Search

Hypoxia is defined as the state in which the availability or delivery of oxygen is insufficient to meet tissue demand. It occurs particularly in aggressive, fast-growing tumours in which the rate of new blood vessel formation (angiogenesis) cannot match the growth rate of tumour cells. Cellular stresses such as hypoxia can cause cells to undergo apoptosis; however some tumour cells adapt to hypoxic conditions and evade apoptosis. Tumour hypoxia has been linked to poor prognosis and to greater resistance to existing cancer therapies. This thesis provides evidence that alterations in alternative splicing patterns of key genes is one method tumour cells adapt to hypoxia. This study confirms a hypoxic-induced change in the alternative splicing of carbonic anhydrase IX (CA IX) following 1% oxygen treatment. CA IX is one of the best studied hypoxia markers, involved in maintaining an intracellular pH that favours tumour cell growth. Furthermore, evidence is provided here that in PC3 cells the regulation of CA IX splicing involves the SAFB1 and PRPF8 splice factors. Additionally, SAFB1 expression is shown to decrease in hypoxia. This study further demonstrates that alternative splicing patterns of previously documented cancer-associated genes are altered in hypoxia. PCR analysis showed that hypoxia significantly altered the alternative splicing of apoptotic-associated genes: caspase-9; Mcl-1; Bcl-x; survivin. The expression of the pro-apoptotic isoforms of the first two genes, and the anti-apoptotic isoforms of the latter two genes were favoured by hypoxia. Furthermore, high-throughput PCR analysis provided evidence of significant changes in the alternative splicing of several other cancer-associated genes in hypoxia: APAF1; BTN2A2; CDC42BPA; FGFR1OP; MBP; PTPN13; PUF60; RAP1GDS1; TTC23; UTRN. Most notably, the pro-oncogenic isoforms of APAF1, BTN2A2 and RAP1GDS1 were favoured in hypoxia. The majority of alternative splicing changes were found in the PC3 cell line. However changes in alternative splicing patterns that mirrored those in the PC3 cell line were also found in the VCaP (CDC42BPA, RAP1GDS1 and UTRN) and PNT2 (BTN2A2, CDC42BPA, FGFR1OP and TTC23) cell lines. The mRNA expression of splice factors (SRSF1, SRSF2, SRSF3, SAM68, HuR and hnRNP A1) and splice factor kinases (CLK1 and SRPK1) were shown to significantly increase in hypoxia. Subsequent experiments provided evidence that CLK1 and SRSF1 protein expression also increased in hypoxia. The phosphorylation of SRSF4 and SRSF5 were demonstrated to increase in hypoxia. However, the phosphorylation of SRSF6 was not. In addition, siRNAs and chemical inhibitors of CLK1 (TG003) and SRPK1 (SPHINX) were used to assess the effect of these splice factor kinases on the subsequent splicing of cancer-associated genes. There were no significant changes to splicing found with SRPK1 siRNA knockdown or SPHINX treatment. However CLK1 siRNA knockdown and TG003 treatment demonstrated a shift in FGFR1OP splicing that mirrored the effect of hypoxia on FGFR1OP splicing. This suggests that CLK1 activity is inhibited in hypoxia. Furthermore, in contrast to previous research CLK1 was found to be localised to the cytoplasm in both normoxia and hypoxia in the PC3 cell line. This work has uncovered factors and provided an insight into mechanisms that are involved in alternative splicing changes in hypoxia in mammalian cell lines. It is hoped that these novel research findings will aid in the understanding of how cells adapt to hypoxia especially in regards to alternative splicing, and may offer future therapeutic targets in hypoxic tumours.

SR proteins are a family of splicing factors that plays important roles in regulating alternative splicing. The function of SR proteins is heavily regulated by the extent of phosphorylation of its C-terminal RS domain. Cdc-2 like kinases (CLKs) are known to hyper-phosphorylate the RS domains and control the splicing functions of SR proteins. CLKs have a folded kinase domain and an N-terminus that is predicted to be disordered. The N-terminus plays vital roles in regulating the function of CLK1 in various aspects. The studies presented here elucidate how the N-terminus modulates three different aspects of CLK1 function: sub-cellular localization, self-association to form oligomers, and substrate phosphorylation. Although prior studies had shown that the N-terminus was important for modulating the nuclear import of CLK1, the mechanism for this change in subcellular localization had largely been uninvestigated. Here, we show that CLK1 lacks a short, classical Nuclear Localization Sequence (NLS), indicating that the nuclear import is not mediated by the classical importin / system. Instead, we show that CLK1 enters the nucleus by forming a complex with its physiological substrate SRSF1, an SR protein prototype, in the cytoplasm and transportin SR-2 (TRN-SR2) imports the kinase-substrate complex into the nucleus. Previous studies from our laboratory had shown that the N-terminus induces oligomerization of CLK1, which helps the kinase select its physiological substrates over non-physiological ones. The nature of the interactions underlying this oligomerization was investigated and our results show that CLK1 oligomerization is driven not only by self-association of the N-terminus (N-N interactions) but also by interactions between the N-terminus and the kinase domain (N-K interactions). While interactions between the N-termini are mediated solely by aromatic residues, interactions between the N-terminus and kinase domain are electrostatic in nature. Lastly, we also investigated the role of the N-terminus in regulating SRSF1 hyper-phosphorylation. Our results show a strong correlation between CLK1 quaternary structure and substrate phosphorylation activity. While substrate binding affinity is solely regulated by the length of the N-terminus, the velocity of hyper-phosphorylation is tightly regulated by the quaternary structure of CLK1 oligomers. Our studies demonstrate that the N-terminus of CLK1 is highly versatile. It is important not only for recognizing a broad range of RS domains for essential SR protein hyper-phosphorylation but also for CLK1 nuclear localization through substrate “piggybacking.”

This thesis details efforts to generate chemical probes for the interrogation of important protein targets: the CLK and DYRK kinases, ceramide synthases (CerS), and the PP2A-inhibitory protein, SBDS. Modulation of activity with chemical probes and protein degradation with PROTACs were investigated. Chapter 1 provides an introduction to chemical probes and PROTACs, introduces the relevant protein targets, and provides background on previous work by the Morris group. Chapters 2-4 are focused on the development of chemical probes and PROTACs for the CLK and DYRK kinases. In Chapter 2, an analysis of crystal structures and inhibitors of CLK1 and DYRK1a resulted in elucidation of the factors that determine selective inhibition. This analysis led to the investigation of 6-azaindole 2.3—previously reported as a DYRK1a-selective inhibitor— and it was found to be a potent inhibitor of CLK1, CLK2, and DYRK1a. Insights gained from the structural analysis were used to design quinoline 2.36, a CLK1 inhibitor with 15-fold selectivity over DYRK1a. Work in Chapter 3 led to improved pharmacophore models for selectivity and potency, which were used to develop novel potent CLK/DYRK inhibitors, including 3-quinoline 3.44, a 1.2 nM inhibitor of DYRK1a with 5.4-fold selectivity over CLK1. Chapter 4 describes a preliminary study of 6-azaindole PROTACs targeted towards CLK1/DYRK1a. A small library of PROTACs based on amide 4.5 were synthesised and initial biological evaluation by our collaborators did not indicate degradation of CLK1 or DYRK1a at 1 µM across multiple treatment times, possibly due to insufficient formation of stable ternary complexes. Chapter 5 outlined investigations with the AAL(S) scaffold. A preliminary investigation of PROTACs for the degradation of CerSs was carried out, with a small library of CerS PROTACs synthesised. These influenced cellular ceramide concentrations, via a yet-to-be-determined mechanism. The second part of Chapter 5 was focused on the elucidation of how AAL(S) binds to the PP2A-inhibitory protein SBDS. This binding site was confirmed using pull-down studies with mutated SBDS proteins. With this structural information improved activators of PP2A can be developed as well as PROTACs targeted towards SBDS. Conclusions and future directions are provided in Chapter 6, and experimental procedures in Chapter 7.