule compound. The relative luminescence value for each compound data point was normalised to the average of the DMSO control wells. Data are plotted against Z score, which is a log2 scaling of the relative luminescence. The cut-off (black dotted line) for a hit that can repress TUBB3 promoter activity was set at Z score -2, which corresponds to a 75% decrease in bioluminescence relative to DMSO control. The cut-off (blue dotted line) for a hit that can enhance TUBB3 promoter activity was set at Z score -1, which corresponds to a 100% increase in bioluminescence relative to DMSO control. Red dots that landed below the black dotted line and above the blue dotted line may be suitable for follow up studies. Data plotted as mean ± SEM of 3 independent experiments.

130

131 screen using random chemical libraries and bioactive libraries were 2.09% and 8.36%, respectively, and hence are within the expected hit range.

3.2.4.1 Secondary screen

In the secondary screen, which focuses on hits identified in the primary screen,

H460/GAPDHp-luc cells were included in addition to H460/TUBB3p-luc cells to minimise non-specific hits. For the secondary screen, hits were selected based on the relative difference between the fold change of GAPDH- and TUBB3 promoter activities.

Based on the experience of the ACRF Drug Discovery Centre at the Children’s Cancer

Institute, arbitrary cut-offs of >0.28 and >1.2 were selected for compounds that repressed and enhanced TUBB3 promoter activity, respectively. From the secondary screen, 28 and

16 small molecule hits repressed and enhanced TUBB3 promoter activity, respectively, with minimal effect on GAPDH promoter activity. Additionally, 3 bioactive hits that are able to repress TUBB3 promoter activity with minimal effect on GAPDH promoter activity were also obtained. No bioactives were observed to demonstrate an enhancing effect on TUBB3 promoter activity. Hits obtained from small molecule and bioactive libraries were selected for follow up ten-point dose response studies and will be discussed separately.

3.2.4.2 Ten-point dose response and cell viability experiments for small molecule

hits

Initially, we will focus on hits obtained from small molecule compound libraries. The bioactive hits will be discussed in the later section, 3.2.4.3. Twenty-eight small molecule hits were tested and selected for their ability to potently repress TUBB3 promoter activity, 132 while having a minimal effect on GAPDH promoter activity. To measure their effectiveness in inhibiting promoter activity, the half maximal inhibitory concentration

(IC50) was determined for each small molecule hit. Both primary and secondary screens were performed at a single dose of 10 μM. To quantitatively measure the difference of a hit compound’s effect on TUBB3 and GAPDH promoter activity, ten-point dose response experiments were performed to calculate their IC50 ratio, using H460/TUBB3p-luc cells and H460/GAPDHp-luc cells. IC50 ratio is obtained by dividing the IC50 in

H460/GAPDHp-luc cells by the IC50 in H460/TUBB3p-luc cells. A hit compound with an

IC50 ratio greater than 1 indicates that its suppressive effect is more pronounced on

TUBB3 promoter activity than GAPDH promoter activity, and vice versa. The 28 IC50 ratios obtained for small molecule hits displayed distinct disparity, either greater than 4 fold inhibitions or smaller than 1 fold (Appendix II), therefore the IC50 ratio cut-off was defined as 4 fold or greater. The two top hits, WECC0018639 (Figure 3.7A; IC50 ratio=

4.13) and WECC0017371 (Figure 3.7B; IC50 ratio= 4.23) resulted in 80 and 70% reduction in TUBB3 promoter activity at a concentration as low as 10 μM and were therefore pursued for further analysis. The suppressive effects of both hits were more prominent on TUBB3 promoter activity than GAPDH promoter activity, which suggests that the repression in TUBB3 promoter activity was not a result of general transcriptional repression.

Next, Alamar blue assays were performed to distinguish the repressing effect of

WECC0017371 on TUBB3 promoter activity from a reduced cell viability. At the concentration of WECC0017371 used in primary and secondary screens (10 μM), the cell viability of H460/TUBB3p-luc cells and H460/GAPDHp-luc cells were 92% and 97%, respectively, relative to non-treated control (Figure 3.8A). This ruled out the possibility that repressed TUBB3- and GAPDH promoter activity was merely a consequence of 133

Figure 3.7 Chemical structures and ten-point dose response experiments of top two small molecule hits, WECC0018639 and WECC0017371.

Initially, primary and secondary screens of small molecules that are able to modify

TUBB3 promoter activity were performed at a single dose of 10 μM. In order to quantitatively measure the hit compound’s effect on TUBB3 and GAPDH promoter activity, ten-point dose response experiments were performed to calculate their IC50 ratio, using H460/TUBB3p-luc cells and H460/GAPDHp-luc cells. (A) Small molecule hit,

WECC0018639 (IC50 ratio= 4.13) elicited potent repression on TUBB3 promoter activity

(red line) and had minimal repressive effect on GAPDH promoter activity (blue line). (B)

Similar effects were observed with WECC0017371 (IC50 ratio= 4.23). Experiments were performed in triplicate and data plotted as mean ± SEM of 3 independent experiments.

134

A H460/GAPDHp-luc cells H460/TUBB3p-luc cells

IC50 ratio= 4.13

WECC0018639 (μM) B

H460/GAPDHp-luc cells H460/TUBB3p-luc cells

IC50 ratio= 4.23

WECC0017371 (μM)

135

Figure 3.8 Effect of small molecule hits, WECC0017371 and WECC0018639 on cell viability in H460/TUBB3p-luc and H460/GAPDHp-luc cells

Primary and secondary drug screens were performed at a single dose of 10 μM. Alamar

Blue assays were performed in H460/TUBB3p-luc and H460/GAPDHp-luc cells to distinguish this repression in TUBB3 and GAPDH promoter activity from a reduction in cell viability. Cells were treated with 10 μM of WECC0017371 and WECC0018639 for

24 h before cell viability (%) was measured in non-treated control (black bar) and treated cells (grey bars). (A) Under 10 μM of WECC0017371 treatment, the cell viability of

H460/TUBB3p-luc cells and H460/GAPDHp-luc cells was 92% and 97%, respectively.

(B) At 10 μM, WECC0018639 reduced cell viability by 20% and 40% in H460/TUBB3p- luc cells and H460/GAPDHp-luc cells, respectively. Cell viability (%) was normalised to non-treated control and plotted as mean ± SEM of 3 independent experiments.

136

137 reduced cell viability, and hence WECC0017371 was prioritised for further analysis. As was shown in WECC0017371, WECC0018639 was also subjected to Alamar Blue studies. At the concentration of WECC0018639 used in primary and secondary screens

(10 μM), the cell growth of H460/TUBB3p-luc cells and H460/GAPDHp-luc cells was

80% and 60%, respectively (Figure 3.8B). This suggests that the marked repression in

TUBB3 promoter activity followed by WECC0018639 treatment was unlikely a consequence of a reduction in cell viability. WECC0018639 was therefore also selected and prioritised for further analysis.

In addition to small molecule hits that repress TUBB3 promoter activity, 16 small molecule hits (Appendix III) were also selected for their ability to increase TUBB3 promoter-luciferase readout, while having a minimal effect on GAPDH promoter activity.

Ten-point dose response experiments were performed to quantitate the effect of hit compounds on TUBB3 and GAPDH promoter activity, using H460/TUBB3p-luc cells and

H460/GAPDHp-luc cells. A hit compound with an IC50 ratio greater than 1 indicates that the enhancing effect is more pronounced on TUBB3 promoter activity than GAPDH promoter activity. The IC50 ratio cut-off was defined as 1 fold or greater. Five hit compounds, including WECC0001587, WECC0022626, WECC0023807,

WECC0024534 and WECC0024767 were obtained. Due to limited time this thesis focused on hit compounds that repressed TUBB3 promoter expression.

3.2.4.3 Bioactive hits

Having narrowed down small molecule hits from the secondary screen, the 3 bioactive hits were investigated next. These compounds, obtained from bioactive libraries are

Reactivation of p53 and Induction of Tumour cell Apoptosis (RITA) (Figure 3.9A),

138

Figure 3.9 Chemical structure of bioactive “hits” RITA, Nemonapride and

GW843682X.

Three bioactives were obtained from 3680 bioactives from primary and secondary screens of TUBB3 promoter activity modifiers. Their chemical structures and known activities in human are described below. (A) RITA or 5,5'-(2,5-Furandiyl)bis-2-thiophenem-ethanol, is a potent inhibitor of p53-MDM2 interaction. (B) Nemonapride or cis-5-Chloro-2- methoxy-4-(methylamino)-N-[2-methyl-1-(phenylmethyl)-3 pyrrolidinyl]benzamide is a potent D2-like antagonists and a 5-HT1A receptor agonist. (C) GW843682X or 5-(5,6-

Dimethoxy-1H-benzimidazol-1-yl)-3-[[2 (trifluoromethyl)phenyl]methoxy]-2- thiophenecarboxamide is a selective inhibitor of polo-like kinase-1 and polo-like kinase-

3 that shows potent inhibition of proliferation in a wide variety of tumour cell lines.

139

140

Nemonapride (Figure 3.9B) and GW843682X (Figure 3.9C). Briefly, RITA has demonstrated apoptotic activity in preliminary studies of various cancer cell lines

(Burmakin et al., 2013; Issaeva et al., 2004; Zhao et al., 2010). RITA binds to wild-type p53, disrupting the interaction between the p53 tumour suppressor and its negative regulator MDM2. RITA subsequently activates p53 and induces apoptosis (Issaeva et al.,

2004; Zhao et al., 2010). The second bioactive hit, Nemonapride is a highly potent dopamine D2-like receptor antagonist and a 5-HT1A receptor agonist. It is used as an antipsychotic for the treatment of schizophrenia in Japan (Bishara and Taylor, 2008).

Moreover, GW843682X is a selective inhibitor of polo-like kinase-1 and -3. It has been shown to potently inhibit cell proliferation in a wide variety of tumour cell lines (Lansing et al., 2007). The 3 bioactive hits were tested and chosen for their ability to repress TUBB3 promoter activity, while having a minimal effect on GAPDH promoter activity.

3.2.4.4 Ten-point dose response and cell viability experiments for bioactive hits

As was shown for the small molecule library hits, similar assays were used to assess the

3 hits from bioactive libraries. To determine the IC50 ratio, the 3 bioactive hits were subjected to ten-point dose response experiments. Figure 3.10A demonstrates that RITA elicits potent repression of TUBB3 promoter activity (IC50 ratio= 7.76) while sparing

GAPDH promoter activity in H460 cells. Although unconfirmed, the repression in

TUBB3 promoter activity is unlikely a result of general transcriptional repression, given that RITA affected GAPDH promoter activity to a much lesser extent than TUBB3 promoter activity. In order to distinguish this repression in promoter activity from a reduced cell viability, Alamar Blue assays were performed in H460/TUBB3p-luc cells and H460/GAPDHp-luc cells. The dose range used in Alamar Blue assays corresponded

141

Figure 3.10 Ten-point dose response experiments of the top 3 bioactive hits, RITA,

Nemonapride and GW843682X, using H460/TUBB3p-luc cells and H460/GAPDHp- luc cells

Primary and secondary screens of bioactives that modulate TUBB3 promoter activity were performed at a single dose at 10 μM. As described for Figure 3.7 (page 135), ten- point dose response experiments were performed to quantitate the effect of hit compounds on TUBB3 and GAPDH promoter activity. The IC50 ratio was then calculated using, IC50 ratio= (IC50 of GAPDH readout)/ (IC50 of TUBB3 readout). Three bioactive hits; RITA,

Nemonaprode and GW843682X were tested and selected from ten-point dose response studies of 308 bioactive hits. (A) RITA caused prominent inhibition in TUBB3 promoter activity (red line), while sparing GAPDH promoter activity (blue line) (IC50 ratio= 7.76).

(B) Nemonapride showed a dose-dependent repressive effect on both TUBB3- (red line) and GAPDH- (blue line) promoter activities, indicating general transcriptional repression, and thus was eliminated from this study. (C) GW843682X repressed TUBB3- and to a lesser extent, GAPDH promoter activity (IC50 ratio= 0.796). Experiments were performed in triplicate and data plotted as mean ± SEM of 3 independent experiments.

142

H460/TUBB3p-luc cells H460/GAPDHp-luc cells IC50 ratio= 7.76

H460/TUBB3p-luc cells H460/GAPDHp-luc cells

H460/TUBB3p-luc cells IC50 ratio= 0.796 H460/GAPDHp-luc cells

143 to that in ten-point response studies. It was demonstrated that at drug concentrations that elicit potent TUBB3 promoter activity inhibition (10 μM or log M= -5), RITA reduced the cell viability of H460/TUBB3p-luc and H460/GAPDHp-luc cells by 30% and 40%, respectively (Figure 3.11A). RITA was prioritised for further studies.

Next, ten-point dose response experiments revealed that Nemonapride repressed TUBB3- and GAPDH promoter activities in a dose-dependent and non-specific manner (Figure

3.10B). Alamar Blue assays demonstrated that Nemonapride treatments resulted in severe reduction in cell viability in both H460/TUBB3p-luc and H460/GAPDHp-luc cells

(Figure 3.11B) and thus was eliminated from the study. As was shown with other bioactive hits, GW843682X demonstrated a marked and dose-dependent inhibitory effect on both TUBB3 and GAPDH promoter activity with an IC50 ratio of 0.796 (Figure 3.11C).

This observation could be the result of general transcriptional repression or a general reduction in cell viability. Alamar Blue assays indicated that at drug concentrations that elicit marked reduction in TUBB3 promoter activity (10-100 μM or log M= -5 to -4), the cell viability (%) of H460/TUBB3p-luc cells and H460/GAPDHp-luc cells was 50% and

40%, respectively (Figure 3.11C). This suggests GW843682X-repressesed cell viability contributes, at least partly, to the marked reduction of TUBB3 and GAPDH promoter activities (Figure 3.11C). Hence, GW843682 was eliminated from this study.

144

Figure 3.11 Effect of top 3 bioactive hits on cell viability of H460/TUBB3p-luc and

H460/GAPDHp-luc cells

The ten-point dose response studies demonstrate the top 3 bioactive hits; RITA,

Nemonapride and GW43682X potently repress TUBB3 promoter activity. To distinguish this repression in promoter activity from a reduction in cell viability, Alamar Blue assays were performed. The dose range used in Alamar Blue assays corresponded to that in ten- point response studies. (A) At a drug concentration that elicits potent TUBB3 promoter activity inhibition (10 μM or log M= -5), RITA reduced cell viability of H460/TUBB3p- luc and H460/GAPDHp-luc cells by 30% and 40%, respectively. (B) Nemonapride treatments resulted in a severe reduction in cell viability in both H460/TUBB3p-luc and

H460/GAPDHp-luc cells. (C) At drug concentrations that elicit marked reduction in

TUBB3 promoter activity (10-100 μM or log M= -5 to -4), the cell viability of

H460/TUBB3p-luc cells and H460/GAPDHp-luc cells were 50% and 40%, respectively.

Experiments were performed in triplicate. Drug concentrations are in log M scale, with data plotted as mean ± SEM of 3 independent experiments.

145

146

3.1 Discussion

NSCLC survival rates are dismal and chemotherapy resistance is the primary cause of treatment failure. The aberrant expression of βIII-tubulin has been strongly implicated in chemoresistance and tumour survival in NSCLC, where it has been identified as a bona fide target for chemosensitisation (Gan et al., 2007; McCarroll et al., 2010). In order to understand how III-tubulin (encoded by TUBB3 gene) is regulated in NSCLC cells we sought to identify chemical small molecules that can modulate its expression. In this chapter, we have developed, validated and conducted a cell-based screen to identify novel small chemical molecules and bioactives that can modulate TUBB3 promoter activity in

NSCLC. Two novel small chemical molecule hits, WECC0018639 and WECC0017371, and one bioactive hit, RITA were obtained from high throughput screening of libraries of diverse chemical small molecules (30,224) and bioactives (3680). All three hits were tested and selected based on their ability to modulate TUBB3 promoter activity. We do not anticipate these hits to be false positive, as reactive, non-drug-like, promiscuous and known frequent hitter scaffolds have been extensively filtered from our small molecule compound libraries (Baell, 2013; Baell and Holloway, 2010; Lackovic et al., 2014). In addition, H460/GAPDHp-luc cells were included in the secondary screen to minimise false positives.

Assays developed for HTS can be divided broadly into biochemical assays and cell-based assays. In traditional biochemical assays, the reaction is homogenous in nature and thus can be miniaturised easily with little variability. However, the activity of a small molecule in reconstituted in vitro assays does not always reflect the same activity in a cellular context, as multiple factors are overlooked. For example, the requirement for cellular co- factors, issues of membrane permeability, off-target effects, and cytotoxicity. In contrast

147 to the traditional one-step biochemical assays, cell-based assays are a more physiologically relevant assay compared to the traditional one-step biochemical assays.

The cell-based screening approach utilised in this study allowed the identification of small molecules and bioactives that can modulate TUBB3 promoter activity in a more physiologically relevant environment, complete with intact regulatory networks and feedback mechanisms. This assay format and shotgun approach is especially suitable for the investigation of TUBB3 modulators as little is known about how TUBB3/βIII-tubulin expression is regulated. Hits could potentially modulate TUBB3 promoter activity at several levels, such as modulating cellular factors that regulate TUBB3 transcriptional activation, interfering with unknown feedback mechanisms, or directly targeting the minimal TUBB3 promoter region and/or repressors and enhancers within the TUBB3 gene. The two small molecule hits and RITA do not share any structural similarities.

Although unclear, this suggests that the mechanisms through which these compounds repress TUBB3 promoter activity could be distinct from one another. Cell-based screens often identify hits that exert their effect in an indirect manner through interactions with cellular factors rather than directly targeting the promoter region. Despite the potential range of mechanisms through which these compounds could be modulating TUBB3 promoter activity, the molecular mechanism by which these hits affect TUBB3 promoter activity is not within the scope of this study. Instead, this thesis will focus on validating these 3 hits by investigating their downstream effect and specificity on TUBB3/βIII- tubulin gene and protein expression in two independent NSCLC cell lines, H460 and

H1299. These molecules may become useful tools to study the biology of how

TUBB3/βIII-tubulin is regulated in NSCLC cells.

148

Chapter 4

Effect of Small Molecule and Bioactive Hits on TUBB3

Gene and βIII-Tubulin Protein Expression

4.1 Introduction

Non-small-cell lung cancer (NSCLC) survival rates are dismal and chemotherapy resistance is the primary cause of treatment failure. The aberrant expression of βIII- tubulin has been strongly implicated in chemoresistance and tumour survival in NSCLC, where it has been identified as a bona fide target for chemosensitisation (Gan et al., 2007;

McCarroll et al., 2010). Currently, there is no commercially available TUBB3/III-tubulin inhibitor and little is known about how this protein is regulated. In an attempt to better target and dissect the regulation of TUBB3/III-tubulin in NSCLC cells, we sought to identify chemical small molecules that can modulate its expression. As described in

Chapter 3, one bioactive hit, RITA and two novel small chemical molecule hits,

WECC0018639 and WECC0017371 were obtained from high-throughput screening of libraries of diverse chemical small molecules and bioactives. All three hits were selected based on their ability to modulate TUBB3 promoter activity, while having a lesser effect on GAPDH promoter activity. This chapter focuses on the molecular and cellular characterisation of these three compounds, evaluating their ability and specificity to modify TUBB3 gene and βIII-tubulin protein expression in NSCLC cells.

149

4.2 Results

4.2.1 Validation of bioactive hit RITA

As discussed in chapter 3 (section 3.2.4.4), RITA was identified as a modulator of TUBB3 promoter activity. RITA has demonstrated apoptotic activity in preclinical studies of various tumour cell lines (as described in section 3.2.4.3). Briefly, RITA binds to wild- type p53, disrupting p53-MDM2 interaction and induces apoptosis in tumour cell lines in vitro and in vivo (Issaeva et al., 2004; Zhao et al., 2010).

Prior to evaluating the effect of RITA on TUBB3 gene and βIII-tubulin protein expression, the subcytotoxic range of RITA in H460 cells was first determined. Subcytotoxic range is determined to identify doses of RITA that can modulate TUBB3 gene and βIII-tubulin protein expression without causing cytotoxicity. Here, subcytotoxic range is defined as a dose that inhibits H460 cell viability by less than or equal to 50% (IC50). H460 cell viability (%) was measured using Alamar Blue assays to determine the subcytotoxic range of RITA at 24, 48 and 72 h. These time points were chosen as the turnover of TUBB3 gene and βIII-tubulin protein transcripts are 20 and 30 h, respectively. The IC50 for RITA was 1.0 μM at all time-points tested and thus a subcytotoxic dose range including 0.05,

0.1, 0.5 and 1.0 μM was chosen for subsequent studies (Figure 4.1A). Next, Trypan blue assays were used to assess the effect of RITA on H460 cell viability and proliferation using 0.05, 0.1, 0.5 and 1 μM of RITA for 48 and 72 h. Data showed that at this dose range RITA has little effect on H460 cell viability, but exerted cytostatic effects at 48 and

72 h post-treatment (Figure 4.1B and C).

To investigate the ability of RITA to alter TUBB3 gene expression in H460 cells over time, RT-PCR was performed to measure TUBB3 gene level in cells treated with

150

Figure 4.1 Dose and time response of RITA on cell viability and cell proliferation in

H460 cells.

To investigate the effect of RITA on TUBB3 gene and βIII-tubulin protein expression, the subcytotoxic dose range of this bioactive compound was first determined in H460 cells.

H460 cells were treated with 0.0001 to 10 μM of RITA and Alamar Blue assays were performed to assess cell viability over 72 h. (A) Cell viability (%) was measured in H460 cells treated with 0.0001, 0.001, 0.005, 0.01, 0.05, 0.1, 0.5, 1.0, 5.0 and 10.0 μM of RITA at 24 (●), 48 (■) and 72 h (▲). In addition to Alamar Blue assays, Trypan blue assays were used to assess effects of RITA on H460 proliferation (%) and cell viability (%). (B)

0.05, 0.1, 0.5 and 1.0 μM of RITA, has little effect on H460 cell viability (●) at 48h, but significantly reduced cell proliferation (■) in a dose-dependent manner. (C) Similar results were observed after RITA treatment at 72 h. Cell viability (%) and cell proliferation (%) were normalised to non-treated control and plotted as mean ± SEM of

3 independent experiments.

151

152 subcytotoxic doses of RITA, including 0.1, 0.5 and 1.0 μM. At 48 h, H460 TUBB3 gene expression remained unchanged (Figure 4.2A). RITA showed a trend in enhancing

TUBB3 gene expression in H460 cells in a dose-dependent manner at 72 h, however, the increase was not statistically significant (Figure 4.2B). To investigate the specificity of this time- and dose-dependent increase in TUBB3 gene expression, RT-PCR was performed to examine the gene levels of TUBB2a, TUBB2b, TUBB2c and TUBB in H460 cells treated with RITA over time. Following 48 h of RITA treatment, gene levels of

TUBB2a, TUBB2b, TUBB2c and TUBB decreased in a dose dependent manner (Figure

4.3). The effects of RITA on TUBB3 were not specific to this isotype and the decrease in

TUBB (p<0.0001) and TUBB2b (p<0.0001) was statistically significant.

At 72 h of RITA treatment, TUBB3 and TUBB2a mRNA levels increased in a dose- dependent manner (Figure 4.4). This increase was statistically significant for TUBB2a

(p<0.001), but not TUBB3. TUBB2b mRNA levels decreased following RITA treatment, while TUBB2c and TUBB mRNA levels remained unchanged. Together, these findings indicate that the effect of RITA is not specific to TUBB3, as it alters the expression of multiple β-tubulin genes. Using western blotting and densitometry analysis, the time- and dose-dependent increase in TUBB3 and TUBB2a mRNA levels following RITA treatment did not lead to increases at the protein level (Figure 4.5). Additionally, βII-, βIV- and total

β-tubulin protein levels in H460 cells remained unchanged, apart from a decrease in total

β-tubulin at 1.0 μM (Figure 4.6). Collectively, the effect of RITA on TUBB3 gene and

βIII-tubulin protein expression is not specific to this β-tubulin isotype. Hence, RITA was eliminated as a candidate for the studies of TUBB3 regulation in NSCLC.

153

Figure 4.2 Effect of RITA on TUBB3 mRNA levels in H460 cells over time.

The initial screen in chapter 3 (section 3.2.4.4) showed that RITA repressed TUBB3 promoter-luciferase expression. To investigate the effect of RITA on endogenous TUBB3 mRNA levels, H460 cells were treated with RITA and the TUBB3 mRNA levels were measured at 48 and 72 h. (A) At 48 h, no change in H460 TUBB3 mRNA level was observed at all doses tested. (B) At 72 h, RITA showed a trend in enhancing TUBB3 mRNA expression in H460 cells in a dose-dependent manner. TUBB3 mRNA levels were normalised to β2 microglobulin (β2M) and were shown relative to non-treated controls.

Gene expression was plotted as a mean ± SEM of 3 independent experiments.

154

155

Figure 4.3 Effect of RITA on expression of β-tubulin genes at 48 h in H460 cells.

Initial RT-PCR analysis indicated that RITA enhanced TUBB3 gene expression in H460 cells in a time- and dose-dependent manner. To determine the specificity of this bioactive compound for TUBB3, H460 cells were treated with RITA and TUBB3, TUBB2a,

TUBB2b, TUBB2c and TUBB mRNA levels were measured at 48 h. TUBB3 mRNA expression data presented here is the same TUBB3 data shown in Figure 4.2 (48h). mRNA levels were normalised to β2M and were shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance,

*, p<0.05; ** p<0.001; ***, p<0.0001 and ****, p<0.00001.

156

RITA 48 h

157

Figure 4.4 Effect of RITA on expression of β-tubulin genes at 72 h in H460 cells.

The initial RT-PCR analysis indicated that 72 h of RITA treatment induced a trend in enhancing TUBB3 mRNA expression. To determine the specificity of this bioactive compound for TUBB3, H460 cells were treated with RITA and mRNA levels of β-tubulin genes were measured at 72 h. At 72 h, TUBB3 and TUBB2a mRNA levels increased in a dose-dependent manner. This increase was statistically significant for TUBB2a, but not

TUBB3. TUBB2b mRNA levels decreased followed by RITA treatments, while TUBB2c and TUBB mRNA levels remained unchanged. Tubulin mRNA levels were normalised to

β2M and were shown relative to non-treated controls. mRNA levels were plotted as mean

± SEM of 3 independent experiments. Statistical significance, *, p<0.05 and ** p<0.001.

TUBB3 mRNA expression data presented here is the same TUBB3 data shown in Figure

4.2 (72h).

158

RITA 72 h

159

Figure 4.5 The effect of RITA treatment on βIII-tubulin protein expression in H460 cells at 48 and 72 h.

RITA treatment demonstrated a time- and dose-dependent increase in TUBB3 mRNA expression in H460 cells. To assess whether this effect is translated at the protein level,

H460 cells were treated with 0.1, 0.5 and 1.0 μM of RITA for 48 and 72 h prior to protein isolation and Western blotting. This figure illustrates representative Western blots and semi-quantitation of βIII-tubulin protein expression in H460 cells after 48 and 72 h of

RITA treatment. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was detected simultaneously as the loading control. Densitometry analysis demonstrated that

βIII-tubulin protein expression remained unchanged after RITA treatment at both time points. βIII-tubulin protein levels were normalised to GAPDH and data are plotted as mean ± SEM of 3 independent experiments.

160

161

Figure 4.6 Effect of RITA on β-tubulin proteins in H460 cells.

In RT-PCR analysis, RITA treatment resulted in a broad and non-specific effect on the level of β-tubulin genes over time. In order to assess the downstream effect of RITA on protein expression of β-tubulin isotypes, H460 cells were treated with 0.1, 0.5 and 1.0 μM of RITA for 72 h prior to protein isolation and Western blotting. This figure shows representative Western blots and semi-quantitation of β-tubulin proteins in H460 cells treated with RITA. No change in βIII-, βII-, βIV- and total β-tubulin protein levels was observed in H460 cells after RITA treatment. Protein expression was normalised to

GAPDH control and plotted as mean ± SEM of 3 independent experiments.

162

RITA 72 h

163

4.2.2 Validation of hit compound WECC0018639

WECC0018639, a small molecule hit identified as described in Chapter 3 (section

3.2.4.2), was investigated next. The cell characterisations carried out for WECC0018639 were similar to those conducted for RITA. Briefly, Alamar Blue and Trypan blue assays were used to determine the subcytotoxic range of WECC0018639 in H460 cells over time.

Next, the potential effects of WECC0018639 on the H460 cell cycle profile and the level of TUBB3 gene and protein expression were assessed. Based on these evaluations, a decision was made on whether WECC0018639 is a suitable candidate for the study of

TUBB3 regulation in NSCLC.

As we needed to identify doses of WECC0018639 that can modulate TUBB3 expression without causing cytotoxicity, the subcytotoxic range of this small molecule in H460 cells was first determined. A dose range of WECC0018639, including 0.0001, 0.001, 0.01, 0.1,

1 and 10 μM was tested over time. The IC50 for WECC0018639 was 10 μM at 24 h and thus a subcytotoxic dose range between 0.0001 and 10 μM was chosen for subsequent studies (Figure 4.7A). The Trypan blue assay indicated that 0.0001 to 10 μM of

WECC0018639 has no effect on H460 cell viability, but inhibited cell proliferation in a time- and dose-dependent manner (Figure 4.7B and C). Together, 0.0001 to 10 μM of

WECC0018639 did not induce cell death, but was cytostatic to H460 cells. Additionally, an increase in Alamar Blue signal was noted in WECC0018639-treated H460 cells over time (Figure 4.7). We cannot exclude possibility that WECC0018639 is unstable over time in culture.

To investigate the potential cause of the cytostatic effects of WECC0018639, flow cytometry was performed to examine the influence of this small molecule on H460 cell cycle profiles. This approach revealed the distribution of cells in major phases of the cell

164

Figure 4.7 Dose and time response of WECC0018639 on cell viability and cell proliferation in H460 cells.

To investigate the effect of WECC0018639 on TUBB3 gene and βIII-tubulin protein expression, the subcytotoxic concentrations of this small molecule were first determined in H460 cells. Alamar Blue assays were performed to assess the effect of WECC0018639 on H460 cell viability (%) compared to non-treated control cells. (A) H460 cells were treated with a range of 0.0001-10 μM of WECC0018639. Alamar Blue assays were performed to measure cell viability (%) of H460 cells at 24 h (●), 48 h (■) and 72 h (▲) post-treatment. In addition to Alamar Blue assays, Trypan blue assays were used to assess the effect of 0.0001-10 μM of WECC0018639 on H460 cell proliferation (%) and cell viability (%) over time. (B) 0.001-10 μM of WECC0018639 has little effect on H460 cell viability (●) at 48 h, but significantly reduced cell proliferation (■) in a dose-dependent manner, exerting a cytostatic effect. (C) Similar results were observed after

WECC0018639 treatment at 72 h. Cell viability (%) and cell proliferation (%) were normalised to non-treated control and plotted as mean ± SEM of 3 independent experiments.

165

166

Figure 4.8 Effect of WECC0018639 on the H460 cell cycle profile over time.

To investigate the potential cause of the cytostatic effect of WECC0018639 on H460 cells, cell cycle analysis was performed. H460 cells were treated with 0.1, 1.0 and 10.0

μM of WECC0018639 for 24, 48 and 72 h prior to staining with propidium iodide and performing cell cycle analysis. Cell cycle analysis examined the sub G1, G0/G1, S, and

G2/M fractions. H460 cell cycle profile remained unchanged over time upon 0.1 (■), 1

(▲) and 10 (▲) μM WECC0018639 treatment in comparison to non-treated control cells

(●). Distribution of cells (%) was plotted as mean ± SEM of 3 independent experiments.

167

168 cycle (G0/G1, S, and G2/M) and made it possible to detect cell death (sub G1). Figure 4.8 shows there was no change in the H460 cell cycle profile following WECC0018639 treatment. Together, 0.0001 to 10 μM of WECC0018639 was subcytotoxic to H460 cells and the cytostatic effect observed was not due to changes in the cell cycle. Based on these data, 0.0001 to 10.0 μM of WECC0018639 was chosen for subsequent expression studies.

Having determined the subcytotoxic doses of WECC0018639, the effect of this small molecule on TUBB3 gene and βIII-tubulin protein expression was next investigated.

Figure 4.9 shows that WECC0018639 significantly enhanced TUBB3 mRNA expression in H460 cells in a time- and dose-dependent manner. TUBB3 upregulation was observed as early as 48 and 72 h after WECC0018639 treatment. A dose-dependent increase in

TUBB3 mRNA levels, up to 1-fold relative to control was observed at 48 h (p<0.0001;

Figure 4.9B). This enhanced expression was even more pronounced at 72 h, with an increase up to 3-fold in TUBB3 mRNA relative to control (p<0.05; Figure 4.9C) following

WECC0018639 treatment. Given that WECC0018639 was chosen based on its repressive effect on TUBB3 promoter-luciferase activity in the initial screen, this finding is unexpected and will be discussed in section 4.3. To investigate the specificity of

WECC0018639-induced TUBB3 enhancement, mRNA levels of TUBB2a, TUBB2b,

TUBB2c and TUBB in H460 cells treated with WECC0018639 were examined. Following

48 h treatment of H460 cells with WECC0018639, TUBB2a, TUBB2c and TUBB mRNA levels remained unchanged, while TUBB2b mRNA levels were reduced in a dose- dependent manner (Figure 4.10). At 0.1 and 1 μM of WECC0018639, TUBB2b mRNA level was significantly reduced by 25% (p<0.05) and 50% (p<0.0001) relative to control, respectively (Figure 4.10). At the highest concentration, 10 μM of WECC00018639 significantly reduced TUBB2b mRNA level by 75% (p<0.00001) relative to control at 48 h. At 72 h, WECC0018639 treatment significantly enhanced TUBB3 and TUBB2a mRNA 169

Figure 4.9 Effect of WECC0018639 on TUBB3 mRNA levels in H460 cells over time.

The initial screen demonstrated that WECC0018639 repressed TUBB3 promoter luciferase expression in H460 cells. To investigate whether this affected the intrinsic levels of TUBB3, mRNA levels of TUBB3 were measured in H460 cells. RT-PCR revealed that WECC0018639 treatment significantly enhanced TUBB3 mRNA levels in a time- and dose-dependent manner. TUBB3 mRNA levels in H460 cells after 24 (A), 48

(B) and 72 h (C) of WECC0018639 treatment is shown. TUBB3 mRNA levels were normalised to β2M gene expression and were shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05; ** indicates p<0.001; and *** indicates p<0.0001.

170

171

Figure 4.10 Effect of WECC0018639 on mRNA levels of β-tubulin genes in H460 cells at 48 h.

Initial RT-PCR analysis indicated that WECC0018639 treatment significantly enhanced

TUBB3 gene expression in a time- and dose-dependent manner. In order to determine the specificity of this compound for TUBB3, H460 cells were treated with 0.01, 0.1, 1 and 10

μM of WECC0018639 for 48 h and mRNA levels of TUBB3, TUBB2a, TUBB2b,

TUBB2c and TUBB were determined. mRNA levels were normalised to β2M gene expression and were shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05;

** indicates p<0.001; and *** indicates p<0.0001. TUBB3 mRNA expression data presented here is the same TUBB3 data shown in Figure 4.9 (48h).

172

WECC0018639 48 h

173

Figure 4.11 Effect of WECC0018639 on mRNA levels of β-tubulin genes in H460 cells at 72 h.

Having demonstrated that TUBB3 and TUBB2b were altered at 48 h following treatment with WECC0018639, the effects on the β-tubulin genes, TUBB3, TUBB2a, TUBB2b,

TUBB2c and TUBB were examined at 72 h of treatment. mRNA levels were normalised to β2M and were shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05;

** indicates p<0.001; and *** indicates p<0.0001. TUBB3 mRNA expression data presented here is the same TUBB3 data shown in Figure 4.9 (72h).

174

WECC0018639 72 h

175 levels in a dose-dependent manner. At 1 and 10 μM, WECC0018639 significantly increased TUBB2a mRNA levels by 150% (p<0.001) and 200% (p<0.0001) relative to control, respectively (Figure 4.11). No change in mRNA levels of other β-tubulin isotypes was observed. Together, WECC0018639 treatment significantly increased TUBB3 and

TUBB2a mRNA levels in a time- and dose-dependent manner. To assess whether this effect was present at the protein level, Western blotting was performed. No change in

βIII-tubulin protein expression was observed after 48 and 72 h of WECC0018639 treatment and thus the enhanced TUBB3 gene expression was not translated at the protein level (Figure 4.12). Due to the limited evidence to support WECC0018639 as a

TUBB3/βIII-tubulin expression modifier, this small molecule was not prioritised for further study in this thesis.

4.2.3 Validation of hit compound WECC0017371

The other small molecule hit identified in chapter 3 (section 3.2.4.2), WECC0017371, was prioritised for further investigation. The characterisations of the cellular effects of

WECC0017371 were similar to those conducted for RITA and WECC0018639. First,

Alamar Blue assays were initially used to determine the subcytotoxic dose range of

WECC0017371 in order to identify a dose range that can modulate TUBB3 gene and βIII- tubulin protein expression without cell death. H460 cells were treated with 0.1, 1, 10, 25,

100 and 250 μM of WECC0017371 and cell viability (%) was measured over 72 h. The

IC50 for WECC0017371 was 100 μM at all time-points tested (Figure 4.13A) and thus a subcytotoxic dose range between 0.1 and 100 μM was chosen for subsequent studies.

Trypan blue assays further showed at this dose range WECC0017371 has little effect on

H460 cell viability, but inhibited cell proliferation in a time- and dose-dependent manner,

176

Figure 4.12 The effect of WECC0018639 treatment on βIII-tubulin protein expression in H460 cells over time.

RT-PCR analysis of WECC0018639 treated H460 cells revealed a potent time- and dose- dependent increase in TUBB3 mRNA expression. To assess whether this effect is translated at the protein level, Western blotting was performed. Representative Western blots and quantitation of βIII-tubulin protein expression in H460 cells, following 48 and

72 h of WECC0018639 treatment are shown. GAPDH was used as a loading control. Data demonstrated that the increase in TUBB3 gene expression was not translated at the protein level, as βIII-tubulin protein expression remained unchanged after 48 and 72 h of

WECC0018639 treatment. βIII-tubulin protein expression was normalised to GAPDH and data are plotted as mean ± SEM of 3 independent experiments.

177

178

Figure 4.13 Dose and time response of WECC0017371 on cell viability and cell proliferation in H460 cells.

To investigate the effect of WECC0017371 on TUBB3 gene and βIII-tubulin protein expression, the subcytotoxic dose range of this small molecule was first determined in

H460 cells. H460 cells were treated with 0.1, 1, 10, 25, 100 and 250 μM of

WECC0017371 and Alamar Blue assays were performed to assess cell viability over 72 h. (A) Cell viability (%) was measured in H460 cells following 0.1, 1, 10, 25, 100 and

250 μM WECC00173710 treatment at 24 h (●), 48 h (■) and 72 h (▲). In addition to

Alamar Blue assays, Trypan blue assays were used to assess the effect of WECC0017371 on H460 cell viability (%) and cell proliferation (%). (B) At 48 h, 1, 10, 25 and 100 μM of WECC0017371 has no effect on H460 cell viability (●), but reduced cell proliferation

(■) in a dose-dependent manner, exerting a cytostatic effect. (C) Similar results were observed after WECC0017371 treatment for 72 h. Cell viability (%) and cell proliferation

(%) were normalised to non-treated control and plotted as mean ± SEM of 3 independent experiments.

179

24h 48h 72h

180 suggesting this small molecule is cytostatic to H460 cells (Figure 4.13B and C). To investigate the potential cause of the cytostatic effects of WECC0017371, flow cytometry was performed to examine the influence of this small molecule on H460 cell cycle profiles. Cell cycle profiles of H460 cells were not affected by WECC0017371 treatment

(Figure 4.14). There was no increase in the sub G1 population over time at concentrations tested, suggesting there was no increase in cell death and 0.1-100 μM of WECC0017371 is subcytotoxic. The cytostatic effect observed was not likely due to alterations in the cell cycle. Based on Alamar Blue, Trypan blue and cell cycle data, 1, 10, 25 and 100 μM of

WECC0017371 were chosen for subsequent expression analysis.

The initial screen demonstrated that the small molecule WECC0017371 repressed TUBB3 promoter luciferase activity. In order to investigate whether WECC0017371 can alter the intrinsic levels of TUBB3, H460 cells were treated and the effects on mRNA and protein expression were examined. To assess the specificity of this small molecule for TUBB3, the expression of different β-tubulin isotypes were also examined. After 48 h of treatment,

WECC0017371 increased TUBB3 mRNA expression in a dose-dependent manner

(Figure 4.15). At 100 μM, WECC0017371 significantly increased TUBB3 mRNA levels by 250% (p<0.0001) relative to control. TUBB2a, TUBB2b, TUBB2c and TUBB mRNA levels remained unchanged at all doses tested. This suggests that the effect of

WECC0017371 on TUBB3 is specific to this β-tubulin isotype at 48 h. At 72 h, the magnitude of the TUBB3 enhancement was greatly reduced, while TUBB2a mRNA levels significantly increased in a dose-dependent manner (p<0.001; Figure 4.16).

WECC0017371 elicited a 100% increase in TUBB and TUBB2c mRNA expression relative to control at 100 μM only, but not lower doses. The mRNA level of TUBB2b remained unchanged for all treatment times and concentrations. Similar analysis was also performed to examine the effect and specificity of WECC0017371 on endogenous βIII- 181

Figure 4.14 Effect of WECC0017371 on the H460 cell cycle profile over time.

To investigate the potential cause of the cytostatic effect of WECC0017371, cell cycle analysis was performed. H460 cells were treated with WECC0017371 at either 10 μM or

100 μM for 24, 48 and 72 h prior to staining with propidium iodide and cell cycle analysis.

Cell cycle analysis examined the sub G1, G0/G1, S, and G2/M fractions. H460 cell cycle remained unchanged over time upon 10 (■) and 100 (▲) μM WECC0017371 treatment in comparison to non-treated control (●). Distribution of cells (%) was plotted as mean ±

SEM of 3 independent experiments.

182

183

Figure 4.15 Effect of WECC0017371 on mRNA levels of β-tubulin genes in H460 cells at 48 h.

The initial screen demonstrated that WECC0017371 repressed TUBB3 promoter luciferase expression. To investigate the effect and specificity of this compound for

TUBB3, mRNA expression of different β-tubulin isotypes were examined in H460 cells.

RT-PCR was performed to quantify mRNA levels of β-tubulin genes relative to non- treated controls. 48 h of WECC0017371 treatment significantly increased TUBB3 mRNA levels in a dose-dependent manner, while TUBB2a, TUBB2b, TUBB2c and TUBB mRNA levels remained unchanged. mRNA levels were normalised to β2M and were shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance, **** indicates p<0.0001.

184

WECC0017371 48 h

185

Figure 4.16 Effect of WECC0017371 on mRNA levels of β-tubulin genes in H460 cells at 72 h.

Given that WECC0017371 treatment significantly enhanced TUBB3 gene expression in a dose-dependent manner at 48 h, the effect and specificity of WECC0017371 on different

β-tubulin genes was also analysed at 72 h post-treatment. RT-PCR analysis was performed to quantify the effect of WECC0017371 on mRNA levels of TUBB3, TUBB2a,

TUBB2b, TUBB2c and TUBB relative to non-treated controls. WECC0017371 treatment induced a small increase in TUBB3 mRNA levels at 100 μM only, though this did not reach statistical significance. TUBB2a mRNA levels increased significantly in a dose- dependent manner. TUBB2b, TUBB2c and TUBB mRNA levels did not significantly changed following 1, 10 and 25 μM of WECC0017371 treatment for 72 h. At 100 μM

WECC0017371 elicited a 1-fold increase in TUBB and TUBB2c mRNA expression relative to control. TUBB2b mRNA levels remained unchanged at all times and concentrations. mRNA levels were normalised to β2M and were shown relative to non- treated controls. Data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * and **, indicates p<0.005; *** indicates p<0.001.

186

WECC0017371 72 h

187 tubulin protein expression. Western blot analysis showed that the increase in TUBB3 gene expression was translated at the protein level. WECC0017371 treatment resulted in a moderate but not significant increase in βIII-tubulin protein expression (Figure 4.17A).

This effect is dose-dependent and specific to βIII-tubulin. No change in other β-tubulin isotypes was observed (Figure 4.17B). This suggests WECC0017371 may be a potential

βIII-tubulin enhancer. Only βII-, βIV, βIII- and total β-tubulin were evaluated.

Next, in order to determine whether the WECC0017371 enhanced TUBB3 gene and βIII- tubulin protein expression was a common phenomenon in lung cancer cells, expression studies on another NSCLC cell line, H1299, were performed. H1299 cells aberrantly express βIII-tubulin at a level similar to H460 cells and was chosen for expression studies.

Prior to assessing the effect of WECC0017371 on TUBB3 gene and βIII-tubulin protein levels, Trypan blue assays were performed to determine the time and dose response of

WECC0017371 on H1299 cell proliferation and viability. The same doses of

WECC0017371 used on H460 cells, 1, 10, 25 and 100 μM, were used on H1299 cells.

Similar to the response in H460 cells, it was demonstrated that WECC0017371 had no effect on H1299 cell viability and induces a moderate cytostatic effect in a dose- dependent manner (Figure 4.18). Having determined subcytotoxic doses of

WECC0017371 in H1299 cells, the effect of the drug on TUBB2 and TUBB3 gene expression was examined, using RT-PCR. Interestingly, as was observed in H460 cells,

WECC0017371 treatment induced an isotype-specific and time- and dose-dependent increase in TUBB3 (Figure 4.19 and Figure 4.20) and TUBB2a (Figure 4.20) mRNA levels in H1299 cells. Specifically, at 25 and 100 μM, 48 h of WECC0017371 treatment significantly increased TUBB3 mRNA expression by 20% (p<0.05) and 50% (p<0.0001) relative to control, respectively. The mRNA levels of TUBB2a, TUBB2b, TUBB2c and

TUBB remained unchanged at all doses tested at 48 h. Further, TUBB3 mRNA levels 188

Figure 4.17 The effect of WECC0017371 treatment on β-tubulin protein expresion in H460 cells at 72 h.

WECC0017371 treatment induced a potent and isotype-specific increase in TUBB3 and

TUBB2a gene expression. To assess the effect and specificity of WECC0017371 on protein expression of different β-tubulin isotypes, H460 cells were treated with 1, 10, 25 and 100 μM of WECC0017371 for 72 h prior to protein isolation and Western blotting.

(A) Representative western blot of βIII-tubulin protein expression in H460 cells following treatment with WECC0017371. GAPDH was used as a loading control. WECC0017371 treatment for 72 h enhanced βIII-tubulin protein expression in H460 cells in a dose- dependent manner. (B) Densitometry analysis and semi-quantitation of βIII-, βII-, βIV- and total β-tubulin protein expression in H460 cells following WECC0017371 treatment for 72 h. Tubulin protein expression was normalised to GAPDH control and plotted as mean ± SEM of 3 independent experiments.

189

190

Figure 4.18 Dose and time response of WECC0017371 on H1299 cell proliferation and cell viability.

Prior to examining the effect of WECC0017371 on TUBB3 gene and βIII-tubulin protein expression in H1299 cells, Trypan blue assays were used to assess the time and dose response of the small molecule on H1299 cell proliferation and cell viability. (A) 1, 10,

25 and 100 μM of WECC0017371 treatment for 48 h had no effect on H1299 cell viability but reduced cell proliferation in a dose-dependent manner. (B) Similar results were observed at 72 h of WECC0017371 treatment. ●- cell viability (%); ■- cell proliferation

(%). Data are normalised to non-treated controls and plotted as mean ± SEM of 3 independent experiments.

191

192

Figure 4.19 Effect of WECC0017371 on mRNA levels of β-tubulin genes in H1299

NSCLC cells at 48 h.

Given that WECC0017371 treatment significantly increased TUBB3 and TUBB2a mRNA levels in H460 cells, we sought to investigate the effect of this drug on another NSCLC cell line, H1299. WECC0017371 treatment for 48 h significantly increased TUBB3 mRNA levels in a dose-dependent manner, while TUBB2a, TUBB2b, TUBB2c and TUBB mRNA levels remained unchanged. mRNA levels were normalised to β2M and are shown relative to non-treated controls. Data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05; **** indicates p<0.0001.

193

WECC0017371 48 h

194

Figure 4.20 Effect of WECC0017371 on mRNA levels of β-tubulin genes in H1299

NSCLC cells at 72 h.

Given that WECC0017371 treatment significantly increased TUBB3 and TUBB2a mRNA levels in H460 cells, we sought to investigate the effect of this drug in another NSCLC cell line, H1299, using RT-PCR. 72 h of WECC0017371 treatment significantly increased

TUBB3 and TUBB2a mRNA levels in a dose-dependent manner, while sparing other β- tubulin isotypes. mRNA levels were normalised to β2M and are shown relative to non- treated controls. Data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * and ** indicates p<0.05; *** indicates p<0.005, and **** indicates p<0.0001.

195

WECC0017371 72 h

196 significantly increased by 20% (p<0.05), 25% (p<0.0001) and 50% (p<0.00001) relative to control after 10, 25 and 100 μM WECC0017371 treatment, respectively, at 72 h.

Treatment of H1299 cells with 25 and 100 μM of WECC0017371 resulted in a significant increase in TUBB2a mRNA level by 50% (p<0.05) and 75% (p<0.01) relative to control, respectively. The mRNA levels of TUBB2b, TUBB2c and TUBB remained unchanged at all doses tested at 72 h.

Additional expression and functional analysis were performed to investigate the effect of

WECC0017371 on βII- and βIII-tubulin expression. To investigate whether the changes in tubulin protein expression resulted in changes in microtubule morphology, immunostaining and confocal microscopy of βII- and βIII-tubulin were performed on

H460 cells treated with WECC0017371. Treatment of H460 cells with 100 μM of

WECC0017371 resulted in potent enhancement of βII-tubulin and βIII-tubulin in both gene and protein studies, without affecting cell viability. Consistent with the Western blot protein data above, βIII-tubulin immunostaining was relatively more prominent in

WECC017371 treated H460 cells than non-treated control cells at both 48 and 72 h

(Figure 4.21A and B). Similar results were obtained for βII-tubulin, where βII-tubulin immunostaining was relatively higher in WEC0017371 treated cells than control cells

(Figure 4.22A and B). In addition, the microtubule morphology of WECC0017371 treated

H460 cells was similar to control cells. These data suggests that WECC0017371 is not a microtubule interfering agent, but instead appears to be a βII-tubulin and βIII-tubulin enhancer. Similar results were obtained with H1299 cells (Figure 4.23A and B and Figure

4.24A and B), suggesting that there is a common mechanism of how WECC0017371 enhances βII- and βIII-tubulin expression in these two NSCLC cell lines.

197

Figure 4.21 Effect of WECC0017371 on microtubule morphology and βIII-tubulin protein expression in H460 cells.

Immunostaining and confocal microscopy were performed (as described in Materials and

Methods section 2.2.9.2) to investigate the effect of WECC0017371 on microtubule morphology and βIII-tubulin protein expression in H460 cells. (A) H460 cells were fixed and immunostained to detect βIII-tubulin and α-tubulin expression 48 h post-

WECC0017371 treatment. βIII-tubulin immunostaining (orange/red) was relatively more prominent in treated than non-treated H460 cells, while α-tubulin immunostaining (green) remained unchanged in both treated and non-treated samples. (B) Similar results were obtained at 72 h post-WECC0017371 treatment, where βIII-tubulin immunostaining

(orange/red) was relatively higher in treated than non-treated H460 cells. Immunostaining of α-tubulin (green) remained unchanged in both treated and non-treated samples.

Nucleus and chromosomes were stained with DAPI (blue). Scale bars-20 μm.

198

199

Figure 4.22 Effect of WECC0017371 on microtubule morphology and βII-tubulin protein expression in H460 cells.

Given that WECC0017371 treatment led to increased TUBB2a gene expression, immunostaining and confocal microscopy were performed to investigate the effect of

WECC0017371 on microtubule morphology and βII-tubulin protein expression in H460 cells. (A) H460 cells were fixed and immunostained to detect βII-tubulin and α-tubulin expression 48 h post-WECC0017371 treatment. βII-tubulin immunostaining (orange/red) was relatively more prominent in treated than non-treated H460 cells, while α-tubulin immunostaining (green) remained unchanged in both treated and non-treated samples.

(B) Similar results were obtained at 72 h post-WECC0017371 treatment, where βII- tubulin immunostaining (orange/red) was relatively higher in treated than non-treated

H460 cells. Immunostaining of α-tubulin (green) remained unchanged. Nucleus and chromosomes were stained with DAPI (blue). Scale bars-20 μm.

200

201

Figure 4.23 Effect of WECC0017371 on microtubule morphology and βIII-tubulin protein expression in H1299 cells.

WECC0017371 treatment caused a prominent increase in βIII-tubulin immunostaining in

H460 cells. In order to determine whether the WECC0017371 enhanced βIII-tubulin expression was a common phenomenon in lung cancer cells, immunostaining and confocal microscopy on an independent NSCLC cell line, H1299, were performed. (A)

H1299 cells were fixed and immunostained to detect βIII-tubulin and α-tubulin expression 48 h post-WECC0017371 treatment. βIII-tubulin immunostaining

(orange/red) was relatively more prominent in treated than non-treated cells, while α- tubulin immunostaining (green) remained unchanged in both treated and non-treated samples. (B) Similar results were obtained at 72 h post-WECC0017371 treatment, where

βIII-tubulin immunostaining (orange/red) was relatively higher in treated than non- treated H1299 cells. Immunostaining of α-tubulin (green) remained unchanged in both treated and non-treated samples. Nucleus and chromosomes were stained with DAPI

(blue). Scale bars-20 μm.

202

203

Figure 4.24 Effect of WECC0017371 on microtubule morphology and βII-tubulin protein expression in H1299 cells.

Given WECC0017371 treatment can specifically increase βII-tubulin protein expression in H460 cells, immunostaining and confocal microscopy were performed to investigate the effect of this small molecule on microtubule morphology and βII-tubulin protein expression in an independent NSCLC cell line, H1299. (A) H1299 cells were fixed and immunostained to detect βII-tubulin and α-tubulin expression 48 h post-WECC0017371 treatment. βII-tubulin immunostaining (orange/red) was relatively more prominent in treated than non-treated cells, while α-tubulin immunostaining (green) remained unchanged in both treated and non-treated samples. (B) Similar results were obtained at

72 h post-WECC0017371 treatment, where βII-tubulin immunostaining (orange/red) was relatively higher in treated than non-treated H1299 cells. Immunostaining of α-tubulin

(green) remained unchanged. Nucleus and chromosomes were stained with DAPI (blue).

Scale bars-20 μm.

204

205

It is well established that stable knockdown of βIII-tubulin significantly sensitises

NSCLC to broad classes of chemotherapeutics in vitro and in vivo (Gan et al., 2011; Gan et al., 2010; Gan et al., 2007; McCarroll et al., 2010). As a functional assay to validate the WECC0017371 enhanced βIII-tubulin expression, we sought to determine whether

WECC0017371 is capable of modifying sensitivity of NSCLC cells to chemotherapeutic drugs. To quantitate changes in drug sensitivity, drug-treated clonogenic assays were performed using cisplatin, vincristine or paclitaxel alone (as previously described in Gan et al., 2007) or in combination with WECC0017371. For clonogenic assays, 25 μM of

WECC0017371 was chosen as this concentration demonstrated a moderate increase in

βIII-tubulin gene and protein expression without markedly compromising H460 cell proliferation. Cisplatin, vincristine and paclitaxel were chosen here as their efficacy has previously been shown to be influenced by βIII-tubulin levels (Gan et al., 2007;

McCarroll et al., 2010). Interestingly, WECC0017371-induced βIII-tubulin expression was functional and led to a significant decrease in sensitivity to cisplatin, paclitaxel and vincristine (Figure 4.25A-C). These data is consistent with our previous studies showing that βIII-tubulin expression reduces sensitivity to broad classes of chemotherapeutic agents (Gan et al., 2007; McCarroll et al., 2010). Given there is limited knowledge on how TUBB3/βIII-tubulin is regulated, WECC0017371 will be a valuable small molecule probe to identify cell-based regulators of TUBB3/βIII-tubulin in NSCLC cells.

206

Figure 4.25 WECC0017371 clonogenic assays in the presence of cisplatin, paclitaxel or vincristine in H460 cells.

Given that βIII-tubulin is known to modulate sensitivity to chemotherapeutic drugs in

NSCLC, we sought to determine whether WECC0017371-induced enhanced βIII-tubulin expression is capable of decreasing sensitivity of H460 cells to conventional chemotherapeutics. Clonogenic assays were performed on H460 cells treated with TBAs or DNA-damaging agents alone, or in combination, with WECC0017371 to determine

H460 drug sensitivity. (A) Clonogenic assays were performed on H460 cells treated with cisplatin alone (●) or in combination with 25 μM WECC0017371 (■). (B) Clonogenic assays were performed on H460 cells treated with paclitaxel alone (●) or in combination with 25 μM WECC0017371 (■). (C) Clonogenic assays were performed on H460 cells treated with vincristine alone (●) or in combination with 25 μM WECC0017371 (■).

Clonogenic survival was expressed as surviving factions (%) (as described in Materials and Methods section 2.2.10) and plotted as mean ± SEM of 3 independent experiments.

Statistics were calculated by comparing the surviving fraction of the drug-treated cells with the non-treated control cells at each drug concentration. Statistical significance, * indicates p<0.05

207

208

4.3 Discussion

In an attempt to understand how TUBB3/βIII-tubulin is regulated in NSCLC cells we sought to identify small molecule and bioactive compounds that can modulate TUBB3 expression. In chapter 3, we identified a bioactive hit, RITA, and two small molecule hits,

WECC0018639 and WECC0017371, based upon their ability to repress TUBB3 promoter activity. In this chapter, we confirmed that TUBB3 promoter repression observed using the three molecule hits was not a result of a reduced cell viability. Further, both small molecule and bioactive hits have no effect on H460 cell viability nor on cell cycle profile.

The effects of these hits on TUBB3 gene and βIII-tubulin protein expression in H460 cells were evaluated. Counterintuitively, RITA enhanced TUBB3 mRNA expression in a time- and dose- dependent manner. This unexpected increase in TUBB3 mRNA level could have been a result of a global increase in β-tubulin level, and hence the effect of RITA on mRNA expression of β-tubulin isotypes was assessed. RITA significantly decreased the mRNA levels of global β-tubulin indicating the effect of RITA is not specific to TUBB3.

Protein levels of β-tubulin isotypes remained unchanged after RITA treatments. This lack of gene to protein translation is not unexpected as multiple studies have found that gene expression of tubulin isotypes does not always correlate with protein expression (Don et al., 2004; Hiser et al., 2006). Hence, it is imperative to quantitate the effect of hit compounds on β-tubulin expression at both gene and protein levels. Collectively, the effect of RITA on TUBB3 and βIII-tubulin is not specific to this β-tubulin isotype and hence, RITA was eliminated as a candidate for the studies of βIII-tubulin regulation in

NSCLC.

Assessments carried out for WECC0018639 were similar to those conducted for RITA.

Briefly, WECC0018639 specifically and significantly enhanced TUBB2a and TUBB3

209 mRNA expression in H460 cells in a time- and dose-dependent manner, while sparing other β-tubulin isotypes. Western blotting showed that WECC0018639 treatments have no effect on βIII-tubulin protein level nor on other β-tubulin isotypes. Due to the limited evidence to support WECC0018639 as a βIII-tubulin expression modifier, this compound was not prioritised for further study in this thesis. On the contrary, WECC0017371 demonstrated the ability to significantly enhance TUBB3 mRNA and βIII-tubulin protein expression in a time- and dose-dependent manner. We showed that WECC0017371 treatments did not alter microtubule morphology but enhanced βIII-tubulin immunostaining in two independent NSCLC cell lines, H460 and H1299. WECC0017371 treatment also increased levels of TUBB2a gene and βII-tubulin immunostraining in H460 and H1299 cells. Collectively, these results suggest two things, first, the WECC0017371 enhanced βII- and βIII-tubulin proteins are incorporated into microtubules; and second,

WECC0017371 is not a microtubule interfering agent, but instead appears to be a βII- and

βIII-tubulin enhancer.

Of particular interest is the novel finding that WECC0017371-enhanced βII- and βIII- tubulin expression is functional and can significantly decrease in vitro sensitivity to broad classes of chemotherapeutic drugs in H460 cells. This raises the question: Is βII-tubulin and/or βIII-tubulin responsible for mediating drug sensitivity to cisplatin, paclitaxel and vincristine? Based on knockdown and rescue experimental findings from studies by Gan et al. (2007) and McCarroll et al. (2010), βIII-tubulin has been shown to play a dominant role in mediating sensitivity to broad classes of chemotherapeutic agents, such as cisplatin, paclitaxel and vincristine in H460 cells. The involvement of βII-tubulin in mediating drug sensitivity to vincristine in WECC0017371-treated H460 cells is likely, as specific knockdown of βII-tubulin in H460 cells significantly enhances sensitivity to

Vinca alkaloids, but not paclitaxel (Gan and Kavallaris, 2008) nor epothilone (Gan et al., 210

2011). It is currently unclear whether βII-tubulin can influence intrinsic sensitivity to cisplatin. In order to clarify whether βII-tubulin plays a dominant role in mediating chemosensitivity in WECC0017371-treated H460 cells, one approach is to perform drug- treated clonogenic assay using βII-tubulin knockdown cells.

A collateral upregulation of TUBB2a mRNA level was consistently observed in RITA,

WECC0018639 and WECC0017371 treated H460 cells, as well as WECC0017371 treated H1299 cells. This collateral increase in βII-tubulin was also reported in human prostate carcinoma cells as a result of overexpression of βIII-tubulin (Ranganathan et al.,

2001). These observations raised the possibility that the two β-tubulin genes share common regulatory motifs, cellular regulators or regulatory pathways. An interrogation of ChIP-Sequencing data for transcription factor binding sites in the TUBB3 and TUBB2a gene promoters in A549 cells was performed (data not shown). Due to the absence of

ChIP-Sequencing data in H460 cells, data from another NSCLC cell line, A549 was examined. Apart from typical SP1 and E2F transcription factor binding sites, no common transcriptional factor binding motifs were identified in TUBB3 or TUBB2a promoters in

A549 cells. In addition, we cannot exclude the possibility that WECC0017371 may also act on trans-regulatory elements, resulting in the activation of both TUBB3 and TUBB2a expression simultaneously.

Given the repressed TUBB3 promoter-luciferase expression observed in our initial screen for all three hits, the time- and dose-dependent TUBB3/βIII-tubulin enhancement is unexpected and counterintuitive. It is believed that the readout provided by the

H460/TUBB3p-luc cells accurately reflects TUBB3 promoter activity, as the minimal promoter region has previously been demonstrated to be essential for its expression

(Dennis et al., 2002). These results demonstrated that TUBB3 promoter is important in

211

TUBB3 regulation but alone it is not sufficient to describe the full complexity of the biology behind it. Hence, TUBB3 regulation is more complex than previously appreciated. It is likely that the effects of these hits are mediated indirectly through a combination of unidentified cellular regulators and upstream/downstream regulatory motifs (enhancers and repressors). In addition, H460/TUBB3p-luc cells do not provide insight into the mechanism of action of these compounds and RT-PCR and Western blot data serve only to illuminate the downstream net effect. Studying the mechanism of action of WECC0017371 may therefore serve as a critical approach in unlocking these alternative mechanisms of TUBB3/βIII-tubulin regulation. This finding has exciting implications, as it indicates the existence of upstream pathways and potentially druggable proteins which may be used to target aberrant TUBB3/βIII-tubulin expression in cancer.

In summary, we detailed the discovery of WECC0017371 as a novel small molecule enhancer of TUBB3/βIII-tubulin from a HTS campaign of 30,224 small molecule chemicals. WECC0017371 was able to enhance endogenous βIII-tubulin expression in two independent NSCLC cell models and significantly reduce sensitivity to TBAs and

DNA-damaging agents. WECC0017371 exhibits the most “drug-like” structure overall compared to WECC0018639 and RITA. Based on our in vitro data, WECC0017371 was chosen for characterisation and prioritised for further development as a TUBB3/βIII- tubulin enhancer. In the hope to expose new therapeutic targets, this small molecule chemical serves as a valuable molecular probe to better understand the regulation of βIII- tubulin in cancer. In the next chapter, we will characterise WECC0017371 and establish structure-activity relationships. Its chemical structure will be optimised to generate compounds with improved specificity and potency compared to the parent molecule.

212

Chapter 5

Structure Activity Relationship Study of

Imidazopyridines and Development of Affinity-

Chromatography Probes for Chemical Proteomics

5.1 Introduction

The aberrant expression of βIII-tubulin has been strongly implicated in chemoresistance and tumour survival in NSCLC, where it has been identified as a bona fide target for chemosensitisation (Gan et al., 2007; McCarroll et al., 2010). Currently, there is no commercially available TUBB3/III-tubulin inhibitor. Little is known about how βIII- tubulin is regulated, making it difficult to target this protein. Therefore, the third aim of this project is to identify cellular regulators of βIII-tubulin in NSCLC cells. To achieve this goal, an affinity chromatography-based proteomics approach was proposed. The novel TUBB3/βIII-tubulin enhancer, WECC0017371, was used as the foundation for development of a molecular probe to identify the intracellular proteins that it binds to effect TUBB3/βIII-tubulin enhancement.

It is rare that a hit compound obtained from a high throughput screening campaign, such as WECC0017371, is suitable for immediate use as a pharmacological tool. Often extensive structural modifications are required to improve potency, efficacy and selectivity of the active molecule. Hence, prior to developing affinity-chromatography probes for chemical proteomics, chemical modifications of WECC0017371 would need to be performed to establish structure-activity relationships (SAR) and to elucidate moieties that are important for activity (as discussed in section 5.2.1 and 5.2.2). Using

213 iterative SAR assessments, the analogue that demonstrated the most potent TUBB/βIII- tubulin enhancing activity could then be developed into an affinity chromatography probe

(section 5.2.3).

The medicinal chemistry described in this chapter was performed by PhD candidate

Elysha Taylor, under the supervision of Associate Professor Jonathan Morris in the

School of Chemistry at UNSW Australia. They explored the role of various functionality of the hit compound and provided the compounds to the author of this thesis, who was responsible for all the biological work described in this chapter. Based on the data obtained, further modifications of the scaffold were made by Taylor and the new analogues tested by the author. The details of the synthesis of this work will not be described or discussed here as the author of this thesis was not involved in this aspect of the project.

5.2 Results

5.2.1 WECC0017371 and a virtual library screen

WECC0017371, the lead candidate for TUBB3/βIII-tubulin enhancement, consists of an imidazopyridine core with two key functional groups: an amide group at position 6 and a phenyl ring at position 2 (Figure 5.1A). Further, the phenyl ring is functionalised with a methoxy group in the 2’-position, which is referred to as 2’-methoxy (Figure 5.1B).

Having validated the lead compound WECC0017371 as a TUBB3/βIII-tubulin enhancer, a structure-based lead optimisation approach was used to determine what functionalities were essential for activity and with the goal of further enhancing the activity of the scaffold.

214

Figure 5.1 Chemical structure of lead compound WECC0017371

The IUPAC name of WECC0017371 is 2-(2-methoxyphenyl)imidazo[1,2-a]pyridine-6- carboxamide. (A) The lead compound WEC0017371 consists of an imidazopyridine core

(positions labelled with red numbers) with an amide group at position 6 and a phenyl ring at position 2. (B) The methoxy group is attached to the phenyl ring in the 2’-position.

Other potential substitution sites at the phenyl ring are the 3’- and the 4’-position.

215

216

Initially, a virtual library screen of 19 million commercially available chemical compounds was performed against the structure of WECC0017371. Results were ranked based on a structural similarity metric, which refers to the similarity of chemical molecules with respect to structural qualities. Unfortunately, no compounds that share strong structural similarities with WECC0017371 were identified. Therefore, a chemical modification approach was used to establish which of the structural features of

WECC0017371 were responsible for the biological activity. By determining the structure-activity relationship (SAR) of WECC0017371, it is anticipated that new analogues can be developed to evoke TUBB3 gene and βIII-tubulin protein enhancement.

Moreover, this SAR analysis will allow us to determine which part of the molecule can be further functionalised so that a chemical proteomic approach can be used to discover the biological target(s) responsible for modulating TUBB3 gene and βIII-tubulin protein enhancement. Identifying the target(s) is critical as once we understand the mechanism of modulation, it should be possible to design new molecules to inhibit βIII-tubulin in cancer cells.

5.2.2 Structure-activity relationship studies of WECC0017371 analogues as

TUBB3/βIII-tubulin enhancers

In an attempt to improve the TUBB3/βIII-tubulin enhancing ability of WECC0017371, we developed SAR strategy. Figure 5.2 shows the initial SAR strategy for

WECC0017371, which includes variation of functionality at position 6 on the imidazopyridine core, variation of the phenyl ring, as well as modification of the imidazopyridine. The reasons for this are as follows:

1. Variation of functionality at amide of the imidazopyridine scaffold will allow us to

217

Figure 5.2 SAR strategy

WECC0017371 consists of three major structural motifs, an amide group, an imidazopyridine core and a phenyl ring. To explore the SAR, variation of substituents at position 6 of imidazopyridine core was undertaken, variation of the phenyl ring, as well as modification of the imidazopyridine scaffold.

218

219

determine whether the amino group is critical to the activity.

2. A substituent is an atom substituted in place of a hydrogen on a parent hydrocarbon

chain. By varying the phenyl ring, particularly the substituents attached, it will

provide us the structural information about the binding of the scaffold.

3. The variarion of the imidazopyridine scaffold will allow us to determine whether the

nitrogens in the ring system are critical to the activity and if they are, will provide us

with some evidence of how the molecule may bind to its target(s).

Using this strategy, Elysha Taylor synthesised 33 analogues and their structures are displayed in Table 5.1.

In order to quantitatively measure the effect of WECC0017371 and its analogues in enhancing TUBB3 gene and βIII-tubulin protein expression, a screening dose was first determined. Based on RT-PCR and Western blotting results (section 4.2.3), 100 μM of

WECC0017371 induced a 1- and 0.5-fold increase in TUBB3 mRNA and βIII-tubulin protein level relative to control, respectively, in H460 cells after 72 h. While it is well documented that the relationship between TUBB3 gene and βIII-tubulin protein level is not always one-to-one (Hiser et al., 2006) (section 4.3), to achieve upregulation of βIII- tubulin protein, we are seeking an increase at the gene level.

Based on the WECC0017371 data, a screening dose is defined as the minimal dose required to induce a 1-fold increase in TUBB3 mRNA level relative to non-treated H460 controls. Further, in an attempt to minimise cellular toxicity and non-specific biological activities, we aim to develop a WECC0017371 analogue that possesses significantly improved activity at a two-digit μM dose. First, the two primary analogues of

WECC0017371, ENTD014 and ENTD003, were generated for the selection of a screening dose. Like WECC0017371, ENTD014 and ENTD003 share the

220

Table 5.1 Structure and internal names of all 33 WECC0017371 analogues

Compound structure Internal ID Compound structure Internal ID WECC0017371 ENTH22

ENTD015 ENTH023

ENTD013 ENTH036

ENTD003 ENTH037

ENTD052 ENTH024

ENTD053 PH003228

ENTD014 ENTC008

ENTH013 ENTF037

ENTH015 ENTF043

ENTE044 ENTE049

ENTE045 ENTH012

ENTH20 ENTG018

ENTH21 ENTG019

ENTF038 ENTF041

ENTF011 ENTG055

ENTF012 ENTG036

ENTF034a ENTH051

221 imidazopyridine core and the phenyl ring at position 2. However, ENTD014 is missing the amide group at position 6 and ENTD003 is missing the 2’-methoxy moiety on the phenyl group (Table 5.1 and Figure 5.3).

As for the determination of a screening dose, H460 cells were treated with a two-digit dose range of 10, 30, 50, 60 and 80 μM of WECC0017371, ENTD014 or ENTD003 for

72 h. RT-PCR was performed to quantify TUBB3 mRNA levels relative to non-treated controls. The minimal dose of each of the analogues required to evoke a 1-fold increase in TUBB3 mRNA relative to control was 50 μM (Figure 5.3). Therefore, WECC0017371 and the two analogues were tested at 50 μM to establish SAR and to identify the most potent TUBB3/βIII-tubulin enhancer.

Having chosen a screening dose, WECC0017371 analogues were grouped and assessed based on different types of modification. For the ease of data presentation, the chemical structures, IUPAC names and a summary of SAR data of all 33 analogues assessed will be presented in separate tables in the following subsections. The activity was reported showing two parameters: % TUBB3 expression relative to the control and % βIII-tubulin expression relative to the control. While both activities were important, only compounds capable of enhancing βIII-tubulin protein expression were selected for further modification. At the gene level, the cut-off for further evaluation was defined as a change to 2-fold the untreated control. Following each modification, the most potent βIII-tubulin protein enhancer was selected and prioritised for further SAR analysis. Additionally, all drug-treated H460 cells were examined by phase-contrast microscopy for compound precipitation and toxicity issues. At 50 μM, none of the analogues tested display solubility or toxicity issues, with the exception of halo analogues discussed in section 5.2.2.4.

222

Figure 5.3 Determination of a screening dose for WECC0017371 and its analogues

In order to quantitate the effect of WECC0017371 and its analogues in enhancing TUBB3 gene and βIII-tubulin protein expression, a screening dose was determined. H460 cells were treated with 10, 30, 50, 60 and 80 μM of WECC0017371 and its two primary analogues, ENTD014 and ENTD003 for 72 h. RT-PCR was performed to quantify

TUBB3 mRNA level relative to non-treated controls. TUBB3 mRNA levels were normalised to β2M and are shown relative to non-treated controls. mRNA levels were plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05.

223

224

5.2.2.1 2’-Methoxy modifications

As methoxy groups are electron-donating groups that can readily form hydrogen bonds with certain amino acids in protein receptors, such as lysine, serine and threonine, we first explored modification of the 2’-methoxy group. In order to explore the importance of the

2’-methoxy group on TUBB3/βIII-tubulin enhancement activity, we kept the amide group, the imidazopyridine core and the phenyl ring constant and studied five methoxy- modified analogues: ENTD015, ENTD013, ENTD052, ENTD053 and ENTD003 (Figure

5.4A). H460 cells were treated with 50 μM of WECC0017371, ENTD015, ENTD013,

ENTD052, ENTD053 and ENTD003 for 72 h and harvested for RT-PCR and Western blot analysis to quantitate changes in TUBB3 mRNA and βIII-tubulin protein level relative to non-treated controls. Table 5.2 summarises the SAR data for 2’-methoxy modifications.

In H460 cells treated with 50 μM WECC0017371, TUBB3 mRNA expression increased

0.2-fold relative to control, while βIII-tubulin protein expression decreased 0.13-fold.

This slight loss of TUBB3/βIII-tubulin enhancing activity is not unexpected, as the minimal dose required for significant WECC0017371 activity is 100 μM (described in section 4.2.3). Changing the methoxy group from position 2’ to 3’ (ENTD015) and 4’

(ENTD013) did not significantly change the endogenous expression of TUBB3 and βIII- tubulin (Figure 5.4B and C). In contrast, deletion of the 2’-methoxy group (ENTD003) strongly reduces βIII-tubulin expression, indicating the potential necessity of the methoxy group (Figure 5.4B and C). Building on this result, we examined the effect of adding a second methoxy group on the phenyl ring. ENTD052 and ENTD053 were synthesised with 2’,4’- and 2’,5’-dimethoxy substituents on the phenyl ring, respectively. In

ENTD052-treated H460 cells, TUBB3 mRNA and βIII-tubulin protein expression increased significantly by 1.34- (Figure 5.4B; p<0.05) and 0.66-fold 225

Figure 5.4 Effect of 50 μM WECC0017371 and methoxy modified analogues on

TUBB3 mRNA and βIII-tubulin protein expression in H460 cells treated for 72 h.

H460 cells were treated with WECC0017371 and 5 methoxy modified analogues for 72 h to assess their effect on TUBB3 gene and βIII-tubulin protein level, using RT-PCR and

Western blotting, respectively. (A) The five methoxy modifications included structural isomers (ENTD015 and ENTD013); deletion of the 2’-methoxy group (ENTD003); and the introduction of a second methoxy group at position 4’ and 5’ (ENTD052 and

ENTD053). (B) RT-PCR analysis of methoxy-modified analogues revealed the methoxy group is important for compound activity, in an additive manner. (C) Representative

Western blots and semi-quantitation of βIII-tubulin protein expression in H460 cells.

GAPDH was used as a loading control. Data demonstrated that the addition of a second methoxy group significantly increased βIII-tubulin protein expression. βIII-tubulin protein expression was normalised to GAPDH and data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05 and ** indicates p<0.01. NT, non-treated.

226

227

Table 5.2 SAR summary of 2’-methoxy modifications of WECC0017371

% TUBB3 % βIII-tubulin expression expression Compound structure Internal ID IUPAC name relative to the relative to the control (%) control (%)

2-(2-methoxyphenyl)imidazo[1,2-a]pyridine- 120.8 87.64 WECC0017371 6-carboxamide

2-(3-methoxyphenyl)imidazo[1,2-a]pyridine- 122.2 111.0 ENTD015 6-carboxamide

2-(4-methoxyphenyl)imidazo[1,2-a]pyridine- 99.49 101.0 ENTD013 6-carboxamide

2-phenylimidazo[1,2-a]pyridine-6- 204.1 68.57 ENTD003 carboxamide

2-(2,4-dimethoxyphenyl)imidazo[1,2- 234.2 166.4 ENTD052 a]pyridine-6-carboxamide

2-(2,5-dimethoxyphenyl)imidazo[1,2- ENTD053 245.6 177.0 a]pyridine-6-carboxamide

228

(Figure 5.4C; p<0.05) to control, respectively. ENTD053-treated H460 cells showed a significantly increased TUBB3 mRNA and βIII-tubulin protein expression, to 1.46-fold

(Figure 5.4B; p<0.05) and 0.77-fold (Figure 5.4C; p<0.01) relative to non-treated controls, respectively. Taken together, these data suggests that methoxy groups are important for TUBB3/βIII-tubulin enhancing activity, in an additive manner.

5.2.2.2 Modification of major functional groups

There are two major functional groups in WECC0017371: the phenyl ring at C2 and the amide group at C6 of the imidazopyridine scaffold. Successive deletion of these functionalities in WECC0017371 was performed to reveal whether these functional groups contribute to TUBB3/βIII-tubulin enhancing activity (Table 5.3). First, ENTD014 was synthesised to assess the importance of the amido group. In ENTD014-treated H460 cells, TUBB3 mRNA expression was significantly enhanced 1.03-fold (Figure 5.5B; p<0.01) relative to control. This result was translated at the protein level, where βIII- tubulin expression was enhanced 0.22-fold relative to control. The significant enhancement indicates that the amide group is not necessary for activity and accordingly, the rest of the analogues in this set of modifications did not have the amido group.

Next, we assessed the necessity of the 2’-methoxy group. In ENTH013, the 2’-methoxy group was removed, while the imidazopyridine core and the phenyl ring were kept constant (Table 5.3). In ENTH013 treated-H460 cells, the TUBB3 mRNA and βIII-tubulin protein expression was enhanced 0.61-fold (Figure 5.5B, p<0.05) and 0.12 fold relative to control, respectively (Figure 5.5C). Removal of the 2’-methoxy group did not result in a complete loss of TUBB3/βIII-tubulin enhancing activity. The enhanced

229

Table 5.3 SAR summary of modification of major functional groups

% TUBB3 % βIII-tubulin Compound expression expression Internal ID IUPAC name structure relative to the relative to the control (%) control (%)

2-(2-methoxyphenyl)imidazo[1,2-a]pyridine-6- 120.8 87.64 WECC0017371 carboxamide

ENTD014 2-(2-methoxyphenyl)imidazo[1,2-a]pyridine 203.1 122.1

ENTH013 2-phenylimidazo[1,2-a]pyridine 167.1 112.5

ENTH015 imidazo[1,2-a]pyridine 131.3 79.53

230

Figure 5.5 Effect of amide group and phenyl ring deletion on TUBB3 mRNA and

βIII-tubulin protein expression in H460 cells.

There are two major functional groups in WECC0017371: the 2’-methoxy phenyl ring at

C2 and the amide group at C6 of the imidazopyridine scaffold. The importance of these moieties in TUBB3/βIII-tubulin enhancing activity was assessed. (A) WECC0017371 and its deletion analogues ENTD014, ENTH013 and ENTH015. (B) RT-PCR analysis of 50

μM WECC0017371, ENTD014, ENTH013 and ENTH015-treated H460 cells revealed the removal of the amide group improves activity, while the phenyl ring is critical for compound activity. (C) Representative Western blots and semi-quantitation of βIII- tubulin protein expression in H460 cells, following 72 h of WECC0017371, ENTD014,

ENTH013 and ENTH015 treatment are shown. GAPDH was used as a loading control.

βIII-tubulin protein expression was normalised to GAPDH and data are plotted as mean

± SEM of 3 independent experiments. Statistical significance, ** indicates p<0.01. NT, non-treated control.

231

232 activity of ENTH013 when compared to WECC0017371 confirms the earlier observation that the amido group is not critical to the potency. However, as ENTH013 has lower activity than ENTD014, it does suggest that the 2’-methoxy group plays a role in the potency.

To assess the importance of the phenyl ring in TUBB3/βIII-tubulin enhancing activity, imidazopyridine (ENTH015) was synthesised. This can be viewed as WECC0017371 without any side-chain functionality (Figure 5.5A). In ENTH015-treated H460 cells, it was shown that TUBB3 mRNA expression was increased by 0.31-fold and βIII-tubulin protein expression was decreased by 0.20-fold relative to control (Figure 5.5B and C;

Table 5.3). Interestingly, ENTH015’s activity was similar to that of WECC0017371, but it was less potent than ENTD014 and ENTH013. Hence, the phenyl ring at position 2 of the imidazopyridine scaffold can be considered critical for TUBB3/βIII-tubulin enhancing activity. From here, given the significantly improved potency observed with removal of the amido group, ENTD014 was prioritised as the primary structure and we screened a series of analogues with substituents of different electronegativity and size on the phenyl substituent, including methoxy and methyl groups (section 5.2.2.3), as well as halo groups

(section 5.2.2.4).

5.2.2.3 Methoxy- and methyl group substituents

Next, ENTD014, ENTE044 and ENTE045 were synthesised to quantitate the effect of strongly electron-donating methoxy substituents at 2’-, 3’-, and 4’-position on activity

(Table 5.4). RT-PCR analysis showed that ENTD014 significantly enhanced TUBB3 mRNA expression by 1.03-fold (Figure 5.6B; Table 5.4; p<0.05) relative to control. In comparison to ENTD014, ENTE044 and ENTE045 increased TUBB3 mRNA

233

Table 5.4 SAR summary of methoxy and methyl group substituents

% TUBB3 expression % βIII-tubulin Compound structure Internal ID IUPAC name relative to the control expression relative to (%) the control (%)

2-(2-methoxyphenyl)imidazo[1,2- 120.8 87.64 WECC0017371 a]pyridine-6-carboxamide

2-(2-methoxyphenyl)imidazo[1,2- 203.1 122.1 ENTD014 a]pyridine

2-(3-methoxyphenyl)imidazo[1,2- 152.8 115.4 ENTE044 a]pyridine

2-(4-methoxyphenyl)imidazo[1,2- ENTE045 162.2 100.5 a]pyridine

175.1 101.3 ENTH20 2-(o-tolyl)imidazo[1,2-a]pyridine

ENTH21 2-(m-tolyl)imidazo[1,2-a]pyridine 267.3 99.29

ENTH22 2-(p-tolyl)imidazo[1,2-a]pyridine 286.2 107.2

234

Figure 5.6 Effect of methoxy and methyl group substituents on TUBB3 mRNA and

βIII-tubulin protein expression in H460 cells at 72 h.

These SAR data suggests that the electron-donating methoxy group is important for

TUBB3/βIII-tubulin enhancing activity. (A) ENTD014, ENTE044, ENTE045,

ENTH020, ENTH021 and ENTH022 were synthesised to quantitate the change in methoxy- and methyl substitutions on activity. (B) RT-PCR analysis of 50 μM ENTD014,

ENTE044 and ENTE045-treated H460 cells revealed that ENTD014 is most active.

ENTH021 and ENTH022 treatment significantly improved TUBB3 enhancing activity.

To assess whether this effect is translated at the protein level, Western blotting was performed. (C) Representative Western blots and semi-quantitation of βIII-tubulin protein expression in H460 cells, following 72 h of compound treatment are shown.

GAPDH was used as a loading control. βIII-tubulin protein expression was normalised to

GAPDH and data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05, ** indicates p<0.01. NT, non-treated control.

235

236 expression to a lesser extent, by 0.53- and 0.62-fold relative to control, respectively

(Figure 5.6B). At the protein level, ENTD014 and ENTE044-treatment induced 0.22- and

0.15-fold βIII-tubulin enhancement relative to control, respectively, while βIII-tubulin expression in ENTE045-treated H460 cells remained unchanged. This result indicates a potency hierarchy of 2’-methoxy > 3’-methoxy > 4’-methoxy. Therefore, ENTD014 was prioritised as a TUBB3/βIII-tubulin enhancing candidate.

Next, we quantitated the TUBB3/βIII-tubulin enhancing activity of methyl substituents.

RT-PCR analysis showed that ENTH021 (3’-methyl) and ENTH022 (4’-methyl) treatment significantly enhanced TUBB3 mRNA expression by 1.67- (p<0.01) and 1.86- fold (p<0.01) relative to control, respectively (Figure 5.6A; Table 5.4). In ENTH020 (2’- methyl)-treated H460 cells, TUBB3 mRNA expression increased 0.75-fold relative to control, however this was not significant (Figure 5.6A; Table 5.4). At the protein level,

βIII-tubulin expression remained unchanged in ENTH020, ENTH021 and ENTH022- treated H460 cells (Figure 5.6B; Table 5.4). Hence, methyl analogues were not prioritised for further study. These data confirms that having an electron-donating group like a methoxy substituent is critical to activity.

5.2.2.4 Halo group substituents

To probe the effect of the substituent further, it was decided to examine substituents that had different electronic and hydrophobic properties. Halogens are known to be substituents that can have profound effects on activity by changing the electronics of a molecule, as halogens are electronegative. Accordingly, a series of halogen-containing analogues were prepared. We quantitated the TUBB3/βIII-tubulin enhancing activity of the chloro and bromo substituents (Table 5.5). Two chloro analogues, ENTH023 (4’-

237

Table 5.5 SAR summary of halo group substituents

% TUBB3 % βIII-tubulin expression expression Compound structure Internal ID IUPAC name relative to the relative to the control (%) control (%)

ENTH023 2-(4-chlorophenyl)imidazo[1,2-a]pyridine 373.2 110.7

ENTH036 2-(3-chlorophenyl)imidazo[1,2-a]pyridine 235.7 123.6

2-(2,4-dichlorophenyl)imidazo[1,2- ENTH037 231.1 114.3 a]pyridine

2-(3,4-dichlorophenyl)imidazo[1,2- ENTH024 638.2 127.4 a]pyridine

2-(4-bromophenyl)imidazo[1,2-a]pyridine- PH003228 N/A N/A 6-carboxamide

ENTC008 6-bromo-2-phenylimidazo[1,2-a]pyridine N/A N/A

6-bromo-2-(2-methoxyphenyl)imidazo[1,2- ENTF037 N/A N/A a]pyridine

238 chloro), ENTH036 (3’-chloro), and two dichloro analogues, ENTH037 (2’,4’-dichloro) and ENTH024 (3’,4’-dichloro) were synthesised. RT-PCR analysis showed that all of these significantly enhanced TUBB3 mRNA expression in H460 cells, ENTH023 by 2.73 fold (p<0.001), ENTH036 by 1.36 fold (p<0.01), ENTH037 by 1.31 fold (p<0.01) and

ENTH024 by 5.38 fold (p<0.00001) relative to control (Figure 5.7B). Western blotting analysis demonstrated that ENTH036 and ENTH024 treatment significantly enhanced

βIII-tubulin expression by 0.24- (Figure 5.7C; p<0.05) and 0.27-fold (Figure 5.7C; p<0.01) of control, respectively. ENTH023 and ENTH037-treated H460 cells also demonstrated a trend in βIII-tubulin upregulation, although this was not significant.

While these results seems promising, it is possible that the observed activity could be a result of compound toxicity. In order to establish comprehensive structure-activity relationships of WECC0017371 and its analogues, all compounds were tested at 50 μM for 72 h. An ideal candidate should produce significantly improved activity in

TUBB3/βIII-tubulin enhancement, without markedly affecting cell growth, cell proliferation or causing cytotoxicity. Here, ENTD053 was used as an example of one of the compounds. As described in section 5.2.2.1, ENTD053 (50μM) showed potent βIII- tubulin enhancing activity (Figure 5.4B and C). Using phase contrast microscopy, it was demonstrated that the morphology and colony size remained unchanged after 72 h of treatment (Figure 5.8B), compared to naïve H460 cells (Figure 5.8A). Despite the positive impact of chloride substituents on TUBB3/βIII-tubulin enhancing activity, these compounds have precipitation and toxicity issues. Phase contrast microscopy demonstrated that ENTH023 and ENTH037 precipitated readily in media, forming crystalline deposits (Figure 5.8C, D and F). This observation suggests H460 cells were not being treated at 50 μM and thus it is difficult to quantitatively compare their activity

239

Figure 5.7 Effect of halo group substituents on TUBB3 mRNA and βIII-tubulin protein expression in H460 cells at 72 h.

The electron-donating methoxy group appears to be important for TUBB3/βIII-tubulin enhancing activity. Using RT-PCR and Western blotting, we quantified the ability of electron-withdrawing halo group substituents in enhancing TUBB3 gene and βIII-tubulin protein expression in H460 cells. (A) Chloro analogues ENTH023, ENTH036, ENTH037 and ENTH024. (B) RT-PCR analysis of 50 μM ENTH023 and ENTH024-treated H460 cells revealed TUBB3 mRNA expression was significantly enhanced, while ENTH036 and ENTH037 also showed some enhancing effect. To investigate whether this effect is translated at the protein level, Western blotting was performed. (C) Representative

Western blots and semi-quantitation of βIII-tubulin protein expression in H460 cells, following 72 h of ENTH023, ENTH036, ENTH037 and ENTH024 treatment are shown.

GAPDH was used as a loading control. βIII-tubulin protein expression was normalised to

GAPDH and data are plotted as mean ± SEM of 3 independent experiments. Statistical significance, * indicates p<0.05 and ** indicates p<0.01. NT, non-treated control.

240

241

Figure 5.8 Phase contrast images of naïve and chloro analogue-treated H460 cells

WECC0017371 analogues were tested at a fixed dose of 50 μM. Naïve and drug treated-

H460 cells were imaged using phase contrast microscopy to identify drug precipitation and toxicity issues. Here, chloro substituents exhibit solubility and toxicity issues. (A)

Non-treated H460 cells with only few floating dead cells (red arrow). (B) H460 cells’ morphology remained unchanged after ENTD053 treatment compared to control. (C)

Cells treated with 50 μM ENTH023 exhibited lower cell density compared to control. (D)

Magnified image of precipitated ENTH023 in media, as indicated by black arrows pointed to crystalline structure. (E) ENTH036-treated H460 cells showed lower cell density than non-treated controls. (F) H460 cells treated with ENTH037, with black arrows indicating crystalline ENTH037 precipitation in media. (G) ENTH024 treatment resulted in extensive cell death, examples indicated by red arrows. Scale bars-150 μm.

242

A B C D

E F G

243 to other analogues. Compared to naïve H460 cells (Figure 5.8A), ENTH023, ENTH036,

ENTH037-treated H460 cells displayed lower cell density after 72 h of growth (Figure

5.8C, E, F and G). Although this was not confirmed with further experiments, this observation could be the result of repressed cell proliferation, given there were few dead cells. In addition, ENTH024 treatment yielded abundant floating dead cells (Figure 5.8G), suggesting this analogue may be toxic to H460 cells at the concentrations used. Hence, it is unclear whether the βIII-tubulin enhancement observed in drug-treated H460 cells was a result of compound activity or non-specific compensatory effects.

Furthermore, three bromo analogues were synthesised: PH003228, ENTC008 and

ENTF037 (Table 5.5), however, at 50 μM, these analogues were so toxic that no H460 cells survived the treatment. Collectively, because of precipitation and toxicity issues, halogen substituents were not prioritised for further studies of TUBB3 regulation in

NSCLC.

To briefly summarise section 5.2.2.3 and 5.2.2.4, the SAR of the phenyl ring substituents with different electronic properties was explored. Methyl analogues demonstrated significantly improved TUBB3 mRNA enhancing activity, however, the effects were not translated at the protein level. Hence, methyl analogues were not considered for further development. Chloro and di-chloro analogues exhibited promising activity, however, due to the solubility and toxicity issues, these were not prioritised for further study. As for the strongly electron-donating methoxy group, ENTD014 treatment showed the most potent

TUBB3/βIII-tubulin enhancing activity, and therefore, was prioritised as the lead candidate for further SAR studies.

244

5.2.2.5 Imidazopyridine modifications

As detailed in the initial strategy, understanding the role of the nitrogen atoms in the imidazopyridine scaffold is crtitical in understanding how these moelcules may be interacting with their target(s). Variation of the imidazopyridine scaffold is summaried in

Table 5.6. As described above, ENTD014 was used as the primary structure for this series of modifications. Nitrogen atoms have a wide variety of functions in drug molecules, including forming hydrogen bonds with amino and hydroxyl groups of proteins, as well as affecting compound solubility, metabolism and hydrophobicity. In an attempt to investigate whether the two nitrogen atoms in the imidazopyridine core at position 1 and

4 are important for compound activity, Elysha Taylor synthesised a variety of heterocyclic analgoues (Figure 5.9A). In ENTE049 and ENTH012, nitrogen atoms at position 1 and position 4 were removed, respectively. In ENTF043 and ENTF038, an additional nitrogen atom was placed at position 8 or position 5, respectively. ENTE049 treatment resulted in a 0.06- and 0.08-fold increase in TUBB3 mRNA and βIII-tubulin protein expression relative to control, respectively (Figure 5.9B and C). ENTH012 treatment induced a significant 1.57-fold increase in TUBB3 mRNA expression (Figure 5.9B; P<0.01) and a

0.10-fold increase in βIII-tubulin expression, relative to control (Figure 5.9C). In comparison to ENTD014, ENTE049 and ENTH012 showed slightly lower βIII-tubulin enhancing activity. This suggests that nitrogen atoms at position 1 and 4 may play a role in compound activity, however, they are unlikely to be critical, as the removal of each did not result in a significant loss of activity (Table 5.6). On the other hand, the addition of another nitrogen atom in ENTF038 and ENTF043 resulted in a loss of βIII-tubulin enhancing activity (Figure 5.9C). Collectively, ENTD014 demonstrated the highest activity of all imidazopyridine analogues tested and hence still remained as the lead candidate.

245

Table 5.6 SAR summary of imidazopyridine modifications

% TUBB3 % βIII-tubulin expression expression Compound structure Internal ID IUPAC name relative to the relative to the control (%) control (%)

2-(2-methoxyphenyl)imidazo[1,2- ENTD014 203.1 122.1 a]pyridine

2-(2-methoxyphenyl)imidazo[1,2- ENTF043 181.9 80.12 a]pyrimidine

ENTE049 2-(2-methoxyphenyl)indolizine 106.6 108.0

ENTH012 2-(2-methoxyphenyl)-1H-indole 257.0 110.2

2-(2-methoxyphenyl)imidazo[1,2- ENTF038 303.6 96.62 b]pyridazine

246

Figure 5.9 Effect of nitrogen modifications in the imidazopyridine core on TUBB3 mRNA and βIII-tubulin protein expression in H460 cells following treatment for 72 h.

Nitrogen atoms have a wide variety of functions (as discussed in section 5.2.2.5). The importance of the two nitrogen atoms in the imidazopyridine core in TUBB3/βIII-tubulin enhancing activity was examined. H460 cells were treated with 50 μM of nitrogen modified analogues. (A) Nitrogen-modified analogues. (B) RT-PCR analysis of 50 μM

ENTD014, ENTH012, ENTF043 and ENTF038-treated H460 cells revealed significantly enhanced TUBB3 mRNA expression, while TUBB3 mRNA levels remained unchanged in ENTE049 treated H460 cells. To assess whether this effect is translated at the protein level, Western blotting was performed. (C) Representative Western blots and semi- quantitation of βIII-tubulin protein expression in H460 cells, following 72 h of nitrogen- modified analogues treatment are shown. GAPDH was used as a loading control. βIII- tubulin protein expression was normalised to GAPDH and data are plotted as mean ±

SEM of 3 independent experiments. Statistical significance, * indicates p<0.05 and ** indicates p<0.01. NT, non-treated control.

247

248

5.2.2.6 Ether modification and additional methoxy substituents

Based on our SAR data, the 2’-methoxy group appears to be important for TUBB3/βIII- tubulin enhancing activity. As described in section 5.2.2.1, the methoxy group is electron- donating and can readily form hydrogen bonds with hydrogen bond donors, for example amino acids, which may contribute to its activity. Given that it is not known how the molecule binds to the protein target, it was decided to examine variations of the methoxy group. Replacing the 2’-methoxy group with a bulkier ether could allow us to gain insight into the drug’s binding mode and whether our lead molecule potentially binds inside a protein binding pocket. If the methoxy moiety is located inside a putative binding pocket, the addition of a bulkier ether group might result in loss of activity through steric hindrance.

Using ENTD014 as the primary structure, the methyl moiety of the 2’-methoxy group was replaced with two bulky groups, either a benzyl or an iso-propyl group, to give

ENTG018 and ENTG019, respectively (Table 5.7). In ENTG018-treated H460 cells,

TUBB3 mRNA expression increased 0.72-fold relative to control (Figure 5.10B), however this was not translated at the protein level and βIII-tubulin expression decreased

0.15-fold relative to control (Figure 5.10C). ENTG019 treatment showed similar activity to ENTG018, where TUBB3 mRNA expression increased 0.53-fold (Figure 5.10B) and

βIII-tubulin expression decreased 0.26-fold relative to control (Figure 5.10C). These data indicate that addition of bulky groups resulted in a reduction in compound activity, implying binding of 2’-methoxy moiety inside the protein binding pocket and with no ability to tolerate a bulky group.

In section 5.2.2.1, we reported that the 2’,5’-dimethoxy (ENTD052) and 2’,4’-dimethoxy

(ENTD053) analogues provide significantly improved TUBB3/βIII-tubulin

249

Table 5.7 SAR summary of top list compounds

% TUBB3 % βIII-tubulin expression expression Compound structure Internal ID IUPAC name relative to the relative to the control (%) control (%)

ENTD014 2-(2-methoxyphenyl)imidazo[1,2-a]pyridine 203.1 122.1

2-(2-(benzyloxy)phenyl)imidazo[1,2- ENTG018 171.5 84.85 a]pyridine

2-(2-isopropoxyphenyl)imidazo[1,2- ENTG019 152.9 73.60 a]pyridine

2-(2,5-dimethoxyphenyl)imidazo[1,2- ENTF011 356.0 127.1 a]pyridine

2-(2,4-dimethoxyphenyl)imidazo[1,2- ENTF012 397.7 121.9 a]pyridine

3-(imidazo[1,2-a]pyridin-2-yl)-4- ENTF034a 311.0 123.2 methoxyphenol

250

Figure 5.10 Effect of ether moiety modification and methoxy addition on TUBB3 mRNA and βIII-tubulin protein expression in H460 cells treated for 72 h.

Based on SAR data, the 2’-methoxy moiety appears to be important for TUBB3/βIII- tubulin enhancing activity. Using bulky group ether modifications, SAR was performed to gain insight into the drug’s binding mode and orientation of our lead molecule inside the protein binding pocket. (A) ENTD014, ENTG018, ENTG19, ENTF011, ENTF012 and ENTF034a were synthesised and structures indicated. (B) RT-PCR analysis of 50 μM

ENTG018 and ENTG019-treated H460 cells revealed reduced TUBB3 mRNA levels compared to ENTD014-treated H460 cells. ENTF011, ENTF012 and ENTF034a all demonstrated superior activity to ENTD014. To assess whether this effect is translated at the protein level, Western blotting was performed. (C) Representative Western blots and quantitation of βIII-tubulin protein expression in H460 cells, following 72 h of ENTD014,

ENTG018, ENTG19, ENTF011, ENTF012 and ENTF034a treatment are shown.

GAPDH was used as a loading control. βIII-tubulin protein expression was normalised to

GAPDH and data are plotted as mean ± SEM of 3 independent experiment. Statistical significance, *indicates p<0.05, *** indicates p<0.001 and **** indicates p<0.0001. NT, non-treated control.

251

C B

252 enhancing activity. With the anticipation that the same SAR could be translated to

ENTD014, ENTF011 (2’,5’-dimethoxy) and ENTF012 (2’,4’-dimethoxy) were synthesised by adding a second methoxy group at position 5’ and 4’ on the phenyl ring, respectively. ENTF011 and ENTF012 treatment significantly enhanced TUBB3 mRNA expression 2.56- (p<0.0001) and 2.98-fold (p<0.0001) relative to control, respectively

(Figure 5.10B and Table 5.7). This result was translated at the protein level, where βIII- tubulin expression was enhanced 0.27- and 0.22-fold relative to control in ENTF011- and

ENTF012-treated H460 cells (Figure 5.10C and Table 5.7). Collectively, ENTF011’s

βIII-tubulin enhancing activity was superior to ENTD014 and ENTD012 and thus this molecule now represents the lead compound for this program.

5.2.3 Development of a TUBB3/βIII-tubulin affinity chromatography probe and control

The biological data obtained from SAR assessments (section 5.2.2) has given insight into the structural requirements essential for ENTF011’s TUBB3/βIII-tubulin enhancing activity. However, its target protein or proteins and mechanism of action are unknown.

Elucidation of these is critical to better understand how TUBB3 and βIII-tubulin expression is regulated in cancer. To achieve this goal, the use of affinity chromatography-based proteomics was proposed.

Affinity chromatography is a technique used to determine protein binding partners of a small molecule from a cell lysate (Figure 5.11) (reviewed in Rylova et al., 2015). Briefly, affinity chromatography probes are immobilised on a solid phase, such as agarose beads, and exposed to a protein extract, such as a cell lysate, and allowed to bind to their target proteins. Proteins that bind non-specifically to the probe or the

253

Figure 5.11 Affinity chromatography and target identification

Affinity chromatography is a technique to elucidate proteins that a small molecule binds to to exert its acitivty. Briefly, a protein extract is collected from cell lysates and incubated with affinity chromatography probes immobilised on a solid phase, such as agarose beads.

Proteins that bind non-specifically to the probe and the matrix are removed by multiple washing steps. Proteins that bind specifically to the affinity chromatography probe are eluted and collected for identification, using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), tryptic digestion and tandem mass spectrometry (MS/MS) analysis.

254

255 matrix are removed by multiple washing steps prior to release of specifically bound proteins by elution with the bioactive compound. This technique is referred to as a pulldown assay. Target proteins are then identified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), tryptic digestion and tandem mass spectrometry (MS/MS) analysis (reviewed in Rylova et al., 2015).

For use in a pulldown assay, the compound of interest needs to be modified to allow its immobilisation on beads. A linker is often used to create space between compound and the agarose beads, such that target-protein binding is not impaired or at least maximised

(Ziegler et al., 2013). Finn and co-workers have developed a robust methodology to irreversibly couple molecules of interest to agarose beads for affinity chromatography, using a “click” reaction between an azide and an alkyne (Punna et al., 2005). To apply this methodology to the system used in this thesis, an appropriately derivatised alkyne analogue and a complementary azide-linker agarose bead were required (Figure 5.12A).

One of the key considerations when designing these coupling partners is the position in which the linker is attached to the analogue. An ideal site for attachment of the linker is one that tolerates variations without significant loss in activity. In section 5.2.2.2, it was shown that the amido group position on the imidazopyridine core is important for compound activity, and hence, addition of linker at this site was not considered. In section

5.2.2.1, it was established that the addition of an extra methoxy group at C5’ of the phenyl substituents provides a potent molecule. Based on this, it was decided to examine whether functionalisation at this point could be used for probe synthesis. ENTF011 was used to synthesise a precursor probe compound, ENTF034a. ENTF034a, displays a 5’-hydroxyl moiety instead of the 5’-methoxy moiety in ENTF011. To assess whether ENTF034a retains activity, RT-PCR and Western blotting were performed to quantitate TUBB3 and

βIII-tubulin expression. In ENTF034a-treated H460 cells, TUBB3 mRNA expression 256 significantly increased 2.11 fold (p<0.001), while βIII-tubulin protein expression increased 0.23 fold (Figure 5.10B and Table 5.7), relative to control. In summary,

ENTF034a retains activity, similar to that of ENTD014, ENTF011 and ENTF012 and hence, was used for the development of affinity chromatography probes.

Therefore, attachment of the linker on the 5’-hydroxyl group was considered to be the most promising approach. An alkyne tail was attached to the hydroxyl group in

ENTF034a to yield ENTF041 (Figure 5.12B). Figure 5.13B shows that ENTF041 treatment significantly increased TUBB3 mRNA expression by 3.44-fold relative to control, indicating that TUBB3 mRNA enhancing activity was retained after the addition of an alkyne tail to the hydroxyl group (Table 5.8; Figure 5.13B). This observation was translated at the protein level, where βIII-tubulin protein expression was enhanced 0.19- fold relative to control in ENTF041-treated H460 cells (Table 5.8; Figure 5.13C).

Compared to ENTF034a, there was a 3.9% drop in βIII-tubulin enhancing activity in

ENTF041-treated H460 cells, however, this difference is negligible. Therefore, ENTF041 was used as a precursor for the development of an affinity chromatography probe.

To distinguish specific binding proteins from non-specific ones in an affinity-based pulldown assay, control experiments are necessary. In theory, the ideal negative control beads would be an inactive compound with the same chemical structure, hydrophobicity, size and charge as the active probe, but without the active site (reviewed in Rylova et al.,

2015). Figure 5.14 illustrates the ideal design of a negative control bead. Proteins that bind to negative control beads only or to both negative control and active probes are classified as non-specific binders. Due to the functional

257

Figure 5.12 Illustration of click reaction and addition of an alkyne tail to lead molecule

Based on SAR studies, ENTF034a was chosen for the development of an affinity chromatography probe for the use in a pulldown assay. (A) A robust methodology using a “click” reaction was developed by Finn and co-workers (Punna et al., 2005), which allows irreversible coupling of an active compound to a terminal azide linker-agarose bead, to form an affinity chromatography probe. (B) To apply this methodology to the system used in this thesis, an appropriately derivatised alkyne analogue was required.

Therefore, a terminal alkyne tail (blue) was added to the 5’-hydorxyl group on ENTF034a

(purple), to yield ENTF041.

258

259

Table 5.8 SAR summary of precursors of affinity chromatography probe and negative bead control

% TUBB3 % βIII- expression tubulin Compound structure Internal ID IUPAC name relative to the expression control (%) relative to the control (%)

O 2-(2-methoxy-5-(prop-2-yn-1- N ENTF041 443.9 119.3 yloxy)phenyl)imidazo[1,2-a]pyridine N MeO tert-butyl (2-(2-(2-(4-((3-(imidazo[1,2-

a]pyridin-2-yl)-4- 613.7 138.6 ENTG036 methoxyphenoxy)methyl)-1H-1,2,3-

triazol-1-

yl)ethoxy)ethoxy)ethyl)carbamate

tert-butyl (2-(2-(2-(4-(phenoxymethyl)- 107.3 91.35 ENTH051 1H-1,2,3-triazol-1-

yl)ethoxy)ethoxy)ethyl)carbamate

ENTG055 (prop-2-yn-1-yloxy)benzene 115.6 111.0

260

Figure 5.13 Effect of precursors of affinity chromatography probe and negative bead control on TUBB3 mRNA and βIII-tubulin protein expression in H460 cells.

For the purpose of an affinity-based pulldown assay, candidates of chromatography probe and negative control were synthesised with a dummy polyethyleneglycol linker azide, to test their biological activity. (A) ENTF041 and ENTG036 were synthesised as precursors of affinity chromatography probes. ENTG055 and ENTH051 are precursors of negative controls. Naïve and drug-treated H460 cells were harvested and RT-PCR and Western blotting were performed to quantitate TUBB3 and βIII-tubulin expression. (B) RT-PCR analysis demonstrated that ENTF041 and ENTG036 maintained significant TUBB3 mRNA enhancing activity. (C) Representative Western blots and semi-quantitation of

βIII-tubulin protein expression in H460 cells, following 72 h of treatment are shown.

GAPDH was used as a loading control. βIII-tubulin protein expression was normalised to

GAPDH and data are plotted as mean ± SEM of 3 independent experiment. Statistical significance, ** indicates p<0.01 and *** indicates p<0.001. NT, non-treated.

261

C B

262

Figure 5.14 Schematic of an affinity chromatography probe and a negative control bead

Affinity chromatography is a technique used to determine the protein binding partners of a small molecule from a cell extract. To distinguish specific binding proteins from non- specific ones in an affinity-based pulldown, control experiments are necessary. An affinity chromatography probe consists of an active compound (purple cross) and a linker

(black line), immobilised on an agarose bead (round green ball). The ideal negative control beads would be equipped with the same linker and beads as the active probe, but contain an inactive compound with similar chemical structure, hydrophobicity, size and charge as the active probe, but without active moieties.

263

Agarose beads Agarose beads

264 importance of the imidazopyridine core and the 2’-methoxy group, both of these moieties were removed from the proposed negative control bead, ENTG055, leaving a phenyl ring with an alkyne group (Table 5.8 and Figure 5.13A). RT-PCR and Western blotting analysis were performed to quantitate TUBB3 mRNA and βIII-tubulin protein expression in ENTG055-treated H460 cells (Table 5.8). After ENTG055-treatment, TUBB3 mRNA and βIII-tubulin expression increased marginally, by 0.16-fold and 0.11-fold relative to control, respectively (Table 5.8 and Figure 5.13B and C). This result indicates ENTG055 has negligible TUBB3/βIII-tubulin enhancing activity and hence was chosen for the development of negative control beads.

In order to synthesise affinity chromatography probes and negative control beads, alkyne analogues of the lead compound and negative control compound need to be linked to agarose beads decorated with a linker and a terminal azide. However, as synthesis of agarose beads with linker and a terminal azide is an expensive process, “dummy” or preliminary versions of affinity chromatography probe and negative bead control were synthesised. These contain the linker with the terminal azide but not the agarose beads, to investigate whether the addition of the linker would impair drug-target interactions.

Alkynes ENTF041 and ENTG055 derivatised with a polyethyleneglycol (PEG2) linker azide were synthesised and resulted in a dummy affinity chromatograph probe, ENTG036

(Figure 5.15A) and a dummy negative control, ENTH051 (Figure 5.15B). H460 cells were treated with 50 μM of ENTG036 and ENTH051 for 72 h. RT- PCR and Western blotting were carried to quantitate TUBB3 mRNA and βIII-tubulin expression, respectively. Figure 5.13B and C showed that TUBB3 mRNA and βIII-tubulin protein expression were enhanced significantly 5.14-fold (p<0.001; Table 5.8) and 0.39-fold

(p<0.05; Table 5.8) in ENTG036-treated H460 cells, relative to control. This confirms that ENTG036 maintained good activity and the PEG2 linker did not 265

Figure 5.15 Coupling of ENTF041 and ENTG055 to dummy PEG2 linkers via click reactions

Due to the expense of synthesising terminal azide linker-agarose beads, a dummy affinity chromatography probe and a dummy negative control were synthesised. They were used to investigate whether the addition of a linker would impair drug-target interactions. (A)

Following Finn and co-worker’s published methodology (2005), a dummy hydrophilic

PEG2 linker was coupled to ENTF041 via a click reaction. The terminal alkyne tail on

ENTF041 formed a 1,2,3-trizaole with the complementary terminal azide group on the

PEG2 linker, yielding ENTG036. The amine group on the linker was protected with a

Boc group to prevent side reactions. (B) A similar procedure was performed with

ENTG055 to yield ENTH051.

266

267 impair drug-target binding. In ENTH051-treated H460 cells, TUBB3 mRNA and βIII- tubulin expression were similar to non-treated control, suggesting the PEG2 linker moiety is suitable for use as a negative control candidate. Therefore, ENTG036 and ENTH051 were chosen as the candidate for the affinity chromatography probe and negative control bead, respectively.

The isotype specificity of these candidates for βIII-tubulin were assessed. H460 cells were treated with 50 μM of ENTH041, ENTG036 and ENTH051 for 72 h. Western blotting and densitometry analysis were performed to quantitate the expression of different β- tubulin isotypes. After 72 h of treatment, ENTH041 significantly enhanced βIII-tubulin expression 0.31-fold (p<0.05; Figure 5.16) relative to control. No change in other β- tubulin isotypes was observed. This suggests that the effect of ENTH041 on βIII-tubulin is specific to this β-tubulin isotype at 72 h. In ENTG036-treated H460 cells, the βIII- tubulin protein level was enhanced significantly 0.49-fold (p<0.001; Figure 5.16) relative to control, while βI-, βIV- and total β-tubulin protein expression remained unchanged.

ENTG036 elicited a 0.19-fold increase in βII-tubulin (Figure 5.16) relative to control.

This is similar to the collateral increase in βII- and βIII-tubulin expression observed in

WECC0017371-treated H460 and H1299 cells (described in chapter 4). In ENTH051- treated H460 cells, the expression of all tested β-tubulin isotypes remained unchanged

(Figure 5.16), suggesting ENTH051 is a suitable negative control.

Having chosen the ideal candidate for affinity chromatography probes and negative control beads, next we assessed the linker moiety. The length and hydrophobicity of the linker are critical to the success of an affinity chromatography probe (reviewed in Rylova et al., 2015). If the linker molecule is too short, drug-target interactions could be compromised as a result of steric hindrance of the agarose bead. If the linker is too

268

Figure 5.16 The effect of dummy affinity chromatography probes and dummy negative controls on β-tubulin protein expression in H460 cells at 72 h.

The effect and specificity of ENTH041, ENTG036, ENTH051 on protein expression of different β-tubulin isotypes. H460 cells were treated with 50 μM of ENTH041,

ENTG036, ENTH051 for 72 h prior to Western blotting. Representative Western blots of different β-tubulin isotypes protein expression levels in H460 cells following treatment with ENTH041, ENTG036, ENTH051 are shown. GAPDH was used as a loading control.

Corresponding densitometry analysis and semi-quantitation of βI-, βII-, βIII-, βIV- and total β-tubulin protein expression in H460 cells following ENTH041, ENTG036,

ENTH051 treatment for 72 h. NT, non-treated. Tubulin protein expression was normalised to GAPDH control and plotted as mean ± SEM of 3 independent experiments.

Statistical significance, * indicates p<0.05 and *** indicates p<0.001.

269

131-tubulin 13'11 -tubulin GAPDH GAPDH e ~ cio- 0~ 'iii~ =~ =i ~ ... ! " ~ ... .,_" .. .._e--g 1 " .!:al . s~ 'E $ 100 e c 0~ ~0 :S.o c c c c =,-0 :o, .. D" ..,., " > ~-; ~~ c.- .,. ='" ~ ~ ... ~"' ~... cc.e t-"'"' ~(j' ~"> ~"""' ~ '<;~ ~ ~ 50j!M 50f1M

13111-tubulin 131V-tubulin GAPDH GAPDH g ~200 § ~ ·c.; 411 en"'.,_ <>13 ll~ ... u i~ 150 e--o 1 .... "' ..cz "i :T" ,- ~:; 100 !:: ... c a.o ea.o e: c c c c =o..,,_ .. 50 ..,.,3! ..,_E~ a> - .. 0 ~i CC.. !!! ... =--e ~.... ,. r:{l' ~~~ +"""" '<"' ~(j ~"" ~">" ' '<;~ "'~ ~ SO.,M

270

Figure 5.17 Synthesis of affinity chromatography probes and negative control beads using PEG4 and C5 linkers

The length and hydrophobicity of the linker are critical to the success of an affinity chromatography probe. A hydrophobic terminal azide linker-agarose bead, PEG4 and a hydrophilic terminal azide linker-agarose bead, C5, were chosen. (A) Using Finn and co- worker’s protocol (2005), the required terminal azide-linker agarose beads were synthesised. The round green balls represent agarose beads. Via a “click” reaction, PEG4- or C5- terminal azide linker-agarose beads formed a 1,2,3-triazole with the terminal alkyne on ENTF041, yielding two affinity chromatography probes. (B) The same procedure was performed using ENTG055 to form negative control beads. K2CO3, potassium carbonate; CuI, copper iodide (catalyst); DIPEA, di-isopropyl ethylamine;

TBTA, tris(benzyltriazolylmethyl)amine; DMF, dimethylformamide (solvent).

271

272 hydrophobic, the linker may aggregate, particularly in aqueous solution, which leads to clustering of the probe and interferes with drug-target binding. To avoid potential problems with linkers, two terminal azide-linker agarose beads, PEG4 and C5, were included in the affinity-based pulldown assay (Figure 5.17). The PEG4-linker consists of four repeating units of ethylene glycol (CH2CH2O). It is stable and its hydrophilicity will prevent aggregation in aqueous solution. C5 consists of four CH2 units and was chosen as a hydrophobic linker. Using Finn and co-workers’ protocol (Punna et al., 2005), the required agarose beads with a terminal azide-linker were synthesised by Elysha Taylor

(Figure 5.17). The synthetic schemes for the affinity chromatography probes and negative control beads are shown in Figure 5.17.

Establishing structure-activity relationships was a time-consuming process, and unfortunately, I was unable to use the affinity chromatography probes to complete affinity-based pulldown assays within the time frame of this PhD course. However, an affinity chromatography-based proteomics study and the subsequent identification of molecular binding partners of the probe compound is now underway.

273

5.3 Discussion

The third aim of this project was to identify cellular factors that regulate βIII-tubulin expression in NSCLC cells. To achieve this goal, WECC0017371 was developed into an affinity chromatography probe for future affinity-based pulldown assays. This may allow the identification proteins that WECC0017371 binds, to exert TUBB3/βIII-tubulin enhancement. To develop the research tools for affinity-based pulldowns, chemical modifications of WECC0017371 were performed to establish structure-activity relationships (SAR) and further enhance its potency, efficacy and selectivity. Briefly, using successive deletions, it was identified that the amido group at position 6 is not essential for the compound to exert its effect on βIII-tubulin levels. The phenyl ring at position 2 was found to be critical for TUBB3/βIII-tubulin enhancing activity and thus, a

SAR of substituents of different electronic nature was explored on the phenyl ring moiety.

Two electron-donating functional groups, methoxy and methyl groups, with and without the ability to form hydrogen bonds with proteins were tested. The weakly electron- donating methyl group did not have a positive impact on βIII-tubulin enhancing activity and was eliminated from further investigation. In contrast, the electron-donating methoxy group was found to be important for TUBB3/βIII-tubulin enhancing activity.

Interestingly, doubling the number of methoxy group on the phenyl ring further boosted the magnitude of TUBB3/βIII-tubulin enhancing activity, with the 2’,5’-dimethoxy substitution providing activity superior to that of the 2’,4’-dimethoxy substitution. One could speculate that this may be the result of the added capacity for hydrogen bond formation or the increased electron-density donated to the phenyl ring or a combination of both. Identification of the molecule’s protein target and subsequent docking studies may give insight into the contribution of each of these factors.

274

Halogens are known to be substituents that can have profound effects on activity by changing the electronics of a molecule as halogens are electronegative. Accordingly, a series of halogen-containing analogues were prepared. All chloro analogues tested demonstrated superior TUBB3/βIII-tubulin enhancing activity. While this result looks promising, all chloro analogues tested exhibited toxicity and/or solubility issues.

Therefore, it is difficult to decipher whether the enhanced TUBB3 mRNA and βIII-tubulin expression was a result of compound activity or an increase in the total β-tubulin pool in response to toxicity. Future quantitative cell viability assays are needed to clarify this observation and solubility issues need to be addressed by additional medicinal chemistry effort. Due to the need for further optimisation, these analogues were not prioritised for the development of affinity chromatography probes.

As part of the SAR assessments, attempts at exploring the drug binding mode of the lead, potentially inside a protein binding pocket, were made. The addition of bulky ether groups resulted in significant loss in compound activity, suggesting the 2’-methoxy moiety binds inside a putative protein binding pocket without the capability to tolerate a bulky group.

It was concluded that the 2’-methoxy moiety is critical for small moelcule-protein interactions and no functional group or linker molecule would be added at this site.

After several rounds of SAR studies, the potency of the original WECC0017371 was improved from 100 μM to 50 μM in H460 cells. ENTF011 was selected for the development of an affinity chromatography probe, as it displayed the greatest

TUBB3/βIII-tubulin enhancing activity of all analogues developed as part of this thesis and a suitable attachment site for an alkyne group at position 5’ of the phenyl ring was identified. The alkyne analogue ENTF041 maintained a similar level of potency and efficacy to that of ENTF011. Western blot results indicated that ENTF041’s activity was

275 specific to βIII-tubulin, as it did not markedly affect other β-tubulin isotypes tested. For these reasons, it was considered that ENTF041 was an excellent candidate to develop an affinity chromatography probe for future affinity-based pulldown assays. Additionally, to identify non-specific protein binding in the assay, the alkyne analogue ENTG055 was synthesised based on our SAR data and chosen as the negative control compound.

Linker length and hydrophobicity are critical to the success of an affinity chromatography probe. To avoid potential problems, two different linkers were utilised, a hydrophilic

PEG4 linker and a short hydrophobic C5 linker. Using click chemistry, these linkers were attached to ENTF041 and ENTG055 to synthesise affinity chromatography probes and negative control beads. Prior to conducting the final pulldown assay, a pilot study confirmed that the addition of a dummy PEG2 linker to ENTF041 (yielding ENTG036) did not impair compound activity. This suggests that the addition of a linker molecule to the alkyne moiety of ENTF041 does not interfere with drug-protein binding and we have chosen an ideal position to attach linker molecules.

Interestingly, the TUBB3/βIII-tubulin enhancing activity of ENTG036 was superior to that of ENTF041. Although this is unlikely, it is possible that the PEG linker contributed to the increased compound activity. Furthermore, H460 cells treated with ENTH051

(negative control compound ENTG055 reacted with PEG2 linker) exhibited βIII-tubulin expression levels similar to that of naïve H460 cells. Hence, it is unlikely that PEG was responsible for the superior TUBB3/βIII-tubulin enhancing activity observed in

ENTG036. However, it cannot be ruled out that the addition of a PEG2 linker to ENTF041 may enhance drug-protein binding efficiency without performing further SAR studies.

Since negative bead controls allow the identification and elimination of proteins that bind

276 non-specifically to agarose beads, linkers and ENTG055, this issue was not explored further as part of this thesis.

Identification of molecular binding partners of the probes designed in this chapter using affinity chromatography-based proteomics and mass spectrometry are currently underway. Unfortunately, due to time constraints, the data was unavailable at the time of this thesis write-up. Instead, in parallel with the SAR and affinity probe generation, a microarray and bioinformatics approach was used to gain potential insight into regulatory mechanisms responsible for induced TUBB3 mRNA expression. The resulting data together with future chemical proteomics data will provide a focused selection of pathways and cellular factors that may be important in TUBB3/βIII-tubulin regulation in

NSCLC cells.

277

Chapter 6

Gene expression changes following treatment with the

TUBB3-enhancing WECC0017371 small molecule

6.1 Introduction

TUBB3 mRNA expression is absent in normal human lung (Leandro-García et al., 2010).

In lung tumours, however, it is aberrantly expressed and is known to play a key role in drug resistance and tumour aggressiveness (reviewed in Karki et al., 2013). In addition, emerging evidence suggests that βIII-tubulin is also involved in the tumourigenic and metastatic potential of tumours (McCarroll et al., 2015a; McCarroll et al., 2010).

However, how increased levels of βIII-tubulin exerts its effect on tumour pathobiology is poorly understood. Investigation of global gene changes in TUBB3 enhancing conditions has enormous potential to identify regulatory pathways underlying βIII-tubulin-mediated pathobiology in NSCLC. To address this, a microarray study was employed to establish a comprehensive temporal gene expression profile for naïve and WECC0017371-treated

H460 cells. The resulting data were analysed to detect coordinated changes in functionally related genes that may provide information on relevant biological pathways. In addition, the regulation of aberrant TUBB3 expression in cancer is poorly understood. The temporal gene changes following treatment with the TUBB3-enhancing WECC0017371 small molecule may provide insight into the molecular mechanism underlying TUBB3 aberrant expression.

278

6.2 Results

6.2.1 Microarray analysis of H460 cells treated with WECC0017371

H460 NSCLC cells, which aberrantly express TUBB3 mRNA expression, were treated with WECC0017371 as a small molecule mimetic to further enhance this phenotype.

Genes of interest in this study are those which have statistically differentially expressed gene expression profiles, either upregulated or downregulated, in comparison to naïve

H460 cells. WECC0017371 was chosen over the SAR-improved analogue, ENTF041, as the biological activity of WECC0017371 has been more thoroughly evaluated (section

4.2.3). Briefly, H460 cells and cells treated with 100 μM of WECC0017371 were harvested at 48 and 72 h after treatment. In an attempt to evaluate differential gene expression in these cells, RNA was isolated. Prior to microarray analysis, integrity of the

RNA was assessed using the Agilent Bioanalyzer RNA Nano 6000. Next, a microarray analysis was performed using the Illumina Human HT 12 Bead Chip Array to simultaneously measure the expression of 25,130 genes. After background correction and normalisation, a moderate t-statistic was applied. Additionally, multiple testing correction, Benjamini Hochberg, was performed to account for false positive discovery due to the large number of genes testsed (Benjamini and Hochberg, 1995). An adjusted p-value <0.01 from Benjamini test-correction indicates statistical significance.

6.2.2 Identification of differentially expressed genes and pathway analyses of naïve and WECC0017371-treated H460 cells at 48 h

In order to identify differentially expressed genes between naïve and WECC0017371- treated H460 cells at 48 and 72 h, a Bayesian analysis of gene expression levels was used.

This includes the determination of the fold change between naïve and WECC0017371- 279 treated samples, as well as multiple testing correction (Benjamini and Hochberg, 1995).

In WECC0017371-treated H460 cells, TUBB3 mRNA expression was upregulated 0.30- fold (adjusted p-value= 0.336) and 1.51-fold (adjusted p-value=0.0001) compared to naïve H460 cells, at 48 and 72 h, respectively. This WECC0017371-induced time- dependent TUBB3 mRNA enhancement was in agreement with our observation in section

4.2.3, whereby it was shown that WECC0017371 increased TUBB3 mRNA levels in a time-dependent manner. Further, to determine a significance cut-off for differentially expressed genes, we considered using both statistical significance (adjusted p-value

<0.01) and biological significance (fold change ≥ |2|). However, an arbitrary fold change

(FC) cut-off value of 2 would require excluding the TUBB3 mRNA data from the 48 h dataset. Since our goal is to establish a comprehensive temporal gene expression profile for naïve and WECC0017371-treated H460 cells, it is important to include data obtained from both 48 and 72 h. Hence, a cut-off of FC ≥ |1.3| was used, as determined by TUBB3 mRNA fold change between naive and WECC0017371-treated H460 cells at 48 h.

Sixteen hundred and one (1601) genes were significantly differentially expressed between naïve and WECC0017371-treated H460 cells at 48 h. Among these, 757 genes were upregulated and 844 genes were downregulated in WECC0017371-treated H460 cells.

Appendix IV shows a list of the top 100 genes that changed in expression in response to

WECC0017371 treatment, with a significance threshold of FC ≥ |1.3|. A summary of the top 20 upregulated and downregulated genes, as determined by fold change between treated and non-treated controls, are listed in Table 6.1. None of these differentially expressed genes have been previously reported to regulate TUBB3 mRNA expression and vice versa. The gene expression of TUBB3 and these differentially expressed genes may be either directly or indirectly modulated by WECC0017371.

280

Table 6.1 Top 40 differentially expressed genes in WECC0017371-treated H460 cells

at 48 h

Fold Adjusted Symbol Gene name change p-value

LINC01239 long intergenic non-protein coding RNA 1239 2.819 6.115E-01

ESM1 endothelial cell-specific molecule 1 2.725 1.643E-02

CXCL8 chemokine (C-X-C motif) ligand 8 2.673 2.086E-03

GJB2 gap junction protein, beta 2, 26kDa 2.474 8.770E-03

FOXQ1 forkhead box Q1 2.194 1.402E-04

NR4A2 nuclear receptor subfamily 4, group A, member 2 2.115 4.197E-03

DHRS2 dehydrogenase/reductase (SDR family) member 2 2.108 8.891E-02

LOC100506123 uncharacterized LOC100506123 2.046 7.388E-02

cytochrome P450, family 4, subfamily V, CYP4V2 1.965 2.243E-02 polypeptide 2

ZNF789 zinc finger protein 789 1.957 1.643E-02

SNURF SNRPN upstream reading frame 1.924 1.119E-01

FZD4 frizzled class receptor 4 1.910 2.086E-03

KLF2 Kruppel-like factor 2 1.864 2.624E-03

NPHP4 nephronophthisis 4 1.802 2.772E-03

STOX1 storkhead box 1 1.794 2.518E-01

SNORA68 small nucleolar RNA, H/ACA box 68 1.789 1.105E-01

HIP1R huntingtin interacting protein 1 related 1.783 1.376E-01

olfactory receptor, family 4, subfamily A, member OR4A16 1.780 7.735E-02 16

TMEM154 transmembrane protein 154 1.757 1.944E-01

NAV2 neuron navigator 2 1.755 3.205E-02

CLCF1 cardiotrophin-like cytokine factor 1 -1.962 5.097E-03

ZFP64 ZFP64 zinc finger protein -2.008 7.735E-02

281

COL16A1 collagen, type XVI, alpha 1 -2.016 1.789E-02

GJA5 gap junction protein, alpha 5, 40kDa -2.031 1.364E-02

TERC telomerase RNA component -2.039 3.747E-02

RNU5A-1 RNA, U5A small nuclear 1 -2.119 2.243E-02

C1QTNF6 C1q and tumor necrosis factor related protein 6 -2.217 2.772E-03

ATP-binding cassette, sub-family A (ABC1), ABCA1 -2.229 5.408E-04 member 1

CYSRT1 cysteine-rich tail protein 1 -2.288 5.097E-03

CD14 CD14 molecule -2.311 7.234E-04

C5orf46 chromosome 5 open reading frame 46 -2.319 4.089E-03

PLAC8 placenta-specific 8 -2.362 3.286E-03

KRT81 keratin 81, type II -2.426 1.402E-04

ANKRD1 ankyrin repeat domain 1 (cardiac muscle) -2.432 3.011E-03

OTOF otoferlin -2.501 2.772E-03

KRT80 keratin 80, type II -2.577 9.678E-03

SOST sclerostin -2.637 3.220E-03

S100A3 S100 calcium binding protein A3 -2.874 3.289E-03

IGFBP3 insulin-like growth factor binding protein 3 -3.262 1.839E-04

SNORA12 small nucleolar RNA, H/ACA box 12 -3.445 1.300E-03

282

With the intention to better understand the biological significance behind these 1601 differentially expressed genes, KEGG pathway mapping was performed to identify molecular pathways that are enriched in these differentially expressed genes. These pathways may be important for TUBB3 mRNA enhancement and/or βIII-tubulin- mediated chemosensitivity in NSCLC. The functional annotation tool, Database for

Annotation, Visualisation and Integrated Discovery (DAVID) was used for KEGG pathway mapping. It was revealed that there were no significantly enriched functionally- related gene groups or molecular pathways at 48 h (adjusted p-value >0.01).

Gene Set Enrichment Analysis (GSEA) is a computational technique that determines whether an a piori set of genes shows statistically significant differences between two biological samples. In addition to Gene Ontology (Go) term and KEGG pathway annotation, GSEA was performed to deduce pathways that may be responsible for

WECC0017371-enhanced TUBB3 mRNA expression in H460 cells. Using GSEA, lists of differentially expressed genes were compared to experimentally determined and computationally curated published sets of genes in the GSEA database. This analysis allows the identification of sets of genes that are statistically over-represented. The GSEA analyses were performed using curated gene sets from four major collections within the molecular signatures database (MSigDB), the C2, C4, C5 and C6 collection. The C2 collection comprises curated gene sets collected from multiple sources such as online pathway databases, published databases and compiled knowledge from experts in the field. The C4 collection includes computational gene sets defined by mining large collections of cancer-oriented microarray data. The C5 collection consists of genes that are annotated by GO term. The C6 collection contains gene sets representing signatures of cellular pathways that are often dysregulated in cancer. In order to minimise the discovery of false enriched gene sets, a q-value threshold of 0.1 for multiple testing 283 correction was applied to all GSEA analyses. A q-value of 0.1 implies that for every 100 significantly enriched gene sets identified, there is a probability that 10 of them are false positives. The q-value was chosen over adjusted p-value for GSEA analyses as it is takes into account the false discovery rate in microarray analysis, making it a more stringent test for statistical significance.

The C2 collection consists of 4725 curated gene sets. These gene sets were filtered to assess only gene sets which have more than 15 genes or fewer than 500 genes to ensure a more accurate enrichment score. Hence, 383 of 4725 gene sets were used in the analysis.

The differentially expressed genes analysed against C2 collection indicated that none of the gene sets were significantly enriched between naïve and WECC0017371-treated

H460 cells at 48 h. As for C4, C5 and C6 collections, no significant gene set enrichment was detected between naïve and WECC0017371-treated H460 cells. Hence, the GSEA result is concordant with the GO term and KEGG pathway results.

6.2.3 Identification of differentially expressed genes and pathway analyses of naïve and WECC0017371-treated H460 cells at 72 h

The bioinformatics analyses carried out for microarray data collected from naïve and

WECC0017371-treated H460 cells at 72 h were similar to those conducted for 48 h

(section 6.2.2). It was revealed that 3332 genes were differentially expressed in

WECC0017371-treated H460 cells when compared to naïve H460 cells at 72 h. Among these differentially expressed genes, 1785 genes were found to be upregulated, while

1547 genes were downregulated. None of these genes has been previously reported to play a role in TUBB3 regulation and vice versa. Appendix V shows a list of top 100 genes that changed in expression in response to WECC0017371 treatment, with a threshold of

284

FC ≥ |1.3|. The top 20 of each upregulated and downregulated genes, as determined by fold change between naïve and WECC0017371-treated H460 cells are listed in Table 6.2.

In order to better understand the biological significance of these 3332 differentially expressed genes, KEGG pathway mapping was used to identify pathways that are enriched in functionally-related differentially expressed genes. It was revealed that

WECC0017371 exposure was associated with significant expression changes in genes involved in the p53 signalling pathway (adjusted p-value<0.00001) and cell cycle

(adjusted p-value <0.01).

Briefly, 33 of 61 p53 signalling pathway genes were differentially expressed in

WECC0017371-treated H460 cells (Figure 6.1). Table 6.3 and Figure 6.2 summarise all genes that were differentially expressed in the p53 signalling pathway. Among them, 13 genes were upregulated and 20 genes were downregulated. Cell cycle arrest genes

(CHK2, TP53, CDKN1A, GADD45), apoptosis genes (IGFBP3, BAX, NOXA, P53AIP,

TNFRSF10B, CYCS, ZMAT3 and SIAH1) and DNA damage repair genes (GADD45,

SESN2, RRM2B) were downregulated (Figure 6.2). Cell cycle genes essential for mitosis were upregulated (CCNB1 and CDK1). Genes that were upregulated from the p53 signalling pathway were primarily cell cycle genes and represent an overlap with the second major pathway identified from the KEGG analysis: cell cycle.

Within the cell cycle pathway, 43 of 82 genes were differentially expressed in

WECC0017371-treated H460 cells compared to naïve H460 cells (Figure 6.3). Table 6.4 and Figure 6.4 summarise all genes that were differentially expressed in the cell cycle pathway. Among these, 16 genes were downregulated and 27 genes were

285

Table 6.2 Top 40 differentially expressed genes in WECC0017371-treated H460 cells at 72 h

Fold Adjusted Symbol Gene name change p-value

INSIG1 insulin induced gene 1 3.807 3.10E-06

HSPA1A heat shock 70kDa protein 1A 3.525 4.98E-05

MSMO1 methylsterol monooxygenase 1 3.381 2.95E-05

NPTX2 neuronal pentraxin II 3.282 2.24E-05

HSPA1B heat shock 70kDa protein 1B 3.191 5.37E-06

3-hydroxy-3-methylglutaryl-CoA synthase HMGCS1 3.042 1.82E-06 1 (soluble)

HSPA8 heat shock 70kDa protein 8 3.001 3.10E-06

phospholysine phosphohistidine inorganic LHPP 2.910 7.17E-04 pyrophosphate phosphatase

ESM1 endothelial cell-specific molecule 1 2.796 1.78E-03

potassium channel, voltage gated modifier KCNF1 2.764 6.79E-05 subfamily F, member 1

RHOU ras homolog family member U 2.736 9.16E-06

C10orf90 chromosome 10 open reading frame 90 2.573 6.88E-04

TUBB3 tubulin, beta 3 class III 2.515 1.13E-04

family with sequence similarity 46, FAM46C 2.496 1.39E-03 member C

ST6 (alpha-N-acetyl-neuraminyl-2,3-beta- ST6GALNAC6 galactosyl-1,3)-N-acetylgalactosaminide 2.454 1.13E-04 alpha-2,6-sialyltransferase 6

MAP6D1 MAP6 domain containing 1 2.429 9.60E-05

STOX1 storkhead box 1 2.428 1.07E-02

CDC27 cell division cycle 27 2.420 9.51E-03

HMGCR 3-hydroxy-3-methylglutaryl-CoA reductase 2.376 1.61E-04

GJB2 gap junction protein, beta 2, 26kDa 2.376 1.39E-03

286

SNORA79 small nucleolar RNA, H/ACA box 79 -3.011 1.16E-04

EPSTI1 epithelial stromal interaction 1 (breast) -3.016 1.41E-05

RAB39B RAB39B, member RAS oncogene family -3.017 3.32E-04

SUCNR1 succinate receptor 1 -3.039 2.98E-04

CBS cystathionine-beta-synthase -3.140 3.35E-06

KRT81 keratin 81, type II -3.206 1.82E-06

ULBP1 UL16 binding protein 1 -3.236 2.75E-05

CLDN7 claudin 7 -3.362 2.98E-04

CARS cysteinyl-tRNA synthetase -3.425 9.00E-05

CLGN calmegin -3.497 2.42E-06

KRT7 keratin 7, type II -3.586 3.86E-04

PDK4 pyruvate dehydrogenase kinase, isozyme 4 -3.614 2.40E-05

ADM2 adrenomedullin 2 -3.796 6.94E-06

DDIT3 DNA-damage-inducible transcript 3 -3.852 2.04E-05

CD226 CD226 molecule -4.012 5.69E-06

NmrA-like family domain containing 1 LOC344887 -4.133 1.82E-06 pseudogene

INHBE inhibin, beta E -5.341 3.77E-06

SNORA12 small nucleolar RNA, H/ACA box 12 -5.388 1.53E-05

ANKRD1 ankyrin repeat domain 1 (cardiac muscle) -5.623 3.35E-06

GDF15 growth differentiation factor 15 -8.097 6.74E-07

287

Figure 6.1 A summary of differentially expressed p53 signalling pathway genes in

WCC0017371-treated H460 cells at 72 h.

Using microarray and bioinformatics analyses, we obtained 3332 differentially expressed genes in WECC017371-treated H460 cells, compared to naïve H460 cells at 72 h. In order to investigate the potential biological impact of these differentially expressed genes,

KEGG pathway mapping was performed and it was revealed that the p53 signalling pathway was significantly enriched in differentially expressed genes. Briefly, 33 of 61 p53 signalling pathway genes were differentially expressed in WECC0017371-treated

H460 cells. Among them, 13 genes were upregulated (red boxes and text) and 20 genes were downregulated (blue boxes and text). P53 signalling pathway network maps were obtained and modified from the KEGG database (Kanehisa and Goto, 2000; Kanehisa et al., 2014). Genes that remained unchanged after drug treatment was coloured in green.

288

289

Table 6.3 Differentially expressed p53 signalling pathway genes KEGG pathway: p53 signalling pathway Pathway fold enrichment: 2.627 p-value: 9.86E-08 Benjamini test-correction: 1.880E-05

Fold Symbol Name chang e CCNE1 cyclin E1 2.10 SFN stratifin 1.80 APAF1 apoptotic peptidase activating factor1 1.71 FAS Fas cell surface death receptor 1.68 CCNE2 cyclin E2 1.64 TP53I3 tumor protein p53 inducible protein 3 1.52 SERPINB5 serpin peptidase inhibitor, cladeB (ovalbumin), member5 1.47 CCNB1 cyclin B1 1.39 TSC2 tuberous sclerosis 2 1.37 BID BH3 interacting domain death agonist 1.33 ZMAT3 zinc finger,matrin-type3 1.33 CDK1 cyclin-dependent kinase1 1.33 STEAP3 STEAP family member 3 , metalloreductase 1.32 CCND3 cyclin D3 1.30 RRM2B ribonucleotide reductase M2B (TP53 inducible) -1.31 TP53AIP1 tumor protein p53 regulated apoptosis inducing protein1 -1.32 SESN2 sestrin 2 -1.333 SIAH1 siah E3 ubiquitin protein ligase1 -1.33 CYCS cytochrome c ,somatic -1.33 TP53 tumour protein p53 -1.34 serpin peptidase inhibitor, clade E SERPINE1 -1.36 (nexin,plasminogenactivatorinhibitortype1), member1 BAX BCL2-associatedXprotein -1.42 SESN3 sestrin3 -1.50 CHEK2 check point kinase2 -1.55 CDK6 cyclin-dependent kinase 6 -1.58 290

CDKN1A cyclin-dependent kinase inhibitor 1A(p21,Cip1) -1.70 GADD45G growth arrest andDNA-damage-inducible, gamma -1.71 CHEK2P2 check point kinase 2 pseudogene 2 -1.72 PMAIP1 phorbol-12-myristate-13-acetate-induced protein1 -1.75 TNFRSF10 tumor necrosis factor receptor superfamily, member10b -1.86 B GADD45B growth arrest and DNA-damage-inducible, beta -2.55 GADD45A growth arrest and DNA-damage-inducible, alpha -2.70 IGFBP3 insulin-like growth factor binding protein3 -2.71

291

Figure 6.2 A heatmap summary of differentially expressed genes in p53 signalling pathway

H460 cells and cells treated with WECC0017371 were harvested at 72 h after treatment.

KEGG pathway mapping identified 33 differentially expressed genes between naïve and

WECC0017371-treated H460 cells of 61 p53 signalling pathway genes. A heatmap summary of three independent experiments, using normalised expression levels clustered and scaled by gene, illustrates 13 upregulated (red) and 20 downregulated genes (blue), relative to naïve H460 cells. The significance of pathway enrichment was evaluated by the Student’s t-test (p-value= 9.86E-08).

292

WECC0017371- Naïve H460 treated H460 cells cells

293

Figure 6.3 A summary of differentially expressed cell cycle genes in WCC0017371- treated H460 cells at 72 h.

Using microarray and bioinformatics analyses, we obtained 3332 differentially expressed genes in WECC017371-treated H460 cells, compared to naïve H460 cells at 72 h. KEGG pathway mapping was performed and revealed that the cell cycle was significantly enriched in differentially expressed genes. Briefly, 43 of 82 cell cycle genes were differentially expressed in WECC0017371-treated H460 cells. Among them, 16 genes were upregulated (red boxes and text) and 27 genes were downregulated (blue boxes and text). Cell cycle network maps were obtained and modified from the KEGG database

(Kanehisa and Goto, 2000; Kanehisa et al., 2014).

294

295

Table 6.4 Differentially expressed cell cycle genes KEGG pathway: cell cycle Pathway fold enrichment: 1.878 p-value: 3.19E-05 Benjamini test-correction: 3.039E-03

Fold Symbol Name change CDC27 cell division cycle27 2.42 CCNE1 cyclinE1 2.11 ORC1 origin recognition complex,subunit1 1.91 SFN stratifin 1.80 CCNE2 cyclinE2 1.72 MAD2L1 MAD2 mitotic arrest deficient-like1(yeast) 1.64 ESPL1 extra spindle pole bodies like1 ,separase 1.59 ORC6 origin recognition complex,subunit6 1.57 tyrosine3-monooxygenase/tryptophan5- monooxygenase YWHAZ 1.52 activation protein,zeta CDC14B cell division cycle 14B 1.49 CDC6 cell division cycle 6 1.44 MCM4 minichromosome maintenance complex component 4 1.44 S-phase kinase-associated protein 2, E3 ubiquitin protein SKP2 1.43 ligase BUB1 BUB1 mitotic check point serine/threonine kinase 1.42 ABL1 ABL proto-oncogene1, non-receptor tyrosine kinase 1.41 BUB1B BUB1 mitotic checkpoint serine/threonine kinase B 1.41 CCNB1 cyclinB1 1.40 BUB3 BUB3 mitotic check point protein 1.37 tyrosine3-monooxygenase/tryptophan5-monooxygenase YWHAH 1.36 activation protein, eta CDC26 cell division cycle 26 1.36 CDC25A cell division cycle 25A 1.34 WEE1 WEE1 G2 checkpoint kinase 1.34 PTTG1 pituitary tumor-transforming 1 1.34 CDC20 cell division cycle 20 1.33

296

tyrosine3-monooxygenase/tryptophan5-monooxygenase YWHAG 1.33 activation protein, gamma CDK1 cyclin-dependent kinase 1 1.33 CCND3 cyclin D3 1.30 CDC16 cell division cycle 16 -1.31 SMAD3 SMAD family member3 -1.32 PKMYT1 protein kinase, membrane associated tyrosine/threonine1 -1.32 ANAPC10 anaphase promoting complex subunit 10 -1.33 TP53 Tumour protein p53 -1.34 ANAPC11 anaphase promoting complex subunit 11 -1.34 CDKN2C cyclin-dependent kinase inhibitor 2C -1.40 CHEK2P2 checkpoint kinase 2 pseudogene2 -1.50 CDC25C cell division cycle 25C -1.52 CHEK2 checkpoint kinase 2 -1.52 CDK6 cyclin-dependent kinase 6 -1.58 CDKN1A cyclin-dependent kinase inhibitor 1A) -1.70 GADD45G growth arrest and DNA-damage-inducible, gamma -1.71 TGFB2 transforming growth factor, beta2 -1.92 GADD45A growth arrest and DNA-damage-inducible, alpha -2.45 GADD45B Growth arrest and DNA-damage-inducible, alpha -2.70

297

Figure 6.4 A heatmap summary of differentially expressed genes in cell cycle

H460 cells and cells treated with WECC0017371 were harvested at 72 h after treatment.

KEGG pathway mapping identified 43 differentially expressed genes between naïve and

WECC0017371-treated H460 cells of 82 cell cycle genes. A heatmap summary of three independent experiments, using normalised expression levels clustered and scaled by gene, illustrates 16 upregulated (red) and 27 downregulated genes (blue), relative to naïve

H460 cells. The significance of pathway enrichment was evaluated by the Student’s t-test

(p-value= 9.86E-08).

298

WECC0017371- Naïve H460 treated H460 cells cells

299 upregulated. Several genes that negatively regulate cell cycle and cell differentiation

(TGFB1, SMAD3, CDKN2C, TNFRSF10B, TP53, CHK2 and GADD45) as well as genes that govern the exit of mitosis (ANAPC6, ANAPC11, ANAPC10) were downregulated

(Figure 6.4). In contrast, mitosis and differentiation-promoting genes (CDK1, CCNB1,

CDC6, ORC and MCM) were upregulated (Figure 6.4).

Similar to the GSEA performed on microarray data collected at 48h, microarray data obtained from naïve and WECC0017371-treated H460 cells at 72 h was compared against gene sets from C2, C4, C5 and C6 collection in the MSigDB. GSEA normalised enrichment scores (NES) provide a measure of relative enrichment between naïve and

WECC0017371-treated H460 cells. Magnitude of the NES correlates with the significance of enrichment. The strongest enriched gene sets often have a NES value between 2 to 3.5. In addition to NES, a q-value of less than 0.1 was also applied to minimise false-positive genes. Interestingly, comparison of microarray data against the

C5 collection revealed a coordinated upregulation of mitosis-related genes in

WECC0017371-treated H460 cells compared to naïve H460 cells (Table 6.5). Together with KEGG pathway data, this suggests that genes upregulated by WECC0017371 were related to both mitosis and proliferation pathways.

300

Table 6.5 Top 20 gene sets identified from the C5 gene ontology (GO) gene sets that are enriched in WECC0017371-treated H460 cells.

q- Direc C5 GO gene sets NES value t-ion SISTER_CHROMATID_SEGREGATION 1.989 0.072 ↑ M_PHASE 1.948 0.071 ↑ MITOTIC_SISTER_CHROMATID_SEGREGATION 1.945 0.049 ↑ AMINO_SUGAR_METABOLIC_PROCESS 1.927 0.045 ↑ GLYCOLIPID_METABOLIC_PROCESS 1.885 0.059 ↑ MICROTUBULE_CYTOSKELETON 1.877 0.055 ↑ DNA_PACKAGING 1.870 0.053 ↑ N_ACETYLTRANSFERASE_ACTIVITY 1.833 0.076 ↑ ACETYLTRANSFERASE_ACTIVITY 1.826 0.071 ↑ CONDENSED_CHROMOSOME 1.811 0.076 ↑ MICROTUBULE_BASED_MOVEMENT 1.806 0.075 ↑ HISTONE_ACETYLTRANSFERASE_ACTIVITY 1.801 0.074 ↑ CHROMOSOMEPERICENTRIC_REGION 1.792 0.078 ↑ MITOSIS 1.782 0.081 ↑ M_PHASE_OF_MITOTIC_CELL_CYCLE 1.768 0.088 ↑ CHROMOSOME_SEGREGATION 1.765 0.085 ↑ SPINDLE 1.763 0.082 ↑ MICROTUBULE_ORGANIZATION_AND_BIOGENESI 1.762 0.078 ↑ S SULFOTRANSFERASE_ACTIVITY 1.762 0.074 ↑ DEOXYRIBONUCLEASE_ACTIVITY 1.758 0.073 ↑

301

6.3 Discussion

In lung tumours, aberrant βIII-tubulin expression is known to play a key role in drug resistance and tumour aggressiveness (reviewed in Karki et al., 2013). As critically reviewed in section 1.3.6 and 1.4.1.1 of this thesis, previous studies have demonstrated a direct functional role for βIII-tubulin in regulating chemosensitivity in NSCLC (Gan et al., 2011; Gan et al., 2010; Gan et al., 2007; McCarroll et al., 2010). In section 4.2.3, we reported a significant decrease in sensitivity to paclitaxel, vincristine and cisplatin in

H460 cells, following treatment with the TUBB3-enhanching WECC0017371 small molecule. In order to investigate the altered drug sensitivity and regulatory mechanisms responsible for aberrant TUBB3 expression in NSCLC cells, a detailed microarray analysis was carried out.

In this study, differentially expressed genes in WECC0017371-treated H460 cells at both

48 and 72 h were identified. The level of these differentially expressed genes, including

TUBB3, may be either directly or indirectly modulated by WECC0017371. Observed expression changes could be due to either: (1) an on-target effect, where WECC0017371 directly binds to and activates cellular regulators of the differentially expressed genes; or

(2) an indirect-targeting effect due to signal propagation after WECC0017371 interacts and activates a signalling pathway; or (3) an off-target effect as a result of non-specific drug-protein interactions. Bioinformatics analyses of these differentially expressed genes identified coordinated changes in functionally related genes and pathways in

WECC0017371-treated H460 cells. The result of this analysis was the identification of two key effected pathways; a strong modulation of genes involved in the cell cycle and p53 signalling pathways. Both of these results will now be discussed including a thorough examination of relevant information from the literature.

302

A direct functional role for βIII-tubulin in regulating chemosensitivity in NSCLC has been demonstrated (Gan et al., 2011; Gan et al., 2010; Gan et al., 2007; McCarroll et al.,

2010). Knockdown of βIII-tubulin sensitises cells to broad classes of chemotherapeutic agents, including cisplatin, vincristine and paclitaxel (Gan et al., 2007; McCarroll et al.,

2010), via enhanced apoptosis induction and reduced tumourigenesis (Gan et al., 2010;

McCarroll et al., 2010), as well as enhanced TBA-induced suppression of microtubule dynamics (Gan et al., 2010; Gan et al., 2007). In this context, WECC0017371-induced upregulation of TUBB3/βIII-tubulin led to decreased sensitivity to cisplatin, paclitaxel and vincristine (section 4.2.3), further supporting a role for this protein in chemosensitivity to NSCLC. The precise mechanism underlying the enhanced apoptosis induction following βIII-tubulin suppression remains unclear. Given these chemotherapeutic agents are structurally and mechanistically distinct from each other,

Gan et al. (2007) suggested that βIII-tubulin may mediate cell susceptibility to drug- induced apoptosis by regulating expression of genes involved in the apoptosis pathway.

The tumour-suppressor protein p53 (encoded by the TP53 gene) is a key regulator of

DNA repair, cell cycle arrest and apoptosis (reviewed in Gudkov and Komarova, 2003;

Viktorsson et al., 2005). Response of NSCLC cells to first-line treatment, cisplatin, is mediated through a p53-dependent apoptotic pathway (Horio et al., 2000; Lai et al.,

2000). The mechanism of apoptosis induction in response to TBA treatment is less well- understood compared to cisplatin, although some authors have suggested a role for p53 and p21 induction (Chen and Horwitz, 2002; Giannakakou et al., 2002; Torres and

Horwitz, 1998), at doses that suppress microtubule dynamics but do not disrupt the structure of the microtubule network (Giannakakou et al., 2002; Giannakakou et al., 2001;

Kim et al., 2012). Multiple clinical and preclinical studies have suggested a role for wild- type p53 in determining chemotherapy-mediated cytotoxicity and cell survival in NSCLC 303 cells. It is thought that wild-type p53 renders tumours sensitive to chemotherapy through the induction of apoptosis (Fujiwara et al., 1994; Fujiwara et al., 2006; Lai et al., 2000;

Osaki et al., 2000; Schuler et al., 2001; Zhang et al., 2000), while p53 mutations and loss of function may lead to resistance (de Vries et al., 2002; Rusch et al., 1995). In fact, the

NSCLC H460 cells used in the current study harbour wild-type p53 and their p53 status has been demonstrated to play a critical role in chemotherapy-induced apoptosis (Lai et al., 2000).

Microtubules are involved in signalling transduction and provide tracks for directional vesicle transport. Studies have demonstrated that the activity of p53 is heavily dependent on microtubules (Giannakakou et al., 2000). p53 is transported along microtubule tracks by dynein proteins to the nucleus (Giannakakou et al., 2000), where it exerts its function.

Further, p53 levels and accumulation are regulated by microtubule dynamics.

Suppression of microtubule dynamics by low doses of TBAs has been shown to favour microtubule trafficking of p53, enhancing its nuclear accumulation (Giannakakou et al.,

2002). Hence, by regulating p53 levels and translocation, microtubules can significantly impact p53-mediated stress response signalling. Furthermore, tubulin isotype composition can influence microtubule-associated motor protein function (Sirajuddin et al., 2014). The carboxyl-terminal tail of βIII-tubulin was shown to influence the function of human kinesin-1 protein, but not the yeast recombinant dynein which was used in the study (Sirajuddin et al., 2014). This study was conducted in yeast with homogenous tubulin, further studies are required in human tumour cells to investigate the influence of

βIII-tubulin on dynein function. Despite this, it is clear that the carboxyl-terminal tail of

βIII-tubulin can play a role in microtubule trafficking events.

304

In the present study, pathway analysis of the microarray data revealed a significant upregulation of βIII-tubulin followed by downregulation of multiple genes involved in the p53 signalling pathway in WECC0017371-treated H460 cells. Specifically, DNA- damage response genes, TP53 and its downstream targets, including genes involved in cell cycle arrest, DNA damage repair and apoptosis were downregulated (section 6.2.3).

Further, clonogenic assays in this thesis showed that WECC0017371-enhanced βIII- tubulin was associated with decreased chemosensitivity in H460 cells (section 4.2.3).

Given the p53 signalling pathway plays a critical role in chemotherapy-induced apoptosis in the H460 cells used in the current study (Lai et al., 2000) and p53 levels and activity are heavily regulated by microtubules (Giannakakou et al., 2000), there is a potential that

WECC0017371-enhanced TUBB3/βIII-tubulin may interact with and dampen p53 signalling pathway activity. This may in turn enhance cell survival by evasion of apoptosis and contribute to a decrease in sensitivity to paclitaxel, vincristine and cisplatin.

However, it is unclear whether and how βIII-tubulin can regulate p53 levels and activity.

One potential explanation is that βIII-tubulin may influence p53 levels and activity by modulating microtubule dynamics. Previous studies have demonstrated that tubulin composition can influence microtubule dynamics. In NSCLC cells, suppression of βIII- tubulin using siRNA decreases microtubule dynamics in the presence of TBA (Gan et al.,

2010). Given that suppression of microtubule dynamics by TBAs can favour microtubule trafficking of p53 to the nucleus (Giannakakou et al., 2002), it is possible that by negatively regulating p53 nuclear transportation, WECC0017371-enhanced βIII-tubulin may hamper the activity of p53 signalling pathway, however direct evidence is required to confirm this.

305

Pathway analysis of the microarray data also revealed a WECC0017371-mediated effect on multiple members of the cell cycle pathway, including a decrease in the WAF1 gene.

The WAF1 gene encodes the cyclin dependent kinase inhibitor p21, a major player in negative cell cycle control (reviewed in Abbas and Dutta, 2009; Dutto et al., 2014).

Preclinical studies have demonstrated an important role of p21 in tumour growth suppression (Martín-Caballero et al., 2001), drug-induced cell cycle arrest (Wei et al.,

2010) and chemosensitivity in multiple tumour cell lines (Kondo et al., 1996; Lincet et al., 2000), including NSCLC (Joshi et al., 1998; Wei et al., 2010). In the current work,

WAF1 was identified as downregulated in WECC0017371-treated H460 cells. Since p53 is a major transcriptional activator of WAF1 (El-Deiry et al., 1993), decreased WAF1 expression in WECC0017371-treated cells is likely a consequence of decreased TP53.

Given the important role of p21 in negative cell cycle control and regulation of tumour growth, we anticipate the decrease in WAF1 following WECC0017371 treatment could result in an impaired induction of this protein in response to paclitaxel, vincristine and cisplatin, permitting evasion of the G1/S checkpoint and cell cycle arrest. Surviving cells may then in turn proliferate and contribute to a dampened response to chemotherapy. This hypothesis requires further direct experimental evidence, as will be discussed in Chapter

This study has provided insight into gene expression and pathway changes in H460 cells after treatment with the small molecule WECC0017371. Due to insufficient time we were unable to validate the effects of WECC0017371 on candidate proteins involved in the cell cycle and p53 signalling pathways. Future validations at the gene and protein level, using

RT-PCR, Western blotting and immunohistochemical staining will be required to examine to what extent WECC0017371 can modulate expression of components of the p53 signalling pathway. Results from this work raised the possibility that 306

WECC0017371-enhanced βIII-tubulin may dampen p53 pathway activity and deregulate cell cycle control, contributing to cell survival and a decreased sensitivity to chemotherapy in NSCLC cells. Future functional experiments are required to determine whether βIII-tubulin can modulate components of the p53 signalling pathway. To address this, protein expression of p53, its upstream regulators and downstream effectors, such as p21, Bax, Bcl-2 and Fas will be quantitated in naïve, WECC0017371-treated and βIII- tubulin knockdown H460 cells, in the presence and absence of chemotherapy. In addition, overexpression and knockdown of βIII-tubulin in H460 cells can also be used to determine the impact of βIII-tubulin level on cell cycle and p53 signalling pathways.

Results of this thesis also provided glimpses of insight into the regulatory mechanisms responsible for aberrant TUBB3 expression in NSCLC cells. Analysis of microarray data strongly indicates a link between a small molecule-induced TUBB3 upregulation and modulation of genes involved in the cell cycle and p53 signalling pathways. Previous studies had shown that p53 signalling can influence microtubule dynamics and remodelling, as well as the expression of tubulin isotypes and MAPs (Galmarini et al.,

2003; Murphy et al., 1996), but not βIII-tubulin (Galmarini et al., 2003). As mentioned earlier, microarray analyses of gene changes in WECC0017371-treated H460 cells revealed that TUBB3 upregulation occurred before the coordinated modulation of cell cycle and p53 pathway genes. Further, WECC0017371 was capable of increasing βIII- tubulin expression in both p53 wild-type H460 cells and p53-null H1299 cells (section

4.2.3). Based on results from this work and previous studies, the small molecule-induced

βIII-tubulin upregulation may not be dependent on changes in p53 expression. Taken together, p53 is unlikely to regulate βIII-tubulin.

307

Overall, this study has provided the global gene and pathway changes in H460 cells, following treatment with the small molecule WECC0017371. Our data raise the possibility that βIII-tubulin may regulate or be regulated with p53 in H460 cells. A detailed understanding of the mechanisms by which βIII-tubulin mediates chemosensitivity in NSCLC will help to identify potential novel therapeutic targets.

308

Chapter 7

Concluding Remarks and Future Directions

Non-small Cell Lung Cancer (NSCLC) survival rates are dismal and chemotherapy resistance is the primary cause of treatment failure. βIII-tubulin (encoded by TUBB3 gene) is a structural component of microtubules. This protein is aberrantly expressed and is associated with chemoresistance and tumour aggressiveness in NSCLC, where it has been identified as a bona fide target for chemosensitisation. Currently, there is no commercially available TUBB3/βIII-tubulin inhibitor. Regulation of βIII-tubulin is poorly understood, making it difficult to target this protein.

A high-throughput screen was developed and conducted to identify small molecule modulators of TUBB3 using a cell-based approach. This approach offered a physiologically-representative model, using the minimal TUBB3 promoter and a luciferase readout, complete with intact regulatory networks and feedback mechanisms.

The small chemical molecule identified in our high throughput screen, WECC0017371, was able to significantly enhance TUBB3/βIII-tubulin expression. Furthermore,

WECC0017371 did not alter microtubule morphology or cell cycle profile, but enhanced

βIII-tubulin immunostaining in two independent NSCLC cell lines, H460 and H1299, compared to control. Importantly, WECC0017371-enhanced βIII-tubulin expression was functional and associated with a significant decrease in in vitro sensitivity to TBAs and

DNA-damaging agents. Chemical modifications of WECC0017371 were performed to establish structure-activity relationships and further enhance its potency, efficacy and selectivity. This work has generated a more potent and specific TUBB3/βIII-tubulin enhancer than WECC0017371 for H460 cells, ENTF041. To the best of our knowledge, this thesis developed the first chemical modulator of βIII-tubulin expression. Future dose- 309 response experiments are required to determine the lowest dose required for ENTF041 to achieve significant TUBB3/βIII-tubulin enhancing activity, to minimise cellular toxicity and non-specific biological activity. Furthermore, the application of ENTF041 to induce

TUBB3 and βIII-tubulin expression in other cancer cell lines requires further characterisation and optimisation.

In order to identify cellular factors that regulate βIII-tubulin expression in NSCLC cells,

ENTF041 was developed into an affinity chromatography probe for future affinity-based pulldown assays. This may allow the identification of proteins that ENTF041 interacts with to exert TUBB3/βIII-tubulin enhancement. Negative control probes were also developed to identify and eliminate proteins that bind non-specifically to the affinity probe. A key focus in the future will be to complete the affinity-based proteomics study, including affinity-based pulldown assay and tandem mass spectrometry, to identify cellular binding partners of ENTF041. Following identification of cellular factors that directly interact with ENTF041, an initial overexpression, knockdown and rescue study can be performed to validate their involvement in the regulation of TUBB3/βIII-tubulin.

Assays will be required to verify whether the candidate protein directly regulates βIII- tubulin, and in turn, whether the βIII-tubulin is functional by treating H460 cells with cisplatin, vincristine or paclitaxel, and quantitating changes in drug sensitivity using clonogenic assays. Further, TUBB3 expression is regulated at the transcriptional level and characterisation of the 5’ flanking region of TUBB3 has revealed its minimal promoter region and several potential regulatory motifs (Dennis et al., 2002). Mechanistic studies aimed at identifying the putative binding sites of the candidate protein on TUBB3 regulatory elements will provide valuable insight into the complex regulatory mechanisms that underlie aberrant TUBB3/βIII-tubulin expression in cancer. One approach is to express truncated regions of the TUBB3 promoter to elucidate regions that 310 are critical to TUBB3 regulation and candidate protein regulatory activity. This may also lead to the identification of novel therapeutic targets of TUBB3/βIII-tubulin regulation.

Results from this study indicated that the H460 cell cycle profile remained unchanged following WECC0017371 treatment (section 4.2.3). However, a significant decrease in cell sensitivity to TBAs and DNA damaging agents was reported in WECC0017371- treated H460 cells. Future assessment of the influence of WECC0017371 treatment on

H460 cell cycle profile with and without TBAs and DNA-damaging agents may assist to understand the basis for the decreased drug sensitivity following WECC0017371 treatment. This thesis provided insight into gene expression and pathway changes in H460 cells following treatment with the small molecule WECC0017371. The in vitro results from this thesis showed a decrease in cell sensitivity to TBAs and DNA damaging agents coordinate with upregulation of TUBB3 and downregulation of p53 signalling pathway genes following WECC0017371 treatment. Previous studies have implicated p53 signalling pathway involvement in chemotherapy-induced apoptosis (Horio et al., 2000;

Lai et al., 2000). Further, βIII-tubulin has been demonstrated to play a direct functional role in regulating chemosensitivity to TBAs and DNA-damaging agents via enhanced apoptosis (McCarroll et al., 2010), although the precise mechanism behind this enhanced apoptosis is not fully understood. Based on the results of this thesis and previous studies, it is possible that WECC0017371 enhanced βIII-tubulin expression may contribute to a decrease in drug sensitivity in NSCLC cells via dampened p53 signalling pathway activity. It would be valuable to determine whether βIII-tubulin can rescue tumour cells from chemotherapy-induced apoptotic and cell cycle arrest signals by modulating the activity of p53 signalling pathway. Initial validations of our microarray data at the gene and protein level, using RT-PCR and Western blotting are required to address whether

WECC0017371 may indeed modulate expression of components of the p53 signalling 311 pathway. A detailed time course study quantifying protein levels of βIII-tubulin and components of the p53 signalling pathway in naïve and WECC0017371-treated H460 cells will allow us to establish whether changes in βIII-tubulin levels occur before or after changes in components of the p53 signalling pathway. Potential candidates include p53, upstream regulators (ATM, Chk2, ATR and Chk1) and downstream effectors (p21, Bax,

PUMA) of the p53 pathway. Using WECC0017371-treated H460 cells, future time course studies quantifying the protein level of βIII-tubulin, phosphop53 (phosphorylation of

Ser15 is indicative of p53 activation), and p21, with or without chemotherapeutic agents, would assist in establishing whether βIII-tubulin can modulate chemosensitivity in

NSCLC cells by regulating the level and activity of the p53 signalling pathway. In addition, overexpression or knockdown of βIII-tubulin in H460 cells may provide useful information on whether this protein can influence the level and activity of the p53 signalling pathway. Furthermore, in vitro work of this thesis showed that WECC0017371 treatment led to significantly enhanced βIII-tubulin expression in both p53 wild-type

H460 cells and p53-null H1299 cells (section 4.2.3). Future clonogenic assays will be used on WECC0017371-treated H460 and H1299 cells to investigate the involvement of p53 in cell sensitivity to chemotherapy.

Overall, this work successfully identified the first chemical modulator of TUBB3/βIII- tubulin expression and developed it into a valuable research tool to probe TUBB3/βIII- tubulin regulation in NSCLC cells and elucidate the molecular mechanisms underlying

βIII-tubulin-mediated pathobiology. Findings arising from this thesis may have important therapeutic implications, as identification of molecular regulators of TUBB3/βIII-tubulin will provide additional avenues to target this tubulin-mediated drug-resistance mechanism in NSCLC. In addition, a detailed understanding of the mechanisms by which

312

βIII-tubulin mediates chemosensitivity in NSCLC may lead to the identification of novel therapeutic targets.

References

Abbas, T., and A. Dutta. 2009. p21 in cancer: intricate networks and multiple activities. Nat Rev Cancer. 9:400-414. Akasaka, K., C. Maesawa, M. Shibazaki, F. Maeda, K. Takahashi, T. Akasaka, and T. Masuda. 2009. Loss of Class III β-Tubulin Induced by Histone Deacetylation Is Associated with Chemosensitivity to Paclitaxel in Malignant Melanoma Cells. J Invest Dermatol. 129:1516-1526. Akhmanova, A., and M.O. Steinmetz. 2008. Tracking the ends: a dynamic protein network controls the fate of microtubule tips. Nat Rev Mol Cell Biol. 9:309-322. Alexander, J.E., D.F. Hunt, M.K. Lee, J. Shabanowitz, H. Michel, S.C. Berlin, T.L. MacDonald, R.J. Sundberg, L.I. Rebhun, and A. Frankfurter. 1991. Characterization of posttranslational modifications in neuron-specific class III beta-tubulin by mass spectrometry. Proc Natl Acad Sci U S A. 88:4685-4689. Alfaro-Aco, R., and S. Petry. 2015. Building the Microtubule Cytoskeleton Piece by Piece. J Biol Chem. 290:17154-17162. Alli, E., J.M. Yang, J.M. Ford, and W.N. Hait. 2007. Reversal of stathmin-mediated resistance to paclitaxel and vinblastine in human breast carcinoma cells. Mol Pharmacol. 71:1233-1240. Altmann, K.-H., and J. Gertsch. 2007. Anticancer drugs from nature-natural products as a unique source of new microtubule-stabilizing agents. Natural product reports. 24:327-357. Amos, L.A. 2011. What tubulin drugs tell us about microtubule structure and dynamics. Seminars in Cell & Developmental Biology. 22:916-926. Andreoli, M., M. Persico, A. Kumar, N. Orteca, V. Kumar, A. Pepe, S. Mahalingam, A.E. Alegria, L. Petrella, L. Sevciunaite, A. Camperchioli, M. Mariani, A. Di Dato, E. Novellino, G. Scambia, S.V. Malhotra, C. Ferlini, and C. Fattorusso. 2014. Identification of the first inhibitor of the GBP1:PIM1 interaction. Implications for the development of a new class of anticancer agents against paclitaxel resistant cancer cells. Journal of medicinal chemistry. 57:7916-7932. Audebert, S., E. Desbruyères, C. Gruszczynski, A. Koulakoff, F. Gros, P. Denoulet, and B. Eddé. 1993. Reversible polyglutamylation of alpha- and beta-tubulin and microtubule dynamics in mouse brain neurons. Molecular Biology of the Cell. 4:615-626. Baell, J.B. 2013. Broad coverage of commercially available lead-like screening space with fewer than 350,000 compounds. Journal of chemical information and modeling. 53:39-55. Baell, J.B., and G.A. Holloway. 2010. New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. Journal of medicinal chemistry. 53:2719-2740. Ballas, N., C. Grunseich, D.D. Lu, J.C. Speh, and G. Mandel. 2005. REST and Its Corepressors Mediate Plasticity of Neuronal Gene Chromatin throughout Neurogenesis. Cell. 121:645-657. 313

Banerjee, A., M.C. Roach, P. Trcka, and R.F. Luduena. 1990. Increased microtubule assembly in bovine brain tubulin lacking the type III isotype of beta-tubulin. J Biol Chem. 265:1794-1799. Banerjee, A., M.C. Roach, K.A. Wall, M.A. Lopata, D.W. Cleveland, and R.F. Ludueña. 1988. A monoclonal antibody against the type II isotype of beta-tubulin. Preparation of isotypically altered tubulin. Journal of Biological Chemistry. 263:3029-3034. Barlesi, F., W. Jacot, P. Astoul, and J.L. Pujol. 2006. Second-line treatment for advanced non-small cell lung cancer: a systematic review. Lung Cancer. 51:159-172. Benjamini, Y., and Y. Hochberg. 1995. Controlling the False Discovery Rate: A Practical and Powerful Approach to Multiple Testing. Journal of the Royal Statistical Society. Series B (Methodological). 57:289-300. Bennouna, J., J.L. Breton, J.M. Tourani, C. Ottensmeier, M. O'Brien, P. Kosmidis, T.E. Huat, M.C. Pinel, C. Colin, and J.Y. Douillard. 2006. Vinflunine -- an active chemotherapy for treatment of advanced non-small-cell lung cancer previously treated with a platinum-based regimen: results of a phase II study. Br J Cancer. 94:1383-1388. Berrieman, H.K., M.J. Lind, and L. Cawkwell. 2004. Do beta-tubulin mutations have a role in resistance to chemotherapy? Lancet Oncol. 5:158-164. Bhat, K.M., and V. Setaluri. 2007. Microtubule-associated proteins as targets in cancer chemotherapy. Clin Cancer Res. 13:2849-2854. Blasinska-Morawiec, M., N. Tubiana-Mathieu, R. Fougeray, M.C. Pinel, and P. Bougnoux. 2013. Phase II study of intravenous vinflunine after failure of first-line vinorelbine based regimen for advanced breast cancer. Breast (Edinburgh, Scotland). 22:58-63. Bobinnec, Y., M. Moudjou, J.P. Fouquet, E. Desbruyères, B. Eddé, and M. Bornens. 1998. Glutamylation of centriole and cytoplasmic tubulin in proliferating non- neuronal cells. Cell Motility and the Cytoskeleton. 39:223-232. Bordji, K., A. Grandval, L. Cuhna-Alves, E. Lechapt-Zalcman, and M. Bernaudin. 2014. Hypoxia-inducible factor-2alpha (HIF-2alpha), but not HIF-1alpha, is essential for hypoxic induction of class III beta-tubulin expression in human glioblastoma cells. FEBS J. 281:5220-5236. Bosch Grau, M., G. Gonzalez Curto, C. Rocha, M.M. Magiera, P. Marques Sousa, T. Giordano, N. Spassky, and C. Janke. 2013. Tubulin glycylases and glutamylases have distinct functions in stabilization and motility of ependymal cilia. J Cell Biol. 202:441-451. Brodin, P., E. DelNery, and E. Soleilhac. 2015. [High content screening in chemical biology: overview and main challenges]. Medecine sciences : M/S. 31:187-196. Burdett, S., L. Rydzewska, J. Tierney, A. Auperin, C. Le Pechoux, T. Le Chevalier, J. Pignon, R. Arriagada, D. Johnson, J. van Meerbeeck, M. Parmar, R. Stephens, L. Stewart, R. Arriagada, A. Auperin, P. Bunn, B. Dautzenberg, D. Gilligan, H. Groen, A. Knuuttila, K. Pisters, R. Rosell, J. Roth, G. Scagliotti, M. Tsuboi, D. Waller, V. Westeel, Y. Wu, and X. Yang. 2014. Preoperative chemotherapy for non-small-cell lung cancer: a systematic review and meta-analysis of individual participant data. Lancet. 383:1561-1571. Burkhart, C.A., M. Kavallaris, and S. Band Horwitz. 2001. The role of beta-tubulin isotypes in resistance to antimitotic drugs. Biochim Biophys Acta. 1471:O1-9. Burmakin, M., Y. Shi, E. Hedstrom, P. Kogner, and G. Selivanova. 2013. Dual targeting of wild-type and mutant p53 by small molecule RITA results in the inhibition of

314

N-Myc and key survival oncogenes and kills neuroblastoma cells in vivo and in vitro. Clin Cancer Res. 19:5092-5103. Caracciolo, V., L. D'Agostino, E. Dráberová, V. Sládková, C. Crozier-Fitzgerald, D.P. Agamanolis, J.-P. de Chadarévian, A. Legido, A. Giordano, P. Dráber, and C.D. Katsetos. 2010. Differential expression and cellular distribution of γ-tubulin and βIII-tubulin in medulloblastomas and human medulloblastoma cell lines. Journal of Cellular Physiology. 223:519-529. Caron, J.M., A.L. Jones, L.B. Rall, and M.W. Kirschner. 1985. Autoregulation of tubulin synthesis in enucleated cells. Nature. 317:648-651. Carrara, L., F. Guzzo, D.M. Roque, S. Bellone, C. Emiliano, E. Sartori, S. Pecorelli, P.E. Schwartz, T.J. Rutherford, and A.D. Santin. 2012. Differential in vitro sensitivity to patupilone versus paclitaxel in uterine and ovarian carcinosarcoma cell lines is linked to tubulin-beta-III expression. Gynecol Oncol. 125:231-236. Caudron, F., E. Denarier, J.C. Thibout-Quintana, J. Brocard, A. Andrieux, and A. Fourest-Lieuvin. 2010. Mutation of Ser172 in yeast beta tubulin induces defects in microtubule dynamics and cell division. PLoS One. 5:e13553. Chang, A. 2011. Chemotherapy, chemoresistance and the changing treatment landscape for NSCLC. Lung Cancer. 71:3-10. Chen, J.-G., and S.B. Horwitz. 2002. Differential Mitotic Responses to Microtubule- stabilizing and -destabilizing Drugs. Cancer Research. 62:1935-1938. Chew, S., R. Balasubramanian, W.M. Chan, P.B. Kang, C. Andrews, B.D. Webb, S.E. MacKinnon, D.T. Oystreck, J. Rankin, T.O. Crawford, M. Geraghty, S.L. Pomeroy, W.F. Crowley, Jr., E.W. Jabs, D.G. Hunter, P.E. Grant, and E.C. Engle. 2013. A novel syndrome caused by the E410K amino acid substitution in the neuronal beta-tubulin isotype 3. Brain. 136:522-535. Chittajallu, D.R., S. Florian, R.H. Kohler, Y. Iwamoto, J.D. Orth, R. Weissleder, G. Danuser, and T.J. Mitchison. 2015. In vivo cell-cycle profiling in xenograft tumors by quantitative intravital microscopy. Nature methods. 12:577-585. Chu, C.-W., F. Hou, J. Zhang, L. Phu, A.V. Loktev, D.S. Kirkpatrick, P.K. Jackson, Y. Zhao, and H. Zou. 2011. A novel acetylation of β-tubulin by San modulates microtubule polymerization via down-regulating tubulin incorporation. Molecular Biology of the Cell. 22:448-456. Cicchillitti, L., R. Penci, M. Di Michele, F. Filippetti, D. Rotilio, M.B. Donati, G. Scambia, and C. Ferlini. 2008. Proteomic characterization of cytoskeletal and mitochondrial class III β-tubulin. Molecular Cancer Therapeutics. 7:2070-2079. Cittelly, D.M., I. Dimitrova, E.N. Howe, D.R. Cochrane, A. Jean, N.S. Spoelstra, M.D. Post, X. Lu, R.R. Broaddus, M.A. Spillman, and J.K. Richer. 2012. Restoration of miR-200c to ovarian cancer reduces tumor burden and increases sensitivity to paclitaxel. Mol Cancer Ther. 11:2556-2565. Cleveland, D.W. 1987. The multitubulin hypothesis revisited: what have we learned? The Journal of Cell Biology. 104:381-383. Cleveland, D.W., M.A. Lopata, R.J. MacDonald, N.J. Cowan, W.J. Rutter, and M.W. Kirschner. 1980. Number and evolutionary conservation of α- and β-tubulin and cytoplasmic β- and γ-actin genes using specific cloned cDNA probes. Cell. 20:95- 105. Cochrane, D.R., E.N. Howe, N.S. Spoelstra, and J.K. Richer. 2010. Loss of miR-200c: A Marker of Aggressiveness and Chemoresistance in Female Reproductive Cancers. J Oncol. 2010:821717. Cochrane, D.R., N.S. Spoelstra, E.N. Howe, S.K. Nordeen, and J.K. Richer. 2009. MicroRNA-200c mitigates invasiveness and restores sensitivity to microtubule- 315

targeting chemotherapeutic agents. Molecular Cancer Therapeutics. 8:1055- 1066. Coulson, J.M., J.L. Edgson, P.J. Woll, and J.P. Quinn. 2000. A Splice Variant of the Neuron-restrictive Silencer Factor Repressor Is Expressed in Small Cell Lung Cancer: A Potential Role in Derepression of Neuroendocrine Genes and a Useful Clinical Marker. Cancer Research. 60:1840-1844. Cowan, N.J., C.D. Wilde, L.T. Chow, and F.C. Wefald. 1981. Structural variation among human beta-tubulin genes. Proc Natl Acad Sci U S A. 78:4877-4881. Cucchiarelli, V., L. Hiser, H. Smith, A. Frankfurter, A. Spano, J.J. Correia, and S. Lobert. 2008. β-tubulin isotype classes II and V expression patterns in nonsmall cell lung carcinomas. Cell Motility and the Cytoskeleton. 65:675-685. Curmi, P.A., O. Gavet, E. Charbaut, S. Ozon, S. Lachkar-Colmerauer, V. Manceau, S. Siavoshian, A. Maucuer, and A. Sobel. 1999. Stathmin and its phosphoprotein family: general properties, biochemical and functional interaction with tubulin. Cell structure and function. 24:345-357. D'Addario, G., and E. Felip. 2009. Non-small-cell lung cancer: ESMO clinical recommendations for diagnosis, treatment and follow-up. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 20 Suppl 4:68-70. Dabydeen, D.A., J.C. Burnett, R. Bai, P. Verdier-Pinard, S.J. Hickford, G.R. Pettit, J.W. Blunt, M.H. Munro, R. Gussio, and E. Hamel. 2006. Comparison of the activities of the truncated halichondrin B analog NSC 707389 (E7389) with those of the parent compound and a proposed binding site on tubulin. Mol Pharmacol. 70:1866-1875. Danza, G., C. Di Serio, F. Rosati, G. Lonetto, N. Sturli, D. Kacer, A. Pennella, G. Ventimiglia, R. Barucci, A. Piscazzi, I. Prudovsky, M. Landriscina, N. Marchionni, and F. Tarantini. 2012. Notch Signaling Modulates Hypoxia-Induced Neuroendocrine Differentiation of Human Prostate Cancer Cells. Molecular Cancer Research. 10:230-238. De Donato, M., M. Mariani, L. Petrella, E. Martinelli, G.F. Zannoni, V. Vellone, G. Ferrandina, S. Shahabi, G. Scambia, and C. Ferlini. 2012. Class III β-tubulin and the cytoskeletal gateway for drug resistance in ovarian cancer. Journal of Cellular Physiology. 227:1034-1041. De Gendt, K., E. Denolet, A. Willems, V.W. Daniels, L. Clinckemalie, S. Denayer, M.F. Wilkinson, F. Claessens, J.V. Swinnen, and G. Verhoeven. 2011. Expression of Tubb3, a Beta-Tubulin Isotype, Is Regulated by Androgens in Mouse and Rat Sertoli Cells. Biology of Reproduction. 85:934-945. de Vries, A., E.R. Flores, B. Miranda, H.M. Hsieh, C.T. van Oostrom, J. Sage, and T. Jacks. 2002. Targeted point mutations of p53 lead to dominant-negative inhibition of wild-type p53 function. Proc Natl Acad Sci U S A. 99:2948-2953. Dennis, K., M. Uittenbogaard, A. Chiaramello, and S.A. Moody. 2002. Cloning and characterization of the 5'-flanking region of the rat neuron-specific Class III β- tubulin gene. Gene. 294:269-277. Derry, W.B., L. Wilson, I.A. Khan, R.F. LudueÃ±a, and M.A. Jordan. 1997. Taxol Differentially Modulates the Dynamics of Microtubules Assembled from Unfractionated and Purified β-Tubulin Isotypes Biochemistry. 36:3554-3562. Ding, L., G. Getz, D.A. Wheeler, E.R. Mardis, M.D. McLellan, K. Cibulskis, C. Sougnez, H. Greulich, D.M. Muzny, M.B. Morgan, L. Fulton, R.S. Fulton, Q. Zhang, M.C. Wendl, M.S. Lawrence, D.E. Larson, K. Chen, D.J. Dooling, A. Sabo, A.C. Hawes, H. Shen, S.N. Jhangiani, L.R. Lewis, O. Hall, Y. Zhu, T. Mathew, Y. Ren, 316

J. Yao, S.E. Scherer, K. Clerc, G.A. Metcalf, B. Ng, A. Milosavljevic, M.L. Gonzalez-Garay, J.R. Osborne, R. Meyer, X. Shi, Y. Tang, D.C. Koboldt, L. Lin, R. Abbott, T.L. Miner, C. Pohl, G. Fewell, C. Haipek, H. Schmidt, B.H. Dunford- Shore, A. Kraja, S.D. Crosby, C.S. Sawyer, T. Vickery, S. Sander, J. Robinson, W. Winckler, J. Baldwin, L.R. Chirieac, A. Dutt, T. Fennell, M. Hanna, B.E. Johnson, R.C. Onofrio, R.K. Thomas, G. Tonon, B.A. Weir, X. Zhao, L. Ziaugra, M.C. Zody, T. Giordano, M.B. Orringer, J.A. Roth, M.R. Spitz, I.I. Wistuba, B. Ozenberger, P.J. Good, A.C. Chang, D.G. Beer, M.A. Watson, M. Ladanyi, S. Broderick, A. Yoshizawa, W.D. Travis, W. Pao, M.A. Province, G.M. Weinstock, H.E. Varmus, S.B. Gabriel, E.S. Lander, R.A. Gibbs, M. Meyerson, and R.K. Wilson. 2008. Somatic mutations affect key pathways in lung adenocarcinoma. Nature. 455:1069-1075. Don, S., N.M. Verrills, T.Y.E. Liaw, M.L.M. Liu, M.D. Norris, M. Haber, and M. Kavallaris. 2004. Neuronal-associated microtubule proteins class III β-tubulin and MAP2c in neuroblastoma: Role in resistance to microtubule-targeted drugs. Molecular Cancer Therapeutics. 3:1137-1146. Donoghue, M., S.J. Lemery, W. Yuan, K. He, R. Sridhara, S. Shord, H. Zhao, A. Marathe, L. Kotch, J. Jee, Y. Wang, L. Zhou, W.M. Adams, V. Jarral, A. Pilaro, R. Lostritto, J.E. Gootenberg, P. Keegan, and R. Pazdur. 2012. Eribulin mesylate for the treatment of patients with refractory metastatic breast cancer: use of a "physician's choice" control arm in a randomized approval trial. Clin Cancer Res. 18:1496-1505. Downing, K.H. 2000. Structural basis for the interaction of tubulin with proteins and drugs that affect microtubule dynamics. Annual review of cell and developmental biology. 16:89-111. Du, J., B. Li, Y. Fang, Y. Liu, Y. Wang, J. Li, W. Zhou, and X. Wang. 2015. Overexpression of Class III beta-tubulin, Sox2, and nuclear Survivin is predictive of taxane resistance in patients with stage III ovarian epithelial cancer. BMC cancer. 15:536. Dumontet, C., S. Isaac, P.J. Souquet, F. Bejui-Thivolet, Y. Pacheco, N. Peloux, A. Frankfurter, R. Luduena, and M. Perol. 2005. Expression of class III beta tubulin in non-small cell lung cancer is correlated with resistance to taxane chemotherapy. Bull Cancer. 92:E25-30. Dumontet, C., and M.A. Jordan. 2010. Microtubule-binding agents: a dynamic field of cancer therapeutics. Nat Rev Drug Discov. 9:790-803. Dumontet, C., M.A. Jordan, and F.F. Lee. 2009. Ixabepilone: targeting betaIII-tubulin expression in taxane-resistant malignancies. Mol Cancer Ther. 8:17-25. Dutto, I., M. Tillhon, O. Cazzalini, L.A. Stivala, and E. Prosperi. 2014. Biology of the cell cycle inhibitor p21CDKN1A: molecular mechanisms and relevance in chemical toxicology. Archives of Toxicology. 89:155-178. Easter, S.S., Jr., L.S. Ross, and A. Frankfurter. 1993. Initial tract formation in the mouse brain. J Neurosci. 13:285-299. Edelman, M.J., and M. Shvartsbeyn. 2012. Epothilones in development for non--small- cell lung cancer: novel anti-tubulin agents with the potential to overcome taxane resistance. Clin Lung Cancer. 13:171-180. Egevad, L., A. Valdman, N.P. Wiklund, P. Seve, and C. Dumontet. 2010. Beta-tubulin III expression in prostate cancer. Scandinavian journal of urology and nephrology. 44:371-377.

317

El-Deiry, W.S., T. Tokino, V.E. Velculescu, D.B. Levy, R. Parsons, J.M. Trent, D. Lin, W.E. Mercer, K.W. Kinzler, and B. Vogelstein. 1993. WAF1, a potential mediator of p53 tumor suppression. Cell. 75:817-825. Fan, F., and K.V. Wood. 2007. Bioluminescent assays for high-throughput screening. Assay and drug development technologies. 5:127-136. Felip, E., R. Rosell, J.A. Maestre, J.M. Rodriguez-Paniagua, T. Moran, J. Astudillo, G. Alonso, J.M. Borro, J.L. Gonzalez-Larriba, A. Torres, C. Camps, R. Guijarro, D. Isla, R. Aguilo, V. Alberola, J. Padilla, A. Sanchez-Palencia, J.J. Sanchez, E. Hermosilla, and B. Massuti. 2010. Preoperative chemotherapy plus surgery versus surgery plus adjuvant chemotherapy versus surgery alone in early-stage non- small-cell lung cancer. J Clin Oncol. 28:3138-3145. Ferlini, C., G. Raspaglio, S. Mozzetti, L. Cicchillitti, F. Filippetti, D. Gallo, C. Fattorusso, G. Campiani, and G. Scambia. 2005. The Seco-Taxane IDN5390 Is Able to Target Class III β-Tubulin and to Overcome Paclitaxel Resistance. Cancer Research. 65:2397-2405. Ferrandina, G., E. Martinelli, G.F. Zannoni, M. Distefano, A. Paglia, C. Ferlini, and G. Scambia. 2007. Expression of class III β-tubulin in cervical cancer patients administered preoperative radiochemotherapy: correlation with response to treatment and clinical outcome. Gynecologic Oncology. 104:326-330. Field, Jessica J., José F. Díaz, and John H. Miller. 2013. The Binding Sites of Microtubule-Stabilizing Agents. Chemistry & Biology. 20:301-315. Fojo, A.T., and M. Menefee. 2005. Microtubule targeting agents: basic mechanisms of multidrug resistance (MDR). Semin Oncol. 32:S3-8. Forde, J.C., A.S. Perry, K. Brennan, L.M. Martin, M.P. Lawler, T.H. Lynch, D. Hollywood, and L. Marignol. 2010. Docetaxel maintains its cytotoxic activity under hypoxic conditions in prostate cancer cells. Urologic Oncology: Seminars and Original Investigations. 30:912-919. Fossella, F.V., R. DeVore, R.N. Kerr, J. Crawford, R.R. Natale, F. Dunphy, L. Kalman, V. Miller, J.S. Lee, M. Moore, D. Gandara, D. Karp, E. Vokes, M. Kris, Y. Kim, F. Gamza, L. Hammershaimb, and t.T.N.S.-C.L.C.S. Group. 2000. Randomized Phase III Trial of Docetaxel Versus Vinorelbine or Ifosfamide in Patients With Advanced Non–Small-Cell Lung Cancer Previously Treated With Platinum- Containing Chemotherapy Regimens. Journal of Clinical Oncology. 18:2354- 2362. Fujiwara, T., E.A. Grimm, T. Mukhopadhyay, W.-W. Zhang, L.B. Owen-Schaub, and J.A. Roth. 1994. Induction of chemosensitivity in human lung cancer cells in vivo by adenovirus-mediated transfer of the wild-type p53 Gene. Cancer Research. 54:2287-2291. Fujiwara, T., S. Tanaka N Fau - Kanazawa, S. Kanazawa S Fau - Ohtani, Y. Ohtani S Fau - Saijo, T. Saijo Y Fau - Nukiwa, K. Nukiwa T Fau - Yoshimura, T. Yoshimura K Fau - Sato, Y. Sato T Fau - Eto, S. Eto Y Fau - Chada, H. Chada S Fau - Nakamura, H. Nakamura H Fau - Kato, and H. Kato. 2006. Multicenter phase I study of repeated intratumoral delivery of adenoviral p53 in patients with advanced non-small-cell lung cancer. Journal of Clinical Oncology. 24:1689- 1699. Galmarini, C.M., K. Kamath, A. Vanier-Viornery, V. Hervieu, E. Peiller, N. Falette, A. Puisieux, M. Ann Jordan, and C. Dumontet. 2003. Drug resistance associated with loss of p53 involves extensive alterations in microtubule composition and dynamics. British Journal of Cancer. 88:1793-1799.

318

Gan, P.P., and M. Kavallaris. 2008. Tubulin-targeted drug action: functional significance of class II and class IVb β-tubulin in vinca alkaloid sensitivity. Cancer Research. 68:9817-9824. Gan, P.P., J.A. McCarroll, F.L. Byrne, J. Garner, and M. Kavallaris. 2011. Specific β- tubulin isotypes can functionally enhance or diminish epothilone B sensitivity in non-small cell lung cancer cells. PLoS ONE. 6:e21717. Gan, P.P., J.A. McCarroll, S.T. Po'uha, K. Kamath, M.A. Jordan, and M. Kavallaris. 2010. Microtubule dynamics, mitotic arrest, and apoptosis: drug-induced differential effects of βIII-tubulin. Molecular Cancer Therapeutics. 9:1339-1348. Gan, P.P., E. Pasquier, and M. Kavallaris. 2007. Class III beta-tubulin mediates sensitivity to chemotherapeutic drugs in non small cell lung cancer. Cancer Res. 67:9356-9363. Gao, S., X. Zhao, B. Lin, Z. Hu, L. Yan, and J. Gao. 2012. Clinical implications of REST and TUBB3 in ovarian cancer and its relationship to paclitaxel resistance. Tumour Biol. 33:1759-1765. Garnham, C.P., and A. Roll-Mecak. 2012. The chemical complexity of cellular microtubules: tubulin post-translational modification enzymes and their roles in tuning microtubule functions. Cytoskeleton (Hoboken, N.J.). 69:442-463. Gascoigne, K.E., and S.S. Taylor. 2008. Cancer cells display profound intra- and interline variation following prolonged exposure to antimitotic drugs. Cancer Cell. 14:111- 122. Gascoigne, K.E., and S.S. Taylor. 2009. How do anti-mitotic drugs kill cancer cells? Journal of Cell Science. 122:2579-2585. Gatzemeier, U., J. von Pawel, M. Gottfried, G.P.M.t. Velde, K. Mattson, F. DeMarinis, P. Harper, F. Salvati, G. Robinet, A. Lucenti, J. Bogaerts, and G. Gallant. 2000. Phase III Comparative Study of High-Dose Cisplatin Versus a Combination of Paclitaxel and Cisplatin in Patients With Advanced Non–Small-Cell Lung Cancer. Journal of Clinical Oncology. 18:3390-3399. Gay, D.A., T.J. Yen, J.T.Y. Lau, and D.W. Cleveland. 1987. Sequences that confer β- tubulin autoregulation through modulated mRNA stability reside within exon 1 of a β-tubulin mRNA. Cell. 50:671-679. Giannakakou, P., M. Nakano, K.C. Nicolaou, A. O'Brate, J. Yu, M.V. Blagosklonny, U.F. Greber, and T. Fojo. 2002. Enhanced microtubule-dependent trafficking and p53 nuclear accumulation by suppression of microtubule dynamics. Proceedings of the National Academy of Sciences of the United States of America. 99:10855- 10860. Giannakakou, P., T. Robey R Fau - Fojo, M.V. Fojo T Fau - Blagosklonny, and M.V. Blagosklonny. 2001. Low concentrations of paclitaxel induce cell type-dependent p53, p21 and G1/G2 arrest instead of mitotic arrest: molecular determinants of paclitaxel-induced cytotoxicity. Oncogene. 20:3806-3813. Giannakakou, P., D.L. Sackett, Y. Ward, K.R. Webster, M.V. Blagosklonny, and T. Fojo. 2000. p53 is associated with cellular microtubules and is transported to the nucleus by dynein. Nature cell biology. 2:709-717. Gigant, B., C. Wang, R.B.G. Ravelli, F. Roussi, M.O. Steinmetz, P.A. Curmi, A. Sobel, and M. Knossow. 2005. Structural basis for the regulation of tubulin by vinblastine. Nature. 435:519-522. Gottesman, M.M., T. Fojo, and S.E. Bates. 2002. Multidrug resistance in cancer: role of ATP-dependent transporters. Nat Rev Cancer. 2:48-58.

319

Grozinger, K., J. Proudfoot, and K. Hargrave. 2006. Discovery and Development of Nevirapine. In Drug Discovery and Development. John Wiley & Sons, Inc. 353- 363. Gudkov, A.V., and E.A. Komarova. 2003. The role of p53 in determining sensitivity to radiotherapy. Nat Rev Cancer. 3:117-129. Haendel, M.A., K.E. Bollinger, and P.W. Baas. 1996. Cytoskeletal changes during neurogenesis in cultures of avain neural crest cells. J Neurocytol. 25:289-301. Hammond, J.W., D. Cai, and K.J. Verhey. 2008. Tubulin modifications and their cellular functions. Curr Opin Cell Biol. 20:71-76. Hari, M., H. Yang, C. Zeng, M. Canizales, and F. Cabral. 2003. Expression of class III beta-tubulin reduces microtubule assembly and confers resistance to paclitaxel. Cell Motil Cytoskeleton. 56:45-56. Harker, W.G., and B.I. Sikic. 1985. Multidrug (pleiotropic) resistance in doxorubicin- selected variants of the human sarcoma cell line MES-SA. Cancer Res. 45:4091- 4096. Hatano, Y., Y. Yamada, K. Hata, S. Phutthaphadoong, H. Aoki, and A. Hara. 2011. Genetic ablation of a candidate tumor suppressor gene, Rest, does not promote mouse colon carcinogenesis. Cancer Science. 102:1659-1664. Hayashi, Y., H. Kuriyama, H. Umezu, J. Tanaka, T. Yoshimasu, T. Furukawa, H. Tanaka, H. Kagamu, F. Gejyo, and H. Yoshizawa. 2009. Class III beta-tubulin expression in tumor cells is correlated with resistance to docetaxel in patients with completely resected non-small-cell lung cancer. Internal medicine (Tokyo, Japan). 48:203- 208. Hegele, A., P. Goebell, U. Matz, and T. Neuhaus. 2014. Monotherapy with intravenous vinflunine in patients with advanced or metastatic urothelial cancer after failure of a platinum-containing regimen: a retrospective analysis of German routine data. Urologia internationalis. 92:174-179. Hida, T., Y. Ariyoshi, M. Kuwabara, T. Sugiura, T. Takahashi, T. Takahashi, K. Hosoda, Y. Niitsu, and R. Ueda. 1993. Glutathione S-transferase pi levels in a panel of lung cancer cell lines and its relation to chemo-radiosensitivity. Japanese journal of clinical oncology. 23:14-19. Hiser, L., A. Aggarwal, R. Young, A. Frankfurter, A. Spano, J.J. Correia, and S. Lobert. 2006. Comparison of beta-tubulin mRNA and protein levels in 12 human cancer cell lines. Cell Motil Cytoskeleton. 63:41-52. Horio, Y., Y. Hasegawa Y Fau - Sekido, M. Sekido Y Fau - Takahashi, J.A. Takahashi M Fau - Roth, K. Roth Ja Fau - Shimokata, and K. Shimokata. 2000. Synergistic effects of adenovirus expressing wild-type p53 on chemosensitivity of non-small cell lung cancer cells. Cancer Gene Therapy. 7:537-544. Horwitz, S.B. 1994. Taxol (paclitaxel): mechanisms of action. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 5 Suppl 6:S3-6. Horwitz, S.B., D. Cohen, S. Rao, I. Ringel, H.J. Shen, and C.P. Yang. 1993. Taxol: mechanisms of action and resistance. Journal of the National Cancer Institute. Monographs:55-61. Howe, E.N., D.R. Cochrane, and J.K. Richer. 2011. Targets of miR-200c mediate suppression of cell motility and anoikis resistance. Breast Cancer Res. 13:R45. Howlader, N., A. Noone, M. Krapcho, J. Garshell, D. Miller, S. Altekruse, C. Kosary, M. Yu, Ruhl, J., Z. Tatalovich, A. Mariotto, D. Lewis, H. Chen, E. Feuer, and K. Cronin. 2014. SEER Cancer Statistics Review, 1975-2011, National Cancer

320

Institute. . Bethesda, MD, http://seer.cancer.gov/csr/1975_2011/, based on November 2013 SEER data submission, posted to the SEER web site, April 2014. Huang da, W., B.T. Sherman, and R.A. Lempicki. 2009a. Bioinformatics enrichment tools: paths toward the comprehensive functional analysis of large gene lists. Nucleic acids research. 37:1-13. Huang da, W., B.T. Sherman, and R.A. Lempicki. 2009b. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nature protocols. 4:44-57. Ikegami, K., S. Sato, K. Nakamura, L.E. Ostrowski, and M. Setou. 2010. Tubulin polyglutamylation is essential for airway ciliary function through the regulation of beating asymmetry. Proc Natl Acad Sci U S A. 107:10490-10495. Issaeva, N., P. Bozko, M. Enge, M. Protopopova, L.G. Verhoef, M. Masucci, A. Pramanik, and G. Selivanova. 2004. Small molecule RITA binds to p53, blocks p53-HDM-2 interaction and activates p53 function in tumors. Nat Med. 10:1321- 1328. Izutsu, N., C. Maesawa, M. Shibazaki, H. Oikawa, T. Shoji, T. Sugiyama, and T. Masuda. 2008. Epigenetic modification is involved in aberrant expression of class III β- tubulin, TUBB3, in ovarian cancer cells. Int J Oncol. 32:1227-1235. Janke, C. 2014. The tubulin code: molecular components, readout mechanisms, and functions. J Cell Biol. 206:461-472. Janke, C., and J.C. Bulinski. 2011. Post-translational regulation of the microtubule cytoskeleton: mechanisms and functions. Nat Rev Mol Cell Biol. 12:773-786. Janssen, A., E. Beerling, R. Medema, and J. van Rheenen. 2013. Intravital FRET imaging of tumor cell viability and mitosis during chemotherapy. PLoS One. 8:e64029. Jiang, Y.Q., and M.M. Oblinger. 1992. Differential regulation of beta III and other tubulin genes during peripheral and central neuron development. Journal of Cell Science. 103:643-651. Jirasek, T., S. Cipro, A. Musilova, M. Kubecova, and V. Mandys. 2009. Expression of class III beta-tubulin in colorectal carcinomas: an immunohistochemical study using TU-20 & TuJ-1 antibody. The Indian journal of medical research. 129:89- 94. Joerger, M., D. deJong, A. Burylo, J.A. Burgers, P. Baas, A.D. Huitema, J.H. Beijnen, and J.H. Schellens. 2011. Tubulin, BRCA1, ERCC1, Abraxas, RAP80 mRNA expression, p53/p21 immunohistochemistry and clinical outcome in patients with advanced non small-cell lung cancer receiving first-line platinum-gemcitabine chemotherapy. Lung Cancer. 74:310-317. Jordan, M.A., K. Kamath, T. Manna, T. Okouneva, H.P. Miller, C. Davis, B.A. Littlefield, and L. Wilson. 2005. The primary antimitotic mechanism of action of the synthetic halichondrin E7389 is suppression of microtubule growth. Mol Cancer Ther. 4:1086-1095. Jordan, M.A., R.L. Margolis, R.H. Himes, and L. Wilson. 1986. Identification of a distinct class of vinblastine binding sites on microtubules. Journal of molecular biology. 187:61-73. Jordan, M.A., D. Thrower, and L. Wilson. 1991. Mechanism of inhibition of cell proliferation by Vinca alkaloids. Cancer Res. 51:2212-2222. Jordan, M.A., R.J. Toso, D. Thrower, and L. Wilson. 1993. Mechanism of mitotic block and inhibition of cell proliferation by taxol at low concentrations. Proc Natl Acad Sci U S A. 90:9552-9556. Jordan, M.A., and L. Wilson. 2004. Microtubules as a target for anticancer drugs. Nat Rev Cancer. 4:253-265. 321

Joshi, U.S., G.P. Chen Yq Fau - Kalemkerian, M.R. Kalemkerian Gp Fau - Adil, M. Adil Mr Fau - Kraut, F.H. Kraut M Fau - Sarkar, and F.H. Sarkar. 1998. Inhibition of tumor cell growth by p21WAF1 adenoviral gene transfer in lung cancer. Cancer Gene Therapy. 5:183-191. Kamath, K., L. Wilson, F. Cabral, and M.A. Jordan. 2005. βIII-tubulin induces paclitaxel resistance in association with reduced effects on microtubule dynamic instability. J Biol Chem. 280:12902-12907. Kanehisa, M., and S. Goto. 2000. KEGG: kyoto encyclopedia of genes and genomes. Nucleic acids research. 28:27-30. Kanehisa, M., S. Goto, Y. Sato, M. Kawashima, M. Furumichi, and M. Tanabe. 2014. Data, information, knowledge and principle: back to metabolism in KEGG. Nucleic acids research. 42:D199-205. Kanojia, D., R.A. Morshed, L. Zhang, J.M. Miska, J. Qiao, J.W. Kim, P. Pytel, I.V. Balyasnikova, M.S. Lesniak, and A.U. Ahmed. 2015. βIII-Tubulin Regulates Breast Cancer Metastases to the Brain. Molecular Cancer Therapeutics. 14:1152- 1161. Karki, R., M. Mariani, M. Andreoli, S. He, G. Scambia, S. Shahabi, and C. Ferlini. 2013. betaIII-Tubulin: biomarker of taxane resistance or drug target? Expert Opin Ther Targets. 17:461-472. Katsetos, C.D., and P. Draber. 2012. Tubulins as therapeutic targets in cancer: from bench to bedside. Curr Pharm Des. 18:2778-2792. Katsetos, C.D., P. Draber, and M. Kavallaris. 2011. Targeting βIII-tubulin in glioblastoma multiforme: from cell biology and histopathology to cancer therapeutics. Anticancer Agents Med Chem. 11:719-728. Katsetos, C.D., E. Dráberová, A. Legido, C. Dumontet, and P. Dráber. 2009. Tubulin targets in the pathobiology and therapy of glioblastoma multiforme. I. class III β- tubulin. Journal of Cellular Physiology. 221:505-513. Katsetos, C.D., E. Draberova, B. Smejkalova, G. Reddy, L. Bertrand, J.P. de Chadarevian, A. Legido, J. Nissanov, P.W. Baas, and P. Draber. 2007. Class III beta-tubulin and gamma-tubulin are co-expressed and form complexes in human glioblastoma cells. Neurochemical research. 32:1387-1398. Katsetos, C.D., M.M. Herman, and S.J. Mörk. 2003. Class III β-tubulin in human development and cancer. Cell Motility and the Cytoskeleton. 55:77-96. Kavallaris, M. 2010. Microtubules and resistance to tubulin-binding agents. Nat Rev Cancer. 10:194-204. Kavallaris, M., C.A. Burkhart, and S.B. Horwitz. 1999. Antisense oligonucleotides to class III β-tubulin sensitize drug-resistant cells to Taxol. Br J Cancer. 80:1020- 1025. Kavallaris, M., D.Y. Kuo, C.A. Burkhart, D.L. Regl, M.D. Norris, M. Haber, and S.B. Horwitz. 1997. Taxol-resistant epithelial ovarian tumors are associated with altered expression of specific beta-tubulin isotypes. J Clin Invest. 100:1282-1293. Kavallaris, M., A.S. Tait, B.J. Walsh, L. He, S.B. Horwitz, M.D. Norris, and M. Haber. 2001. Multiple Microtubule Alterations Are Associated with Vinca Alkaloid Resistance in Human Leukemia Cells. Cancer Research. 61:5803-5809. Kelley, M.J., S. Li, and D.H. Harpole. 2001. Genetic Analysis of the β-Tubulin Gene, TUBB, in Non-Small-Cell Lung Cancer. Journal of the National Cancer Institute. 93:1886-1888. Kim, J.H., E.-K. Yoon, H.-J. Chung, S.-Y. Park, K.-M. Hong, C.-H. Lee, Y.-S. Lee, K. Choi, Y. Yang, K. Kim, and I.-H. Kim. 2012. p53 acetylation enhances Taxol- induced apoptosis in human cancer cells. Apoptosis. 18:110-120. 322

Kinoshita, K., I. Arnal, A. Desai, D.N. Drechsel, and A.A. Hyman. 2001. Reconstitution of physiological microtubule dynamics using purified components. Science. 294:1340-1343. Koh, Y., B. Jang, S.W. Han, T.M. Kim, D.Y. Oh, S.H. Lee, C.H. Kang, D.W. Kim, S.A. Im, D.H. Chung, Y.T. Kim, T.Y. Kim, Y.W. Kim, J.H. Kim, D.S. Heo, and Y.J. Bang. 2010. Expression of class III beta-tubulin correlates with unfavorable survival outcome in patients with resected non-small cell lung cancer. J Thorac Oncol. 5:320-325. Köhler, J., S. Schäfer-Preuss, and D. Buttgereit. 1996. Related enhancers in the intron of the beta1 tubulin gene of Drosophila melanogaster are essential for maternal and CNS-specific expression during embryogenesis. Nucleic acids research. 24:2543- 2550. Kohonen-Corish, M.R., H. Qin, J.J. Daniel, W.A. Cooper, L. Rivory, B. McCaughan, M.J. Millward, and R.J. Trent. 2002. Lack of beta-tubulin gene mutations in early stage lung cancer. Int J Cancer. 101:398-399. Kondo, S., Y. Barna Bp Fau - Kondo, Y. Kondo Y Fau - Tanaka, G. Tanaka Y Fau - Casey, J. Casey G Fau - Liu, T. Liu J Fau - Morimura, R. Morimura T Fau - Kaakaji, J.W. Kaakaji R Fau - Peterson, B. Peterson Jw Fau - Werbel, G.H. Werbel B Fau - Barnett, and G.H. Barnett. 1996. WAF1/CIP1 increases the susceptibility of p53 non-functional malignant glioma cells to cisplatin-induced apoptosis. Oncogene. 13:1279-1285. Kreisler, A., P.L. Strissel, R. Strick, S.B. Neumann, U. Schumacher, and C.M. Becker. 2010. Regulation of the NRSF/REST gene by methylation and CREB affects the cellular phenotype of small-cell lung cancer. Oncogene. 29:5828-5838. Kwak, E.L., Y.J. Bang, D.R. Camidge, A.T. Shaw, B. Solomon, R.G. Maki, S.H. Ou, B.J. Dezube, P.A. Janne, D.B. Costa, M. Varella-Garcia, W.H. Kim, T.J. Lynch, P. Fidias, H. Stubbs, J.A. Engelman, L.V. Sequist, W. Tan, L. Gandhi, M. Mino- Kenudson, G.C. Wei, S.M. Shreeve, M.J. Ratain, J. Settleman, J.G. Christensen, D.A. Haber, K. Wilner, R. Salgia, G.I. Shapiro, J.W. Clark, and A.J. Iafrate. 2010. Anaplastic lymphoma kinase inhibition in non-small-cell lung cancer. N Engl J Med. 363:1693-1703. Lackovic, K., G. Lessene, H. Falk, K.J. Leuchowius, J. Baell, and I. Street. 2014. A perspective on 10-years HTS experience at the Walter and Eliza Hall Institute of Medical Research - eighteen million assays and counting. Comb Chem High Throughput Screen. 17:241-252. Lacroix, B., J. van Dijk, N.D. Gold, J. Guizetti, G. Aldrian-Herrada, K. Rogowski, D.W. Gerlich, and C. Janke. 2010. Tubulin polyglutamylation stimulates spastin- mediated microtubule severing. J Cell Biol. 189:945-954. Lai, S.L., J. Perng Rp Fau - Hwang, and J. Hwang. 2000. p53 gene status modulates the chemosensitivity of non-small cell lung cancer cells. J Biomed Sci. 7:64-70. Lansing, T.J., R.T. McConnell, D.R. Duckett, G.M. Spehar, V.B. Knick, D.F. Hassler, N. Noro, M. Furuta, K.A. Emmitte, T.M. Gilmer, R.A. Mook, Jr., and M. Cheung. 2007. In vitro biological activity of a novel small-molecule inhibitor of polo-like kinase 1. Mol Cancer Ther. 6:450-459. Leandro-García, L.J., S. Leskelä, L. Inglada-Pérez, I. Landa, A.A. de Cubas, A. Maliszewska, I. Comino-Méndez, R. Letón, Á. Gómez-Graña, R. Torres, J.C. Ramírez, S. Álvarez, J. Rivera, C. Martínez, M.L. Lozano, A. Cascón, M. Robledo, and C. Rodríguez-Antona. 2012. Hematologic β-Tubulin VI Isoform Exhibits Genetic Variability That Influences Paclitaxel Toxicity. Cancer Research. 72:4744-4752. 323

Leandro-García, L.J., S. Leskelä, I. Landa, C. Montero-Conde, E. López-Jiménez, R. Letón, A. Cascón, M. Robledo, and C. Rodríguez-Antona. 2010. Tumoral and tissue-specific expression of the major human β-tubulin isotypes. Cytoskeleton. 67:214-223. Lee, K.M., D. Cao, A. Itami, P.M. Pour, R.H. Hruban, A. Maitra, and M.M. Ouellette. 2007. Class III β-tubulin, a marker of resistance to paclitaxel, is overexpressed in pancreatic ductal adenocarcinoma and intraepithelial neoplasia. Histopathology. 51:539-546. Leskelä, S., L.J. Leandro-García, M. Mendiola, J. Barriuso, L. Inglada-Pérez, I. Muñoz, B. Martínez-Delgado, A. Redondo, J. de Santiago, M. Robledo, D. Hardisson, and C. Rodríguez-Antona. 2011. The miR-200 family controls β-tubulin III expression and is associated with paclitaxel-based treatment response and progression-free survival in ovarian cancer patients. Endocrine-Related Cancer. 18:85-95. Levallet, G., E. Bergot, M. Antoine, C. Creveuil, A.O. Santos, M. Beau-Faller, F. de Fraipont, E. Brambilla, J. Levallet, F. Morin, V. Westeel, M. Wislez, E. Quoix, D. Debieuvre, F. Dubois, I. Rouquette, J.L. Pujol, D. Moro-Sibilot, J. Camonis, and G. Zalcman. 2012. High TUBB3 expression, an independent prognostic marker in patients with early non small cell lung cancer (NSCLC) treated by preoperative chemotherapy, is regulated by K-Ras signaling pathway. Molecular Cancer Therapeutics. 11:1203-1213. Lewis, S.A., M.E. Gilmartin, J.L. Hall, and N.J. Cowan. 1985. Three expressed sequences within the human beta-tubulin multigene family each define a distinct isotype. J Mol Biol. 182:11-20. Lilenbaum, R.C., J.E. Herndon, M.A. List, C. Desch, D.M. Watson, A.A. Miller, S.L. Graziano, M.C. Perry, W. Saville, P. Chahinian, J.C. Weeks, J.C. Holland, and M.R. Green. 2005. Single-Agent Versus Combination Chemotherapy in Advanced Non–Small-Cell Lung Cancer: The Cancer and Leukemia Group B (study 9730). Journal of Clinical Oncology. 23:190-196. Lim, E., G. Harris, A. Patel, I. Adachi, L. Edmonds, and F. Song. 2009. Preoperative versus postoperative chemotherapy in patients with resectable non-small cell lung cancer: systematic review and indirect comparison meta-analysis of randomized trials. J Thorac Oncol. 4:1380-1388. Lincet, H., L. Poulain, J.S. Remy, E. Deslandes, F. Duigou, P. Gauduchon, and C. Staedel. 2000. The p21cip1/waf1 cyclin-dependent kinase inhibitor enhances the cytotoxic effect of cisplatin in human ovarian carcinoma cells. Cancer Letters. 161:17-26. Lindemann, C.B., and K.A. Lesich. 2010. Flagellar and ciliary beating: the proven and the possible. J Cell Sci. 123:519-528. Liu, B., E. Staren, T. Iwamura, H. Appert, and J. Howard. 2001. Taxotere resistance in SUIT Taxotere resistance in pancreatic carcinoma cell line SUIT 2 and its sublines. World journal of gastroenterology : WJG. 7:855-859. Liu, L., E.E. Geisert, A. Frankfurter, A.J. Spano, C.X. Jiang, J. Yue, I. Dragatsis, and D. Goldowitz. 2007. A transgenic mouse Class-III β tubulin reporter using yellow fluorescent protein. genesis. 45:560-569. Lobert, S., B. Jefferson, and K. Morris. 2011. Regulation of β-tubulin isotypes by micro- RNA 100 in MCF7 breast cancer cells. Cytoskeleton. 68:355-362. Locher, H., K.E. de Rooij, J.C.M.J. de Groot, R. van Doorn, N.A. Gruis, C.W.G.M. Löwik, S.M. Chuva de Sousa Lopes, J.H.M. Frijns, and M.A. Huisman. 2013. Class III β-tubulin, a novel biomarker in the human melanocyte lineage. Differentiation. 85:173-181. 324

Lopus, M., G. Smiyun, H. Miller, E. Oroudjev, L. Wilson, and M.A. Jordan. 2015. Mechanism of action of ixabepilone and its interactions with the betaIII-tubulin isotype. Cancer Chemother Pharmacol. 76:1013-1024. Lorenz, W.W., R.O. McCann, M. Longiaru, and M.J. Cormier. 1991. Isolation and expression of a cDNA encoding Renilla reniformis luciferase. Proc Natl Acad Sci U S A. 88:4438-4442. Ludueña, R.F. 1998. Multiple forms of tubulin: different gene products and covalent modifications. Int Rev Cytol. 178:207-275. Ludueña, R.F. 2013. A hypothesis on the origin and evolution of tubulin. Int Rev Cell Mol Biol. 302:41-185. Ludueña, R.F., M.C. Roach, P.P. Trcka, M. Little, P. Palanivelu, P. Binkley, and V. Prasad. 1982. beta 2-Tubulin, a form of chordate brain tubulin with lesser reactivity toward an assembly-inhibiting sulfhydryl-directed cross-linking reagent. Biochemistry. 21:4787-4794. Lynch, T.J., D.W. Bell, R. Sordella, S. Gurubhagavatula, R.A. Okimoto, B.W. Brannigan, P.L. Harris, S.M. Haserlat, J.G. Supko, F.G. Haluska, D.N. Louis, D.C. Christiani, J. Settleman, and D.A. Haber. 2004. Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med. 350:2129-2139. Maas, C., D. Belgardt, H.K. Lee, F.F. Heisler, C. Lappe-Siefke, M.M. Magiera, J. van Dijk, T.J. Hausrat, C. Janke, and M. Kneussel. 2009. Synaptic activation modifies microtubules underlying transport of postsynaptic cargo. Proc Natl Acad Sci U S A. 106:8731-8736. Maemondo, M., A. Inoue, K. Kobayashi, S. Sugawara, S. Oizumi, H. Isobe, A. Gemma, M. Harada, H. Yoshizawa, I. Kinoshita, Y. Fujita, S. Okinaga, H. Hirano, K. Yoshimori, T. Harada, T. Ogura, M. Ando, H. Miyazawa, T. Tanaka, Y. Saijo, K. Hagiwara, S. Morita, and T. Nukiwa. 2010. Gefitinib or chemotherapy for non- small-cell lung cancer with mutated EGFR. N Engl J Med. 362:2380-2388. Mariani, M., S. Shahabi, S. Sieber, G. Scambia, and C. Ferlini. 2011. Class III -Tubulin (TUBB3): More than a Biomarker in Solid Tumors? Current Molecular Medicine. 11:726-731. Mariani, M., G.F. Zannoni, S. Sioletic, S. Sieber, C. Martino, E. Martinelli, C. Coco, G. Scambia, S. Shahabi, and C. Ferlini. 2012. Gender influences the class III and V beta-tubulin ability to predict poor outcome in colorectal cancer. Clin Cancer Res. 18:2964-2975. Martello, L.A., P. Verdier-Pinard, H.J. Shen, L. He, K. Torres, G.A. Orr, and S.B. Horwitz. 2003. Elevated levels of microtubule destabilizing factors in a Taxol- resistant/dependent A549 cell line with an alpha-tubulin mutation. Cancer Res. 63:1207-1213. Martín-Caballero, J., J.M. Flores, P. García-Palencia, and M. Serrano. 2001. Tumor Susceptibility of p21Waf1/Cip1-deficient Mice. Cancer Research. 61:6234-6238. Matsumoto, Y., M. Oka, A. Sakamoto, F. Narasaki, M. Fukuda, H. Takatani, K. Terashi, K. Ikeda, J. Tsurutani, S. Nagashima, H. Soda, and S. Kohno. 1997. Enhanced expression of metallothionein in human non-small-cell lung carcinomas following chemotherapy. Anticancer Res. 17:3777-3780. McCarroll, J.A., P.P. Gan, R.B. Erlich, M. Liu, T. Dwarte, S.S. Sagnella, M.C. Akerfeldt, L. Yang, A.L. Parker, M.H. Chang, M.S. Shum, F.L. Byrne, and M. Kavallaris. 2015a. TUBB3/betaIII-tubulin acts through the PTEN/AKT signaling axis to promote tumorigenesis and anoikis resistance in non-small cell lung cancer. Cancer Res. 75:415-425. 325

McCarroll, J.A., P.P. Gan, M. Liu, and M. Kavallaris. 2010. βIII-tubulin is a multifunctional protein involved in drug sensitivity and tumorigenesis in non- small cell lung cancer. Cancer Research. 70:4995-5003. McCarroll, J.A., G. Sharbeen, J. Liu, J. Youkhana, D. Goldstein, N. McCarthy, L.F. Limbri, D. Dischl, G.O. Ceyhan, M. Erkan, A.L. Johns, A.V. Biankin, M. Kavallaris, and P.A. Phillips. 2015b. BetaIII-tubulin: a novel mediator of chemoresistance and metastases in pancreatic cancer. Oncotarget. 6:2235-2249. McGrogan, B.T., B. Gilmartin, D.N. Carney, and A. McCann. 2008. Taxanes, microtubules and chemoresistant breast cancer. Biochim Biophys Acta. 1785:96- 132. Michelini, E., L. Cevenini, L. Mezzanotte, A. Coppa, and A. Roda. 2010. Cell-based assays: fuelling drug discovery. Analytical and Bioanalytical Chemistry. 398:227- 238. Mitsudomi, T., S. Morita, Y. Yatabe, S. Negoro, I. Okamoto, J. Tsurutani, T. Seto, M. Satouchi, H. Tada, T. Hirashima, K. Asami, N. Katakami, M. Takada, H. Yoshioka, K. Shibata, S. Kudoh, E. Shimizu, H. Saito, S. Toyooka, K. Nakagawa, and M. Fukuoka. 2010. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. Lancet Oncol. 11:121-128. Mok, T.S., Y.L. Wu, S. Thongprasert, C.H. Yang, D.T. Chu, N. Saijo, P. Sunpaweravong, B. Han, B. Margono, Y. Ichinose, Y. Nishiwaki, Y. Ohe, J.J. Yang, B. Chewaskulyong, H. Jiang, E.L. Duffield, C.L. Watkins, A.A. Armour, and M. Fukuoka. 2009. Gefitinib or carboplatin-paclitaxel in pulmonary adenocarcinoma. N Engl J Med. 361:947-957. Moody, S.A., M.S. Quigg, and A. Frankfurter. 1989. Development of the peripheral trigeminal system in the chick revealed by an isotype-specific anti-beta-tubulin monoclonal antibody. J Comp Neurol. 279:567-580. Moriceau, G., A. Vallard, R. Rivoirard, B. Mery, S. Espenel, J. Langrand-Escure, M. Ben Mrad, G. Wang, P. Diao, C. Pacaut, A. Guillot, O. Collard, P. Fournel, and N. Magne. 2015. Real-World Vinflunine Outcomes in Bladder Cancer in a Single- Institution Study: Moving Beyond Clinical Trials. Clinical genitourinary cancer. 13:588-592. Mozzetti, S., C. Ferlini, P. Concolino, F. Filippetti, G. Raspaglio, S. Prislei, D. Gallo, E. Martinelli, F.O. Ranelletti, G. Ferrandina, and G. Scambia. 2005. Class III beta- tubulin overexpression is a prominent mechanism of paclitaxel resistance in ovarian cancer patients. Clin Cancer Res. 11:298-305. Mozzetti, S., R. Iantomasi, I. De Maria, S. Prislei, M. Mariani, A. Camperchioli, S. Bartollino, D. Gallo, G. Scambia, and C. Ferlini. 2008. Molecular Mechanisms of Patupilone Resistance. Cancer Research. 68:10197-10204. Murphy, M., A. Hinman, and A.J. Levine. 1996. Wild-type p53 negatively regulates the expression of a microtubule-associated protein. Genes Dev. 10:2971-2980. Näär, A.M., P.A. Beaurang, K.M. Robinson, J.D. Oliner, D. Avizonis, S. Scheek, J. Zwicker, J.T. Kadonaga, and R. Tjian. 1998. Chromatin, TAFs, and a novel multiprotein coactivator are required for synergistic activation by Sp1 and SREBP-1a in vitro. Genes & Development. 12:3020-3031. Nair, R.R., H. Avila, X. Ma, Z. Wang, M. Lennartz, B.G. Darnay, D.D. Boyd, and C. Yan. 2008. A Novel High-Throughput Screening System Identifies a Small Molecule Repressive for Matrix Metalloproteinase-9 Expression. Molecular Pharmacology. 73:919-929. 326

Nakakura, E.K., D.N. Watkins, K.E. Schuebel, V. Sriuranpong, M.W. Borges, B.D. Nelkin, and D.W. Ball. 2001a. Mammalian Scratch: a neural-specific Snail family transcriptional repressor. Proc Natl Acad Sci U S A. 98:4010-4015. Nakakura, E.K., D.N. Watkins, V. Sriuranpong, M.W. Borges, B.D. Nelkin, and D.W. Ball. 2001b. Mammalian Scratch participates in neuronal differentiation in P19 embryonal carcinoma cells. Brain Res Mol Brain Res. 95:162-166. Nettles, J.H., H. Li, B. Cornett, J.M. Krahn, J.P. Snyder, and K.H. Downing. 2004. The Binding Mode of Epothilone A on α,ß-Tubulin by Electron Crystallography. Science. 305:866-869. Nicoletti, M.I., G. Valoti, P. Giannakakou, Z. Zhan, J.-H. Kim, V. Lucchini, F. Landoni, J.G. Mayo, R. Giavazzi, and T. Fojo. 2001. Expression of β-Tubulin Isotypes in Human Ovarian Carcinoma Xenografts and in a Sub-Panel of Human Cancer Cell Lines from the NCI-Anticancer Drug Screen: Correlation with Sensitivity to Microtubule Active Agents. Clinical Cancer Research. 7:2912-2922. Niethammer, P., I. Kronja, S. Kandels-Lewis, S. Rybina, P. Bastiaens, and E. Karsenti. 2007. Discrete states of a protein interaction network govern interphase and mitotic microtubule dynamics. PLoS Biol. 5:e29. Niwa, S., H. Takahashi, and N. Hirokawa. 2013. beta-Tubulin mutations that cause severe neuropathies disrupt axonal transport. The EMBO journal. 32:1352-1364. Nogales, E. 2001. Structural insight into microtubule function. Annual review of biophysics and biomolecular structure. 30:397-420. Nogales, E., S.G. Wolf, and K.H. Downing. 1998. Structure of the [alpha][beta] tubulin dimer by electron crystallography. Nature. 391:199-203. Okouneva, T., O. Azarenko, L. Wilson, B.A. Littlefield, and M.A. Jordan. 2008. Inhibition of centromere dynamics by eribulin (E7389) during mitotic metaphase. Mol Cancer Ther. 7:2003-2011. Okuda, K., H. Sasaki, C. Dumontet, O. Kawano, H. Yukiue, T. Yokoyama, M. Yano, and Y. Fujii. 2008. Expression of excision repair cross-complementation group 1 and class III β-tubulin predict survival after chemotherapy for completely resected non-small cell lung cancer. Lung Cancer. 62:105-112. Omar, M., F. Hansmann, R. Kreutzer, M. Kreutzer, G. Brandes, and K. Wewetzer. 2013. Cell Type- and Isotype-Specific Expression and Regulation of β-Tubulins in Primary Olfactory Ensheathing Cells and Schwann Cells In Vitro. Neurochemical Research. 38:981-988. Osaki, S., K. Nakanishi Y Fau - Takayama, X.H. Takayama K Fau - Pei, H. Pei Xh Fau - Ueno, N. Ueno H Fau - Hara, and N. Hara. 2000. Alteration of drug chemosensitivity caused by the adenovirus-mediated transfer of the wild-type p53 gene in human lung cancer cells. Cancer Gene Therapy. 7:300-307. Paez, J.G., P.A. Janne, J.C. Lee, S. Tracy, H. Greulich, S. Gabriel, P. Herman, F.J. Kaye, N. Lindeman, T.J. Boggon, K. Naoki, H. Sasaki, Y. Fujii, M.J. Eck, W.R. Sellers, B.E. Johnson, and M. Meyerson. 2004. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science. 304:1497-1500. Panda, D., H.P. Miller, A. Banerjee, R.F. Luduena, and L. Wilson. 1994. Microtubule dynamics in vitro are regulated by the tubulin isotype composition. Proc Natl Acad Sci U S A. 91:11358-11362. Parker, A.L., M. Kavallaris, and J.A. McCarroll. 2014. Microtubules and their role in cellular stress in cancer. Frontiers in oncology. 4:153. Pepe, A., L. Sun, I. Zanardi, X. Wu, C. Ferlini, G. Fontana, E. Bombardelli, and I. Ojima. 2009. Novel C-seco-taxoids possessing high potency against paclitaxel-resistant

327

cancer cell lines overexpressing class III beta-tubulin. Bioorganic & medicinal chemistry letters. 19:3300-3304. Pereira, D.A., and J.A. Williams. 2007. Origin and evolution of high throughput screening. British journal of pharmacology. 152:53-61. Pesch, B., B. Kendzia, P. Gustavsson, K.H. Jockel, G. Johnen, H. Pohlabeln, A. Olsson, W. Ahrens, I.M. Gross, I. Bruske, H.E. Wichmann, F. Merletti, L. Richiardi, L. Simonato, C. Fortes, J. Siemiatycki, M.E. Parent, D. Consonni, M.T. Landi, N. Caporaso, D. Zaridze, A. Cassidy, N. Szeszenia-Dabrowska, P. Rudnai, J. Lissowska, I. Stucker, E. Fabianova, R.S. Dumitru, V. Bencko, L. Foretova, V. Janout, C.M. Rudin, P. Brennan, P. Boffetta, K. Straif, and T. Bruning. 2012. Cigarette smoking and lung cancer--relative risk estimates for the major histological types from a pooled analysis of case-control studies. Int J Cancer. 131:1210-1219. Pfister, D.G., D.H. Johnson, C.G. Azzoli, W. Sause, T.J. Smith, S. Baker Jr, J. Olak, D. Stover, J.R. Strawn, A.T. Turrisi, and M.R. Somerfield. 2004. American Society of Clinical Oncology Treatment of Unresectable Non–Small-Cell Lung Cancer Guideline: Update 2003. Journal of Clinical Oncology. 22:330-353. Pignon, J.P., H. Tribodet, G.V. Scagliotti, J.Y. Douillard, F.A. Shepherd, R.J. Stephens, A. Dunant, V. Torri, R. Rosell, L. Seymour, S.G. Spiro, E. Rolland, R. Fossati, D. Aubert, K. Ding, D. Waller, and T. Le Chevalier. 2008. Lung adjuvant cisplatin evaluation: a pooled analysis by the LACE Collaborative Group. J Clin Oncol. 26:3552-3559. Pillai, R.S., S.N. Bhattacharyya, and W. Filipowicz. 2007. Repression of protein synthesis by miRNAs: how many mechanisms? Trends Cell Biol. 17:118-126. Pittenger, M.F., and D.W. Cleveland. 1985. Retention of autoregulatory control of tubulin synthesis in cytoplasts: demonstration of a cytoplasmic mechanism that regulates the level of tubulin expression. J Cell Biol. 101:1941-1952. Ploussard, G., S. Terry, P. Maillé, Y. Allory, N. Sirab, L. Kheuang, P. Soyeux, N. Nicolaiew, E. Coppolani, B. Paule, L. Salomon, S. Culine, R. Buttyan, F. Vacherot, and A. de la Taille. 2010. Class III β-Tubulin Expression Predicts Prostate Tumor Aggressiveness and Patient Response to Docetaxel-Based Chemotherapy. Cancer Research. 70:9253-9264. Poirier, K., Y. Saillour, N. Bahi-Buisson, X.H. Jaglin, C. Fallet-Bianco, R. Nabbout, L. Castelnau-Ptakhine, A. Roubertie, T. Attie-Bitach, I. Desguerre, D. Genevieve, C. Barnerias, B. Keren, N. Lebrun, N. Boddaert, F. Encha-Razavi, and J. Chelly. 2010. Mutations in the neuronal ss-tubulin subunit TUBB3 result in malformation of cortical development and neuronal migration defects. Human molecular genetics. 19:4462-4473. Portyanko, A., P. Kovalev, J. Gorgun, and E. Cherstvoy. 2009. βIII-tubulin at the invasive margin of colorectal cancer: possible link to invasion. Virchows Arch. 454:541- 548. Pratesi, G., D. Laccabue, C. Lanzi, G. Cassinelli, R. Supino, M. Zucchetti, R. Frapolli, M. D'Incalci, E. Bombardelli, P. Morazzoni, A. Riva, and F. Zunino. 2003. IDN 5390: an oral taxane candidate for protracted treatment schedules. Br J Cancer. 88:965-972. Prislei, S., E. Martinelli, M. Mariani, G. Raspaglio, S. Sieber, G. Ferrandina, S. Shahabi, G. Scambia, and C. Ferlini. 2013. MiR-200c and HuR in ovarian cancer. BMC Cancer. 13:72. Punna, S., E. Kaltgrad, and M.G. Finn. 2005. "Clickable" agarose for affinity chromatography. Bioconjugate chemistry. 16:1536-1541. 328

Raff, E.C., J.D. Fackenthal, J.A. Hutchens, H.D. Hoyle, and F.R. Turner. 1997. Microtubule Architecture Specified by a β-Tubulin Isoform. Science. 275:70-73. Rami-Porta, R., V. Bolejack, D.J. Giroux, K. Chansky, J. Crowley, H. Asamura, and P. Goldstraw. 2014. The IASLC lung cancer staging project: the new database to inform the eighth edition of the TNM classification of lung cancer. J Thorac Oncol. 9:1618-1624. Rana, T.M. 2007. Illuminating the silence: understanding the structure and function of small RNAs. Nat Rev Mol Cell Biol. 8:23-36. Ranganathan, S., C.A. Benetatos, P.J. Colarusso, D.W. Dexter, and G.R. Hudes. 1998a. Altered beta-tubulin isotype expression in paclitaxel-resistant human prostate carcinoma cells. Br J Cancer. 77:562-566. Ranganathan, S., D.W. Dexter, C.A. Benetatos, A.E. Chapman, K.D. Tew, and G.R. Hudes. 1996. Increase of βIII- and βIVa-tubulin isotopes in human prostate carcinoma cells as a result of estramustine resistance. Cancer Research. 56:2584- 2589. Ranganathan, S., D.W. Dexter, C.A. Benetatos, and G.R. Hudes. 1998b. Cloning and sequencing of human βIII-tubulin cDNA: induction of βIII isotype in human prostate carcinoma cells by acute exposure to antimicrotubule agents. Biochimica et Biophysica Acta (BBA) - Gene Structure and Expression. 1395:237-245. Ranganathan, S., R.A. McCauley, D.W. Dexter, and G.R. Hudes. 2001. Modulation of endogenous beta-tubulin isotype expression as a result of human beta(III)cDNA transfection into prostate carcinoma cells. Br J Cancer. 85:735-740. Raspaglio, G., I. De Maria, F. Filippetti, E. Martinelli, G.F. Zannoni, S. Prislei, G. Ferrandina, S. Shahabi, G. Scambia, and C. Ferlini. 2010. HuR Regulates β- Tubulin Isotype Expression in Ovarian Cancer. Cancer Research. 70:5891-5900. Raspaglio, G., F. Filippetti, S. Prislei, R. Penci, I. De Maria, L. Cicchillitti, S. Mozzetti, G. Scambia, and C. Ferlini. 2008. Hypoxia induces class III beta-tubulin gene expression by HIF-1α binding to its 3' flanking region. Gene. 409:100-108. Raspaglio, G., M. Petrillo, E. Martinelli, D.D. Li Puma, M. Mariani, M. De Donato, F. Filippetti, S. Mozzetti, S. Prislei, G.F. Zannoni, G. Scambia, and C. Ferlini. 2014. Sox9 and HIF-2alpha regulate TUBB3 gene expression and affect ovarian cancer aggressiveness. Gene. 542:173-181. Redeker, V., N. Levilliers, J.M. Schmitter, J.P. Le Caer, J. Rossier, A. Adoutte, and M.H. Bre. 1994. Polyglycylation of tubulin: a posttranslational modification in axonemal microtubules. Science. 266:1688-1691. Reiman, T., R. Lai, A.S. Veillard, E. Paris, J.C. Soria, R. Rosell, M. Taron, S. Graziano, R. Kratzke, L. Seymour, F.A. Shepherd, J.P. Pignon, and P. Seve. 2012. Cross- validation study of class III beta-tubulin as a predictive marker for benefit from adjuvant chemotherapy in resected non-small-cell lung cancer: analysis of four randomized trials. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 23:86-93. Retz, M., P. de Geeter, P.J. Goebell, U. Matz, W. de Schultz, and A. Hegele. 2015. Vinflunine in routine clinical practice for the treatment of advanced or metastatic urothelial cell carcinoma - data from a prospective, multicenter experience. BMC cancer. 15:455. Riihimaki, M., A. Hemminki, M. Fallah, H. Thomsen, K. Sundquist, J. Sundquist, and K. Hemminki. 2014. Metastatic sites and survival in lung cancer. Lung Cancer. 86:78-84. Risinger, A.L., F.J. Giles, and S.L. Mooberry. 2009. Microtubule dynamics as a target in oncology. Cancer Treatment Reviews. 35:255-261. 329

Risinger, A.L., E.M. Jackson, L.A. Polin, G.L. Helms, D.A. LeBoeuf, P.A. Joe, E. Hopper-Borge, R.F. Luduena, G.D. Kruh, and S.L. Mooberry. 2008. The taccalonolides: microtubule stabilizers that circumvent clinically relevant taxane resistance mechanisms. Cancer Res. 68:8881-8888. Ritchie, M.E., B. Phipson, D. Wu, Y. Hu, C.W. Law, W. Shi, and G.K. Smyth. 2015. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic acids research. 43:e47. Robinson, D.R., T. Sherwin, A. Ploubidou, E.H. Byard, and K. Gull. 1995. Microtubule polarity and dynamics in the control of organelle positioning, segregation, and cytokinesis in the trypanosome cell cycle. The Journal of Cell Biology. 128:1163- 1172. Roque, D., N. Buza, M. Glasgow, S. Bellone, I. Bortolomai, S. Gasparrini, E. Cocco, E. Ratner, D.-A. Silasi, M. Azodi, T. Rutherford, P. Schwartz, and A. Santin. 2014. Class III β-tubulin overexpression within the tumor microenvironment is a prognostic biomarker for poor overall survival in ovarian cancer patients treated with neoadjuvant carboplatin/paclitaxel. Clin Exp Metastasis. 31:101-110. Rosell, R., E. Carcereny, R. Gervais, A. Vergnenegre, B. Massuti, E. Felip, R. Palmero, R. Garcia-Gomez, C. Pallares, J.M. Sanchez, R. Porta, M. Cobo, P. Garrido, F. Longo, T. Moran, A. Insa, F. De Marinis, R. Corre, I. Bover, A. Illiano, E. Dansin, J. de Castro, M. Milella, N. Reguart, G. Altavilla, U. Jimenez, M. Provencio, M.A. Moreno, J. Terrasa, J. Munoz-Langa, J. Valdivia, D. Isla, M. Domine, O. Molinier, J. Mazieres, N. Baize, R. Garcia-Campelo, G. Robinet, D. Rodriguez- Abreu, G. Lopez-Vivanco, V. Gebbia, L. Ferrera-Delgado, P. Bombaron, R. Bernabe, A. Bearz, A. Artal, E. Cortesi, C. Rolfo, M. Sanchez-Ronco, A. Drozdowskyj, C. Queralt, I. de Aguirre, J.L. Ramirez, J.J. Sanchez, M.A. Molina, M. Taron, and L. Paz-Ares. 2012. Erlotinib versus standard chemotherapy as first- line treatment for European patients with advanced EGFR mutation-positive non- small-cell lung cancer (EURTAC): a multicentre, open-label, randomised phase 3 trial. Lancet Oncol. 13:239-246. Rosell, R., G. Scagliotti, K.D. Danenberg, R.V. Lord, G. Bepler, S. Novello, J. Cooc, L. Crino, J.J. Sanchez, M. Taron, C. Boni, F. De Marinis, M. Tonato, M. Marangolo, F. Gozzelino, F. Di Costanzo, M. Rinaldi, D. Salonga, and C. Stephens. 2003. Transcripts in pretreatment biopsies from a three-arm randomized trial in metastatic non-small-cell lung cancer. Oncogene. 22:3548-3553. Rossi, A., P. Chiodini, J.-M. Sun, M.E.R. O'Brien, C. von Plessen, F. Barata, K. Park, S. Popat, B. Bergman, B. Parente, C. Gallo, C. Gridelli, F. Perrone, and M. Di Maio. 2014. Six versus fewer planned cycles of first-line platinum-based chemotherapy for non-small-cell lung cancer: a systematic review and meta-analysis of individual patient data. The Lancet Oncology. 15:1254-1262. Rouzier, R., R. Rajan, P. Wagner, K.R. Hess, D.L. Gold, J. Stec, M. Ayers, J.S. Ross, P. Zhang, T.A. Buchholz, H. Kuerer, M. Green, B. Arun, G.N. Hortobagyi, W.F. Symmans, and L. Pusztai. 2005. Microtubule-associated protein tau: a marker of paclitaxel sensitivity in breast cancer. Proc Natl Acad Sci U S A. 102:8315-8320. Rubin, C.I., and G.F. Atweh. 2004. The role of stathmin in the regulation of the cell cycle. J Cell Biochem. 93:242-250. Rudd, R.M., N.H. Gower, S.G. Spiro, T.G. Eisen, P.G. Harper, J.A.H. Littler, M. Hatton, P.W.M. Johnson, W.M.C. Martin, E.M. Rankin, L.E. James, W.M. Gregory, W. Qian, and S.M. Lee. 2005. Gemcitabine Plus Carboplatin Versus Mitomycin, Ifosfamide, and Cisplatin in Patients With Stage IIIB or IV Non-Small-Cell Lung

330

Cancer: A Phase III Randomized Study of the London Lung Cancer Group. Journal of Clinical Oncology. 23:142-153. Rusan, N.M., C.J. Fagerstrom, A.-M.C. Yvon, and P. Wadsworth. 2001. Cell Cycle- Dependent Changes in Microtubule Dynamics in Living Cells Expressing Green Fluorescent Protein-α Tubulin. Molecular Biology of the Cell. 12:971-980. Rusch, V., D. Klimstra, E. Venkatraman, J. Oliver, N. Martini, R. Gralla, M. Kris, and E. Dmitrovsky. 1995. Aberrant p53 Expression Predicts Clinical Resistance to Cisplatin-based Chemotherapy in Locally Advanced Non-Small Cell Lung Cancer. Cancer Research. 55:5038-5042. Rylova, G., T. Ozdian, L. Varanasi, M. Soural, J. Hlavac, D. Holub, P. Dzubak, and M. Hajduch. 2015. Affinity-based methods in drug-target discovery. Current drug targets. 16:60-76. Ryu, H., J. Lee, K. Zaman, J. Kubilis, R.J. Ferrante, B.D. Ross, R. Neve, and R.R. Ratan. 2003. Sp1 and Sp3 Are Oxidative Stress-Inducible, Antideath Transcription Factors in Cortical Neurons. The Journal of Neuroscience. 23:3597-3606. Saillour, Y., L. Broix, E. Bruel-Jungerman, N. Lebrun, G. Muraca, J. Rucci, K. Poirier, R. Belvindrah, F. Francis, and J. Chelly. 2014. Beta tubulin isoforms are not interchangeable for rescuing impaired radial migration due to Tubb3 knockdown. Human molecular genetics. 23:1516-1526. Šale, S., R. Sung, P. Shen, K. Yu, Y. Wang, G.E. Duran, J.-H. Kim, T. Fojo, P.J. Oefner, and B.I. Sikic. 2002. Conservation of the Class I β-Tubulin Gene in Human Populations and Lack of Mutations in Lung Cancers and Paclitaxel-resistant Ovarian Cancers1. Molecular Cancer Therapeutics. 1:215-225. Sandler, A.B., J. Nemunaitis, C. Denham, J. von Pawel, Y. Cormier, U. Gatzemeier, K. Mattson, C. Manegold, M.C. Palmer, A. Gregor, B. Nguyen, C. Niyikiza, and L.H. Einhorn. 2000. Phase III Trial of Gemcitabine Plus Cisplatin Versus Cisplatin Alone in Patients With Locally Advanced or Metastatic Non–Small-Cell Lung Cancer. Journal of Clinical Oncology. 18:122. Sasaki, T., and P.A. Janne. 2011. New strategies for treatment of ALK-rearranged non- small cell lung cancers. Clin Cancer Res. 17:7213-7218. Saussede-Aim, J., E.-L. Matera, C. Ferlini, and C. Dumontet. 2009a. β3-Tubulin is induced by estradiol in human breast carcinoma cells through an estrogen-receptor dependent pathway. Cell Motility and the Cytoskeleton. 66:378-388. Saussede-Aim, J., E.-L. Matera, S. Herveau, R. Jean-Pierre, C. Ferlini, and C. Dumontet. 2009b. Vinorelbine Induces β3-Tubulin Gene Expression through an AP-1 Site. Anticancer Research. 29:3003-3009. Schuler, M., R. Herrmann, J.L.P. De Greve, A.K. Stewart, U. Gatzemeier, D.J. Stewart, L. Laufman, R. Gralla, J. Kuball, R. Buhl, C.P. Heussel, F. Kommoss, A.P. Perruchoud, F.A. Shepherd, M.A. Fritz, J.A. Horowitz, C. Huber, and C. Rochlitz. 2001. Adenovirus-Mediated Wild-Type p53 Gene Transfer in Patients Receiving Chemotherapy for Advanced Non–Small-Cell Lung Cancer: Results of a Multicenter Phase II Study. Journal of Clinical Oncology. 19:1750-1758. Sederholm, C. 2002. Gemcitabine versus gemcitabine/carboplatin in advanced non-small cell lung cancer: preliminary findings in a phase III trial of the Swedish Lung Cancer Study Group. Semin Oncol. 29:50-54. Sequist, L.V., D.W. Bell, T.J. Lynch, and D.A. Haber. 2007. Molecular Predictors of Response to Epidermal Growth Factor Receptor Antagonists in Non–Small-Cell Lung Cancer. Journal of Clinical Oncology. 25:587-595. Seve, P., and C. Dumontet. 2005. Chemoresistance in non-small cell lung cancer. Current medicinal chemistry. Anti-cancer agents. 5:73-88. 331

Seve, P., and C. Dumontet. 2008. Is class III beta-tubulin a predictive factor in patients receiving tubulin-binding agents? Lancet Oncol. 9:168-175. Seve, P., and C. Dumontet. 2010. Class III beta tubulin expression in nonsmall cell lung cancer. Revue des maladies respiratoires. 27:383-386. Seve, P., S. Isaac, O. Tredan, P.J. Souquet, Y. Pacheco, M. Perol, L. Lafanechere, A. Penet, E.L. Peiller, and C. Dumontet. 2005a. Expression of class III beta-tubulin is predictive of patient outcome in patients with non-small cell lung cancer receiving vinorelbine-based chemotherapy. Clin Cancer Res. 11:5481-5486. Seve, P., J. Mackey, S. Isaac, O. Tredan, P.J. Souquet, M. Perol, R. Lai, A. Voloch, and C. Dumontet. 2005b. Class III beta-tubulin expression in tumor cells predicts response and outcome in patients with non-small cell lung cancer receiving paclitaxel. Mol Cancer Ther. 4:2001-2007. Seve, P., T. Reiman, and C. Dumontet. 2010. The role of betaIII tubulin in predicting chemoresistance in non-small cell lung cancer. Lung Cancer. 67:136-143. Shalli, K., I. Brown, S.D. Heys, and A.C. Schofield. 2005. Alterations of β-tubulin isotypes in breast cancer cells resistant to docetaxel. FASEB journal : official publication of the Federation of American Societies for Experimental Biology. 19:1299-1301. Shannon, P., A. Markiel, O. Ozier, N.S. Baliga, J.T. Wang, D. Ramage, N. Amin, B. Schwikowski, and T. Ideker. 2003. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 13:2498- 2504. Shaw, A.T., D.W. Kim, K. Nakagawa, T. Seto, L. Crino, M.J. Ahn, T. De Pas, B. Besse, B.J. Solomon, F. Blackhall, Y.L. Wu, M. Thomas, K.J. O'Byrne, D. Moro-Sibilot, D.R. Camidge, T. Mok, V. Hirsh, G.J. Riely, S. Iyer, V. Tassell, A. Polli, K.D. Wilner, and P.A. Janne. 2013. Crizotinib versus chemotherapy in advanced ALK- positive lung cancer. N Engl J Med. 368:2385-2394. Shaw, A.T., and B. Solomon. 2011. Targeting anaplastic lymphoma kinase in lung cancer. Clin Cancer Res. 17:2081-2086. Sheetz, M.P., E.R. Steuer, and T.A. Schroer. 1989. The mechanism and regulation of fast axonal transport. Trends in neurosciences. 12:474-478. Shen, Y., and L.C. Yu. 2008. Potential protection of curcumin against hypoxia-induced decreases in beta-III tubulin content in rat prefrontal cortical neurons. Neurochemical Research. 33:2112-2117. Shi, W., A. Oshlack, and G.K. Smyth. 2010. Optimizing the noise versus bias trade-off for Illumina whole genome expression BeadChips. Nucleic acids research. 38:e204. Shibazaki, M., C. Maesawa, K. Akasaka, S. Kasai, S. Yasuhira, K. Kanno, I. Nakayama, T. Sugiyama, G. Wakabayasi, T. Masuda, and N. Mori. 2012. Transcriptional and post-transcriptional regulation of betaIII-tubulin protein expression in relation with cell cycle-dependent regulation of tumor cells. Int J Oncol. 40:695-702. Siegel, R., E. Ward, O. Brawley, and A. Jemal. 2011. Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA Cancer J Clin. 61:212-236. Sirajuddin, M., L.M. Rice, and R.D. Vale. 2014. Regulation of microtubule motors by tubulin isotypes and post-translational modifications. Nature cell biology. 16:335- 344. Sleiman, S.F., B.C. Langley, M. Basso, J. Berlin, L. Xia, J.B. Payappilly, M.K. Kharel, H. Guo, J.L. Marsh, L.M. Thompson, L. Mahishi, P. Ahuja, W.R. MacLellan, D.H. Geschwind, G. Coppola, J.r. Rohr, and R.R. Ratan. 2011. Mithramycin Is a 332

Gene-Selective Sp1 Inhibitor That Identifies a Biological Intersection between Cancer and Neurodegeneration. The Journal of Neuroscience. 31:6858-6870. Smith, I.E., M.E.R. O’Brien, D.C. Talbot, M.C. Nicolson, J.L. Mansi, T.F. Hickish, A. Norton, and S. Ashley. 2001. Duration of Chemotherapy in Advanced Non– Small-Cell Lung Cancer: A Randomized Trial of Three Versus Six Courses of Mitomycin, Vinblastine, and Cisplatin. Journal of Clinical Oncology. 19:1336- 1343. Smyth, G.K. 2004. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Statistical applications in genetics and molecular biology. 3:Article3. Sobin, L., and C. Wittekind. 2002. TNM Classification of Malignant Tumours. 6th ed. John Wiley & Sons, New Jersey. Socinski, M.A., M.J. Schell, A. Peterman, K. Bakri, S. Yates, R. Gitten, P. Unger, J. Lee, J.-H. Lee, M. Tynan, M. Moore, and M.S. Kies. 2002. Phase III Trial Comparing a Defined Duration of Therapy Versus Continuous Therapy Followed by Second- Line Therapy in Advanced-Stage IIIB/IV Non–Small-Cell Lung Cancer. Journal of Clinical Oncology. 20:1335-1343. Soda, M., Y.L. Choi, M. Enomoto, S. Takada, Y. Yamashita, S. Ishikawa, S. Fujiwara, H. Watanabe, K. Kurashina, H. Hatanaka, M. Bando, S. Ohno, Y. Ishikawa, H. Aburatani, T. Niki, Y. Sohara, Y. Sugiyama, and H. Mano. 2007. Identification of the transforming EML4-ALK fusion gene in non-small-cell lung cancer. Nature. 448:561-566. Solomon, B.J., T. Mok, D.W. Kim, Y.L. Wu, K. Nakagawa, T. Mekhail, E. Felip, F. Cappuzzo, J. Paolini, T. Usari, S. Iyer, A. Reisman, K.D. Wilner, J. Tursi, and F. Blackhall. 2014. First-line crizotinib versus chemotherapy in ALK-positive lung cancer. N Engl J Med. 371:2167-2177. Song, Y., L.L. Kirkpatrick, A.B. Schilling, D.L. Helseth, N. Chabot, J.W. Keillor, G.V. Johnson, and S.T. Brady. 2013. Transglutaminase and polyamination of tubulin: posttranslational modification for stabilizing axonal microtubules. Neuron. 78:109-123. Sontag, C.A., J.T. Staley, and H.P. Erickson. 2005. In vitro assembly and GTP hydrolysis by bacterial tubulins BtubA and BtubB. J Cell Biol. 169:233-238. Spigel, D.R., F. Anthony Greco, D.M. Waterhouse, D.L. Shipley, J.D. Zubkus, M.J. Bury, C.D. Webb, L.L. Hart, V.G. Gian, J.R. Infante, H.A. Burris, 3rd, and J.D. Hainsworth. 2012. Phase II trial of ixabepilone and carboplatin with or without bevacizumab in patients with previously untreated advanced non-small-cell lung cancer. Lung Cancer. 78:70-75. Spira, A., and D.S. Ettinger. 2004. Multidisciplinary Management of Lung Cancer. New England Journal of Medicine. 350:379-392. Spiro, S.G., R.M. Rudd, R.L. Souhami, J. Brown, D.J. Fairlamb, N.H. Gower, L. Maslove, R. Milroy, V. Napp, M.K.B. Parmar, M.D. Peake, R.J. Stephens, H. Thorpe, D.A. Waller, and P. West. 2004. Chemotherapy versus supportive care in advanced non-small cell lung cancer: improved survival without detriment to quality of life. Thorax. 59:828-836. Stengel, C., S.P. Newman, M.P. Leese, B.V. Potter, M.J. Reed, and A. Purohit. 2010. Class III beta-tubulin expression and in vitro resistance to microtubule targeting agents. Br J Cancer. 102:316-324. Stewart, B.W.W., C. P. 2014. World Cancer Report 2014. International Agency for Research on Cancer Nonserial Publication.

333

Sullivan, I., J. Salazar, M. Majem, C. Pallarés, E. del Río, D. Páez, M. Baiget, and A. Barnadas. 2014. Pharmacogenetics of the DNA repair pathways in advanced non- small cell lung cancer patients treated with platinum-based chemotherapy. Cancer Letters. 353:160-166. Sullivan, K.F., and D.W. Cleveland. 1986. Identification of conserved isotype-defining variable region sequences for four vertebrate beta tubulin polypeptide classes. Proceedings of the National Academy of Sciences. 83:4327-4331. Sun, S., J.H. Schiller, and A.F. Gazdar. 2007. Lung cancer in never smokers [mdash] a different disease. Nat Rev Cancer. 7:778-790. Surrey, T., F. Nedelec, S. Leibler, and E. Karsenti. 2001. Physical properties determining self-organization of motors and microtubules. Science. 292:1167-1171. Swami, U., I. Chaudhary, M.H. Ghalib, and S. Goel. 2012. Eribulin -- a review of preclinical and clinical studies. Crit Rev Oncol Hematol. 81:163-184. Tamura, D., T. Arao, T. Nagai, H. Kaneda, K. Aomatsu, Y. Fujita, K. Matsumoto, M.A. De Velasco, H. Kato, H. Hayashi, S. Yoshida, H. Kimura, Y. Maniwa, W. Nishio, Y. Sakai, C. Ohbayashi, Y. Kotani, Y. Nishimura, and K. Nishio. 2013. Slug increases sensitivity to tubulin-binding agents via the downregulation of betaIII and betaIVa-tubulin in lung cancer cells. Cancer Med. 2:144-154. Theodorakis, N.G., and D.W. Cleveland. 1992. Physical evidence for cotranslational regulation of beta-tubulin mRNA degradation. Mol Cell Biol. 12:791-799. Tischfield, M.A., and E.C. Engle. 2010. Distinct alpha- and beta-tubulin isotypes are required for the positioning, differentiation and survival of neurons: new support for the 'multi-tubulin' hypothesis. Biosci Rep. 30:319-330. Toma, J.G., H. El-Bizri, F. Barnabe-Heider, R. Aloyz, and F.D. Miller. 2000. Evidence that helix-loop-helix proteins collaborate with retinoblastoma tumor suppressor protein to regulate cortical neurogenesis. J Neurosci. 20:7648-7656. Tommasi, S., A. Mangia, R. Lacalamita, A. Bellizzi, V. Fedele, A. Chiriatti, C. Thomssen, N. Kendzierski, A. Latorre, V. Lorusso, F. Schittulli, F. Zito, M. Kavallaris, and A. Paradiso. 2007. Cytoskeleton and paclitaxel sensitivity in breast cancer: The role of β-tubulins. International Journal of Cancer. 120:2078- 2085. Torres, K., and S.B. Horwitz. 1998. Mechanisms of Taxol-induced Cell Death Are Concentration Dependent. Cancer Research. 58:3620-3626. Toso, R.J., M.A. Jordan, K.W. Farrell, B. Matsumoto, and L. Wilson. 1993. Kinetic stabilization of microtubule dynamic instability in vitro by vinblastine. Biochemistry. 32:1285-1293. Travis, W.D., E. Brambilla, and G.J. Riely. 2013. New pathologic classification of lung cancer: relevance for clinical practice and clinical trials. J Clin Oncol. 31:992- 1001. Tsourlakis, M.C., P. Weigand, K. Grupp, M. Kluth, S. Steurer, T. Schlomm, M. Graefen, H. Huland, G. Salomon, T. Steuber, W. Wilczak, H. Sirma, R. Simon, G. Sauter, S. Minner, and A. Quaas. 2014. βIII-tubulin overexpression is an independent predictor of prostate cancer progression tightly linked to ERG fusion status and PTEN deletion. The American journal of pathology. 184:609-617. Tsurutani, J., T. Komiya, H. Uejima, H. Tada, N. Syunichi, M. Oka, S. Kohno, M. Fukuoka, and K. Nakagawa. 2002. Mutational analysis of the β-tubulin gene in lung cancer. Lung Cancer. 35:11-16. Tuszynski, J.A., E.J. Carpenter, J.T. Huzil, W. Malinski, T. Luchko, and R.F. Luduena. 2006. The evolution of the structure of tubulin and its potential consequences for

334

the role and function of microtubules in cells and embryos. The International journal of developmental biology. 50:341-358. Ueda, K., C. Cardarelli, M.M. Gottesman, and I. Pastan. 1987. Expression of a full-length cDNA for the human "MDR1" gene confers resistance to colchicine, doxorubicin, and vinblastine. Proc Natl Acad Sci U S A. 84:3004-3008. Uittenbogaard, M., and A. Chiaramello. 2002. Constitutive overexpression of the basic helix-loop-helix Nex1/MATH-2 transcription factor promotes neuronal differentiation of PC12 cells and neurite regeneration. J Neurosci Res. 67:235- 245. Uppuluri, S., L. Knipling, D.L. Sackett, and J. Wolff. 1993. Localization of the colchicine-binding site of tubulin. Proceedings of the National Academy of Sciences of the United States of America. 90:11598-11602. van Dijk, J., K. Rogowski, J. Miro, B. Lacroix, B. Edde, and C. Janke. 2007. A targeted multienzyme mechanism for selective microtubule polyglutamylation. Mol Cell. 26:437-448. Vansteenkiste, J., D. De Ruysscher, W.E. Eberhardt, E. Lim, S. Senan, E. Felip, and S. Peters. 2013. Early and locally advanced non-small-cell lung cancer: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 24 Suppl 6:vi89-98. Vansteenkiste, J., P.N. Lara, T. Le Chevalier, J.-L. Breton, P. Bonomi, A.B. Sandler, M.A. Socinski, C. Delbaldo, B. McHenry, D. Lebwohl, R. Peck, and M. Edelman. 2007. Phase II Clinical Trial of the Epothilone B Analog, Ixabepilone, in Patients With Non-Small-Cell Lung Cancer Whose Tumors Have Failed First-Line Platinum-Based Chemotherapy. Journal of Clinical Oncology. 25:3448-3455. Verdier-Pinard, P., F. Wang, L. Martello, B. Burd, G.A. Orr, and S.B. Horwitz. 2003. Analysis of tubulin isotypes and mutations from taxol-resistant cells by combined isoelectrofocusing and mass spectrometry. Biochemistry. 42:5349-5357. Verhey, K.J., and J.W. Hammond. 2009. Traffic control: regulation of kinesin motors. Nat Rev Mol Cell Biol. 10:765-777. Viktorsson, K., L. De Petris, and R. Lewensohn. 2005. The role of p53 in treatment responses of lung cancer. Biochem Biophys Res Commun. 331:868-880. Vilmar, A.C., E. Santoni-Rugiu, and J.B. Sorensen. 2010. ERCC1 and histopathology in advanced NSCLC patients randomized in a large multicenter phase III trial. Annals of oncology : official journal of the European Society for Medical Oncology / ESMO. 21:1817-1824. Vilmar, A.C., E. Santoni-Rugiu, and J.B. Sorensen. 2011. Class III beta-tubulin in advanced NSCLC of adenocarcinoma subtype predicts superior outcome in a randomized trial. Clin Cancer Res. 17:5205-5214. Wang, J.Z., and F. Liu. 2008. Microtubule-associated protein tau in development, degeneration and protection of neurons. Progress in neurobiology. 85:148-175. Wang, Y., J.A. Sparano, S. Fineberg, L. Stead, J. Sunkara, S.B. Horwitz, and H.M. McDaid. 2013. High expression of class III beta-tubulin predicts good response to neoadjuvant taxane and doxorubicin/cyclophosphamide-based chemotherapy in estrogen receptor-negative breast cancer. Clinical Breast Cancer. 13:103-108. Wehbe, H., C.M. Kearney, and K.G. Pinney. 2005. Combretastatin A-4 Resistance in H460 Human Lung Carcinoma Demonstrates Distinctive Alterations in β-Tubulin Isotype Expression. Anticancer Research. 25:3865-3870. Wei, J., J. Zhao, M. Long, Y. Han, X. Wang, F. Lin, J. Ren, T. He, and H. Zhang. 2010. p21WAF1/CIP1 gene transcriptional activation exerts cell growth inhibition and 335

enhances chemosensitivity to cisplatin in lung carcinoma cell. BMC cancer. 10:632-632. Westbrook, T.F., E.S. Martin, M.R. Schlabach, Y. Leng, A.C. Liang, B. Feng, J.J. Zhao, T.M. Roberts, G. Mandel, G.J. Hannon, R.A. DePinho, L. Chin, and S.J. Elledge. 2005. A Genetic Screen for Candidate Tumor Suppressors Identifies REST. Cell. 121:837-848. Wolff, M., S. Kredel, J. Wiedenmann, G.U. Nienhaus, and R. Heilker. 2008. Cell-based assays in practice: cell markers from autofluorescent proteins of the GFP-family. Comb Chem High Throughput Screen. 11:602-609. Wozniak, A.J., J.J. Crowley, S.P. Balcerzak, G.R. Weiss, C.H. Spiridonidis, L.H. Baker, K.S. Albain, K. Kelly, S.A. Taylor, D.R. Gandara, and R.B. Livingston. 1998. Randomized trial comparing cisplatin with cisplatin plus vinorelbine in the treatment of advanced non-small-cell lung cancer: a Southwest Oncology Group study. Journal of Clinical Oncology. 16:2459-2465. Wu, M.H., S.B. Huang, and G.B. Lee. 2010. Microfluidic cell culture systems for drug research. Lab on a chip. 10:939-956. Yan, T., K.O. Skaftnesmo, L. Leiss, L. Sleire, J. Wang, X. Li, and P.O. Enger. 2011. Neuronal markers are expressed in human gliomas and NSE knockdown sensitizes glioblastoma cells to radiotherapy and temozolomide. BMC cancer. 11:524. Yen, T.J., D.A. Gay, J.S. Pachter, and D.W. Cleveland. 1988. Autoregulated changes in stability of polyribosome-bound beta-tubulin mRNAs are specified by the first 13 translated nucleotides. Mol Cell Biol. 8:1224-1235. Yoshida, N., K. Haga, and T. Haga. 2003. Identification of sites of phosphorylation by G-protein-coupled receptor kinase 2 in beta-tubulin. European journal of biochemistry / FEBS. 270:1154-1163. Zhai, Y., P.J. Kronebusch, P.M. Simon, and G.G. Borisy. 1996. Microtubule dynamics at the G2/M transition: abrupt breakdown of cytoplasmic microtubules at nuclear envelope breakdown and implications for spindle morphogenesis. The Journal of Cell Biology. 135:201-214. Zhang, H., B. Wang H Fau - Lai, C. Lai B Fau - Zhang, X. Zhang C Fau - Yang, and X. Yang. 2000. Effects of exogenous wild-type p53 gene expression on sensitivity to chemotherapeutic agents in human lung cancer cell. Zhonngquo Fei Ai Za Zhi. 3:166-169. Zhang, H.L., L. Ruan, L.M. Zheng, D. Whyte, C.M. Tzeng, and X.W. Zhou. 2012a. Association between class III beta-tubulin expression and response to paclitaxel/vinorebine-based chemotherapy for non-small cell lung cancer: a meta- analysis. Lung Cancer. 77:9-15. Zhang, J.-H., T.D.Y. Chung, and K.R. Oldenburg. 1999. A Simple Statistical Parameter for Use in Evaluation and Validation of High Throughput Screening Assays. Journal of Biomolecular Screening. 4:67-73. Zhang, L., S. Ma, X. Song, B. Han, Y. Cheng, C. Huang, S. Yang, X. Liu, Y. Liu, S. Lu, J. Wang, S. Zhang, C. Zhou, X. Zhang, N. Hayashi, and M. Wang. 2012b. Gefitinib versus placebo as maintenance therapy in patients with locally advanced or metastatic non-small-cell lung cancer (INFORM; C-TONG 0804): a multicentre, double-blind randomised phase 3 trial. The Lancet Oncology. 13:466- 475. Zhao, C.Y., L. Szekely, W. Bao, and G. Selivanova. 2010. Rescue of p53 function by small-molecule RITA in cervical carcinoma by blocking E6-mediated degradation. Cancer Res. 70:3372-3381. 336

Zheng, W.E., H. Chen, S.F. Yuan, L.L. Wu, W. Zhang, H.Y. Sun, and W.J. Chen. 2012. Overexpression of betaIII-tubulin and survivin associated with drug resistance to docetaxel-based chemotherapy in advanced gastric cancer. Journal of B.U.ON. : official journal of the Balkan Union of Oncology. 17:284-290. Zhou, C., Y.L. Wu, G. Chen, J. Feng, X.Q. Liu, C. Wang, S. Zhang, J. Wang, S. Zhou, S. Ren, S. Lu, L. Zhang, C. Hu, Y. Luo, L. Chen, M. Ye, J. Huang, X. Zhi, Y. Zhang, Q. Xiu, J. Ma, and C. You. 2011. Erlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): a multicentre, open-label, randomised, phase 3 study. Lancet Oncol. 12:735-742. Zhou, J., H.-B. Shu, and H.C. Joshi. 2002. Regulation of tubulin synthesis and cell cycle progression in mammalian cells by γ-tubulin-mediated microtubule nucleation. Journal of Cellular Biochemistry. 84:472-483. Ziegler, S., V. Pries, C. Hedberg, and H. Waldmann. 2013. Target identification for small bioactive molecules: finding the needle in the haystack. Angewandte Chemie (International ed. in English). 52:2744-2792.

337

Appendices

Appendix I

Figure 0.1 Human TUBB3 promoter sequence cloned in the pLightSwitch_Prom_puro_ RenSP vector

The insert size is 975bp. The promoter sequence contains the first 838 bp of human TUBB3 promoter sequence from the transcriptional start site (underlined bold C), a translational start site (bold underlined ATG) and 117 bp of 5’untranslated region.

TTCCTTCTCTGACCAGTAAACGCATCTCCAGTCGCTCCCCCGAGGCCCCCCTGCACGCAGTCT CCGACCTCAGAGAGTGAGTCAGATAGAAGCCGCTGCCCACCCTCCCCCACCCTGGCCGGGTT CTTGCCTGGCTTGTGCGTCCTTGTGACAATTCCTGACACGGATGCGCACCGGGCAGTGGCAC GTCCAGGTGCTGTGCGCGCCCGGGCCCACAAAGCTCCTTTGGACGCTCATGGCAGCCCCTTG GTGGAGGAGGTCGGGTGCCCACCCCCTCCCCGCCCACTGCGGAAGCCGGCGACCCACGGAG CTCGCTCTCGGCCGCCGCCACCCCTCTGTGTTCGCGCCCTTCCGAGCTCTGATCCGACGCTTT GTTTCTTCTCAGTGGGTTCAGGGCCTGGGCCAGCCTTTACCTACCTCCCCCACCCAAAACCGG CAAAAGCTCAGAGCACCTTGTCTGCCAAAAGACAGGGAGCTGGGATGGTGCGGGTTGGTCT CTAAACCGGCGTGGGGAAAAAAGACCCTCCGTACAAAGCCGCAGGGTGGGGCTGTCGCAAG GGCGGAACCGAGAGGGTAGCTGGGGGCGGGGTTCCCAGGGCCAAGAGGGGCCATTGTCCTC CCTGGAGCCCGGCGCCCCCACAGCCAGCTCCTCTGGGAGACAGCCCCTCCTTTCGAATGCGC GGGGCCCTCAGACCGCGCCCGGCCCAGCGCTGGGGGATCCTTGGCTGCGGGAGGGGCGCCG CATTGCGCGCGGCGGGCGGGGACGCGCGGTGCGGAGCCTGCGGGCCGGGCGGGGCTCTGCG GCGGCGCCTCCCGATTGGCCACCCGCGGTGACATCAGCCGATGCGAAGGGCGGGGCCGCGG CTATAAGAGCGCGCGGCCGCGGTCCCCGACCCTCAGCAGCCAGCCCGGCCCGCCCGCGCCC GTCCGCAGCCGCCCGCCAGACGCGCCCAGTATGAGGGAGATCGTGCACAT

Figure 0.2 Human GAPDH promoter sequence cloned in the pLightSwitch_Prom_puro_ RenSP vector. The insert size is 1063 bp and contains the first 991bp of human GAPDH promoter sequence from the translational start site (bold underlined ATG).

GGGATTGTCTGCCCTAATTATCAGGTCCAGGCTACAGGGCTGCAGGACATCGTGACCTTCCG TGCAGAAACCTCCCCCTCCCCCTCAAGCCGCCTCCCGAGCCTCCTTCCTCTCCAGGCCCCCAG TGCCCAGTGCCCAGTGCCCAGCCCAGGCCTCGGTCCCAGAGATGCCAGGAGCCAGGAGATG GGGAGGGGGAAGTGGGGGCTGGGAAGGAACCACGGGCCCCCGCCCGAGGCCCATGGGCCC CTCCTAGGCCTTTGCCTGAGCAGTCCGGTGTCACTACCGCAGAGCCTCGAGGAGAAGTTCCC CAACTTTCCCGCCTCTCAGCCTTTGAAAGAAAGAAAGGGGAGGGGGCAGGCCGCGTGCAGC CGCGAGCGGTGCTGGGCTCCGGCTCCAATTCCCCATCTCAGTCGTTCCCAAAGTCCTCCTGTT TCATCCAAGCGTGTAAGGGTCCCCGTCCTTGACTCCCTAGTGTCCTGCTGCCCACAGTCCAGT CCTGGGAACCAGCACCGATCACCTCCCATCGGGCCAATCTCAGTCCCTTCCCCCCTACGTCG GGGCCCACACGCTCGGTGCGTGCCCAGTTGAACCAGGCGGCTGCGGAAAAAAAAAAGCGGG GAGAAAGTAGGGCCCGGCTACTAGCGGTTTTACGGGCGCACGTAGCTCAGGCCTCAAGACC TTGGGCTGGGACTGGCTGAGCCTGGCGGGAGGCGGGGTCCGAGTCACCGCCTGCCGCCGCG CCCCCGGTTTCTATAAATTGAGCCCGCAGCCTCCCGCTTCGCTCTCTGCTCCTCCTGTTCGAC AGTCAGCCGCATCTTCTTTTGCGTCGCCAGGTAAGACGGGCGGAGAGAAACCCGGGAGGCT AGGGACGGCCTGAAGGCGGCAGGGGCGGGCGCAGGCCGGATGTGTTCGCGCCGCTGCGGGG TGGGCCCGGGCGGCCTCCGCATTGCAGGGGCGGGCGGAGGACGTGATGCGGCGCGGGCTGG GCATGGAGGCCTGGTGGGGGAGGGGAGGGGAGGCGTGTGTGTCGGCCGGGGCCACTAGGC GCTCACTGTTCTCTCC

338

Figure 0.3 Random promoter sequence cloned in the pLightSwitch_Prom_puro_RenSP vector. The random promoter sequence has an insert size of 966 bp and is consisted of a random scramble sequence that does not encode any functional protein in mammalian cells. CTCTTCCTGCCAAAGGGTTAGTTCCTGGTCCCCCAGGGTGGGGAGGCCCTAG AAAGGCCCTGAGCTTTCACAAGTTTAGGTGCTTTAAGTCTGGGGCTCCCAGG GTGCTGGGGAGCAGACCAGTGGGAGTTTAGTGGGCAGAGTCCTGCCTGGCT CACCCGGCCCCATCCGTCTCCGGTGGGCTCTTCTGCTGCCGGTTGTACAGGA AGAAGACAGTGAGCACAGCAACTAGGATGAGGAACACGGCCACGGTGCCC GCCAGGAGCCTGGCTGCACTTCCCAGACCCCTCTGGGGGCGGGGCTTTTCTG GAAACAAAAGACAGGGAAGAGGATTAGAGAGAGAAACTCAAGTGACCCAA AGAGAGTCAGAGATGTGAGAGAGTCAGACCCTGAGGCAGTACTCTGCCTTT GACATTGCACCCCCAGCTCCCCTCCCCATGGTGCTCCAGTTCTAGGAGGGGC TCCCAGCCGGTGTGGGATGGAGGCTCCTCCTCTAGTTCCCTATGACCCTGTA ACCCCACACCCTGCTCTGATCCTGGGATCTGGGGCATGGAGGTAGAACAAG GGTGAGGGTATCTCCTGGACAGGCAGCAGGTGCTGAGATGGGGTCCCCACC CCGTCATTCCATTTCTGGCCCTGCAGGGAAGATTCCTGTTGAGGAGTTTAAC TTCTTTCCCCTCCCTGTGCAATTAAACAAGCAAATGAACTAATGGGTGCTTC ACCCTCCCAGCTTGGGACAGAGGGTGAGGATCAGGAGGGAAAGGCAGGAG CTGGGAGGGTCGATGAAGAAGGGGCACCTGGGGGTGCCTCCCGGGAAACC AAGCCTCTTCAGTACAGATAGGTGAAAAGCCCAGAGGGCTGTGGTCTAAGG TGAGGGGGCTGAGCCCACCTTCTGGACCCACCTGGTCATGGACCCAGAAGA TTGCTGGGCTTGTCGAGGTATTTCACAGTCTGAGCTGGAGACAC

Figure 0.4 pLightSwitch_Prom_puro_RenSP vector construct sequence. The vector construct is 4762 bp in length, containing a multiple cloning region (underlined), Renilla SP luciferase reporter gene sequence (synthetic Renilla luciferase; bold), a SV40 late poly(A) region, a synthetic puromycin resistance expression cassette (italic underlined), ColE1-derived plasmid replication origin, synthetic Beta-lactamase (Ampr) coding region, and a synthetic poly(A) signal/transcriptional pause site.

GGCCTAACTGGCCGGTACCTGAGCTCTTACGCGTGCTAGCCCGGGCTCGAGATCTGCGATCT AAGTAAGCTTAACTAAGTAAGGCATTCCGGTACTGTTGGTAAAGCCACCATGGCTTCCAAG GTGTACGACCCGGAGCAGCGCAAGAGGATGATCACCGGCCCTCAGTGGTGGGCTCGG TGCAAGCAGATGAACGTGCTCGACTCCTTCATCAACTACTACGACAGCGAGAAACATG CGGAGAACGCCGTGATCTTCCTCCACGGCAACGCCGCTTCCTCCTACCTGTGGCGCCA CGTCGTGCCCCACATCGAGCCCGTCGCCCGGTGCATCATCCCTGATCTGATCGGGATG GGGAAGAGCGGGAAGAGCGGCAACGGCAGCTACCGCCTGCTCGACCACTACAAGTAC CTCACCGCCTGGTTCGAGCTGCTGAACCTCCCCAAGAAGATCATCTTTGTGGGCCACG ACTGGGGCGCTTGTCTCGCTTTTCACTACTCCTACGAGCACCAGGATAAGATCAAGGC TATCGTGCATGCTGAGAGCGTCGTGGACGTGATCGAGTCCTGGGACGAGTGGCCCGAT ATCGAGGAGGATATTGCTCTGATCAAGTCCGAGGAGGGCGAGAAGATGGTCCTGGAGA ATAACTTCTTCGTGGAGACTATGCTGCCTAGCAAGATCATGCGCAAGCTGGAGCCCGA GGAGTTCGCTGCTTACCTGGAGCCCTTCAAGGAGAAGGGCGAGGTCAGAAGACCAACC CTCAGCTGGCCTCGGGAGATCCCTCTGGTCAAGGGCGGGAAGCCGGACGTGGTGCAG ATCGTCCGGAACTACAACGCCTACCTGCGCGCCAGCGACGACCTGCCTAAGATGTTCA TCGAGTCCGACCCCGGCTTCTTCAGCAACGCTATCGTGGAGGGCGCCAAGAAGTTCCC CAACACCGAGTTCGTGAAGGTGAAGGGCCTCCACTTCTCCCAAGAGGACGCCCCTGAT GAGATGGGGAAGTACATCAAGAGCTTCGTCGAGCGCGTCCTCAAGAACGAGCAGAATT CTCACGGCTTCCCTCCCGAGGTGGAGGAGCAGGCCGCCGGCACCCTGCCCATGAGCTG 339

CGCCCAGGAGAGCGGCATGGATAGACACCCTGCTGCTTGCGCCAGCGCCAGGATCAAC GTCTAATCTAGAGTCGGGGCGGCCGGCCGCTTCGAGCAGACATGATAAGATACATTGATGA GTTTGGACAAACCACAACTAGAATGCAGTGAAAAAAATGCTTTATTTGTGAAATTTGTGATG CTATTGCTTTATTTGTAACCATTATAAGCTGCAATAAACAAGTTAACAACAACAATTGCATTC ATTTTATGTTTCAGGTTCAGGGGGAGGTGTGGGAGGTTTTTTAAAGCAAGTAAAACCTCTAC AAATGTGGTAAAATCGATAAGGATCCGTACCCGGTCACCTCTCTGATCTGCGCATGTGCTGGG CTACGCGCGGGCGCAAGCGCCAAGAGCGGCTGCGTCTATGGTCATGACGTCTGACAGAGCGTCC ACCCGTCTTCGACAGGACTCTATGGTTCTTACGCGCGCAGACAGACCGCCTATATAAGCCATGCGC AGGCGGAGGAGCGCCTCTTTCCCTTCGGTGTGGGGAGCAAGCGCAGTTGTCGTCTCTTGCGGTG CCGTCGCTGGTTCTCACACCTTTTAGGTCTGTTCTCGTCTTCCATGACCGAGTACAAGCCCACGGT GCGCCTCGCCACCCGCGACGACGTCCCCCGGGCCGTACGCACCCTCGCCGCCGCGTTCGCCGA CTACCCCGCCACGCGCCACACCGTCGACCCGGACCGCCACATCGAGCGGGTCACCGAGCTGCAA GAACTCTTCCTCACGCGCGTCGGGCTCGACATCGGCAAGGTGTGGGTCGCGGACGACGGCGCCG CGGTGGCGGTCTGGACCACGCCGGAGAGCGTCGAAGCGGGGGCGGTGTTCGCCGAGATCGGCC CGCGCATGGCCGAGTTGAGCGGTTCCCGGCTGGCCGCGCAGCAACAGATGGAAGGCCTCCTGGC GCCGCACCGGCCCAAGGAGCCCGCGTGGTTCCTGGCCACCGTCGGCGTCTCGCCCGACCACCA GGGCAAGGGTCTGGGCAGCGCCGTCGTGCTCCCCGGAGTGGAGGCGGCCGAGCGCGCCGGGG TGCCCGCCTTCCTGGAGACCTCCGCGCCCCGCAACCTCCCCTTCTACGAGCGGCTCGGCTTCACC GTCACCGCCGACGTCGAGGTGCCCGAAGGACCGCGCACCTGGTGCATGACCCGCAAGCCCGGTG CCTGACTGCAGGATCCAGACATGATAAGATACATTGATGAGTTTGGACAAACCACAACTAGAATGCA GTGAAAAAAATGCTTTATTTGTGAAATTTGTGATGCTATTGCTTTATTTGTAACCATTATAAGCTGCAA TAAACAAGTTAACAACAACAATTGCATTCATTTTATGTTTCAGGTTCAGGGGGAGGTGTGGGAGGTT TTTTAAAGCAAGTAAAACCTCTACAAATGTGGTAGTCGACCGATGCCCTTGAGAGCCTTCAACCC AGTCAGCTCCTTCCGGTGGGCGCGGGGCATGACTATCGTCGCCGCACTTATGACTGTCTTCTT TATCATGCAACTCGTAGGACAGGTGCCGGCAGCGCTCTTCCGCTTCCTCGCTCACTGACTCG CTGCGCTCGGTCGTTCGGCTGCGGCGAGCGGTATCAGCTCACTCAAAGGCGGTAATACGGTT ATCCACAGAATCAGGGGATAACGCAGGAAAGAACATGTGAGCAAAAGGCCAGCAAAAGGC CAGGAACCGTAAAAAGGCCGCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTGACGAGC ATCACAAAAATCGACGCTCAAGTCAGAGGTGGCGAAACCCGACAGGACTATAAAGATACCA GGCGTTTCCCCCTGGAAGCTCCCTCGTGCGCTCTCCTGTTCCGACCCTGCCGCTTACCGGATA CCTGTCCGCCTTTCTCCCTTCGGGAAGCGTGGCGCTTTCTCATAGCTCACGCTGTAGGTATCT CAGTTCGGTGTAGGTCGTTCGCTCCAAGCTGGGCTGTGTGCACGAACCCCCCGTTCAGCCCG ACCGCTGCGCCTTATCCGGTAACTATCGTCTTGAGTCCAACCCGGTAAGACACGACTTATCG CCACTGGCAGCAGCCACTGGTAACAGGATTAGCAGAGCGAGGTATGTAGGCGGTGCTACAG AGTTCTTGAAGTGGTGGCCTAACTACGGCTACACTAGAAGAACAGTATTTGGTATCTGCGCT CTGCTGAAGCCAGTTACCTTCGGAAAAAGAGTTGGTAGCTCTTGATCCGGCAAACAAACCAC CGCTGGTAGCGGTGGTTTTTTTGTTTGCAAGCAGCAGATTACGCGCAGAAAAAAAGGATCTC AAGAAGATCCTTTGATCTTTTCTACGGGGTCTGACGCTCAGTGGAACGAAAACTCACGTTAA GGGATTTTGGTCATGAGATTATCAAAAAGGATCTTCACCTAGATCCTTTTAAATTAAAAATG AAGTTTTAAATCAATCTAAAGTATATATGAGTAAACTTGGTCTGACAGCGGCCGCAAATGCT AAACCACTGCAGTGGTTACCAATGCTTAATCAGTGAGGCACCTATCTCAGCGATCTGTCTAT TTCGTTCATCCATAGTTGCCTGACTCCCCGTCGTGTAGATAACTACGATACGGGAGGGCTTA CCATCTGGCCCCAGCGCTGCGATGATACCGCGAGAACCACGCTCACCGGCTCCGGATTTATC AGCAATAAACCAGCCAGCCGGAAGGGCCGAGCGCAGAAGTGGTCCTGCAACTTTATCCGCC TCCATCCAGTCTATTAATTGTTGCCGGGAAGCTAGAGTAAGTAGTTCGCCAGTTAATAGTTT GCGCAACGTTGTTGCCATCGCTACAGGCATCGTGGTGTCACGCTCGTCGTTTGGTATGGCTTC ATTCAGCTCCGGTTCCCAACGATCAAGGCGAGTTACATGATCCCCCATGTTGTGCAAAAAAG CGGTTAGCTCCTTCGGTCCTCCGATCGTTGTCAGAAGTAAGTTGGCCGCAGTGTTATCACTCA TGGTTATGGCAGCACTGCATAATTCTCTTACTGTCATGCCATCCGTAAGATGCTTTTCTGTGA CTGGTGAGTACTCAACCAAGTCATTCTGAGAATAGTGTATGCGGCGACCGAGTTGCTCTTGC CCGGCGTCAATACGGGATAATACCGCGCCACATAGCAGAACTTTAAAAGTGCTCATCATTGG AAAACGTTCTTCGGGGCGAAAACTCTCAAGGATCTTACCGCTGTTGAGATCCAGTTCGATGT AACCCACTCGTGCACCCAACTGATCTTCAGCATCTTTTACTTTCACCAGCGTTTCTGGGTGAG CAAAAACAGGAAGGCAAAATGCCGCAAAAAAGGGAATAAGGGCGACACGGAAATGTTGAA TACTCATACTCGTCCTTTTTCAATATTATTGAAGCATTTATCAGGGTTACTAGTACGTCTCTC AAGGATAAGTAAGTAATATTAAGGTACGGGAGGTATTGGACAGGCCGCAATAAAATATCTT TATTTTCATTACATCTGTGTGTTGGTTTTTTGTGTGAATCGATAGTACTAACATACGCTCTCCA TCAAAACAAAACGAAACAAAACAAACTAGCAAAATAGGCTGTCCCCAGTGCAAGTGCAGGT GCCAGAACATTTCTCT

340

Appendix II The following table illustrates structures and IC50 ratios of 28 hit compounds identified in the ten-point dose response experiment of the 30K small compound library. Please refer to section 3.2.4.2 for detailed description.

Compound Structure IC50 IC50 ratio ID GAPDH TUBB3 (GAPDH readout readout IC50/ TUBB3 assay assay IC50) WECC- O 1.43 3.60 0.40 N

S 0001549 N

WECC- 2.91 5.18 0.56 0001696 S N

N O

O WECC- 6.26 9.96 0.63 O 0002640 S N N O

WECC- 3.58 5.67 0.63 O O

0002641 N N

N O WECC- F 0.72 1.27 0.57 0002675

N Z O

N S N O O WECC- O 1.79 2.64 0.68

O 0003197 S O N

S N

WECC- 2.30 3.71 0.62 O 0004165 S N N

341

WECC- 20.06 12.55 1.60 0011277 N O E N S

WECC- O 0.66 1.02 0.64 O

N 0013536 E

WECC- N 1.82 3.63 0.50

N 0014000 E N

S O

O F F WECC- 0.93 1.58 0.59 N 0014349 N O

WECC- 23.49 14.44 1.63 0017313 N

WECC- O 24.40 5.77 4.23 0017371 N N N O WECC- O 3.61 0.87 4.13 0018639

O WECC- 16.03 15.26 1.05

O 0022568 N O

O N S O S WECC- O 0.63 1.15 0.55

N 0023257 O N O O

WECC- 2.17 4.77 0.45 0023771 O N

O O

S WECC- S 0.44 0.72 0.61

0025058 N

O O N O O

342

WECC- S 2.48 3.23 0.77 Cl

0077243 O N N O N N WECC- N 1.27 2.19 0.58 N

0081203 N O

WECC- N 18.95 12.27 1.54

0081582 S N

O N

WECC- N 15.22 9.26 1.64 0093348

N N O

N N S

WECC- N 3455125.14 582471939. 0.01 N 0098927 24

N N

WECC- N 0.69 1.08 0.64 0099078 N N

WECC- S N 2.65 7.00 0.38 0099090

N N O

WECC- O 5.64 7.75 0.73 O N

0116589 O

Z WECC- N 5.32 11.15 0.48 0118298 O N O N

WECC- O 0.01 22.56 0.00 0119079

N F O F

343

Appendix III

The following table illustrates structures and IC50 ratios of 16 hit compounds identified in the ten-point dose response experiment of the 30K small compound library. Please refer to section 3.2.4.2 for detailed description.

Compound Structure IC50 IC50 ratio ID TUBB3 GAPDH (GAPDH readout readout IC50/ TUBB3 assay assay IC50) WECC- O 5.21 4.06 0.78 S N S O

N 0001587 N O WECC- 0.04 undefined undefined

0003647 N N O O S N O

O N WECC- S 0.31 0.34 1.10 S Cl O 0004514 N O O

WECC- 2.08 10.42 5.01 Cl 0004515 O S O N S O N

WECC- O N 0.38 0.30 0.79 O S N 0008214 O

WECC- O 0.24 0.48 2 N 0022626

O O N S

Cl WECC- O 2.18 4.85 2.22 N 0022629

O N S

O S

WECC- O 0.26 0.46 1.77 N 0022631

O N S

344

WECC- O 5.57 20 3.59 N 0023807

O O N S

F O WECC- Z 2.45 20 8.16 N E 0024534 N

O N S Cl O

WECC- O O 5.17 10.69 2.07 S Z N

0024767 O N O

WECC- O 2.40 4.56 1.9 Cl 0024986 N S N N N

N WECC- O 1.99 3.04 1.53 O 0081431 S N

WECC- O 1.79 0.78 0.44

S 0110171 O N

C l

F F

O WECC- N 3.25 1.80 0.55

0117455 O S S O O WECC- F 1.38 2.37 1.72

0119863 O

N S N

Appendix IV The following table consists the top 100 differentially expressed genes in WECC0017371-treated H460 cells at 48 h.

Fold Adjusted p- Symbol Name change value (FC) LINC01239 long intergenic non-protein coding RNA 1239 6.115E-01 2.819 ESM1 endothelial cell-specific molecule 1 1.643E-02 2.725 CXCL8 chemokine (C-X-C motif) ligand 8 2.086E-03 2.673 GJB2 gap junction protein, beta 2, 26kDa 8.770E-03 2.474 FOXQ1 forkhead box Q1 1.402E-04 2.194

345

NR4A2 nuclear receptor subfamily 4, group A, member 2 4.197E-03 2.115 DHRS2 dehydrogenase/reductase (SDR family) member 2 8.891E-02 2.108 ESM1 endothelial cell-specific molecule 1 8.770E-03 2.096 LOC100506123 uncharacterized LOC100506123 7.388E-02 2.046 NR4A2 nuclear receptor subfamily 4, group A, member 2 1.281E-01 2.009 CXCL8 chemokine (C-X-C motif) ligand 8 4.452E-03 1.979 CYP4V2 cytochrome P450, family 4, subfamily V, polypeptide 2 2.243E-02 1.965 ZNF789 zinc finger protein 789 1.643E-02 1.957 SNURF SNRPN upstream reading frame 1.119E-01 1.924 FZD4 frizzled class receptor 4 2.086E-03 1.910 KLF2 Kruppel-like factor 2 2.624E-03 1.864 NPHP4 nephronophthisis 4 2.772E-03 1.802 STOX1 storkhead box 1 2.518E-01 1.794 SNORA68 small nucleolar RNA, H/ACA box 68 1.105E-01 1.789 HIP1R huntingtin interacting protein 1 related 1.376E-01 1.783 OR4A16 olfactory receptor, family 4, subfamily A, member 16 7.735E-02 1.780 TMEM154 transmembrane protein 154 1.944E-01 1.757 NAV2 neuron navigator 2 3.205E-02 1.755 FAM73B family with sequence similarity 73, member B 1.796E-01 1.753 MMP23B matrix metallopeptidase 23B 1.563E-01 1.752 TFPI2 tissue factor pathway inhibitor 2 1.146E-02 1.746 OGT O-linked N-acetylglucosamine (GlcNAc) transferase 1.671E-01 1.745 ZNF695 zinc finger protein 695 6.398E-02 1.734 AMDHD1 amidohydrolase domain containing 1 3.220E-03 1.715 solute carrier family 25 (mitochondrial carrier; phosphate SLC25A24 8.891E-02 1.713 carrier), member 24 NKD2 naked cuticle homolog 2 (Drosophila) 2.153E-02 1.710 ATG12 autophagy related 12 3.843E-01 1.701 HSPA13 heat shock protein 70kDa family, member 13 1.379E-01 1.695 LOC100131289 uncharacterized LOC100131289 3.272E-01 1.692 CD55 molecule, decay accelerating factor for complement CD55 1.885E-02 1.686 (Cromer blood group) GRAMD1B GRAM domain containing 1B 5.097E-03 1.684 FAM234A family with sequence similarity 234, member A 8.770E-03 1.683 RDM1 RAD52 motif containing 1 2.698E-01 1.682 FKBP11 FK506 binding protein 11, 19 kDa 5.097E-03 1.681 CLK3 CDC-like kinase 3 2.212E-01 1.673 REEP1 receptor accessory protein 1 8.891E-02 1.669 NPM1 nucleophosmin (nucleolar phosphoprotein B23, numatrin) 4.615E-01 1.659 TICRR TOPBP1-interacting checkpoint and replication regulator 1.174E-01 1.656 COG3 component of oligomeric golgi complex 3 1.342E-01 1.652 FAM207A family with sequence similarity 207, member A 1.921E-01 1.650 C4A complement component 4A (Rodgers blood group) 1.944E-01 1.648 ADAM10 ADAM metallopeptidase domain 10 3.884E-01 1.647 TXNIP thioredoxin interacting protein 5.097E-03 1.644

346

ADARB1 adenosine deaminase, RNA-specific, B1 1.620E-01 1.638 GOLGA1 golgin A1 8.729E-02 1.638 CD226 CD226 molecule 1.880E-02 -1.766 PHEX phosphate regulating endopeptidase homolog, X-linked 3.458E-02 -1.770 KRT80 keratin 80, type II 6.202E-03 -1.771 AXL AXL receptor tyrosine kinase 2.699E-02 -1.772 potassium channel, voltage gated KQT-like subfamily Q, KCNQ2 2.997E-02 -1.772 member 2 FBXL13 F-box and leucine-rich repeat protein 13 5.527E-02 -1.777 ATP-binding cassette, sub-family G (WHITE), member 2 ABCG2 5.038E-02 -1.779 (Junior blood group) PLAC8 placenta-specific 8 7.553E-02 -1.780 IL9 interleukin 9 2.381E-01 -1.783 CLDN7 claudin 7 1.119E-01 -1.789 SH2D3A SH2 domain containing 3A 1.658E-01 -1.790 CD14 CD14 molecule 6.643E-03 -1.800 SCARNA23 small Cajal body-specific RNA 23 5.038E-02 -1.819 CRYAB crystallin, alpha B 4.614E-04 -1.825 HRCT1 histidine rich carboxyl terminus 1 7.126E-03 -1.839 BMP5 bone morphogenetic protein 5 7.735E-02 -1.845 FGF2 fibroblast growth factor 2 (basic) 1.552E-01 -1.860 STMN3 stathmin-like 3 6.902E-02 -1.869 IFI44L interferon-induced protein 44-like 1.643E-02 -1.872 TMEM27 transmembrane protein 27 7.735E-02 -1.881 DNAAF3 dynein, axonemal, assembly factor 3 7.388E-02 -1.882 NMNAT2 nicotinamide nucleotide adenylyltransferase 2 6.580E-02 -1.889 PHLDA3 pleckstrin homology-like domain, family A, member 3 6.987E-02 -1.905 NEFH neurofilament, heavy polypeptide 9.634E-03 -1.918 PDK4 pyruvate dehydrogenase kinase, isozyme 4 1.643E-02 -1.918 GLIPR1 GLI pathogenesis-related 1 2.144E-02 -1.923 CES1 carboxylesterase 1 2.242E-02 -1.923 ZNF467 zinc finger protein 467 2.772E-03 -1.935 KRT7 keratin 7, type II 9.517E-02 -1.940 CLCF1 cardiotrophin-like cytokine factor 1 5.097E-03 -1.962 ZFP64 ZFP64 zinc finger protein 7.735E-02 -2.008 COL16A1 collagen, type XVI, alpha 1 1.789E-02 -2.016 GJA5 gap junction protein, alpha 5, 40kDa 1.364E-02 -2.031 TERC telomerase RNA component 3.747E-02 -2.039 RNU5A-1 RNA, U5A small nuclear 1 2.243E-02 -2.119 C1QTNF6 C1q and tumor necrosis factor related protein 6 2.772E-03 -2.217 ABCA1 ATP-binding cassette, sub-family A (ABC1), member 1 5.408E-04 -2.229 CYSRT1 cysteine-rich tail protein 1 5.097E-03 -2.288 CD14 CD14 molecule 7.234E-04 -2.311 C5orf46 chromosome 5 open reading frame 46 4.089E-03 -2.319 PLAC8 placenta-specific 8 3.286E-03 -2.362

347

KRT81 keratin 81, type II 1.402E-04 -2.426 ANKRD1 ankyrin repeat domain 1 (cardiac muscle) 3.011E-03 -2.432 OTOF otoferlin 2.772E-03 -2.501 KRT80 keratin 80, type II 9.678E-03 -2.577 SOST sclerostin 3.220E-03 -2.637 S100A3 S100 calcium binding protein A3 3.289E-03 -2.874 IGFBP3 insulin-like growth factor binding protein 3 1.668E-04 -3.204 SNORA12 small nucleolar RNA, H/ACA box 12 1.839E-04 -3.445

Appendix V The following table consists the top 100 differentially expressed genes in WECC0017371-treated H460 cells at 72 h.

Adjusted Fold Symbol Gene p-value change INSIG1 insulin induced gene 1 3.10E-06 3.807 HSPA1A heat shock 70kDa protein 1A 4.98E-05 3.525 MSMO1 methylsterol monooxygenase 1 2.95E-05 3.381 NPTX2 neuronal pentraxin II 2.24E-05 3.282 HSPA1B heat shock 70kDa protein 1B 5.37E-06 3.191 HMGCS1 3-hydroxy-3-methylglutaryl-CoA synthase 1 (soluble) 1.82E-06 3.042 HSPA8 heat shock 70kDa protein 8 3.10E-06 3.001 phospholysine phosphohistidine inorganic LHPP 7.17E-04 2.910 pyrophosphate phosphatase ESM1 endothelial cell-specific molecule 1 1.78E-03 2.796 HSPA8 heat shock 70kDa protein 8 3.35E-06 2.792 potassium channel, voltage gated modifier subfamily F, KCNF1 6.79E-05 2.764 member 1 RHOU ras homolog family member U 9.16E-06 2.736 MSMO1 methylsterol monooxygenase 1 2.18E-05 2.714 C10orf90 chromosome 10 open reading frame 90 6.88E-04 2.573 HSPA8 heat shock 70kDa protein 8 3.80E-05 2.554 TUBB3 tubulin, beta 3 class III 1.13E-04 2.515 FAM46C family with sequence similarity 46, member C 1.39E-03 2.496 ST6 (alpha-N-acetyl-neuraminyl-2,3-beta-galactosyl- ST6GALNAC6 1,3)-N-acetylgalactosaminide alpha-2,6- 1.13E-04 2.454 sialyltransferase 6 MAP6D1 MAP6 domain containing 1 9.60E-05 2.429 STOX1 storkhead box 1 1.07E-02 2.428 CDC27 cell division cycle 27 9.51E-03 2.420 HMGCR 3-hydroxy-3-methylglutaryl-CoA reductase 1.61E-04 2.376 GJB2 gap junction protein, beta 2, 26kDa 1.39E-03 2.376 LPIN1 lipin 1 4.31E-05 2.358 ACYP2 acylphosphatase 2, muscle type 6.95E-03 2.351 IDI1 isopentenyl-diphosphate delta isomerase 1 1.71E-04 2.336 DHCR24 24-dehydrocholesterol reductase 1.11E-03 2.330

348

TOR1AIP2 torsin A interacting protein 2 5.56E-04 2.317 TUBB tubulin, beta class I 1.78E-04 2.299 TUBB8 tubulin, beta 8 class VIII 1.19E-04 2.295 SHCBP1 SHC SH2-domain binding protein 1 1.39E-03 2.280 ERCC6L excision repair cross-complementation group 6-like 3.92E-03 2.278 HSPA2 heat shock 70kDa protein 2 2.98E-04 2.264 ESM1 endothelial cell-specific molecule 1 5.69E-04 2.257 OGT O-linked N-acetylglucosamine (GlcNAc) transferase 7.70E-05 2.252 phosphatidylinositol glycan anchor biosynthesis, class PIGL 6.00E-04 2.245 L ERCC6L excision repair cross-complementation group 6-like 5.09E-04 2.239 EFNA4 ephrin-A4 2.16E-04 2.233 SH3BGR SH3 domain binding glutamate-rich protein 1.88E-02 2.231 TUBB4B tubulin, beta 4B class IVb 1.51E-04 2.230 LRRC26 leucine rich repeat containing 26 5.30E-05 2.225 NR4A2 nuclear receptor subfamily 4, group A, member 2 3.38E-04 2.225 TAF5 RNA polymerase II, TATA box binding protein TAF5 2.69E-04 2.223 (TBP)-associated factor, 100kDa REEP1 receptor accessory protein 1 1.31E-03 2.219 MIR1282 microRNA 1282 2.86E-01 2.196 TMEM91 transmembrane protein 91 1.49E-04 2.187 HELZ2 helicase with zinc finger 2, transcriptional coactivator 5.57E-03 2.159 NTSR1 neurotensin receptor 1 (high affinity) 1.69E-03 2.148 MERTK MER proto-oncogene, tyrosine kinase 4.44E-04 2.147 FAM172A family with sequence similarity 172, member A 1.43E-02 2.123 IGFBP3 insulin-like growth factor binding protein 3 9.85E-05 -2.597 TBC1D19 TBC1 domain family, member 19 2.28E-04 -2.629 solute carrier family 6 (neurotransmitter transporter, SLC6A9 5.09E-04 -2.639 glycine), member 9 WARS tryptophanyl-tRNA synthetase 4.10E-05 -2.652 RCOR2 REST corepressor 2 5.09E-04 -2.655 SMOC1 SPARC related modular calcium binding 1 5.11E-03 -2.675 KRT86 keratin 86, type II 6.91E-05 -2.679 FBXL13 F-box and leucine-rich repeat protein 13 3.32E-04 -2.680 ACOT2 acyl-CoA thioesterase 2 9.82E-04 -2.685 PSPH phosphoserine phosphatase 2.92E-05 -2.687 GADD45A growth arrest and DNA-damage-inducible, alpha 5.49E-05 -2.696 IGFBP3 insulin-like growth factor binding protein 3 5.04E-05 -2.706 SLPI secretory leukocyte peptidase inhibitor 7.54E-04 -2.719 TERC telomerase RNA component 7.51E-04 -2.725 CTGF connective tissue growth factor 5.45E-05 -2.733 WARS tryptophanyl-tRNA synthetase 2.04E-05 -2.734 FGG fibrinogen gamma chain 5.45E-05 -2.751 DEFB1 defensin, beta 1 2.10E-05 -2.755 IFRD1 interferon-related developmental regulator 1 6.95E-05 -2.759 SLC48A1 solute carrier family 48 (heme transporter), member 1 1.02E-04 -2.761 349

CARS cysteinyl-tRNA synthetase 1.49E-04 -2.794 SCARNA16 small Cajal body-specific RNA 16 5.61E-04 -2.821 potassium channel, voltage gated subfamily E KCNE4 2.75E-05 -2.841 regulatory beta subunit 4 WISP2 WNT1 inducible signaling pathway protein 2 1.14E-03 -2.855 SERTAD4 SERTA domain containing 4 2.16E-04 -2.887 SNHG1 small nucleolar RNA host gene 1 1.12E-05 -2.954 ACOT2 acyl-CoA thioesterase 2 2.75E-05 -2.961 PPP1R15A protein phosphatase 1, regulatory subunit 15A 3.59E-06 -2.964 PAPPA pregnancy-associated plasma protein A, pappalysin 1 2.16E-04 -2.966 SCARNA23 small Cajal body-specific RNA 23 1.87E-04 -2.968 SNORA79 small nucleolar RNA, H/ACA box 79 1.16E-04 -3.011 EPSTI1 epithelial stromal interaction 1 (breast) 1.41E-05 -3.016 RAB39B RAB39B, member RAS oncogene family 3.32E-04 -3.017 SUCNR1 succinate receptor 1 2.98E-04 -3.039 CBS cystathionine-beta-synthase 3.35E-06 -3.140 KRT81 keratin 81, type II 1.82E-06 -3.206 ULBP1 UL16 binding protein 1 2.75E-05 -3.236 CLDN7 claudin 7 2.98E-04 -3.362 CARS cysteinyl-tRNA synthetase 9.00E-05 -3.425 CLGN calmegin 2.42E-06 -3.497 KRT7 keratin 7, type II 3.86E-04 -3.586 PDK4 pyruvate dehydrogenase kinase, isozyme 4 2.40E-05 -3.614 ADM2 adrenomedullin 2 6.94E-06 -3.796 DDIT3 DNA-damage-inducible transcript 3 2.04E-05 -3.852 CD226 CD226 molecule 5.69E-06 -4.012 LOC344887 NmrA-like family domain containing 1 pseudogene 1.82E-06 -4.133 INHBE inhibin, beta E 3.77E-06 -5.341 SNORA12 small nucleolar RNA, H/ACA box 12 1.53E-05 -5.388 ANKRD1 ankyrin repeat domain 1 (cardiac muscle) 3.35E-06 -5.623 GDF15 growth differentiation factor 15 6.74E-07 -8.097

350