FUNCTIONAL CHARACTERISATION OF ASAP1 IN HCC PROGRESSION AND METASTASIS

Acknowledgements This thesis would not have been possible if not for the kind guidance from Dr Dan Yock Young to whom I owe a debt of gratitude. I would like to extend my heartfelt gratitude to my mentor, Dr Tony Wong K.F., for his tutelage and patience during this time and Dr Zhou Lei for his professional expertise and unwavering support in the in vivo work. 2 Table of contents Declaration 1 Acknowledgements 2 Summary 4 Materials & Methods 5 Introduction 12 Results 21 Discussion 42 References 46 3 Summary The aim of the study was to perform a functional characterisation of ASAP1, a gene that has been implicated in a broad number of cancers and reported to be an important regulator of tumour cell invasion and metastasis, in the context of HCC. Its contribution towards cancer metastasis will be addressed by in vitro as well as in vivo work. Coupled with HCC clinical data, we will demonstrate the utility of ASAP1 as a biomarker or even as a therapeutic drug target in the treatment of HCC metastasis. 4 Materials & Methods Cell culture Human HCC cell line SK-Hep1 and normal human hepatocyte cell line LO2 were grown in 10-cm diameter or six well tissue culture plates (BD) and maintained in Dulbecco’s modified Eagle’s medium (DMEM) (Invitrogen) with high glucose and supplemented with 10% heat inactivated fetal bovine serum (Invitrogen), at 37̊C in a humidified atmosphere of 5% CO2. Plasmid construct and transfection The commercial cDNA clone of the human ASAP1 ORF (Origene) was generated by transforming TOP10 competent E.coli cells (Invitrogen). The amplified plasmids were subsequently purified using the Wizard Plus SV Minipreps DNA Purification System (Promega) according to the supplier’s instructions. Cells plated into six-well plates were transfected with 4000 – 60000 ng of plasmid DNA in Lipofectamine 2000 reagent (Invitrogen) based on the manufacturer’s recommendations. Establishment of knockdown cells For RNA interference assays, small interfering RNA (siRNA) duplex targeting ASAP1 (sense: 5’-GGAUAUCAGUAUUGACAAAtt-3’; antisense: UUUGUCAAUACUGAUAUCCat-3’) was purchased from 5’- Ambion. Mock(Lipofectamine 2000) and control siRNA (ON-TARGETplus nontargeting control siRNA) from Dharmacon were used as controls. To examine the siRNA-mediated knock down effect on ASAP1 expression, SK-Hep1 cells were seeded in six-well plates in antibiotic-free complete DMEM. After 16 5 hours, the cells were transfected with 40 – 100 nmol/L siRNA in Lipofectamine 2000. Briefly, a complex of Lipofectamine 2000 and siRNA in Opti-MEM (500 µl) was gently added to each well. After eight hours, the transfection mixture was replaced with fresh antibiotic-free complete medium. Mock and the control siRNA were added in separate wells. The cells were subsequently used for downstream assays 24 – 48 hours post transfection. Quantitative real-time PCR Total RNA from the cells was isolated using the RNeasy Mini kit (Qiagen). 0.5 µg of total RNA was reversibly transcribed into the first-strand cDNA synthesis using poly dT12-18 primers and Superscript II reverse transcriptase (Invitrogen). cDNA products were diluted fivefold for subsequent PCR amplification. For quantitative real-time PCR (qPCR)analysis, cDNA aliquots were amplified using the RT2 SYBR Green qPCR Mastermix (Qiagen) following the manufacturer’s instructions and the expression of target genes was analysed using the ABI PRISM 7000 Sequence Detection System (Applied Biosystems). The cycling conditions consisted of 10 minutes at 95 ̊C for the initial heat activation ,40 cycles of 95 ̊C for 15 seconds and 60 ̊C for 1 minute, followed by a melting curve analysis. Ct value was measured during the exponential amplification phase, and the amplification plots were analysed using the SDS software (Applied Biosystems). Relative expression level (defined as the fold change) of the target gene is given by 2-∆∆Ct (∆Ct = ∆Cttarget-∆CtGAPDH; ∆∆Ct=∆CtASAP1 cDNA clone or ASAP1siRNA or control siRNA - ∆Ctmock) and normalised to the fold change detected in the corresponding control cells, which was defined as 1.0. All reactions were performed in triplicates. Primer sequences are listed as follows: 6 for GAPDH, forward 5’- GCTCTCTGCTCCTCCTGTTCGACA-3’ and CTGAGCGATGTGGCTCGGCT-3’; reverse ASAP1,forward GAGGCTCCCCCTCTGCCTCC-3’ and GTGGGCTGGGAGGGTCGGAT-3’; MMP-2, AGAAGGATGGCAAGTACGGCTTCT-3’ and AGTGGTGCAGCTGTCATAGGATGT-3’; MMP-9, TACCACCTCGAACTTTGACAGCGA-3’ GCCATTCACGTCGTCCTTATGCAA-3’; CCTATGCAGGGGTGGTCAAC-3’ reverse and β-catenin, and CGACCTGGAAAACGCCATCA-3’; E-cadherin, TGCCCAGAAAATGAAAAAGG-3’ and GTGTATGTGGCAATGCGTTC-3’; N-cadherin, ACAGTGGCCACCTACAAAGG-3’ and CCGAGATGGGGTTGATAATG-3’; Fibronectin, TCCCTCGGAACATCAGAAAC-3’ and CAGTGGGAGACCTCGAGAAG-3’; Vimentin, GAGAACTTTGCCGTTGAAGC-3’ and 5’5’5’- forward 5’- reverse 5’- forward reverse forward reverse forward reverse forward reverse forward reverse forward reverse 5’5’5’5’5’5’5’5’5’5’5’5’- GCTTCCTGTAGGTGGCAATC-3’. Gene expressions are expressed as mean ± S.E. Western Blot Cells were washed in ice-cold PBS and lysed in RIPA cell lysis buffer constituted with a protease inhibitor cocktail (Roche). Cell lysates were fractionated at 14,000 rpm and the protein concentration was determined using the Bradfordassay (Bio-Rad). Equal quantities of cell lysateswere mixed with 4 x loading dye, boiled for 10 minutes and resolved in 10% SDS7 polyacrylamide (PAGE) gels before being transferred onto polyvinylidene difluoride (PVDF) membranes (Millipore). Membranes were incubated in 5% non-fat dry milk for an hour followed by an overnight incubation at 4 ̊C with primary antibodies against ASAP1 (1:2500; Abcam) and β-actin (1:2000; Sigma-Aldrich). The membranes were subsequently washed with tris-buffered saline – 0.05% Tween-20 (TBST) prior to incubation with goat anti-mouse or rabbit horseradish peroxidase (HRP) – conjugated secondary antibodies (1:10000) for one hour at room temperature. The immunocomplexes were detected using ECL Prime Western blotting Detection System (Amersham) after exposure to X-ray film. Cell proliferation assay Transfected and mock control cells were seeded at a density of 20,000 cells per well in 48-well plates (200 µl/well) in complete medium containing 10% fetal bovine serum. Cell proliferation was monitored by counting cells daily using the hemocytometer. To corroborate the cell count data, proliferation was also assayed by incubating the cells with the alamarBlue cell viability reagent (Invitrogen) and fluorescence was routinely measured following two hours of incubation. Data presented are expressed as means and the error bars represent S.E.; n = 3. Cell adhesion assay Post transfection, cells from six-well plates were trypsinised and resuspended in serum-free medium. 500 µl of suspended cells (2 x 105 per well) were plated into 24-well plates coated with collagen (Inamed). The cells were incubated in 5% CO2 at 37 ̊C and washed twice with PBS to remove non8 adherent cells. The adherent cells were trypsinised at 40 minutes, 80 minutes and 2 hours post seeding. Cells collected at the indicated time points were counted using the hemocytometer. Data presented are expressed as means and the error bars represent S.E., n = 4. Transwell migration assay Post transfection, 200 µl of cells resuspended in serum-free medium were seeded (5 x 104 per well) into 8-µm pore Transwell filter inserts in 24-well plates (Corning). Complete medium containing 10% fetal bovine serum in the lower chamber of the well was used as a chemoattractant. After 48 hours of incubation at 37 ̊C, the inserts were removed. Cells that had passed through the filter onto the lower surface of the membrane as well as cells that had migrated to the bottom of the well were trypsinised and counted using the hemocytometer. Each control was carried out in triplicate wells per plate and each set of experiment was repeated thrice.Results are expressed as means and the error bars represent S.E., n = 3. Wound healing assay Cells were cultured in six-well plates and upon reaching confluence, the medium was removed and sterile pipette tips were drawn across the wells to produce clean 1-mm-wide wounds. Thewounded cell monolayerswere subsequently rinsed with PBS to remove cellular debris andincubated at 37 ̊C in complete media with reduced serum content (5% fetal bovine serum). Images of the scratch wound field were taken immediately, 24 and 48 hours post wounding under an inverted microscope at 40 X magnification.Wound closure at the indicated time points was measured using the Fiji software to 9 estimate the extent of cell migration and proliferation. Percentage wound healing data are expressed as means. The error bars represent S.E.; n = 3. Tumour xenograft mouse model Female NSG mice were housed under standard conditions and cared for according tothe institutional guidelines for animal care. In our cirrhotic mouse model, hepatic fibrogenesis and cirrhosis is induced with the administration of thioacetamide (200 mg/L) in the drinking water of the mice from the age of three months. In this study, NSG mice were orally administered with thioacetamide for five months to induce cirrhosis. Briefly, cirrhotic mice were anaesthetized with isoflurane inside the laminar flow cabinet. The abdomen was swabbed with ethanol and the spleen was exteriorised through a small left flank incision. 5 x 105 cells in 0.9% saline solution (300 µl) were introduced into the spleen slowly using a 27-gauge needle. The spleen was returned to the abdominal cavity and the incision site was closed with suture. After two weeks, the mice were euthanized with mouse anaesthesia (Ketamine, Medetomidine, and saline)by intraperitoneal injection. Each experimental group (mock-transfected and ASAP1siRNA-transfected SK-Hep1 cells) consisted of three mice. The spleen, liver and lungs were excised and embedded in paraffin. Serial sections (4µm) of the tissue were stained with hematoxylin and eosin (H&E) to visualise the tumour architecture. Statistics Unless otherwise indicated, numerical data are presented as the mean ± S.E. of three independent experiments.The SPSS statistical package for Windows (version 18; SPSS) was used for the data analysis. The clinicopathological 10 features in ASAP1-positive and -negative patients were compared using Pearson χ2 test for categorical variables and independent Student’s t test for continuous data. Differences in mRNA expression across groups were analysed using ANOVA. A p value less than 0.05 was considered statistically significant. 11 Introduction Liver cancer is the fifth most commonly diagnosed cancer and the third leading cause of cancer-related deaths worldwide[1].Hepatocellular carcinoma (HCC), one of the most aggressive and lethal liver malignancies, remains refractory to conventional therapeutic interventions in spite of recent advances in treatment strategies. Carcinogenesis of HCC is a complex multistep process involving multiple genetic aberrations. As the molecular basis for hepatocellular carcinogenesis remains largely unknown and given the bleak prognosis facing patients with advanced stages of the disease,further insight into the molecular mechanisms underlying HCC development and progression is necessary for improving prognosis. The development of new drug targets and reliable diagnostic biomarkers for tumour detection will culminate inimproved disease survival and management. The challenge of developing more specific and potent therapeutics for the treatment of HCC has always been compounded by the heterogeneous phenotypes manifested by the disease. These phenotypes are the result of genetic factors such as chromosomal aberrations, epigenetic modifications and copy number variations[2]. Like most types of genetic alterations, copy number variations (CNV) are associated with disease susceptibility and they have been reported to influence the phenotype and prognosis of the disease[3]. Through genomic DNA genotyping and mRNA profiling of HCC clinical samples, a CNV catalogue was generated (unpublished data) for a genomewide copy number analysis. Subsequently, a model delineating the mechanism by which CNV affects the HCC transcriptome was established. Based on the 12 derived model, six candidate genes postulated to drive HCC progression were shortlisted. Not only do these genes haveroles in various cellular processes contributing to tumour development and survival, they were also found to correlate with the patients’ overall and disease-free survivals. From a clinical perspective, the identification of oncogenic driver genes in HCC will contribute to more specific and potent therapeutics. This study will focus on the functional characterisation of one of them, ASAP1, an 8q24 gene encoding a multi-domain adaptor protein with ADP-ribosylation factor (Arf)GTPaseactivating protein (GAP) activity. Clinically, ASAP1 has been implicated in a broad range of tumours and it has been reported to be an important regulator of tumour cell motility, invasiveness and metastasis. This study aims to explore the role of ASAP1 in the HCC context and its contribution towards metastasis, a typical feature often associated with poor survival outcomes in patients despite having undergone surgical resection. Arf GAPs are a family of proteins that inactivate Arfs by inducing the hydrolysis of GTP bound to Arf[4]. These proteins contain a conserved Arf GAP domain bearing a zinc finger motif which mediates the catalysis[5]. The substrate, Arf.GTP, is involved in the regulation of membrane trafficking and actin reorganisation[6,7]. Apart from their Arf-dependent and Arf-independent effects on actin, Arf GAPs are also responsible for the regulation of membrane remodelling that accompanies actin polymerisation. In addition to the Arf GAP domain which contributes to one of their catalytic functions, Arf GAPs also contain protein-protein interaction and lipid interaction domains. Given the structurally diverse nature of these multi-domain proteins, Arf GAPs are 13 involved in cellular activities like migration and movement, which depend largely on the remodelling of the membrane and the actin cytoskeleton. Arfs belong to a family of GTP binding proteins which includes Arf-like (Arl) proteins, Ard and Arp [7, 8]. These proteins are highly conserved throughout all eukaryotes examined to date and were first identified as co-factors for cholera toxin-catalysed ADP ribosylation of Gs [9]. The primary physiological function of Arfs is to regulate endocytosis and vesicle trafficking by controlling the interaction of vesicle coat proteins with intracellular organelle membranes 12] [10- . Until recently, Arfs have been associated with the regulation of cytoskeletal remodelling [13-16]. Like other G proteins, Arfs are inactive in their GDP bound state and become activated when binding GTP. Cycling between their GDP- and GTP-bound states is accomplished by the interaction with effector proteins which mediates Arf function. For instance, Arf.GDP interaction with guanine nucleotide exchange factors facilitates the release of bound GDP to promote the formation of Arf.GTP.Conversely, inactivation is mediated by GAPs which recognise Arf.GTP and inducethe hydrolysis of bound GTP to GDP.As Arfs exhibit extremely slow intrinsic undetectableintrinsic GTPase activity nucleotide [17-19] exchange rates and , accessory proteins like guanine nucleotide exchange factors and Arf GAPs, which stimulate GTP binding and hydrolysis, are necessary for their function. Six mammalian isoforms of the Arf protein have been identified, of which Arf2 has been lost in humans. According to their structural similarities, these proteins are divided into three groups: class I (Arf1-3), class II (Arf4, 5) and 14 class III (Arf6). Arf1, Arf3, Arf5 and Arf6 are typically used as representatives of their classes. Among them, Arf1 and Arf6 have been the most extensively studied. Arf1 functions primarily at the perinuclear regions where it is involved in the secretory processes. Given that it predominantly localises to the internal membranes of the Golgi apparatus, Arf1 has been implicated in endoplasmic reticulum-to-Golgi transport, Golgi function, transport from the trans-Golgi network and transport in the endocytic pathway. Apart from affecting membrane transport in the Golgi apparatus and the endocytic compartment, Arf1 has also been reported to regulate focal adhesions and recruit paxillin to focal adhesions (FAs) [11, 20-28]. Arf6 localises to the cell membraneand endosomes, where it is known to affect endocytosis, phagocytosis, recycling of plasma membrane components including cell surface receptors, as well as actin-cytoskeletal remodelling at the cell periphery which includes the formation of filopodia, actin-rich protrusions and membrane ruffles [16,25,29-37]. There are 24 genes in the human genome which encode proteins bearing the Arf GAP domain[38]. ASAPs belong to a subgroup of Arf GAP proteins with a pleckstrin homology (PH) domain immediately N-terminal and ankyrin (ANK) repeats immediately C-terminal of the Arf GAP domain. Other prominent features of ASAPs include the Bin, amphiphysin and Rvs 161 and 167 (BAR) domain at the N-terminus and the proline-rich domain containing three proline-rich SH3 binding motifs located near the C-terminus.Two members of this group, ASAP1 and ASAP2, were named for the Src homology 3 (SH3) domain located at the C-terminus[39-41]. Despite lacking an 15 SH3 domain, ASAP3 is similar in terms of the phylogenetic analysis as well as the BAR, PH, Arf GAP, ANK repeat and proline-rich domains. Supporting evidence from bothin vitro and in vivo studies hasdemonstrated that Arf GAPs exhibit substrate specificity. Likewise, ASAPs are characterised by their use of specific Arf isozymes as substrates. ASAPs act on Arf1 and Arf5 in preference to Arf6 (Arf5 ≥ Arf1 > Arf6) as substrates in vitro, with more than a 200 fold difference between Arf1 and Arf6 [39, 40] . In vivo ASAP1 overexpression has been found to reduce Arf1.GTP levels while increasing Arf6.GTP levels [42]. To demonstrate this specificity, a point mutant of ASAP1 which could bind Arf-GTP but not induce GTP hydrolysis, was reported to co-localise with Arf1[43]. Most Arf GAPsare known to regulate cytoskeletal structures during the coordinated remodelling of the membrane and the actin cytoskeleton. These dynamic structures include focal adhesions (FAs), invadopodia, podosomes, lamellipodia and circular dorsal ruffles (CDRs)[25, 42, 44]. FAs, invadopodia and podosomes exist as points of contact between the actin cytoskeleton and the extracellular matrix (ECM)[45-53]. FAs consist of a coalescence of integrins, transmembrane heterodimeric receptors which bind to the substratum on the extracellular face of the plasma membrane and associate with proteins such as focal adhesion kinase (FAK), vinculin, paxillin, integrin-linked kinase and talin on the cytoplasmic side. Integrins also link bundles of actin filaments called stress fibers to the ECM[54-56]. Consequently, FAs appear as linear structures located behind the cell edge at the end of these stress fibers[57-61]. 16 Invadopodia are present in invasive cancer cell types and exist as points of attachment to the ECM. Podosomes are physiologically important for migration and invasion. Apart from integrins, invadopodia and podosomes also contain cortactin, polymerised actin and matrix metalloproteinase (MMP)[52,53]. Lamellipodia are projections from the cell edge that facilitate in driving cellular movement[57,61]. CDRs are ring-like structures extending from the dorsal surface of the cell [51, 62, 63] and their roles include mediating endocytosis of activated growth factor receptors and three-dimensional cell movement. ASAP1 has been reported to associate with four of these cytoskeletal structures in the remodelling of cell architecture. These structures are namely FAs, CDRs, invadopodia and podosomes. The regulatory role of ASAPs in actin remodelling was first established when ASAP1 and ASAP2 were found to bind FAK and Src, two non-receptor tyrosine kinasesthat are associated with the turnover of FAs [39-41, 44, 64] . Subsequently, ASAP1 was reported to bind to multiple proteins with regulatory or structural functions in the actin cytoskeleton. The cellular localisation of ASAP1 is consistent with its role in FA turnover and cell movement. ASAP1 cycles among the focal adhesions, the cell edge[25, 42, 44] and possibly endosomes at different points of maturation (Hirsch, D.S. and Randazzo, P.A. unpublished). Targeting ASAP1 to FAs involves the binding of its SH3 domain tothe proline-rich motif in FAK [44]and its proline-rich domain interaction with adaptor proteins like Crk (chicken tumour virus 10 regulator of kinase) and/or CrkL (Crk-like) [39, 65].Src binds to the proline-rich type I motifs on ASAP1 through its SH3 domain [39].The SH3 domain of ASAP1 mediates binding to proline-rich motifs in Pyk2 and the EH 17 domain-containing protein POB1[40, 44, 66] . POB1 can simultaneously bind to RalBP (Ral Binding Protein) which contains a RhoGAP domain [66] . The complex of ASAP1, POB1 and RalBP is hypothesized to link membrane trafficking with actin remodelling. ASAP1 has been found to bind cortactin [67] , an important component of invadopodia and podosomes[68]. ASAP1 also binds to CD2AP [69] and localises to membrane ruffles [70, 71] . Domains that target ASAP1 to CDRs have not yet been determined. FAs and membrane ruffles are associated in events leading to changes incell shape, movement and migration. Actin stress fibres bind to focal adhesions at the opposite cellular poles. The FAs and actin stress fibres are disassembled in order for cells to change shape or move[72-74]. This process is regulated by two Rho family proteins, Cdc42 and Rac, and by the two tyrosine kinases, Src and FAK[60, 75-80]. After FA dissociation, some of the proteins move into membrane ruffles at the edge of the cell where they begin to assemble under the influence of Cdc42, Rac, Src and FAK, into focal complexes. Activation of RhoA results in the formation of FAs from these complexes and the association of actin stress fibers with the FAs[57, 78, 81]. While the role of Arf GAPs in FAs has yet to be characterised, several speculations include: serving as scaffolds for proteins involved in actin regulation; trafficking of actin regulators; direct targets of the signalling cascades affecting the remodelling of the cytoskeleton. CDR formation accompanies the dissolution of FAs when fibroblasts are treated with PDGF [25, 48] NIH 3T3 fibroblasts [25] . ASAP1 overexpression decreases CDR number in and the expression of an ASAP1 mutant deficient in GAP activity has been shown to increase CDR number 18 [43] . Taken together, these findings indicate that the GAP activity of ASAP1 is involved in the dissolution of CDRs. As for invadopodia and podosome, it has been shown that ASAP1 suppression prevents their formation. Unlike CDR, formation of invadopodia and podsomes do not require GAP activity. It has been shown that the SH3 domain as well as a tyrosine residue on ASAP1 which gets phosphorylated by an activated Src is required instead (Bharti, S., Inoue, H., Bharti, K., Hirsch, D.S., Nie, Z., Yoon, H.-Y., Artym, V., Yamada, K.M., Mueller, S.C., Barr, V.A. and Randazzo, P.A., unpublished data). It is speculated that in the formation of invadopodia, the BAR domain of ASAP1 serves to either induce membrane tabulation [19] or recruit ASAP1 to the sites of membrane tubulation, where it functions as a scaffold protein. In line with [67] and bundles actin [68, 83] . Although ASAP1 this hypothetical model, ASAP1 binds to cortactin filaments, providing a nidus for actin polymerisation overexpression has no visible effect on invadopodia and podosome formation, reducing ASAP1 expression has been shown to inhibit their formation. It has been reported that ASAP1 employs different mechanisms to regulate different cytoskeletal structures. For example, ASAP1 overexpression has been found to inhibit cell spreading, prevent the organisation of paxillin and FAK during cell spreading and abrogate membrane ruffling in response to PDGF [25, 44] migration[42, . ASAP1 overexpression has also been shown to accelerate cell 82] . Paradoxically, silencing of ASAP1 expression by RNA interferenceinduces the same effect as overexpression on paxillin dissociation from FAs[44, 69] . These effects on FAs arein part attributed to the Arf GAP domain as it was observed that an ASAP1 mutant lacking GAP activity was less efficient in cell spreading and migration as compared to the wild type[25, 42, 19 65] . In addition, a mutant lacking the SH3 domain was also found to have inhibited cell spreading and FA organisation[44]. Although the molecular mechanism regulating FA turnover has not yet been defined, it is speculated that ASAP1 functions as a scaffold protein and targets cargoes of regulatory proteins to the sites for actin and membrane remodelling. ASAP1 may even regulate the membrane trafficking event which moves cytoskeletal regulators to their sites of action. Clinically, elevated ASAP1 expression has been implicated in a broad range of tumours and it is an important regulator of tumour cell motility, invasiveness and metastasis[42, 66, 67, 69, 84-86]. ASAP1 is amplified and overexpressed in uveal melanoma[87], as well as colorectal[88] and prostate [86] carcinomas. It has also been shown to be functionally linked to breast cancer metastasis[67].ASAP1 has also been reported to correlate with poor metastasis-free survival and prognosis in colorectal cancer patients[89]. Its pro-metastatic influence through increased cell motility and invasiveness has been the focus in clinical research so far. While the Arf-GAP activity is responsible for general cell motility, ASAP1 overexpression has been demonstrated to promote cell migration in response to PDGF and IGF-1 [42] . On the other hand, mislocalisation or siRNA-mediated silencing of the gene inhibits cell migration and retards EGFdependent chemotaxis[69]. Although overexpression has been found to inhibit cell spreading and membrane ruffling induced by PDGF[25, 42, 44, 90], retardation of cell spreading was also seen in the mislocalisation or siRNA-mediated suppression of ASAP1[69]. These findingsreflect the complex functional and structural biology of ASAP1 which warrantsfurther researched given that it is an important regulator of tumour invasiveness and metastasis. 20 Results Parameter ASAP1 protein expression, n (%) Negative Positive All Gender Male Female Age (years) ≤ 55 ˃ 55 Hepatitis B surface Ag Negative Positive Serum AFP log10 (ng/ml) Tumour size (cm) ≤5 ˃5 Venous infiltration Absent Present Tumour stage Stage I - II Stage III - IV Disease-free survival (months) Overall survival (months) p-value 179 49 90 (50.3) 24 (49.0) 89 (49.7) 25 (51.0) 0.872 118 110 50 (42.4) 64 (58.2) 68 (57.6) 46 (41.8) 0.017 31 197 18 (58.1) 96 (48.7) 13 (41.9) 101 (51.3) 0.334 2.18 ± 1.37 1.92 ± 1.34 2.44 ± 1.36 0.04 86 131 43 (50.0) 63 (48.1) 43 (50.0) 68 (51.9) 0.783 114 113 64 (56.1) 49 (43.4) 50 (43.9) 64 (56.6) 0.054 102 125 59 (57.8) 54 (43.2) 43 (42.2) 71 (56.8) 0.028 26.42 ± 28.1 26.69 ± 27.9 26.15 ± 28.41 0.886 37.38 ± 28.55 37.8 ± 29.31 36.97 ± 27.89 0.827 TABLE 1. Clinicopathological correlation of ASAP1 expression in HCC. Values denote n except for serumAFP,disease-free survival and overall survival (mean ± SD). Significant differences are shown in bold. 21 ASAP1 expression is up-regulated in HCC The cDNA microarray data was used to examine ASAP1 expression in 228 pairs of HCC and adjacent non-tumour liver tissues. Out of these cases, 114 pairs were positive for ASAP1 expression. ASAP1 overexpression in the HCC tissue was associated with age (p value = 0.017), serum AFP level (p value 0.04), and tumour stage (p value = 0.028). A slight association with venous infiltration (p value = 0.054) was also observed. However, ASAP1 overexpression did not correlate with gender, Hepatitis B surface antigen (HBsAg) and tumour size (Table 1). qPCR was used to detect the expression levels of ASAP1 in the high metastatic potential HCC cell line SK-Hep1 and the normal hepatocyte cell line LO2. ASAP1 mRNA level was significantly increased in SK-Hep1 as compared to LO2 (Figure 8A, p value < 0.05). 22 FIGURE 1. Modulation of ASAP1 expression in SK-Hep1 cell line. Cells were transiently transfected with different concentrations of control siRNA, ASAP1siRNAand ASAP1 cDNA clones. RNA and protein were harvested from the cells 48 hours post transfection and they were analysed by qPCR and western blot respectively. (A) ASAP1 mRNA levels were down-regulated by ASAP1siRNA while up-regulation was achieved with ASAP1 cDNA. *** indicates significantly different from mock control, pvalue ˂ 0.001. (B) ASAP1 protein levels were down-regulated by ASAP1siRNA and (C) protein levels were augmented by ASAP1 cDNA as demonstrated by western blot. 23 ASAP1 suppression and augmentation in HCC cells RNA interference mediated by small interfering RNA (siRNA) is a powerful and useful tool for characterising gene function. To evaluate the effects of ASAP1 expression in HCC progression, ASAP1 targeting siRNA (ASAP1siRNA) was used to knockdown ASAP1 expression in the HCC cell line SK-Hep1 and the effect of such manipulations on cell behaviours was studied. Using siRNA concentrations ranging from 40 to 100 nM, reductions in ASAP1 gene expression were demonstrated by qPCR (Figure 1A) and western blot (Figure 1B). In contrast, the control siRNA used had no effect on decreasing ASAP1 expression. These data indicate that the siRNA is capable of silencing ASAP1 expression in HCC cells. Since the aforementioned concentrations were capable of inhibiting ASAP1 expression, 40 nM was the minimal concentration used for all subsequent experiments to mitigate the risk of off-target effects typically associated with the use of high siRNA concentrations. For the purpose of completeness, commercial complementary DNA (cDNA) clones of the ASAP1 open reading frame (ORF) were used to investigate the effects of ASAP1 overexpression in the study as well. Due to the minute amounts of the commercial clone, the plasmids were amplified in competent E.coli cells and purified prior to transfection. ASAP1 overexpression was demonstrated by qPCR (Figure 1A) and western blot (1C) at different plasmid concentrations. 24 FIGURE 2. ASAP1 promotes cell proliferation. SK-Hep1 cells were transiently transfected with either control siRNA (40 nM) or ASAP1siRNA (40 nM) or ASAP1 cDNA(4000 ng) or were mock-transfected as a control. 24 hours after transfection, the cells were trypsinised and subsequently seeded at a density of 20,000 cells per well in 48-well plates. (A) Cell numbers were determined daily using a hemocytometer. (B) AlamarBlue assay was performed in parallel to corroborate the cell count data. The assay was 25 performed every 24 hours.The error bars represent S.E.; n = 3. * indicates significantly different form the mock control, p value < 0.05. **indicates significantly different form the mock control, p value < 0.01. 26 ASAP1 promotes cell proliferation in HCC cells While enhanced expression levels of ASAP3 have been reported to increase the proliferation of NIH3T3 fibroblasts[97], ASAP1 expression has been reported to have no effect on the proliferation or apoptosis rates in rat pancreatic and prostatic cells[89]. To determine if ASAP1 expression affects cell proliferation in the HCC context, SK-Hep1 cells were transfected at the aforementioned conditions. The cells were subsequently cultured for up to 6 days. Cell numbers were quantified manually using the hemocytometer. Thedecline in the cell numbers observed towards the end of the assay is likely attributed to theloss of cell viability caused by overgrowth.Significant differences in growth rates were not observed across the controls (Figure 2A). As a second measure of cell proliferation, a parallel study involving the same batch of cells wasperformed by incubating the cells with the alamarBlue reagent. This reagent contains an oxidation-reduction indicator dye that undergoes a colorimetric change in response to cellular metabolic reduction and functions as an indicator for measuring cell viability. Fluorescence values were therefore used to corroborate the manual cell count data. Interestingly,a significant statistical difference was foundbetween the plasmid and the mock controls(Figure 2B). Although ASAP1 overexpression was observed to promote cell proliferation in HCC cells, attenuation of ASAP1 expression did not achieve the opposite effect. Nonetheless, the study could be repeated at a higher siRNA concentration to determine whether ASAP1 suppression inhibits cell proliferation. 27 FIGURE 3. Transient overexpression and suppressionof ASAP1 inhibits cell adhesion of SK-Hep1 cells to collagen. SK-Hep1 cells were transiently transfected with either control siRNA (60 nM)or ASAP1siRNA (60 nM) or ASAP1 cDNA (4000 ng) or were mock-transfected as a control. 48 hours after transfection, the cells were trypsinised and replated on collagen for the indicated amounts of time. At each time point, non-bound cells were washed away and adherent cells were trypsinised. Cell numbers were determined using a hemocytometer. The error bars represent S.E.; n = 4. * indicates significantly different from mock control, p value < 0.05. ** indicates significantly different from mock control, p value < 0.01. *** indicates significantly different from mock control, p value < 0.001. 28 FIGURE 4. Transient ASAP1 overexpression enhances cell motility. (A) Cell migration rates were compared via wound-healing assays performed on SKHep1 cells. Cells were transiently transfected with either ASAP1siRNA (100 nM) or ASAP1 cDNA (6000 ng) or were mock-transfected as a control. 24 hours after transfection, the surface of the confluent cell monolayer was scratched. Microscopic examinationof the wound fields was recorded at 0, 24, and 48 hours after the wounds were made.(A) Representative images of three independent wound-healing assays. (B) Quantitative analyses of three independent wound-healing experiments. (C) Cells were transiently transfected with either control siRNA (60 nM) or ASAP1siRNA (60 nM) or ASAP1 cDNA (4000 ng) or were mock-transfected as a control. The cells were seeded into 24-well plates containing 8-µm membrane inserts and incubated for 48 hours. Migrated cells were trypsinised and cell numbers were determined using the hemocytometer. Columns represent the mean and the error bars denote S.E.; n = 3. 29 ASAP1 regulates the adhesive and motile properties of HCC cells Since ASAP1 expression has been reported to promote tumour cell motility and is associated with the cancer metastasis phenotype, several assays were performed to address the effect of ASAP1 gene modulation on these properties. To investigate the effect of ASAP1 on the adhesive properties of HCC cells, transfected and mock control cells were replated onto collagen coated plates and incubated for 40, 80 and 120 minutes. The number of adherent cells were subsequently quantified manually using the hemocytometer. As evidenced in Figure 3, both attenuation and augmentation of ASAP1 expression resulted in a significant retardation of cell attachment on collagen. These data demonstrate the importance of ASAP1 function in the cellular response to collagen. To determine the effect of ASAP1 on cell motility, monolayers of SK-Hep1 cells which had been transfected with either ASAP1siRNA or ASAP1 cDNA or were mock-transfected as a control were wounded. The migration of cells into the scratch wound field was quantified as a percentage of wound closure at 0, 24 and 48 hours. Although no significant differences were observed across the three controls in the quantitative analysis, there was an appreciable difference in the means amongst the controls (Figure 4B). These data suggest that ASAP1 augmentation increased the rate of wound closure, a collective effect of both cell proliferation and migration while ASAP1 suppression exhibited an inhibitory effect. To substantiate the wound healing data, the Transwell migration assay was performed using SK-Hep1 cells as well. This assay was carried out in the presence of a 10% FBS gradient. As shown in Figure 4C, SK-Hep1 cells that were transfected with ASAP1 cDNA exhibited 30 a higher motility than the control siRNA- and mock-transfected cells. Although the differences across the controls were short of being statistically significant, an appreciable difference in the means between the plasmid and mock controls suggest that ASAP1 overexpression enhances cell motility in the presence of a chemotactic gradient. 31 FIGURE 5. ASAP1 abrogation inhibits the invasive process of SK-Hep1 cells and tumorigenesis in NSG mice. SK-Hep1 cells were transiently transfected with ASAP1siRNA (60 nM) or mock-transfected as a control. 24 hours post transfection, the cells were pooled and resuspended in 0.9% saline solution for intrasplenic transplantation into thioacetamide-treated mice. (A) Remnant cells for the transplant were harvested for RNA and down-regulation of ASAP1 expression was ascertained by qPCR. ** indicates significantly different from mock control, p value ˂ 0.01. (B) Number of tumour colonies quantified from photomicrographs of H & E stained liver serial sections. Mean values with SD are denoted for each group (n = 3). The statistical value between mock and ASAP1siRNA groups was calculated to be significant (p value = 0.046) using the Mann-Whitney test. 32 FIGURE 5. (C) Tumours induced in the spleen (bottom left) and liver (bottom right) by mock-transfected and ASAP1siRNA-transfected SK-Hep1 cells 2 weeks after the transplantation. (D) Photomicrographs of H & E stained serial sections of the spleen for both the mock and ASAP1siRNA controls. Original magnification, x200. (E) Photomicrographs of H & E stained serial sections of the lung for both the mock and ASAP1siRNA controls. Images of E1 and E2 are shown in higher magnifications in E3 and E4 respectively. Original magnification, x 100 (E1 and E2); x 200 (E3); x400 (E4). 33 FIGURE 5. (F) Photomicrographs of H & E stained serial sections of the liver for both the mock and ASAP1siRNA controls. Arrows indicate the location of tumour colonies. Original magnification, x100. 34 ASAP1 depletion inhibits the invasive process of HCC cells and tumorigenesis in NSG mice To investigate the in vivo effects of ASAP1 expression on tumourigenesis, ASAP1siRNA-transfected and mock-transfected SK-Hep1 cells were injected into the spleen of thioacetamide-treated mice. Remaining cells left from the transplant were harvested for RNA and qPCR was performed to ascertain the successful siRNA-mediated knock down of ASAP1 expression (Figure 5A, p value [...]... its SH3 domain [39].The SH3 domain of ASAP1 mediates binding to proline-rich motifs in Pyk2 and the EH 17 domain-containing protein POB1[40, 44, 66] POB1 can simultaneously bind to RalBP (Ral Binding Protein) which contains a RhoGAP domain [66] The complex of ASAP1, POB1 and RalBP is hypothesized to link membrane trafficking with actin remodelling ASAP1 has been found to bind cortactin [67] , an... C-terminal of the Arf GAP domain Other prominent features of ASAPs include the Bin, amphiphysin and Rvs 161 and 167 (BAR) domain at the N-terminus and the proline-rich domain containing three proline-rich SH3 binding motifs located near the C-terminus.Two members of this group, ASAP1 and ASAP2, were named for the Src homology 3 (SH3) domain located at the C-terminus[39-41] Despite lacking an 15 SH3 domain,... 44] and possibly endosomes at different points of maturation (Hirsch, D.S and Randazzo, P.A unpublished) Targeting ASAP1 to FAs involves the binding of its SH3 domain tothe proline-rich motif in FAK [44 ]and its proline-rich domain interaction with adaptor proteins like Crk (chicken tumour virus 10 regulator of kinase) and/ or CrkL (Crk-like) [39, 65].Src binds to the proline-rich type I motifs on ASAP1. .. face of the plasma membrane and associate with proteins such as focal adhesion kinase (FAK), vinculin, paxillin, integrin-linked kinase and talin on the cytoplasmic side Integrins also link bundles of actin filaments called stress fibers to the ECM[54-56] Consequently, FAs appear as linear structures located behind the cell edge at the end of these stress fibers[57-61] 16 Invadopodia are present in invasive... suppression and augmentation in HCC cells RNA interference mediated by small interfering RNA (siRNA) is a powerful and useful tool for characterising gene function To evaluate the effects of ASAP1 expression in HCC progression, ASAP1 targeting siRNA (ASAP1siRNA) was used to knockdown ASAP1 expression in the HCC cell line SK-Hep1 and the effect of such manipulations on cell behaviours was studied Using siRNA... domain which contributes to one of their catalytic functions, Arf GAPs also contain protein-protein interaction and lipid interaction domains Given the structurally diverse nature of these multi-domain proteins, Arf GAPs are 13 involved in cellular activities like migration and movement, which depend largely on the remodelling of the membrane and the actin cytoskeleton Arfs belong to a family of GTP binding... results in the formation of FAs from these complexes and the association of actin stress fibers with the FAs[57, 78, 81] While the role of Arf GAPs in FAs has yet to be characterised, several speculations include: serving as scaffolds for proteins involved in actin regulation; trafficking of actin regulators; direct targets of the signalling cascades affecting the remodelling of the cytoskeleton CDR... data) It is speculated that in the formation of invadopodia, the BAR domain of ASAP1 serves to either induce membrane tabulation [19] or recruit ASAP1 to the sites of membrane tubulation, where it functions as a scaffold protein In line with [67] and bundles actin [68, 83] Although ASAP1 this hypothetical model, ASAP1 binds to cortactin filaments, providing a nidus for actin polymerisation overexpression... levels of ASAP1 in the high metastatic potential HCC cell line SK-Hep1 and the normal hepatocyte cell line LO2 ASAP1 mRNA level was significantly increased in SK-Hep1 as compared to LO2 (Figure 8A, p value < 0.05) 22 FIGURE 1 Modulation of ASAP1 expression in SK-Hep1 cell line Cells were transiently transfected with different concentrations of control siRNA, ASAP1siRNAand ASAP1 cDNA clones RNA and protein... regulator of tumour cell motility, invasiveness and metastasis This study aims to explore the role of ASAP1 in the HCC context and its contribution towards metastasis, a typical feature often associated with poor survival outcomes in patients despite having undergone surgical resection Arf GAPs are a family of proteins that inactivate Arfs by inducing the hydrolysis of GTP bound to Arf[4] These proteins ... membrane and associate with proteins such as focal adhesion kinase (FAK), vinculin, paxillin, integrin-linked kinase and talin on the cytoplasmic side Integrins also link bundles of actin filaments... motility and invasiveness in vitro and stimulation of metastasis formation in vivo To the best of our knowledge, this study is novel in investigating the role and mechanisms of ASAP1 in context of HCC. .. ankyrin (ANK) repeats immediately C-terminal of the Arf GAP domain Other prominent features of ASAPs include the Bin, amphiphysin and Rvs 161 and 167 (BAR) domain at the N-terminus and the proline-rich