TIA (T-cell intracellular antigen) proteins function as DNA/RNA trans-acting regulators to expand transcriptome and proteome diversity in mammals. In the present paper we report that the stable silencing of TIA1 and/or TIAR/TIAL1 (TIA1-related/like protein 1) expression in HeLa cells enhances cell proliferation, anchorage-dependent and -independent growth and invasion. HeLa cells lacking TIA1 and/or TIAR generate larger and faster-growing epithelial tumours with high rates of proliferation and angiogenesis in nude mice xenografts. Protein array analysis of a collection of human tumours shows that TIA1 and TIAR protein expression is down-regulated in a subset of epithelial tumours relative to normal tissues. Our results suggest a link between the epigenetic control exerted by TIA proteins and the transcriptional and post-transcriptional regulation of a subset of specific genes involved in tumour progression. Taken together, these results are consistent with a role for TIA proteins as growth/tumour-suppressor genes.
TIA1 (T-cell intracellular antigen 1) and TIAR/TIAL1 (TIA1-related/like) proteins are DNA/RNA-binding proteins involved in transcriptional and post-transcriptional regulation of eukaryotic gene expression . These proteins are structurally similar to hnRNPs (heterogeneous ribonucleoproteins) and encompass three RNA-recognition motifs and one C-terminal glutamine-rich domain [2,3]. TIA proteins have been shown to control nuclear and cytoplasmic regulatory events including gene transcription, alternative pre-mRNA splicing and turnover, and translation of cellular mRNAs. In the nucleus, they modulate the rates of gene transcription by interacting with DNA [4,5] and RNA polymerase II . TIA proteins also regulate alternative pre-mRNA splicing of approx. 15% of alternative cassette human exons, as well as 5′ splice-site selection of Alu exons, through binding to uracil-rich intronic stretches and facilitating atypical 5′ splice-site recognition by U1 small nuclear ribonucleoprotein [7–12]. In the cytoplasm, TIA proteins regulate the stability and translation of some mRNAs through their binding to adenosine-, uracil- and cytidine-rich sequences located in 3′ untranslated regions [13–19].
Consequently, TIA proteins have a pleiotropic role in the control of cell physiology. Indeed, roles for TIA proteins have been described in the regulation of apoptosis , cell death caused by viruses , viral replication , inflammation  and cell response to metabolic stress . Furthermore, mice lacking either TIA1 or TIAR show higher rates of embryonic lethality [14,25].
In the present paper, we show that TIA proteins control several pathways involved in different aspects of tumorigenesis. The down-regulation of TIA protein expression results in increased cell proliferation, tumour growth and invasion. Furthermore, we show that these proteins are down-regulated in a variety of tumours of epithelial origin. These results highlight the role of TIA proteins as tumour suppressors.
Cell culture, transfection, RNAi (RNA interference) and Western blot analyses
Adherent HeLa cells were cultured under standard conditions [1,10]. The TIA1 and TIAR shRNA–pSUPER (shRNA is short hairpin RNA) constructs were generated with the pSUPER RNAi System (OligoEngine), using specific sequences as described previously [1,10,26–28]. In fact, we had tested previously three different siRNAs (small interfering RNAs) against different regions of the genes for each of the two proteins , with similar silencing effects. Furthermore, we had generated mutant versions of the proteins resistant to the silencing siRNAs to exclude off-target RNAi effects . These are specific sequences against TIA1 and TIAR mRNAs and are used in the present study. Briefly, to knockdown TIA1 and TIAR, we have followed previously described methods (OligoEngine) to generate TIA1 and TIAR shRNA–pSUPER or empty pSUPER constructs from pSUPER.retro vectors. Sequences coding for shRNAs were inserted as double-stranded oligonucleotides followed by a short spacer, the reverse complement of the sense strand and five thymidines as an RNA polymerase III transcriptional stop signal into pSUPER vectors using the BglII and XhoI sites. The target sequences for the TIA1 and TIAR genes were 5′-AAGCTCTAATTCTGCAACTCT-3′ and 5′-AACCATGGAATCAACAAGGAT-3′ respectively, as described previously . To recapitulate the results of the silencing of the TIA proteins on cell proliferation and growth, two new siRNAs against TIA1 (5′-AACAACTAATGCGTCAGACTTTT-3′) and TIAR (5′-AAGTCCTTATACTTCAGTTGTTC-3′) mRNAs were used (Supplementary Figure S1 at http://www.BiochemJ.org/bj/435/bj4350337add.htm). The configuration of the constructs was verified by DNA sequencing. The TIA1 and TIAR shRNA–pSUPER constructs, and the empty pSUPER vector as a control, were transfected into the cell line HeLa (1 μg of DNA per 1×106 cells). Stable cell lines expressing empty, TIA1 and/or TIAR shRNA–pSUPER constructs were selected by plating for 3 weeks according to the manufacturer's instructions (OligoEngine). Clonal lines were isolated with cloning cylinders and verified by Western blotting.
Proliferation rates, cell-cycle analyses, anchorage-dependent and -independent growth, and invasion assays
Measurement of cell proliferation by a direct count of cell number for 4–6 days, and MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide] cell proliferation assays were carried out as described previously . Cell-cycle progression was analysed by flow cytometry and propidium iodide staining, and colony formation was assessed both on plastic and in soft agar as described previously . In vitro invasiveness was evaluated using Matrigel-coated invasion chambers with an 8 μm pore size (BD Biosciences). Briefly, 1×104 cells per 100 μl were allowed to migrate for 24 h through the Matrigel-coated membranes using 20% fetal bovine serum as a chemoattractant. The cells on the upper surface of the membrane were scraped, and the cells that reached the lower surface of the membrane were stained with 0.1% Crystal Violet, counted and then photographed under a light microscope at ×100 magnification.
Xenograft-tumour development in nude mice, HE (haematoxylin and eosin) staining and immunohistochemistry
Subconfluent HeLa cells stably transfected with empty vector, TIA1 and/or TIAR shRNAs (1–2×106 cells per site in 200 μl of PBS working solution) were injected subcutaneous into each rear flank in female nude mice (nu/nu; Harlam Ibérica). Tumour growth was monitored and then the mice were killed. Eight mice were used per experimental condition. Formalin-fixed paraffin-embedded tumour samples were stained with HE as described previously . For immunohistochemical analyses, the tumour samples were incubated with MIB1 (Master Diagnostica) at a dilution of 1:80, anti-E-cadherin (Master Diagnostica) at a dilution of 1:20, anti-VEGF (vascular endothelial growth factor; Santa Cruz Biotechnology) at a dilution of 1:400, and anti-TIA1, anti-TIAR and anti-HuR (Hu antigen R; Santa Cruz Biotechnology) at dilutions of 1:200, as described previously . The samples were visualized by light microscopy and photographed at 100× magnification. All of the experiments were performed with 8–12-week-old female mice and were conducted according to the guidelines of the Committee on Animal Experimentation of the Centro de Biología Molecular Severo Ochoa and Universidad Autónoma de Madrid.
Protein expression profiling using a human tumour array
A protein array (BioChain), consisting of 47 pairs of tumour samples and their respective normal tissues spotted as duplicates, was probed with anti-TIA1, anti-TIAR, anti-HuR, anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase; Santa Cruz Biotechnology) and anti-PTB (polypyrimidine tract binding protein)/hnRNP (BB7, generously given by Professor Chris Smith, Department of Biochemistry, University of Cambridge, U.K.). The blots were processed using the ECL (enhanced chemiluminescence) technique (GE Healthcare).
Knockdown of TIA proteins increases cell proliferation, growth and invasion
Knockdown of TIA1 and TIAR was performed using shRNA–pSUPER constructs and verified by Western blotting (see the Experimental section and Figure 1A). The comparison of control clones (transfected with empty vector) with three independent stably transfected clones for each condition revealed 60–90% depletions of both TIA1 and TIAR proteins in both the single- and double-knockdown experiments (Figures 1A–1C and Supplementary Figure S1). α-Tubulin protein levels, used as a control for shRNA specificity, were not significantly affected. These results agree with previous TIA1/TIAR siRNA-mediated depletion experiments in HeLa cells [1,10,26–28].
Knockdown of TIA proteins activates cell proliferation
The transient depletion of TIA1 and/or TIAR proteins in HeLa cells increased cell proliferation and growth ( and Supplementary Figure S1). To confirm these observations in stably transfected cell lines, we analysed the total cell number and metabolic activity. Reduced TIA1 and/or TIAR protein levels were correlated with proliferation in independent clones (Figures 1A–1C). An increase in the proportion of cells in G2/M-phase correlated with TIA1 and/or TIAR depletion (Figure 1D), thus supporting a role for these proteins in the control of cell proliferation. It is also noteworthy that depletion of TIA1 and TIAR significantly reduced the percentage of apoptotic cells (identified by the sub G1-phase peak; Figure 1D). For subsequent experiments the cell clones with stable knockdown of TIA1 and/or TIAR, denoted in Figures 1(A)–1(D) with (1), were chosen.
Many transformed cells are characterized by their capacity to grow in both an anchorage-dependent and -independent fashion and their invasive ability. To address the effect on the invasiveness of TIA1 and/or TIAR depletion, an investigation into the colony-forming capacity of the knockdown cells, on both plastic and soft agar plates, was performed. As shown in Figure 2, TIA1-, TIAR- and TIA1/R-depletion significantly increased the colony-forming capacity of the cells compared with the corresponding controls (Figures 2A and 2B). The invasiveness of these cells was also tested in the chemoinvasion assay, in which the cells move through a reconstituted basement membrane of Matrigel, coated on top of Transwell filters. The invasion potential of stable TIA1-, TIAR- and TIA1/R-depleted clones was significantly higher than that of control cells (Figure 2C). These results strongly support a role for TIA proteins in anchorage-dependent and -independent cell growth and invasion.
Depletion of TIA proteins triggers anchorage-dependent and -independent growth and invasion
Depletion of TIA proteins in HeLa cells triggers tumour growth potential in vivo
Tumour growth following TIA-protein knockdown was analysed in nude mice xenografts. TIA1-, TIAR- or TIA1/R-depleted cells, and their corresponding controls, were injected into the right and left hind leg respectively of female nude mice. Palpable tumours developed within 1–2 weeks, at which time animals were killed and the tumours analysed (Figure 3). Comparisons with the corresponding control tumours revealed a clear increase in tumour size in TIAR- (seven out of eight) and TIA1/R-depleted (eight out of eight) clones, whereas tumour size was more heterogeneous for the TIA1-depleted clone, showing only an evident increase in tumour size in five out of eight mice (Figure 3A). Interestingly, the depletion of both TIA1 and TIAR caused maximum tumour growth, thus suggesting a certain redundancy in the functions of both genes in the control of cell homoeostasis.
Silencing of TIA proteins supports tumour cell growth in vivo
Knockdown persistence was verified by immunohistochemistry of the isolated tumours using antibodies against TIA1 (Figure 4A) and TIAR (Figure 4B), or by reverse transcription–PCR (results not shown). As expected, the tumours from nude mice inoculated with TIA1-, TIAR- or TIA1/R-depleted HeLa cells had lower expression levels of the corresponding TIA1 or TIAR proteins compared with control tumours, which had focal and cytoplasmic expression of both proteins (Figures 4A and 4B; compare panel A with B and D). Formalin-fixed paraffin-embedded tumours were stained with HE or analysed by immunohistochemistry with MIB1 (to evaluate the proliferation index), anti-E-cadherin (to test epithelial nature and cell adhesion), anti-VEGF (to assess angiogenesis) or anti-HuR (to evaluate another RNA-binding protein). The histological features of all of the tumours indicated poorly differentiated carcinomas. The epithelial cells of the tumours had large nucleoli and nuclei, and displayed abundant events of mitosis (Figure 4C, and Supplementary Figures S2 and S3, panels A and B at http://www.BiochemJ.org/bj/435/bj4350337add.htm). The cells with reduced TIA1 and/or TIAR expression gave rise to large tumours, with a higher number of mitoses and a greater degree of necrosis, than the control group and only featured tumour cells around the blood vessels. Furthermore, immunohistochemistry analysis revealed that these tumour cells were highly proliferative, with higher binding to MIB1, undetectable E-cadherin expression and an extensive expression of VEGF, whereas control tumour cells had a lower proliferation index (MIB1), a lower expression of VEGF and focally maintained E-cadherin expression in the cell membrane (Figure 4C, and Supplementary Figures S2 and S3 panels C–H). The nuclear expression of HuR was similar in the tumours from the different experimental groups (Figure 4C, and Supplementary Figures S2 and S3 panels I and J). These findings suggest that the reduction in TIA1 and/or TIAR expression increases the in vivo tumorigenicity of HeLa cells.
Immunohistochemical characterization of xenograft tumours derived from TIA1- and/or TIAR-depleted HeLa cells
Down-regulation of TIA proteins in human cancers
To assess further the role of TIA proteins in tumorigenesis, we examined the expression of TIA proteins, and other RNA-binding proteins, in a panel of normal and tumour-derived tissues (Figure 5 and Supplementary Figure S4 at http://www.BiochemJ.org/bj/435/bj4350337add.htm). TIA1 and/or TIAR proteins were down-regulated in many tumour types including adrenal gland, lung, ovary, pancreas, parotid gland, skin, small intestine, stomach, thymus and uterus for TIA1, and brain, breast, colon, duodenum, lung, lymphoma, ovary, pancreas, skin, thymus, thyroid gland, ureter and uterus for TIAR (Figure 5 and Supplementary Figure S4). A subset of these tumours also showed variable expression of HuR and PTB, two closely related RNA-binding proteins, whereas the GAPDH metabolic protein showed characteristic expression levels according to the glycolytic potential of the different tumours. These results suggest that the expression of TIA proteins can be specifically down-regulated in several human cancers of epithelial origin.
Down-regulation of TIA proteins in human tumours
Association of TIA proteins with the repression of malignant cell phenotypes
The results of the present study show that the stable knockdown of TIA1 and/or TIAR in HeLa cells leads to increased cell proliferation, invasiveness and tumour growth. These clones produced larger and faster-growing tumours following injection into nude mice. In contrast, cells that stably overexpress TIA1 (ten out of ten clones) or TIAR (nine out of ten clones) are not viable and undergo apoptosis (results not shown, and ). We also showed that a subset of human tumours had a reduced expression of TIA1 and/or TIAR. Previous results showed that mutant mice lacking either TIA1 or TIAR have a hyperinflammatory phenotype due to the overexpression of cytokines [13,14,23]. Epidemiological and experimental data suggest a close connection between inflammation and cancer [30,31]. Thus many inflammatory mediators are important for the growth and survival of premalignant cells and activate oncogenic transcription factors such as NF-κB (nuclear factor κB)  or oncogenes such as MYC, which can also initiate the inflammatory response . Interestingly, these oncogenic factors can be regulated by TIA proteins at the transcriptional and translational levels [1,33]. Tumour-associated inflammation can suppress anti-tumour immune responses and cause tumour-specific immune cells to become tumorigenic . The inflammatory response can also stimulate tumour angiogenesis, invasion and metastasis . Mice with a disruption in the tiar gene develop ovarian sex cord stromal tumours . Furthermore, TIA proteins modulate alternative splicing of WT1 (Wilm's tumour-suppressor gene)  and the neurofibromatosis 1 gene NF1 , whose protein isoforms have been implicated in the development of childhood kidney cancer and fibrosarcoma respectively. In the same vein, TIA1 expression in tumours correlates strongly with responsiveness to immunotherapy for melanoma patients . Thus down-regulation of TIA proteins could play a role in oncogenesis by modulating the genes involved in neoplastic and malignant transformation processes to evade the immune system and to enhance the growth and survival of cancer cells [37–39].
To our knowledge, TIA1 proteins have never been reported to be mutated in cancer. Human cancers are caused by both genetic and epigenetic alterations. Progressive acquisition of mutations in oncogenes (gain-of-function mutations) or tumour-suppressor genes (loss-of-function mutations) might act in concert with epigenetic events, such as the functional down-regulation of TIA proteins, to provide transformed cell clones with a competitive growth advantage. TIA proteins regulate the transcription, alternative splicing, stability and/or translation of many human genes with potential roles in tumour development and progression (Supplementary Figure S5 at http://www.BiochemJ.org/bj/435/bj4350337add.htm and the References section of the Supplementary Material). Thus the tumour-suppressor role of TIA proteins could influence mechanisms such as the control of cell proliferation, apoptosis, angiogenesis, invasion and metastasis, and evasion of immune recognition (Figure S5). Further studies on the anti-oncogenic role of TIA proteins are warranted.
We thank José Manuel Sierra and Juan Manuel Zapata for support and encouragement. We are also grateful to Juan Valcárcel and Chris Smith for providing reagents.
haematoxylin and eosin
Hu antigen R
v-myc myelocytomatosis viral oncogene homologue (avian)
polypyrimidine tract binding protein
small interfering RNA
short hairpin RNA
T-cell intracellular antigen 1
vascular endothelial growth factor A
María Dolores Ludeña and José Izquierdo conceived the research and designed all the experiments. María Dolores Ludeña, José Alcalde, Isabel Carrascoso, Raquel Reyes and José Izquierdo carried out the experiments. María Dolores Ludeña and José Izquierdo wrote the paper.
This work was supported by the Fondo de Investigaciones Sanitarias-FEDER [grant number PI051605]; the Ministerio de Ciencia e Innovación-FEDER [grant number BFU2008–00354]; and the Agencia Estatal Consejo Superior de Investigaciones Científicas [grant number 200920I108]. The Centro de Biología Molecular ‘Severo Ochoa’ receives an institutional grant from the Fundación Ramón Areces, Spain.