BINDING OF FUS TO TMPRSS2-ERG FUSION mRNA IN PROSTATE CANCER
Rueangsit Inthong Master of Science thesis Master´s Degree Programme in Biomedicine
University of Eastern Finland Faculty of Health Sciences School of Medicine
13.6.2020
University of Eastern Finland, Faculty of Health Sciences, School of Medicine Master´s Degree Programme in Biomedicine
Rueangsit Inthong: Binding of FUS to TMPRSS2-ERG fusion mRNA in prostate cancer Master of Science thesis; 35 pages, 2 appendices
Supervisors: Leena Latonen, Docent & Mauro Scaravilli, Ph.D.: Institute of Biomedicine 13.6.2020
Keywords: FUS, TMPRSS2-ERG fusion gene, androgen receptor, prostate cancer
Abstract
Prostate cancer is a cancer type most occuring among male northern Europeans. The cancer presents many genetic alterations, including DNA copy number changes, gene mutations and gene rearrangements. The most common alteration found in patients is the fusion gene between the androgen-regulated TMPRSS2 and the oncogene ERG. FUS is a RNA-binding protein that has been found to bind TMPRSS2 mRNA in LNCaP prostate cancer cells. This binding presents in the area involved in the TMPRSS2-ERG fusion. The aim of this study is to verify the binding of FUS to wild-type TMPRSS2 mRNA transcript, and to test whether FUS also binds TMPRSS2-ERG fusion mRNA transcripts. Furthermore, we aim to understand whether FUS affects or regulates the expression of TMPRSS2-ERG fusion gene product at the protein level.
Prostate cancer cell lines LNCaP and VCaP were used for RNA immunoprecipitation (RIP), RT-qPCR and western blotting, whereas PC-3 were used as ERG overexpression model.
LNCaP and VCaP were chosen due to their high expression of androgen receptor, as well as due to their lack and presence of TMPRSS2-ERG fusion gene expression, respectively. As a result, we confirmed that FUS binds TMPRSS2 mRNA transcript and found that the protein is also able to bind TMPRSS2-ERG fusion mRNA transcript. In contrast, the overexpression of ERG in PC-3 cells was not successful. The conclusion of the study is that FUS binds both TMPRSS2 and TMPRSS2-ERG mRNA transcripts. Additonal studies are required to determine whether this binding regulates the expression of TMPRSS2-ERG fusion gene protein product.
Abbreviations
ALS amyothrophic lateral sclerosis AR androgen receptor
cDNA complementary DNA
CRPC castration-resistant prostate cancer Ct threshold cycle
DNA deoxyribonucleic acid
ERG ETS-related erythroblastosis virus E26 oncogene homolog ETS E-26 transformation-specific
FTLD frontotemporal lobar degeneration FUS fused in sarcoma
lncRNA long non-coding RNA mRNA messenger RNA NT non-transfected control NTC no-template control
PSA prostate-specific antigen RIP RNA immunoprecipitation RNA ribonucleic acid
RT-qPCR quantitative reverse transcription polymerase chain reaction siFUS siRNA against FUS
siRNA small interfering RNA
TMPRSS2 transmembrane protease serine 2
1
Introduction
Cancer is a complex disease with heterogeneity, different types, different etiology and many unknown underlying mechanisms. It can occur at many sites in the human body, including prostate gland in the male reproductive system. Prostate cancer heterogeneity, the role of human immune response during cancer development and progrssion, cancer modeling, hormone therapy and machine learning classification are now hot topics for researchers to study about prostate cancer. However, there are other research topics that are being investigated, including fusion gene alterations and functional changes found in the cancer cell. TMPRSS2- ERG fusion gene is the most recurrent alteration found in approximately 50% of prostate cancer patients. The main focus in this master thesis is to assess whether or not the RNA-binding protein called FUS can bind the mRNA transcript of TMPRSS2-ERG fusion gene. The functional aspect of this binding will also be discussed in the context of prostate cancer, along with the molecular mechanism of the fusion gene alteration in prostate cancer.
Theoretically, cancer is often referred to as an abnormal state of the cells, in which they divide uncontrollably. Tumors are classified by the tissue of the origin, for example, carcinoma is the cancer of epithelial cells, sarcoma the cancer of bone or connective tissue, leukemia and lymphoma the cancer of blood cell precursors. Physicians or pathologists need to be able to distinguish between benign (non-invasive) and malignant (invasive) tumors for futher appropriate classification using histology. However, only tumor classification is not enough to treat cancer patients. Successful detection of chromosomal rearrangements and genome-wide expression profiling can benefit classification of cancers. Together, these tools can be important and useful for choosing suitable prognostic indicators and development of new effective therapeutic targets for cancer patients (Strachan, 2011).
Based on Hannahan and Weinberg, any invasive cancer is likely to depend on cells that have acquired six basic capabilities, referred to as the hallmarks of cancer. These capabilities are 1) independence of external growth signals, 2) insensitivity to external anti-growth signals, 3) ability to avoid apoptosis, 4) ability to replicate indefinitely, 5) ability to trigger angiogenesis and vascularize, and 6) ability to invade tissues and establish secondary tumors (Hanahan &
Weinberg, 2011; Strachan, 2011). Hannahan and Weinberg also propose two new emerging hallmarks: deregulating cellular energetics and avoiding immune response and two new
2 enabling characteristics: genome instability and mutation, and tumor-promoting inflammation (Hanahan & Weinberg, 2011). A normal epithelial cell can evolve to an invasive cancer cell by six or seven successive mutations, on average. This evolution can occur by enhanced proliferation with selective growth advantage expanding a clone population for successive mutations in multi-step tumorigenesis, or by increased genetic instability at the level of DNA or chromosome to increase the rate of mutations. Each successive mutations produces more diverse cell population, until one cell acquire a change or a combination of changes, along with acquiring the six capabilities proposed by Hannahan and Weinberg, which gives the growth advantage to the cell. Eventually, the invasive cancer cell grows, expands, and develops into a tumor (Strachan, 2011).
During the evolution of the cancer cell, the most important genes that are the targets of mutations can be generally divided into two categories: oncogenes and tumor suppressor genes. Oncogenes are genes whose normal activity promotes cell proliferation and whose gain- of-function mutations in tumors will lead to excessively or inappropriately active gene. Under normal circumstances, they have a role in controlling the growth signaling pathways, such as PDGFB controlling secretion of growth factors like platelet-derived growth factor, MYC controlling DNA-binding transcription factors, and KRAS regulating the components of intracellular signal transduction systems such as the binding and cleavage of guanosine nucleotides. Non-mutant proto-oncogenes can be activated to oncogenes by gain-of-function amplification, point mutation, and translocation creating new chimeric genes or rearranging into different transcriptionally active chromatin regions. Tumor suppressor genes are genes whose products act to limit normal cell proliferation. In other words, their main function is to keep the cells under control by suppressing inappropriate cell divisions, maintaining genome integrity and promoting apoptosis of abnormal cells. In addition to inactivating genetic alterations, tumor suppressor genes can be silenced epigenetically by the methylation of CpG islands promoters. Therefore, those genes whose promoters contain CpG islands can be silenced by methylation, such as RASSF1A and HIC1. Example of tumor suppressor genes is RB1 gene encoding pRB protein, which serves as the guardian of the restriction-point gate by the inactivation of the transcription factor E2F by its unphosphorylated form to control the cell cycle clock before entering the S phase and gene transcription. Another example is TP53 gene encoding the transcription factor p53, which functions as the guardian of the genome by preventing inappropriate cell cycling (Strachan, 2011; Weinberg, 2014).
3 Prostate cancer is a cancer of the male prostate gland. The prostate is a single gland in the size of a peach pit or a walnut. It is located below or inferior to the urinary bladder and around the urethra (Figure 1). It is one of the exocrine organs in the male reproductive system, and it weighs between 20 to 25 grams. The function of the prostate gland is to secrete up to one-third of the volume of the semen. The semen is a milky, slightly acidic fluid that contains many beneficial substances for the vitality of the sperm, such as citrate, fibrinolysin, hyaluronidase, acid phosphatase, and prostate-specific antigen (PSA). The fluid helps activate the sperm through several ducts to the prostatic urethra during the ejaculation, which is defined as the contraction of the prostatic smooth muscle (Marieb, 2010).
Figure 1. Male human reproductive system. The figure shows the location of prostate gland in male human body.
(Adapted from https://www.britannica.com/science/human-reproductive-system).
4 In 2018, prostate cancer was the type of cancer most occurred in males in northern Europe.
There were 1,276,106 new cases (7.1% of all cancer sites) and 626,679 deaths (3.8% of all cancer sites) for prostate cancer globally. For men around the world, the incidence for prostate cancer was 13.5%. The age-standardized incidence rate of prostate cancer for northern European countries was 85.7 per 100,000 males, the second to Australia and New Zealand, whereas the age-standardized mortality rate was 13.5 per 100,000 males (Bray et al, 2018). It is often diagnosed as local in prostate neoplasia or advanced in metastatic prostate cancer (Damber & Aus, 2008; Wang et al, 2018). The incidence of prostate cancer is linked to familial inheritance of genes, such as hereditary prostate cancer locus-1 (HPC1), EPAC2 and vitamin D receptor gene. Other inherited alterations found in prostate cancer are genetic variants at chromosome 8q24 in European and African populations, genetic polymorphism in genes for the androgen receptors and enzymes involved in androgen metabolism like 5-reductase type 2 and steroid hydroxylase. Also, diet, pattern of sexual behavior, alcohol consumption, and occupational ultraviolet radiation exposure affect prostate cancer incidence (Damber & Aus, 2008).
The most common genetic alteration in prostate cancer is the fusion gene called TMPRSS2- ERG, which will be introduced in details further (Tomlins et al, 2005). In addition, there are other genetic and epigenetic predispositions. Other common genetic alterations include the amplification or mutations of the androgen receptor, loss of the androgen receptor corepressor NCOR1/2 and abnormal expression of the protein variants. PTEN is also altered by deletion and mutation. Moreover, there are alterations found in genes involved in DNA repair pathways, such as mutations in BRCA1, BRCA2, ATM, ATR, and RAD51. Neuroendocrine metastatic castration-resistant prostate cancer clearly shows alterations in RB1 and TP53 genes (Taylor et al, 2010; TCGA, 2015; Wang et al, 2018). Another pathway that is altered in prostate cancer is, for instance, TGF-/SMAD4 pathway controlling cell proliferation (Grasso et al, 2012).
Additionally, The Cancer Genome Atlas Research Network (2015) revealed that 74% of all prostate cancer tumors could be classified into seven major subtypes based on oncogenic drivers ERG, ETV1, ETV4, and FLI1, and based on mutations in SPOP, FOXA1 and IDH1.
These subtypes were also correlated with clusters, including iClusters based on computationally-derived individual characterization platforms, mRNA clusters based on ETS fusion, miRNA clusters based on difference in microRNA expression between ETS positive and negative tumors, and methylation clusters based on the most variably hypermethylated
5 CpGs in tumors. For the methylation cluster, unsupervised clustering was used to show DNA methylation changes to classify the primary tumor with molecular distinction (TCGA, 2015).
It can be seen that bioinformatics approach can help researchers to understand the cancer alterations using computational classification of tumor types and progression (Barillot et al, 2013).
Daily intakes of lycopenes, oral selenium and vitamin E are considered as protective agents for prostate cancer. Moreover, it was also reported that chemoprevention with finasteride, a 5- reductase inhibitor, could decrease the severity of prostate cancer by reducing the rate of cancer by 24.8% over 7 years. The treatments of primary or clinically localized prostate cancer are such as active monitoring, radical prostatectomy with conventional and robot-assisted laparoscopic techniques, radiotherapy, high-dose brachytherapy with iridium-192, high- intensity focused ultrasound, and cryosurgery. 50% of primary prostate cancer patients will develop advanced prostate cancer if they are screened within 9 years. Advanced prostate cancer can be treated by androgen deprivation therapy (ADT) by blocking the androgen pathway (Damber & Aus, 2008). Unfortunately, resistance to this therapy occurs, leading to castration- resistant prostate cancer (CRPC) (Wang et al, 2018). Therefore, CRPC refers to disease progression in patients that have undergone ADT, whose laboratory results show either continuous rising of serum PSA levels, the progression of pre-existing disease, or the appearance of new metastases (Gomella et al, 2014).
For prostate cancer patients, early diagnoses, accurate prognostic biomarkers, and essential therapeutic approaches are important for effective treatments. Nowadays, there are novel screening and diagnostic tools to detect prostate tumor, such as molecular imaging of prostate- specific membrane antigen (PSMA) on cells membrane, traditional serum PSA, urinary Prostate Cancer Antigen 3 (PCA3) biomarker, urinary TMPRSS2-ERG marker, and Prostate Health Index (PHI) score calculated from three PSA subforms (Kretschmer & Tilki, 2017; Tian et al, 2018). Some of these diagnostic markers, like serum PSA and urinary PCA3, can also work as prognostic markers. Besides, prognostic markers to identify the outcome, recovery and recurrence of the cancer include Gleason grading score, copy number alterations, and molecular tumor markers; for instance, sustained angiogenesis and metastasis (Martin et al, 2011). Lastly, new drug targets to treat prostate cancer patients, especially in advanced stage, are small molecules involved in signaling, DNA repair, or epigenetic pathways: poly ADP- ribose polymerase (PARP), Hedgehog pathway, and the phosphatidylinositol-4,5-bisphosphate
6 3-kinase/AKT/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway (Saad et al, 2019).
Androgen receptor (AR) is a transcription factor (TF) that belongs to the group of steroid hormone receptors in the nuclear receptor superfamily and can be activated by endocrine ligands (Carlberg, 2016; Sutinen et al, 2017). TFs can be classified to constitutively active TFs and signal-dependent TFs. AR belongs to signal-dependent class as one of the endocrine nuclear receptors, along with other nuclear receptors: for example, vitamin D receptor and thyroid hormone receptor (Carlberg, 2016). It is most highly expressed in epididymis and basal and luminal parts of epithelial cells in human prostate gland. The structure of AR contains N- terminal activating domain, DNA-binding domain, hinge region, and C-terminal ligand- binding domain, which is similar to the structures of other steroid hormone receptors, such as glucocorticoid receptor and progesterone receptor. The expression of the gene encoded for AR is regulated by androgens and responds to androgens differently in various tissues and cell types. The transcriptional regulation function of AR is involved with other TFs, such as ERG and key pioneer factor like forkhead box A1 (FOXA1), which helps in gene activation and repression activity of AR. Furthermore, it also requires coregulators to regulate its target genes and is controlled by post-translational modifications (PTMs), for example, phosphorylation, ubiquitination and small ubiquitin-related modification (SUMOylation) (Sutinen et al, 2017).
The signaling of AR is important in the development and function of the prostate, and genomic alterations of the signaling result in castration-resistant form of prostate cancer (Wang et al, 2018). Prostate cancer cell lines that have the expression of AR are called androgen-responsive cell lines.
The TMPRSS2-ERG fusion gene is the most common recurrent genetic alteration found in approximately 50% of prostate cancer cases. The transmembrane protease serine 2 (TMPRSS2) gene is an androgen-responsive gene that encodes for a transmembrane serine protease and is expressed in many tissues, such as in the apical part of membrane in secretory epithelial cells, in the glands’ lumen of the glands and in the basal cells, but it is mostly expressed in the normal prostate glands’ epithelial cells (Adamo & Ladomery, 2016; Kumar-Sinha et al, 2008).
Notably, TMPRSS2 gene expression is remarkably higher in prostate cancer and hyperplasia than in normal prostate gland (Kumar-Sinha et al, 2008). It functions as a sodium absorption regulator in the epithelium of human airways, as an activator of influenza and in the replication of severe acute respiratory syndrome (SARS). It is also involved with prostate carcinogenesis
7 and works as an activator of the protease-activated receptor 2, playing a partial role in a signal transduction pathway related to inflammation, metastasis and invasion (Adamo & Ladomery, 2016).
In prostate cancer, the translocation partner of TMPRSS2 gene called ETS-related or v-ets erythroblastosis virus E26 oncogene homolog (ERG) gene is an oncogene, which is a member of E26 transformation-specific or erythroblast transformation-specific (ETS) family of transcription factors. ERG acts as a key regulator of many physiological mechanisms, for example, cell proliferation, differentiation, angiogenesis, and vascular inflammation. It also regulates apoptosis, pluripotent hematopoietic stem cells and endothelial cells homeostasis, definitive hematopoiesis, and normal hematopoietic stem cell function, and maintains normal peripheral blood platelet number (Adamo & Ladomery, 2016; Wang et al, 2017).
The fusion of the TMPRSS2 gene to oncogenic ERG gene of ETS family genes usually occurs by the replacement of 5’ ends of ERG with the 5’ untranslated region of TMPRSS (21q22.2) gene (Figure 2). The most common variants of this fusion gene is the variants including TMPRSS2 exon 1 or 2 and ERG exon 2, 3, 4 or 5. Other less frequent variants include TMPRSS2 exon 4 or 5 and ERG exon 4 or 5, or TMPRSS2 exon 2 and inverted ERG exon 6-4. These variants are useful in postulating the prognostic outcomes of each patients by detection of rearrangement and ERG overexpreesion using RT-PCR, DNA sequencing and fluorescence in situ hybridization (FISH) (Kumar-Sinha et al, 2008). Additionally, functional characterization of the fusion gene showed that the gene could be regulated by androgens. The study of gene expression signature of ETS-overexpressing prostate cancer to investigate the functional consequences of TMPRSS2-ERG fusion gene expression is focused on prostate specificity, androgen responsiveness and the mechanistic role of the fusion gene. From the study, it can be concluded that when treated with androgen, ERG was expressed in prostate cancer cell lines that had fusion gene, such as VCaP. In contrast, ERG was not found to be expressed in prostate cancer cells that did not have fusion gene, such as LNCaP cell lines (Kumar-Sinha et al, 2008;
Tomlins et al, 2005). Furthermore, there are other fusions between TMPRSS2 gene and ETS family genes as well, such as TMPRSS-ETV1, TMPRSS-ETV4, TMPRSS-ETV5 and between other genes and ETS family genes, such as SLC45A3-ETV5 (Kumar-Sinha et al, 2008).
8
Figure 2. The fusion of the androgen-regulated TMPRSS2 gene and the oncogene ERG. TMPRSS2-ERG fusion (top) generates the most frequent fusion transcripts (down): the fusion of exon 1 in TMPRSS2 to exon 4 and 2 in ERG and the fusion of exon 0 in TMPRSS2 to exon 2 in ERG, respectively. T; exon of TMPRSS2 gene, E; exon of the oncogene ERG. (Adapted from Filella & Foj, 2016).
The role of TMPRSS2-ERG fusion gene in prostate cancer has also been studied extensively.
Delliaux and colleagues found that the expression of the fusion gene in bone-metastasized prostate cancer increased the properties of bone cell-like phenotype acquisition, or the osteomimicry, and also increased the formation of osteoblastic lesion, while decreased the bone resorption (Delliaux et al, 2018). The postitive status of the fusion gene could also enhance the nerve density in prostate tumor, and ERG transcription factor could target mRNAs of NRP/PLXN/SEMA family network, which might affect the innervation by increasing the nerve density (Hänze et al, 2020). Moreover, when there is TMPRSS2-ERG fusion gene present in prostate tumor, the distinct intratumoral androgen profiles will be shown in patients, such as altered intratumoral dihydrotestosterone (DHT) to testosterone ratio and altered androgen response (Knuuttila et al, 2018). TMPRSS2-ERG fusion gene is also found to be associated with non-communicable disease, such as obesity and diabetes. Interestingly, Graff and her coworkers revealed that taller height was positively correlated with the risk of TMPRSS2-ERG fusion gene with ERG-positive prostate cancer, whereas obesity was negatively correlated with the risk of developing the fusion gene with ERG-positive (Graff et al, 2018). Accordingly, it was also confirmed that the frequency of the fusion gene formation could be decreased by the exposure of insulin and insulin-like growth factor-1 (IGF-I), but the number of fusion gene
9 could be increased by hyperglycemia and insulin-like growth factor binding protein-2 (IGFBP- 2) (Holly et al, 2017).
Figure 3. The structure of FUS protein. The structure includes several functional domains, for instance, two
“prion-like” domains. The figure also shows 15 exons and 526 amino acids. QGSY-rich;
glutamine/glycine/serine/tyrosine-rich, RGG; arginine/glycine-rich regions, E; nuclear export signal, RRM; RNA recognition motif, ZnF; zinc-figer motifs, L; atypical nuclear localization signal (Aulas & Vande Velde, 2015).
(Adapted from Nolan et al, 2016).
Fused in sarcoma (FUS) protein is a RNA-binding protein encoded by the FUS gene. It is also referred to as translocated in liposarcoma (TLS) protein. FUS gene is a part of FET family genes, along with the other two genes: EWSR1 and TAF15 (Lindén et al, 2019). The structure of the protein contains functional domains (Figure 3), such as a QGSY-rich region, arginine methylated RNA-binding regions (RGG), RNA recognition motifs, two “prion-like” domains, and a higly-conserved C-terminal nuclear localization signal (NLS) at exon 15. Many known mutations in ALS are found in NLS at exon 15 (Nolan et al, 2016). FUS is involved in neurodegenerative disorders, such as amyotrophic lateral sclerosis (ALS) and frontotemporal lobar degeneration (FTLD), where it forms abnormal, toxic, prion-like protein aggregation, resulting from FUS gene mutations (Deng et al, 2014; Ederle & Dormann, 2017). FUS protein is involved in many cellular processes, including transcription, pre-mRNA splicing and miRNAs and lncRNAs processing functions in the nucleus, and mRNA trafficking and mRNA translation functions in the cytosol (Ederle & Dormann, 2017). According to Gerstberger et al, diseases related with the function of RNA-binding proteins should be correlated with tissue- specific expression in the pathological context of the proteins. This context can be explained by tissue-specific expression of important RNA targets and cofactor of RNA-binding proteins.
10 It can be inferred that these interacting RNA targets, in this case TMPRSS2-ERG fusion gene, may turn out to be better predictors of disease pathologies and mechanisms caused by these RNA-binding proteins (Gerstberger et al, 2014).
FUS is often considered as a multi-functional RNA-binding protein. The work by Bronisz et al showed that FUS could act as a co-activator of the microphthalmia-associated transcription factor (MITF) to regulate osteoclasts differentiation. The phosphorylation of MITF is important for the formation of stable components for co-activators, such as CBP/p300 and BRG1, leading to target gene transcription (Bronisz et al, 2014). Aside from DNA and RNA, FUS binds proteins, such as ribonucleotide complexes. Kamelgran and colleagues revealed that FUS binds overlapping ribonucleotide complexes, including spliceosomal components, IMP1-dependent ribonucleoprotein granules, RNA transporting granules, and stress granules (Kamelgarn et al, 2016). FUS protein is also associated with RNA polymerase II (RNAPII) promoters, reflecting that the protein could regulate transcription (Tan et al, 2012).
Other functions of FUS protein have also been studied. As a FET protein family members, FUS bound with DDIT3 and other two protein members could bind SWI/SNF chromatin remodelling complex and deregulate its activity. The data could help reflecting the tumor development and the common pathogenic mechanism for FET-associated tumors (Lindén et al, 2019). Together with another FET protein members Ewing sarcoma breakpoint region 1 (EWS; also abbreviated as EWSR1), FUS can interact with poly(A)-signal downstreaming transcribed genes, which is important for encoding proteins for RNA regulatory mechanisms.
The study demonstrated that FUS and EWS can target genes that are involved in cellular pathways regulation, and these target genes can be used as potential mediators of disease- associated cellular functions (Luo et al, 2015). In the latest publication by Imperatore and coworkers, it was shown that FUS can recognize the G quadruplex structures in mRNAs of two neuronal cells, PSD-95 and Shank1. The structures are important for synaptic plasticity and maintenance (Imperatore et al, 2020).
Ward et al found that FUS had an important role in cell cycle progression and cellular proliferation, such as in mitosis and meiosis stages of prophase (Ward et al, 2014). FUS interacts with AR and increases the transcriptional activity of AR in prostate cancer cells. In addition, it was found that when there was less FUS, the levels of androgen-induced genes were reduced, and the proliferation of prostate cancer cell lines was limited. FUS was also found to
11 be recruited to AREs in response to androgen. These results implied that FUS could act as a co-activator of AR (Haile et al, 2011). According to the work of Brooke et al, it was shown that FUS is downregulated by androgen signalling. FUS can regulate proteins important for cell cycle progression. Moreover, the expression of FUS was found to be reduced in advanced- staged prostate cancer. These results implied that loss of FUS might result in increased androgen signaling and higher growth of prostate cancer cells, and that FUS might function as a putative tumor suppressor (Brooke et al, 2011).
Additionally, FUS protein has a role in other types of cancer. In acute myeloid leukemia, FUS- ERG fusion oncoprotein was found to bind non-promoters regions in a complex containing other ETS factors, like RUNX1 and RNAPII. The fusion protein was found to be involved in abnormal regulation of all-trans retinoic acid (ATRA) signalling response and inhibit the myeloid cell lineage differentiation (Sotoca et al, 2016). In glioma, FUS was bound by the long non-coding RNA (lncRNA) called rhabdomyosarcoma 2-associated transcript (RMST). It was shown that FUS SUMOylation was modified and enhanced by RMST to suppress the glioma.
These results reflected that RMST could be used as a prognostic marker for unfavourable outcome in glioma (Liu et al, 2020). In colorectal cancer, FUS was found to interact with lncRNA called ITIH4 anti-sense RNA 1 (ITIH4-AS1) to activate JAK/STAT3 tumorigenic signalling, suggesting that ITIH4-AS1 could be a regulator in colorectal cancer progression (Liang et al, 2019). Lastly, in gastric cancer, FUS partly activated the lncRNA LBX1 through the positive regulation of LBX1. Another lncRNA called LBX2 antisense RNA 1 (LBX2-AS1) is also involved in the activation of the lncRNA LBX1 via FUS. The positive feedback loop involving lncRNA and FUS (LBX2-AS1/miR-219a-2-3p/FUS/LBX2 axis) was purposed, which might affect the growth of gastric cancer cell and apoptosis abilities of the cancer cell (Yang et al, 2019).
From a recent proteomic profile analysis of samples from benign prostate cancer, untreated prostate cancer and CRPC using mass spectrometry by Latonen and her colleagues, the main overrepresented protein class is nucleic acid binding proteins, which includes mainly RNA- binding proteins (Latonen et al, 2018). Unpublished work by Leena Latonen’s research group has also recently revealed a binding of FUS protein to TMPRSS2 mRNA in dihydrotestosterone (DHT)-treated androgen-responsive LNCaP cell lines in the area that is present also in most TMPRSS2-ERG fusion gene transcripts. Therefore, to fill in gaps in the knowledge of biomedicine research field, this project aims to assess whether the functional FUS protein can
12 regulate TMPRSS2-ERG fusion transcript in prostate cancer cells, providing a new insight in the role of RNA-binding proteins in prostate cancer. In the long run, the project can benefit human health and well-being by the discovery of novel molecular mechanism in prostate cancer.
This master thesis is focused on the role of FUS protein in prostate cancer. The main research question is, whether this protein affects the regulation of the TMPRSS2-ERG fusion gene expression in prostate cancer cells. The study has four main aims: 1) to verify the binding of FUS protein to wild-type TMPRSS2 mRNA transcript, 2) to test whether FUS binds TMPRSS2- ERG fusion mRNA transcript, 3) to assess possible effects of FUS binding to wild-type and fusion products at their expression level, and 4) to determine possible effects of this potential binding to the proliferation of prostate cancer cells. The laboratory experiments for the thesis project, including RNA immunoprecipitation (RIP), RT-qPCR and western blotting, were performed in eight-month timeframe, during June 2019 to February 2020.
13
Materials and methods
Cell culture
For daily upkeep, HeLa, LNCaP, VCaP, and PC-3 cell lines were maintained by performing cell splitting and plating twice a week or when necessary. For long-term storage of cells, cryotubes of cell lines were prepared and stored in - 80C. The culture medium for HeLa and VCaP cell lines were DMEM (Lot # 0000618582, BioWhittaker), supplemented with 55 ml of 10% Fetal Bovine Serum (FBS), 5.5 ml of penicillin and streptomycin (Pen-strep), 5.5 ml of L-glutamine, and 2 ml of the selection marker Geneticine. The culture medium for PC-3 and LNCaP cell lines were RPMI-1640 (Lot # 0000717807, BioWhittaker), supplemented with the same reagents as for HeLa and VCaP cell lines. Cell lines were incubated at 37C.
Lysate preparation
If the lysates were needed, cells scraping on ice using spatula and 1x cold PBS was performed.
20 ml of Triton lysis buffer were prepared using 1 ml of 50 mM Tris-HCl pH 7.5, 600 l of 150 mM NaCl and 100 l of 0.5% Triton X-100, with supplementation of 1 – 1.5 ml of dissolved 1x Protease inhibitor cocktail tablet (Roche, Lot # 38653200), 8 ml of dissolved 0.25 mM Dithiothreitol (DTT) and 57.4 ml of 0.25 mM PMSF dissolved in anhydrous isopropanol.
For western blotting, cell pellets were suspended in 200 – 400 l of Triton lysis buffer. Lysates were then sonicated three times, 30 seconds on and off. After the sonication, lysates were centrifuged at 16000 g for 10 minutes. When lysates were ready, protein concentration measurement was performed using 6 standard dilutions of 3% BSA (Sigma Aldrich, Lot # SLCB1974) in 2-fold dilution: 5, 2.5, 1.25, 0.625, 0.3125 and 0.15625 mg/ml or µg/µl. 25 l of Reagent A (Bio-Rad, Cat. #5000113) supplemented with Reagent S (Bio-Rad, Cat. #500- 0115) and 200 l of Reagent B (Bio-Rad, Cat. #5000114) were added to each well, respectively. Protein concentration measurement was performed in Tecan Infinite machine using the wavelength of 690 nm. Lysates were stored at - 20C.
Reverse transfection using jetPEI®
To transfect plasmids into cell lines, reverse transfection using jetPEI(Polyplus transfection, Lot # 17052C1L) was performed using PC-3 and HeLa cells transfected with ERG and control plasmids (Supplementary figure 1). Transfection Mix reaction was calculated and prepared using DNA plasmids and jetPEI, according to the PolyPlus transfection protocol: 5 g of
14 plasmids were added to cell lines in 5 ml plate. The transfection mixes were prepared in l ml of Opti-MEM, which was also used as a mock-treatment in non-transfected controls. After the DNA-jetPEI complexes were added to the cell plates, cell plates were incubated at 37C and proceed to western blotting and RNA extraction protocols at the timepoint indicated (2-3 days).
Western blotting
To assess levels of the protein expression, western blotting was performed using 7.5% and 10%
SDS-PAGE gels and lysates from every cell line. SDS-PAGE gels were casted from water, 1.5 M of Tris buffer for lower gels and 0.5 M for upper gels, 10% SDS, TEMED and 10% APS.
Both lower and upper gels for SDS-PAGE gels were polymerized for 30 minutes. To prepare lysates for gel loading, lysates were mixed with 2x or 4x Laemmli loading buffer in an equal volume. Lysates ready for loading were incubated at 95C for exactly 5 minutes. Equal amount of samples were loaded into each well. Then, SDS-PAGE gels were runned at 160 - 200V for approximately 45 minutes to 1 hour. Proteins were transferred to nitrocellulose membranes at 250 mA for 1 hour, and the membranes were blocked with 10 ml of 3% BSA (Sigma Aldrich, Lot # SLCB1974) for approximately 45 minutes to 1 hour.
In first and second probing, primary and secondary antibodies were probed to membranes for 1 hour. First probing was performed using 10 ml of antibodies against ERG (EPR3864, Abcam), FUS (4H11, Santa Cruz Biotechnology) and AR (441, Santa Cruz Biotechnology) as the primary antibodies, and 10 ml of anti-mouse (invitrogen, Lot # 2043839) and anti-rabbit (invitrogen, Lot # 2030230) as the secondary antibodies. Second probing was performed using 10 ml of pan-actin as the loading control and 10 ml of anti-mouse as the secondary antibody.
Antibodies were diluted with 10 ml of 3% BSA: anti-ERG, anti-FUS, anti-AR, and pan-actin were diluted 1:1000, anti-mouse and anti-rabbit (invitrogen) were diluted 1:10000. After probing, signals were detected with 100 – 200 l of ECL reaction (Bio-Rad, Cat. # 170-5060) and exposed to chemiluminescent light using ChemiDoc machine (Bio-Rad).
RNA immunoprecipitation (RIP)
For RIP experiments, 100 ml of RIP lysis buffer was prepared using 7.5 ml of 150 mM KCl, 2.5 ml of 25 mM Tris pH 7.4, 800 l of 5 mM EDTA prepared from 1 tablet of 7x Protease Inhibitor Cocktail (PIC) (Roche, Lot # 41353800) dissolved in 1.5 ml of nuclease-free water,
15 500 l of 0.5% mM DTT and 500 l of 0.5% NP40 (Sigma-Aldrich, Lot # BCCB0819) with 15 l of 100 U/ml RNase inhibitor (Invitrogen, Lot # 00764447) added for daily use as 10 - 15 ml of complete RIP lysis buffer (Rinn et al, 2007).
To prepare lysates, cell plates were washed with 2 – 3 ml of cold PBS. Then, 500 – 1000 l of cold RIP lysis buffer was added to plates. Plates were incubated on ice for 30 minutes. After that, cells were scraped and pipetted into 2-ml Eppendorf tubes for shearing. Next, DNA was sheared using a syringe and needle 10 – 20 times. Tubes were then centrifuged at 12000 g for 10 minutes in + 4C. After DNA shearing and centrifugation, the lysates were divided into two new tubes, 100 l for wach tubes: a tube for RNA extraction stored in - 80C and an original lysate tube stored in - 20C for western blotting.
The lysates were also divided for each 400 l in two more tubes, for adding precipitating (FUS), control antibodies (IgG) and GammaBind sepharose in immunoprecipitation steps, repectively.
Anti-FUS A300-294A (Bethyl Laboratories, Lot # A300-294A-2) was used as a precipitating antibody, and IgG antibody (Bethyl Laboratories, Lot # P120-101-12) was used as a control.
2-5 l of anti-FUS precipitationg antibody was added to FUS tubes. 1-2 l of IgG control antibody was added to IgG tubes. Each tubes were tightly sealed and incubated at 4C for 2 hours or overnight on rotator. After the incubation, 40 l of GammaBind sepharose (GE Healthcare, Lot # 10244030) was added to each tubes. Similarly, each tubes were sealed and incubated at 4C for 1 hour on rotator. GammaBind sepharose was used to bind antibody and pull down the antibody bound to FUS and the transcripts complexes as beads.
After the incubation and brief centrifugation, supernatants in FUS and IgG tubes were transferred to two remaining lysate tubes. Beads in each tubes were washed with 500 l of RIP buffer for three times. When beads were already washed twice, 50 l of each mix were transferred to two immunoprecipitated (IP) tubes: FUS-IP and IgG-IP tubes. After beads were already washed for three times and briefly centrifuged, RNA extraction was performed.
Remaining lysate and IP tubes were stored at - 20C for western blotting.
RNA extraction
To purify RNA transcripts from RIP, RNA extraction was performed using 850 l of TriReagent (invitrogen, Lot # 00733484) per samples. RNA was extracted by the addition of
16 170 l of chloroform, 425 l of isopropanol (2-propanol), and 800 l of 75% ethanol, incubations at room temperature and centrifugations. RNA was resuspended in nuclease-free water. To determine RNA concentration, 1 l of extracted RNA was used. RNA concentration and purity were measured using NanoDrop One machine. RNA samples were stored at - 80C.
cDNA synthesis
For RT-qPCR, cDNA synthesis of RNA samples was performed according to Thermo Fischer protocol using the Maxima First Strand cDNA synthesis kit (Thermo Scientific, Lot # 00679700), according to manufacturer’s instructions. Master Mix reaction was calculated and prepared using 10x dsDNase buffer, dsDNase, template RNA (total RNA samples) and nuclease-free water (Hyclone water), according to manufacturer’s instructions. 10 l of reaction were added to 96-well plates, mixed up and down several times. Incubation was performed either in T3 thermocycler or T100 thermocycler. cDNA samples were stored in the freezer at - 20C.
Primer testing
To verify the best annealing temperature for each forward and reverse primer, primer testing was performed. cDNA samples used for primer testing were from LNCaP, VCaP, HeLa and PC-3 samples pooled together. Firstly, forward and reverse primers were diluted 1:10 with nuclease-free water (Hyclone water). Master Mix reaction was prepared using Hot Start Taq DNA Polymerase and buffer (NEB M0495S), supplemented with 10 mM dNTP mix (BN- 1006-08), 1x forward and reverse primers (Merck, Lot # SY190612973, SY190611219, SY190612973, SY19017942, SY191138273) and Hyclone water, according to manufacturer’s instructions. Two primers for TMPRSS2-ERG transcripts were used: ERG R4 and R6 exon primers. It should be noted that all forward primers target exon 1 of TMPRSS2, whereas reverse primers target different exons of ERG. The sequences of each primer pairs are found in Table 1. 9 l of cDNA samples were added to Master Mix reaction. 25 l of Master Mix reaction with cDNA samples were pipetted to each well in 96-well plates. Primer testing was performed using traditional PCR machine (Bio-Rad). To determine the optimal annealing temperature, temperatures were set to a gradient program (58 – 65C). The annealing temperatures are also found in Table 1.
17 RT-qPCR
RT-qPCR (Thermo Scientific, Lot # 00812490) of cDNA samples was performed according to the protocol. 1x Master Mix reaction was prepared using 1x SYBR Green Master Mix, 1x forward and reverse primers (Merck, Lot # SY190612973, SY190611219, SY190612973, SY19017942, SY191138273) and nuclease-free water (Hyclone water), according to manufacturer’s instructions. TMPRSS2 primer and the same primers for TMPRSS2-ERG transcripts were used, whose sequences can be found in Table 1. cDNA samples were diluted 1:10 with Hyclone water. 20 l of Master Mix reaction and 2 l of diluted cDNA samples were pipetted to each well in 96-well plate. RT-qPCR was performed and analyzed in qPCR machine (Roche).
Table 1. Sequences of primers pairs, their targets and annealing temperatures from primer tesing experiments.
Name of primers
Forward primers Targets Reverse primers Targets Annealing temperatures (C)
ERG R4 exon primer
CAGGAGGCGG
AGGCGGA Exon 1 of
TMPRSS2 TAACTCTGCGC
TCGTTCGTG Exon 5 of ERG 58
ERG R6 exon primer
CAGGAGGCGG AGGCGGA
Exon 1 of TMPRSS2
AGAGAAGGAT GTCGGCGTTG
Exon 4 of ERG 58
TMPRSS2
primer CTGGTGGCTGA
TAGGGGATA TMPRSS2 CAGCCCCATTG
TTTTCTTGT TMPRSS2 58
Agarose gel electrophoresis
For the verification of PCR result, agarose gel electrophoresis was performed. Samples were prepared using 6x Tritrak loading dye (Thermo Scientific, Lot # 00728082), sterile water and 10000x SYBR Green (Invitrogen, Lot # 36631A), diluted 1:100 with TE pH 8.0. Marker Mix was prepared using 1 part of loading dye, 1 part of marker and 4 parts of sterile water. Ready marker for loading the gel was prepared with 9 l of Marker Mix and 1 l of 100x SYBR Green. For loading to agarose gel, 4 l of Marker Mix and 6 l of samples were loaded to lanes.
Then, gels were runned at 80 – 100V for approximately 45 minutes to 1 hour. Finally, gels were exposed to chemiluminescent light and imaged by ChemiDoc machine (Bio-Rad).
18
Results
In this thesis project, we used three prostate cancer cell lines: LNCaP, VCaP and PC-3. The AR expression status of these cell lines was initially confirmed: VCaP with the highest expression, followed by LNCaP, and PC-3 with no expression of AR (Figure 4). It can also be seen in Figure 4 that HeLa has no AR expression, as well. The expression of AR in LNCaP and VCaP cells leads to binding of AR to the androgen responsive elements (ARE) in androgen-regulated genes, such as TMPRSS2 gene and TMPRSS2-ERG fusion gene. This results in more transcription and expression of these genes in the cells. Therefore, VCaP cells can express both TMPRSS2 gene and the TMPRSS2-ERG fusion gene. In addition, the ERG protein expression status of prostate cancer cell lines used in the study was also confirmed:
VCaP with the highest expression and others with less expression (Figure 5).
Figure 4. The AR expression status of the cell lines used in the study. The image clearly shows the band in androgen receptor (AR) molecular position (87 kDa). Pan-actin (45 kDa) is a loading control for the image.
Figure 5. The ERG protein expression status of the cell lines used in the study. The image clearly shows the band in VCaP lane at ERG molecular position (55 kDa). Pan-actin (45 kDa) is the loading control for the image.
INTRO: Blot 33
LNCaP PC-3 HeLa VCaP
AR 87 kDa
Pan-actin 45 kDa
19 FUS was found to bind TMPRSS2 in a large-scale eCLIPseq assay performed in LNCaP cells.
RNA immunoprecipitation (RIP) is a technique which helps identifying the interaction between RNA and protein using antibodies against the protein of interest. Antibodies against the specific proteins binds proteins, and specific RNA-protein complexes are immunoprecipiated. RNA from the complexes can be isolated and detected by PCR-based methods, hybridization, massive sequencing and computational bioinformatics for creating the binding map of RNA- binding protein and identifying the protein binding sites (Gagliardi & Matarazzo, 2016). In the thesis project, we used RIP assay to confirm that there is the binding of FUS to TMPRSS2 mRNA transcript in LNCaP cells, and to test whether FUS also binds TMPRSS2-ERG fusion mRNA transcript in VCaP cells.
For RT-qPCR, we used three lysates from RIP experiment: original lysate and two immunoprecipitated tubes. 1) The original lysate tube refers to the tube in which the lysate was collected after DNA shearing. 2) Immunoprecipitated tubes refer to tubes in which washed immunoprecipitated beads from control IgG and FUS tubes were collected: control IgG immunoprecipiated (IgG-IP) and FUS immunoprecipitated (FUS-IP) beads. Ct values are shown in bar graph for interpretating if FUS protein binds exact specific RNA transcripts.
For western blotting, we used five lysates from RIP experiment: original lysate, two remaining lysates (control IgG and FUS) and two immunoprecipitated tubes. 1) The original lysate tube refers to the tube in which the lysate was collected after DNA shearing. 2) Remaining lysates tubes refer to tubes in which supernatants from control IgG and FUS tubes were collected without touching immunoprecipitated beads. 3) Immunoprecipitated tubes refer to tubes in which washed immunoprecipitated beads from control IgG and FUS tubes were collected:
control IgG immunoprecipiated (IgG-IP) and FUS immunoprecipitated (FUS-IP) beads. The blot was imaged to show that the correct protein was pulled down in RIP experiment.
FUS binds TMPRSS2 mRNA transcript in LNCaP cells
To verify whether or not FUS protein binds TMPRSS2 mRNA transcript, RIP experiment was performed with LNCaP and VCaP cells using antibody against FUS and IgG antibody to pull down the protein and RNA transcript complexes. To validate if TMPRSS2 mRNA transcript was actually bound to FUS and to determine the enrichment of the transcript from threshold cycle (Ct) value, RT-qPCR experiment was performed with primers for TMPRSS2 mRNA transcript. Lastly, western blotting was performed to verify the pulling down of FUS protein in
20 the lysates. LNCaP cells were used because of their expression status of AR, and PC-3 cells were used as negative controls because they do not express AR (Figure 4).
There was more TMPRSS2 mRNA transcript bound by FUS than control IgG, as shown in Panel A of Figure 6. From the figure, Ct value of FUS-IP lysate was lower than Ct value of control IgG-IP lysate because the amount of starting cDNA materials from TMPRSS2 mRNA transcripts in FUS-IP lysate was higher than in control IgG-IP lysate. As the amount of starting cDNA in FUS-IP was high, it took less qPCR cycle numbers in the amplification curve to reach above the threshold line earlier. Therefore, higher amount of TMPRSS2 mRNA transcript in FUS-IP lysate gave less Ct value. This result can be interpreted that there are more TMPRSS2 mRNA transcripts bound by FUS compared to control IgG, as Ct value of FUS-IP was lower than of control IgG-IP in LNCaP cells. It should also be noted that the no-template control (NTC) was not detected with signal in qPCR result. As expected, agarose gel electrophoresis result showed that all bands were in expected size (data not shown), indicating that TMPRSS2 qPCR experiment was working specifically.
It can also be noticed that there were Ct values of PC-3 cells lysates from RIP as well (Panel A in Figure 6). From the graph of Ct values in Panel A, it can be seen that Ct values of every PC-3 RIP samples were higher than those of every LNCaP RIP samples, because PC-3 RIP samples had less starting TMPRSS2 mRNA transcript than LNCaP RIP samples. This resulted from less expression of AR in AR-negative PC-3 cells (Figure 4), leading to less binding of AREs and less TMPRSS2 mRNA transcription and gene expression in the cells. Therefore, it can be concluded that in PC-3 cells the detected levels of TMPRSS2 mRNA transcripts were significantly lower. This can explain why we could not detect such a difference of Ct values between IgG-IP and FUS-IP in PC-3 cells.
FUS protein was correctly pulled down in RIP experiment, as indicated in Panel B in Figure 6.
There was a clear band in FUS-IP lane, which was the precipitated antibody against FUS protein bound to the protein and mRNA transcript complexes, compared to the faint band in IgG-IP lane, which was the precipitated IgG bound to other unspecific mRNA transcripts in the lysate. In summary, there was a binding of FUS protein to TMPRSS2 transcript in LNCaP cells.
21
Figure 6. Binding of FUS to TMPRSS2 mRNA transcript. Panel A shows RT-qPCR result with Ct values in LNCaP and PC-3 cells. Panel B shows the image of western blot in LNCaP cells lysates. It can be seen that there is a binding of FUS protein to TMPRSS2 mRNA transcript in lysates. For RT-qPCR experiment, PC-3 lysates are used as negative controls in this experiment. IgG-IP; IgG immunoprecipitated lysate, FUS-IP; FUS immunoprecipitated lysate.
A
B
Aim 1: Blot 42
Remaining lysates
Original lysate IgG FUS IgG-IP FUS-IP
FUS 75 kDa
Pan-actin 45 kDa
22 FUS binds TMPRSS2-ERG fusion mRNA transcript in VCaP cells
To test whether FUS also binds the TMPRSS2-ERG fusion mRNA transcript, RIP experiment was performed with VCaP cells expressing the fusion transcript, followed by RT-qPCR to identify exact specific mRNA transcript to which FUS binds, and western blotting to verify the pulling down of the exact protein. VCaP cells were used because they are the only prostate cancer cell line in this study that expresses TMPRSS2-ERG fusion mRNA transcript, driven by the highest expression of AR (Figure 4).
Since VCaP cells may also express native TMPRSS2 and ERG transcripts in addition to the fusion transcript, we further designed two set of primers recognizing specifically the fusion transcript in VCaP cells by harboring the 5’ primer site on TMPRSS2 by the forward primer and 3’ primer site on ERG by two reverse primers (Figure 7). These reverse primers are called ERG R4 and R6 exon primers. According to Table 1, the ERG R4 exon primer binds exon 5 of ERG, whereas the ERG R6 exon primer binds exon 4 of ERG, respectively. The forward primer binds exon 1 of TMPRSS2. It should be noted that the same forward primer is used in both two sets of primer for the fusion transcript (Table 1). Agarose gel electrophoresis using VCaP lysates showed that ERG R4 and R6 primers worked specifically and recognized the primer sites in the fusion transcript (Figure 8). Supplementary figure 2 shows how ERG R4 and R6 primers were selected amother other two candidates.
More TMPRSS2-ERG fusion mRNA transcript was bound by FUS protein, compared to control IgG, according to Panel A in Figure 9. Both Ct values of FUS-IP samples from ERG R4 and R6 exon primers were lower than those of IgG-IP samples, indicating that there was high amount of starting TMPRSS2-ERG fusion transcript in the samples.
Western blotting confirmed that FUS was correctly pulled down in RIP experiment. The band in FUS-IP lane could also be seen in the western blot image (Panel B in Figure 9), although the band was a little faint than the band in Panel A from Figure 6. Similarly, it is verified that the right protein was pulled down in RIP experiment. Therefore, it can also be concluded that there was FUS binding to TMPRSS2-ERG fusion transcript in VCaP cells.
23
Figure 7. The locations of each exons in TMPRSS2 and ERG genes. The figure shows the site of exon 1 in TMPRSS2 and exon 4 and 5 in the oncogene ERG structure, which are pointed by black arrows. (Adapted from Shafi et al, 2013).
Figure 8. Agarose gel electrophoresis of two specific reverse primers for fusion transcripts. ERG R4 and R6 primers bind different sites in ERG exons. The bands in VCaP original (Lane 3), IgG (Lane 6) and FUS (Lane 7) lanes are specific to ERG exons. PC-3 lanes (Lane 2, 4 and 5) were used as negative controls. R4; ERG R4 exon primer, R6; ERG R6 exon primer, NTC; no-template control.
24
Figure 9. Binding of FUS to TMPRSS2-ERG fusion mRNA transcript in VCaP cells. Panel A shows RT-qPCR result with Ct values of reverse primers for fusion transcript in VCaP cells: ERG R4 and R6 exon primers. Panel B shows the image of western blot in VCaP cells lysates. It can be seen that there is a binding of FUS protein to TMPRSS2-ERG RNA transcript in lysates. IgG-IP; IgG immunoprecipitated lysate, FUS-IP; FUS immunoprecipitated lysate.
A
B
Aim 2: Blot 43
Remaining lysates
Original lysate IgG FUS IgG-IP FUS-IP
FUS 75 kDa
Pan-actin 45 kDa
25 Overexpression of ERG in PC-3 cells
To determine whether FUS binds TMPRSS2 and TMPRSS2-ERG fusion RNA transcripts specifically through TMPRSS2 part, we wanted to test if FUS binds ERG mRNA. For this, ERG plasmid (Supplementary figure 1) was transfected to PC-3 cells that do not express ERG endogenously. RT-qPCR experiment and western blotting was performed to verify the overexpression of ERG plasmid in the cells before RIP experiment was to be performed. VCaP cell was used as the positive control in these experiments due to high expression of ERG protein in VCaP cells (Figure 5). Other cell lines, including PC-3 and HeLa, have less expression of ERG, as seen in Figure 5. Therefore, PC-3 and HeLa cells can be used as suitable models for overexpression of ERG.
Unfortunately, no ERG expression could be detected in the transfected PC-3 cells. Figure 10 shows the image of western blot from PC-3 cells lysates transfected with ERG plasmid, control plasmid, and non-transfected control, respectively. Each band cannot be clearly seen at ERG protein molecular position in the blot (Figure 10). This might result from the inefficient plasmid transfection. We also attempted to transfect ERG plasmid in HeLa cells, but unfortunately, the transfection did not work efficiently in these cells either. To sum up, the transfection of ERG plasmid in PC-3 cell model was not successful, and thus the RIP assay testing FUS binding to wild-type ERG mRNA could not be performed.
Figure 10. Overexpression of ERG in PC-3 cells. The figure shows the western blot image from PC-3 cells lysates.
Faint bands can be seen at ERG molecular position (55 kDa). The band in VCaP lane serves as a positive control.
NT; non-transfected control.
26
Discussion
FUS is an RNA-binding protein, which is not only involved in neurodegenerative disorders like ALS and FTLD, but also in cancer biology. In the study of proteomic profile in many types of prostate cancer samples, it has been found that the most overrepresented group is RNA- binding protein (Latonen et al, 2018). In addition, the unpublished work has also revealed that FUS binds TMPRSS2 mRNA in DHT-treated LNCaP cells. TMPRSS2-ERG fusion gene is the most recurrent fusion found in approximately 50% of prostate cancer patients, and is involved with cell function and prostate cancer progression and prognosis. The fusion gene is used for detection and therapeutic approach for prostate cancer patients (Kumar-Sinha et al, 2008).
Therefore, it is important to find out if FUS binding has any effects on or regulates the expression of TMPRSS2-ERG fusion gene.
This study aims to answer whether FUS protein can affect or regulate the expression of TMPRSS2-ERG fusion gene by attempting to accomplish four aims: 1) to verify FUS binding to wild-type TMPRSS2 mRNA transcript, 2) to test FUS binding to TMPRSS2-ERG fusion mRNA transcript, 3) to determine if the binding to FUS affects the expression level of wild- type and fusion product, and 4) to assess if potential changes in expression levels of ERG protein caused by FUS binding affect the proliferation of prostate cancer cells.
The first aim was accomplished by using RIP experiments in LNCaP cells to pull down TMPRSS2 mRNA transcript bound by FUS. Then, RT-qPCR and western blotting were performed to verify actual FUS binding to the wild-type transcript. We revealed that FUS bound TMPRSS2 mRNA transcript in LNCaP cells (Figure 6). This result is consistent with the unpublished work by the research group that FUS binds to TMPRSS2 mRNA in DHT-treated androgen-responsive LNCaP cell lines. Additionally, the result also suggests that RIP assay is reliable, because there is an enrichment of mRNA transcript in RT-qPCR result and FUS band can be seen in FUS-IP lane in the blot image, according to Figure 6.
We were able to successfully accomplish the second aim by performing RIP experiment in VCaP cells to pull down fusion transcript-and-FUS protein complex and performing RT-qPCR and western blotting to identify exact specific mRNA transcripts to which FUS protein binds and to check the presence of FUS protein. We showed that FUS also bound TMPRSS2-ERG fusion mRNA transcript in VCaP cells (Figure 9). The result is coherent with the result from the unpublished work by Latonen research group, which shows FUS binding to TMPRSS2
27 mRNA on the area which is included in TMPRSS2-ERG fusion transcripts. Moreover, it can be implied that RIP is a reliable method, due to the enrichment of the fusion transcript in RT- qPCR and FUS band seen in FUS-IP band in Figure 9.
Previous findings have also shown that FUS protein can bind RNAs in prostate cancer. Feng and colleagues reveal that FUS is interacted by the circular RNA called circ0005276 in prostate cancer cell lines: PC-3 and DU145 (Feng et al, 2019). From the analysis of regulatory networks of lncRNAs and the prediction of their potential RNA-binding proteins, it is found that the potential RNA-binding protein that can interact with lncRNAs RP11-33A14.1, RP11-423H2.3 and LAMTOR5-AS1 is FUS protein. These three lncRNAs are the most differentially expressed lncRNAs in prostate biopsy tissue (Li et al, 2020). In other types of cancer, the interaction of FUS to RNAs has also been found; for example, the binding of FUS to lncRNA SOX2OT in pancreatic cancer (Chen et al, 2019a) and the positive correlation of FUS expression in the presence of circular RNA circRNA_0000285 (Chen et al, 2019b) and lncRNA DLX6-AS1 in cervical cancer (Tian et al, 2019). There is also a binding of FUS to RNAs in other diseases as well, such as in cardiac hypertrophy. The protein can bind lncRNA CTBP1- AS2 in cardiomyocyte (Luo et al, 2019).
As a multi-functional RNA-binding protein, FUS has been found to have a role in various important processes, such as regulation of cellular proliferation (Ward et al, 2014), cell cycle progression (Brooke et al, 2011), and enhancement of AR transcriptional activity (Haile et al, 2011). Previous findings regarding to the binding of FUS to RNAs in cancer and other diseases have also revealed new roles of the protein: the transcriptional regulation of X-linked inhibitor of apoptosis protein (XIAP) to promote tumorigenesis and prostate cancer development (Feng et al, 2019), effect to prostate cancer progression through the prediction of three most differentially expressed lncRNAs in prostate tumor (Li et al, 2020), and a suppressor of CCND1 and p27 expression regulation in pancreatic cancer cells (Chen et al, 2019a). The recent finding in master thesis confirms that there was a binding of FUS protein to wild-type TMPRSS2 and TMPRSS2-ERG fusion mRNAs. This FUS binding may decrease the proliferation of prostate cancer cells. Additional studies are still required to assess whether this binding has an effect or regulates the expression of the fusion gene protein product.
In the future, the results in this thesis may be vital for the development of new survival prognostic and therapeutic approaches. Since FUS binding may affect prostate cancer cells
28 proliferation, FUS may work as a prognostic biomarker and drug target to detect and treat prostate cancer patients as early as possible. In addition to their binding partners, RNA-binding proteins can be good candidates to support the treatment of patients, as the abnormal regulation of these proteins may have an effect to many cancer characteristics. Moreover, the alteration of RNA-binding proteins expression in cancers could lead to aberrant interaction of their binding targets and the formation of incorrect complexes, affecting post-translational modifications and cell phenotypes (Hong, 2017). Therefore, RNA-binding proteins that have altered expression in tumor samples may be utilized as potential candidates for prognostic biomarkers and therapeutic targets (Galante et al, 2009; Hong, 2017). Additionally, there are many attempts to discover more novel prognostic and therapeutic approaches for cancer patients, using genome-scale metabolic models (GDMs) in prostate cancer to validate promising drug candidates (Turanli et al, 2019), transcriptional regulators in prostate cancer stem cells (Civenni et al, 2019), and integrated bioinformatics analysis of differentially expressed RNA-binding proteins profile in breast cancer multiomics data (Wang et al, 2019).
There have been some articles proposing RNA-bindng proteins as biomarkers and therapeutic approaches. Because the overexpression of RNA-binding motif protein 3 (RBM3) strongly associates with the early recurrence of PSA in more than 11,000 prostate cancer tissue samples, the analysis of RBM3 can be useful as prognostic biomarker alone or in combination with others (Grupp et al, 2013). Another RNA-binding protein called PSF can be the potential therapeutic target, due to its function associated with the development of abnormal splicing mechanism in CRPC and the important link with AR (Takayama et al, 2017). Furthermore, Musashi2 (MSI2) is found to be the AR upregulator in androgen-sensitive prostate cancer (ASPC) and CRPC, as it can directly bind 3’ untranslated region (3’ UTR) of AR mRNA and regulate mRNA stability of AR and protein expression. This indicates that MSI2 can be a new therapeutic target by inhibiting the protein (Zhao et al, 2020).
In addition, ERG R4 and R6 exon primers were working specifically to their sites in VCaP cells. In other words, they were able to specifically recognize the correct primer sites (Figure 8). Prior to the selection of these primers, agarose gel electrophoresis of all primers candidates for ERG exons were performed (Supplementary figure 2). ERG R4 and R6 exon primers were selected because they were specific to ERG exon 5 and 4, respectively. Agarose gel electrophoresis in Figure 8 revealed that these selected primers worked specifically in recognizing exon 5 and 4 of ERG gene in VCaP cells. The structure of ERG is consisted of 15
29 – 17 exons, alternatively. ERG can be transcribed to other several splice variants (Owczarek et al, 2004; Sreenath et al, 2011). These splice variants from ERG can be found in two major types, and may be used as prostate cancer biomarkers (Hu et al, 2008). In TMPRSS2-ERG fusion, exon 4 of ERG (the 218nt exon) can be maintained as the main partner of TMPRSS2, and exon 5 of ERG is referred to as a minor isoform (Zammarchi et al, 2013). The use of primers for exon 4 of ERG is seen in the study of TMPRSS2-ERG fusion detection using RT- PCR in urine sediment after digital rectal examination (DRE), yielding a successful detection of fusion transcript with 37% of sensitivity and 93% of specificity (Hessels et al, 2007).
Unfortunately, the last two aims were not achieved; the result for the third aim was inconclusive. We were not able to transfect ERG plasmid neither in PC-3 cells nor in HeLa cells (Figure 10). Inefficient plasmid transfection might result from inappropriate or ineffective plasmid concentration, inappropriate transfection reagents, the degradation of the plasmid after the transfection, and excess toxicity in cell lines. The solutions for future studies are to use different cell lines, different plasmid concentration, different transfection reagent, and to better optimize the reagent to plasmid ratio.
As a future development, the effect of FUS binding to TMPRSS2 and TMPRSS2-ERG mRNA transcripts on the prostate cancer cells proliferation should be studied. These prospective research projects should consider using different cell lines and transfection methods in their protocols. Moreover, there is sill a lack of insights on how FUS binds wild-type TMPRSS2 and fusion mRNA transcripts. Additional studies may be performed using bioinformatics approach, such as an RNA-binding protein predictor (RBPPred) to predict RNA and RNA-binding proteins interactions from sequences using the support vector maching (SVM) (Zhang & Liu, 2017), or RNA Bioinformatics Center within the German Network of Bioinformatics Infrastructure for prediction of binding motifs (Backofen et al, 2017).
Conclusion
This master thesis confirms that FUS protein can bind wild-type TMPRSS2 mRNA transcript in LNCaP cells and can also bind TMPRSS2-ERG fusion mRNA transcript in VCaP cells. The results from this study are useful for future research, studying possible effects of FUS binding in the proliferation of prostate cancer cells. Such increased understanding of these molecular mechanisms of prostate cancer may, in the prospect years, benefit patient survival prognosis and therapeutic approaches, such as drug discovery and development.
30
Acknowledgements
The thesis project was performed as a part of a study supported by the grant from the Academy of Finland. Moreover, I would like to express my gratitude to my supervisors, Leena Latonen and Mauro Scaravilli, for giving me this opportunity to work on this project for my master thesis, helping me with daily lab works and giving valuable comments and feedback on the research plan and the thesis. I also would like to say thank you to my family and friends for their encouragements that have been keeping me motivated and inspired to finish my thesis in a given timeframe. Last but not least, I would like to appreciate myself who does not give up when feeling down and tired, and achieves his goal within two years.
