Summary
Tumor necrosis factor-alpha (TNF-α) plays important roles in chronic inflammation-associated tumorigenesis but the mechanisms involved remain poorly understood. Previously, we reported that high levels of FAT10 led to chromosomal instability that is mediated by an abbreviated mitotic phase. Here, we show that TNF-α induces FAT10 gene expression through TNF receptor 1 (TNFR1) and activates the NF-κB pathway in HCT116 and SW620 cells. TNF-α treatment also leads to an abbreviated mitotic phase that can be reversed by inhibiting FAT10 expression. This abbreviated mitotic phase is correlated with a TNF-α-induced reduction in the kinetochore localization of MAD2 during prometaphase which, again, can be reversed by inhibiting FAT10 gene expression. There is greater variability of chromosome numbers in HCT116 and SW620 cells treated with TNF-α than in untreated cells, which can be reversed by the introduction of short hairpin RNA (shRNA) against FAT10. The more stable chromosome numbers in HCT116 cells expressing FAT10 shRNA can revert to greater variability with the addition of a mutant FAT10 that is not recognized by the FAT10 shRNA. Upon TNF-α stimulation, higher cell death is observed when FAT10 expression is inhibited by shRNA. These data strongly suggest that FAT10 plays an important role in mediating the function of TNF-α during tumorigenesis by inducing cell cycle deregulation and chromosomal instability, and by inhibiting apoptosis.
Introduction
Ever since Rudolf Virchow first observed in the nineteenth century that tumors often arose at sites of chronic inflammation, there is now increasing epidemiological as well as molecular evidence to further suggest the association between chronic inflammation and tumorigenesis (Colotta et al., 2009; Maeda and Omata, 2008; Marx, 2004; Schafer and Werner, 2008). Inflammation has been implicated in several cancers including liver (Szabo et al., 2007), colorectal (Itzkowitz and Yio, 2004; Lakatos and Lakatos, 2008; Terzic et al., 2010), gastric (McNamara and El-Omar, 2008), lung (Engels, 2008), breast (Hojilla et al., 2008) and cervical (Hiraku et al., 2007) cancers. Anti-inflammatory drugs have been found to be effective in reducing the risk of development of certain cancers as well as in reducing the mortality caused by these cancers (Arrieta et al., 2006; Kumada, 2002; Mantovani et al., 2008).
The molecular mechanisms through which chronic inflammation promotes carcinogenesis is currently under intense investigation (Allavena et al., 2008). One inflammation-associated molecule that has been implicated to play crucial roles in the development of cancer is the cytokine, tumor necrosis factor-alpha (TNF-α), whose levels are elevated in the serum during inflammation (Arnott et al., 2004; Moore et al., 1999; Popivanova et al., 2008). TNF-α is a pleiotropic proinflammatory cytokine that plays important yet diverse roles in cell survival, proliferation, differentiation and death. TNF-α was also demonstrated to cause cytogenetic changes such as telomere shortening, DNA breaks, DNA end-to-end fusions and numerical chromosomal instability (CIN) or aneuploidy (Beyne-Rauzy et al., 2004; Li et al., 2007; Yan et al., 2006). TNF-α plays dual but paradoxical roles in cancer both as a tumor promoter and as an anti-tumor cytokine (Balkwill, 2009). The tumor-promoting activities of TNF-α are thought to be mediated by TNF receptor-1 (TNFR1), which is found primarily on the tumor and stroma cells, whereas TNF receptor-2 (TNFR2) is generally present on the leukocyte infiltrate (Balkwill, 2009).
Binding of TNF-α to the TNFR1 activates the NF-κB signaling pathway, which is a major mediator of the tumor-promoting activities of TNF-α (Balkwill, 2009). NF-κB is a transcription factor that can be activated by more than 150 stimuli including TNF-α (Pahl, 1999). In non-activated cells, NF-κB dimers are found in the cytoplasm but they translocate to the nucleus upon activation to promote the expression of more than 150 target genes primarily involved in the immune response, including numerous transcription factors (Pahl, 1999), hence making NF-κB the central mediator of the immune response.
One of the gene targets of TNF-α is FAT10 or di-ubiquitin (UBD), an 18 kDa protein whose expression is synergistically inducible by TNF-α and interferon-γ (IFN-γ) (Raasi et al., 1999). The expression of FAT10 was also found to be cell-cycle regulated (Lim et al., 2006) and is also negatively regulated by p53 (Zhang et al., 2006). FAT10 belongs to the ubiquitin-like modifier (UBL) family of proteins and contains two ubiquitin-like moieties fused in tandem (Jentsch and Pyrowolakis, 2000). At the N-terminus, FAT10 is 29% identical to ubiquitin, whereas at the C-terminus, it is 36% identical to ubiquitin. FAT10 conserves the free di-glycine motif of ubiquitin at the C-terminus, which might play roles in conjugating to yet-to-be-identified target proteins similar to ubiquitin (Chiu et al., 2007; Raasi et al., 2001). The FAT10 conjugation cascade was reported to involve the E1 enzyme UBA6 (E1-L2, UBE1L2) (Chiu et al., 2007) and the UBA6-specific E2 enzyme, USE1 (Aichem et al., 2010). This enzyme was also suggested to be a substrate of FAT10 as it is auto-FAT10lyated only in cis but not in trans. Artificial fusion of FAT10 to the long-lived proteins GFP and dihydrofolate reductase (DHFR) leads to the degradation of the fusion proteins (Hipp et al., 2005; Hipp et al., 2004), which is independent of poly-ubiquitylation (Schmidtke et al., 2009) but can be accelerated by the UBL-UBA domain protein, NEDD8 ultimate buster 1-long (NUB1L) (Hipp et al., 2005; Hipp et al., 2004). FAT10 has also been reported to interact noncovalently with MAD2 (Liu et al., 1999) and HDAC6 (Kalveram et al., 2008). Nonetheless, the functions of FAT10 remain unclear. FAT10-deficient mice exhibit minimal phenotype change except that their lymphocytes are more sensitive to endotoxin-induced apoptosis (Canaan et al., 2006).
FAT10 was reported to be upregulated during various viral infections including Kaposi sarcoma-associated herpesvirus infected cells (Hong et al., 2004) as well as renal tubular reticular cells (RTEC) in HIV-associated neuropathy (HIVAN), where FAT10 interacts with the HIV accessory protein Vpr to mediate Vpr-induced cell death (Ross et al., 2006; Snyder et al., 2009).
Significantly, FAT10 was found to be highly upregulated in the tumors of hepatocellular carcinoma (HCC) and other gastrointestinal cancers (Ji et al., 2009; Lee et al., 2003; Lukasiak et al., 2008), non-small cell lung cancer (Heighway et al., 2002) as well as mantle cell lymphoma (Martinez et al., 2003). FAT10 was even recently implicated to be a HCC stem-cell marker (Oliva et al., 2010b) as well as an epigenetic marker for liver preneoplasia in a drug-primed mouse model of tumorigenesis (Oliva et al., 2008). FAT10 was also one of the genes that was found to be overexpressed in a carcinogen-induced rat stomach cancer model (Yamashita et al., 2002).
Upregulation of FAT10 gene expression in cancers was found to correlate with the expression of proinflammatory cytokines, implicating potential roles of FAT10 in inflammation-induced tumorigenesis (Lukasiak et al., 2008). However, it remains unclear how FAT10 promotes cancer development. Nonetheless, FAT10 was found to interact with MAD2, a spindle checkpoint protein, during mitosis, and high levels of FAT10 protein in cells led to increased mitotic non-dysjunction and CIN (Ren et al., 2006). This effect was mediated by an abbreviated mitotic phase and the reduction in the kinetochore localization of MAD2 during the prometaphase stage of the cell cycle (Ren et al., 2006). Since both TNF-α (Beyne-Rauzy et al., 2004) and FAT10 (Ren et al., 2006) can induce aneuploidy in cells and aneuploidy has been proposed to play an important role in carcinogenesis (Weaver and Cleveland, 2006) based on various experimental (Babu et al., 2003; Baker et al., 2004; Michel et al., 2001) and clinical (Hitzler and Zipursky, 2005) observations, it is important to investigate whether FAT10 might play important roles in TNF-α-associated tumorigenesis.
Here, we demonstrate that FAT10 expression is upregulated by TNF-α through the NF-κB pathway. FAT10 also mediates the effect of TNF-α in inducing abbreviated mitosis and aneuploidy in cells. Additionally, high levels of FAT10 were found to protect cells from TNF-α-induced apoptosis.
Results
Endogenous FAT10 expression is induced through TNF-α and the NF-κB pathway
FAT10 gene expression in the human colon cancer MIN (microsatellite instability) cell-line, HCT116 (Dunican et al., 2002; Silkworth et al., 2009), is more highly induced by TNF-α alone than by IFN-γ alone but its expression can be synergistically induced when both cytokines are simultaneously introduced (Fig. 1A and supplementary material Fig. S1). Notably, whereas TNF-α significantly increased FAT10 expression in IFN-γ-pretreated cells, IFN-γ did not increase FAT10 expression in TNF-α-pretreated HCT116 cells (Fig. 1A), probably due to the upregulation of TNF-α receptor levels by IFN-γ in these cells, similar to the effect of IFN-γ on A2780 and Caov-3 cells that was previously reported (Kost et al., 1999).
As evident in Fig. 1B, the degree of FAT10 induction by TNF-α and/or IFN-γ varies in different cell lines. IFN-γ alone can induce detectable FAT10 protein expression only in the human liver cell line HepG2, similar to previous observations in mouse liver Hepa 1–6 cells (Oliva et al., 2010a). TNF-α alone can induce detectable FAT10 protein expression only in the CIN SW620 cells (Bertholon et al., 2003; Silkworth et al., 2009) (Fig. 1B). However, FAT10 expression was not detectable by western blot analyses when HEK293, HCT116, and the immortalized non-tumorigenic liver cell line THLE-3 (Pfeifer et al., 1993) were treated with IFN-γ or TNF-α alone. Nonetheless, using real-time RT-PCR, more FAT10 transcripts were observed in HCT116 cells induced with TNF-α alone than with IFN-γ alone (Fig. 1A). Hence, these data suggest that different cytokines affect different cells differently, probably due to the difference in the expression of the different receptors in the different cells. Consistent with our previous observations (Ren et al., 2006), the endogeneous FAT10 that is induced by TNF-α and IFN-γ localizes to the nucleus of HCT116 cells (Fig. 1C).
In this study, we wanted to elucidate the role of FAT10 in the TNF-α pathway. Because TNF-α plays a more dominant role than IFN-γ in inducing FAT10 expression (Fig. 1A) in HCT116 cells, where the role of FAT10 in CIN was previously reported (Ren et al., 2006), we focused primarily on the role of FAT10 in modulating only the TNF-α-induced phenotype in HCT116 cells to facilitate the clearer interpretation of data. To determine the generality of the observations, we also compared some of the phenotypes observed in the MIN HCT116 cells with those of another TNF-α pathway dominant but CIN cell line, SW620. As shown in Fig. 1D, the induction of FAT10 expression in HCT116 cells by TNF-α is transient, with FAT10 expression decreasing 6 hours after the withdrawal of TNF-α.
Of the two receptors, TNFR1 and TNFR2 (Baker and Reddy, 1998; Chen and Goeddel, 2002), that TNF-α can act through to exert its effects, TNFR1 was found to be the receptor that TNF-α utilizes to modulate the expression of FAT10 in HCT116 cells. The effect of TNF-α on FAT10 gene expression was only greatly diminished in cells containing small interfering RNA (siRNA) against TNFR1 but not TNFR2 (Fig. 2A). We found that TNF-α modulated FAT10 expression through the NF-κB pathway. As shown in Fig. 2B, FAT10 expression in HCT116 cells containing control siRNA (siCTR) can be induced by TNF-α, whereas TNF-α induction of FAT10 expression is significantly inhibited in cells containing siRNA against FAT10 (siFAT10a and siFAT10b, having different target sequences, see supplementary material Table S2) or against NF-κB subunit p65 protein (sip65). A similar phenotype was also observed in the SW620 cell-line, where inactivation of the NF-κB pathway using siRNA against p65 led to the inhibition of TNF-α induced FAT10 expression (supplementary material Fig. S2).
TNF-α induces endogenous FAT10 expression. (A) FAT10 expression is synergistically induced by TNF-α and IFN-γ. HCT116 cells were grown for 6 hours with either TNF-α or IFN-γ alone, or for 15 hours with IFN-γ then 6 hours with TNF-α, or for 15 hours with TNF-α then 6 hours with IFN-γ. RNA was then extracted from the cells. RT-PCR was performed and transcript levels estimated using real-time-RT PCR. (B) FAT10 expression is induced by TNF-α in various cell lines. Cells were treated with or without TNF-α and IFN-γ for 9 hours and harvested for western blot analysis. (C) Endogenous FAT10 induced by TNF-α localizes in the nucleus. Cells were grown with or without 50 ng/ml TNF-α and 50 IU/ml IFN-γ for 9 hours. Fluorescent immunostaining was performed to visualize FAT10 (green), which localizes to the nucleus. β-actin (red) showed the cytoplasm and DAPI (blue) was used to show nuclear localization. Scale bar: 10 μm. (D) Induction of FAT10 by TNF-α is transient. Cells were cultured with TNF-α for 12 hours and then further grown in the presence or absence of TNF-α for the stipulated time periods before the cells were harvested, RNA extracted and RT-PCR performed.
FAT10 mediates the effect of TNF-α on regulation of mitosis
As high levels of FAT10 in cells might lead to an abbreviated mitosis (Ren et al., 2006), we investigated whether TNF-α can influence mitosis through the induction of FAT10. To observe mitotic progression in live cells, we stably expressed H2B–GFP fusion protein in HCT116 cells to monitor the status of DNA (supplementary material Fig. S3A). Real time RT-PCR analysis showed that TNF-α induces FAT10 expression in these stable cells (supplementary material Fig. S3B). Mitotic duration was determined from the time of DNA condensation in cells with intact nuclear membrane to the time when the chromosomes began to segregate. HCT116 cells required ~45 minutes (46.0±3.4 minutes) to complete mitosis under normal culture conditions (supplementary material Fig. S3C and Fig. 3A, top panel). However, when these cells were treated with TNF-α for 3 hours, the mitotic duration was reduced by ~10 minutes (35.1±4.3 minutes) (Fig. 3A). To evaluate whether FAT10 mediates the effect of TNF-α on mitosis, HCT116 cell lines stably expressing either control shRNAs (CTRi cells) or FAT10 shRNAs (FAT10i cells) were engineered (supplementary material Fig. S4A). These stable cells are karyotypically similar (supplementary material Fig. S4C). As evident in supplementary material Fig. S4B, FAT10 expression is significantly suppressed in cells stably expressing shRNA against FAT10 (FAT10ia and FAT10ib) compared with cells stably expressing control shRNAs (CTRia and CTRib) upon TNF-α treatment. Clones CTRib and FAT10ib were then stably transfected with H2B–GFP (H2BGFP-CTRib and H2BGFP-FA10ib) to monitor their mitotic profiles. These cells were then treated with 50 ng/ml TNF-α for 6 hours before the mitotic duration was determined. No difference in the duration was observed between H2BGFP-CTRib cells (49.5±4.4 minutes) and H2BGFP-FAT10ib cells (49.0±6.2 minutes) in the absence of TNF-α. However, in the presence of TNF-α, the mitotic duration in H2BGFP-CTRib (36.1±5.0 minutes) was ~10 minutes shorter than in H2BGFP-FAT10ib cells (46.6±6.5 minutes) (Fig. 3B).
TNF-α reduces the kinetochore aggregation of MAD2 during prometaphase through FAT10
Because TNF-α affected mitotic duration through FAT10, and FAT10 was reported to influence mitosis by interfering with the activities of the checkpoint protein MAD2 (Ren et al., 2006), we proceeded to investigate whether TNF-α can similarly affect MAD2 functions during mitosis through FAT10. HCT116 cells were transfected with plasmids encoding the MAD2–EGFP fusion protein (Ren et al., 2006) and cultured for five passages before MAD2–EGFP-expressing fluorescent cells were enriched using the FACSAria II cell sorter (BD Bioscience). In untreated cells, the MAD2–EGFP fusion protein aggregated at the kinetochore during prometaphase. When the cells were treated with TNF-α, the MAD2–EGFP fluorescent signals become more diffused, suggesting that the MAD2 protein had been delocalized from the kinetochore (Fig. 4A). To determine whether FAT10 mediated this delocalization of MAD2 from the kinetochore, siRNAs against control or FAT10 were transfected into the MAD2–EGFP-expressing cells by electroporation. In the absence of TNF-α treatment, the mean percentage of cells exhibiting clear aggregation of MAD2 at the kinetochore during prometaphase was ~77% and 73% for siCTR-treated and siFAT10a-treated cells, respectively. When these cells were treated with TNF-α, there was a significant reduction in the percentage of siCTR-treated cells (24%) showing aggregation of MAD2 at the kinetochores compared with siFAT10a-treated cells (56%) or untreated cells (>70%) (Fig. 4B). These results suggest that the redistribution of MAD2 away from the kinetochores during prometaphase under TNF-α treatment is mediated through FAT10.
TNF-α induces FAT10 expression through the TNFR1–NF-κB signaling pathway. (A) TNF-α induces FAT10 expression though TNFR1 but not TNFR2. HCT116 cells were electroporated with siRNAs against FAT10, TNFR1 or TNFR2 and grown for 24 hours. The cells were then cultured in medium with or without TNF-α for 6 hours before RNA extraction and RT-PCR. Asterisk indicates that TNFR2 gene-specific primers were utilized for the reverse transcription of TNFR2. (B) TNF-α induces FAT10 expression though activated NF-κB. Cells were electroporated with siRNAs against FAT10 or p65 and grown for 24 hours. They were then cultured with or without TNF-α for 6 hours before RNA extraction and real time RT-PCR. Left: Data represent mean ± s.d. of five technical replicates. Similar trends were observed in another biological replicate. *P<0.001 comparing cells containing either siFAT10 or sip65 with those containing siCTR. Right: Inhibition of NF-κB subunit p65 expression by its specific siRNA was confirmed by western blot.
FAT10 mediates the effect of TNF-α on chromosomal stability
We also investigated whether the abbreviated mitosis induced by TNF-α led to chromosomal instability. HCT116 cells at 10% confluency were grown in 50 ng/ml TNF-α-supplemented medium for 24 hours before TNF-α was removed and cells were allowed to grow until confluence. The cells were then passaged and TNF-α again added for another 24 hours, after the passaged cells became attached. This process was repeated for about ten passages (~30 doublings). The cells were then harvested for determination of chromosome number. Greater than 80% of untreated cells but only ~40% of cells chronically treated with TNF-α retained the modal chromosome number of 40–49. Most of the TNF-α-treated cells exhibited abnormal chromosome number (Fig. 5A), similar to what was observed in FAT10-overexpressing HCT116 cells (Ren et al., 2006). To evaluate whether FAT10 plays a role in mediating the effect of TNF-α in inducing CIN in HCT116 cells, HCT116 cells stably expressing either control shRNAs (CTRia and CTRib cells) or FAT10 shRNAs (FAT10ia and FAT10ib cells) (supplementary material Fig. S4) were used. Karyotyping and comparative genomic hybridization (CGH) microarray analyses revealed that CTRi and FAT10i cells have similar chromosome numbers (supplementary material Fig. S4C) and cytogenetic profiles (supplementary material Fig. S4D). These stable cells were then mock-treated or treated with TNF-α for ten passages as described above and the cytogenetic changes examined. Similar cytogenetic profiles with normal modal chromosome numbers of 40–49 were observed in mock-treated HCT116 cells stably expressing either control (CTRia and CTRib cells) or FAT10 (FAT10ia and FAT10ib cells) shRNAs (Fig. 5B, left panel). However, under chronic TNF-α treatment, more control shRNA-expressing stable cells (CTRia and CTRib cells) were found to exhibit abnormal chromosome numbers. Only ~40% of these cells retained normal chromosome numbers of 40–49 (Fig. 5B, right panel). However, the majority (~70%) of cells stably expressing FAT10 shRNA (FAT10ia and FAT10ib cells), in which the expression of FAT10 is suppressed, display normal modal chromosome numbers of 40–49 after chronic TNF-α treatment (Fig. 5B, right panel). Similar observations were made for the SW620 cell line. Upon chronic TNF-α treatment, a greater percentage (~70%) of SW620 cells stably expressing FAT10 (SW620FAT10i cells), resulting in less induction of the FAT10 expression, showed relatively normal chromosome numbers than the same cells stably expressing control shRNA (SW620CTRi cells) (43%) (Fig. 5C).
TNF-α accelerates mitosis through FAT10. Cells stably transfected with H2B–GFP were grown with or without TNF-α for 6 hours and live cell fluorescence microscopy was employed to visualize the mitotic process in cells. Mitosis duration was determined from the time when DNA condensation commenced but the nuclear membrane remained intact to the time when sister chromosomes started to segregate. (A) TNF-α induces abbreviated mitotic phase in HCT116 cells. Left: Data represent mean ± s.d. of mitosis duration in 20 cells expressing H2B–GFP; *P<0.05. Right: Mitotic progression in a representative cell of each group. (B) FAT10 mediates the effect of TNF-α in regulation of mitotic duration. H2B–GFP cells stably expressing control and FAT10 shRNAs were examined. Left: Data are expressed as mean ± s.d. of mitosis duration in ten cells. *P<0.05 comparing stable H2B–GFP CTRib cells induced by TNF-α with other cells or treatment. Right: Mitosis progression in a representative cell of each group. Scale bars: 10 μm.
To ascertain that the effect of TNF-α on chromosome stability is due directly to FAT10 and not through other genes that the FAT10 shRNA might also influence, a construct expressing a mutant FAT10 that was resistant to the FAT10 shRNA was engineered (supplementary material Fig. S5). This mutant FAT10-expressing plasmid and a vector control were then transfected into HCT116 cells stably expressing the FAT10 shRNA (supplementary material Fig. S4) and the cells grown in TNF-α supplemented medium for 24 hours. Thereafter, the cells were grown in normal medium until the next subculture. The subcultured cells were then again transfected with either the vector control or the mutant FAT10 gene and the cells again treated with TNF-α for 24 hours. These transfection and treatment procedures were repeated ten times before cells were harvested for karyotype analysis. As evident in Fig. 5D, ~77% of control vector-transfected HCT116 cells stably expressing FAT10 shRNA still retained normal chromosome number. However, only <50% of mutant FAT10-transfected HCT116 stably expressing shRNA against FAT10 retained normal chromosome number (Fig. 5D). This result strongly suggested that FAT10 is a direct mediator of the effect of TNF-α on chromosomal instability.
FAT10 facilitates the resistance of cells against TNF-α-induced cell death
In addition to causing CIN through FAT10 (Fig. 5), TNF-α stimulation has previously been shown to induce apoptotic cell death (Aggarwal, 2003). As numerical CIN might facilitate the resistance of HCT116 cells against DNA damage-induced cell death (Castedo et al., 2006), we evaluated whether FAT10 plays a role in facilitating this resistance to TNF-α-induced cell-death. HCT116 cells stably expressing either control or FAT10 shRNA were either mock-treated or treated with TNF-α for 24 h before the apoptosis profiles were examined. Similar apoptosis profiles were observed when both types of cells were mock-treated (5–6% apoptosis). Interestingly, a significantly greater percentage of FAT10 shRNA-expressing cells (18-23%) underwent apoptosis compared with control shRNA-expressing cells (12–13%) (Fig. 6A).
FAT10 mediates the delocalization of MAD2 away from the kinetochores during mitosis caused by TNF-α. Cells transfected with MAD2–EGFP were grown with or without TNF-α for 6 hours, paraformaldehyde fixed and then visualized using confocal microscopy. Early mitosis was determined according to the extent of condensation of DNA stained with DAPI. (A) TNF-α treatment causes the delocalization of MAD2 in HCT116 cells. (B) FAT10 mediates the effect of TNF-α on MAD2 localization during mitosis. MAD2–EGFP-expressing HCT116 cells were transfected with siCTR or siFAT10 before TNF-α treatment. Cells showing clear MAD2–EGFP aggregation at the kinetochores were considered positive cells. In upper panel, each column represents mean ± s.d. of data obtained from three independent counts, each count including at least 20 prometaphase cells; *P<0.01. Inset: RT-PCR for FAT10 and β-actin expression. Lane 1, siCTR, no TNF-α; lane 2, siFAT10a, no TNF-α; lane 3, siCTR, 50 ng/ml TNF-α; lane 4, siFAT10a, 50 ng/ml TNF-α. Lower panel shows representative images of each group of cells. Scale bars: 10 μm.
The percentage of cells expressing either control or FAT10 shRNA that resisted TNF-α induced cell-death was then evaluated. The number of surviving cells upon mock or TNF-α treatment were counted at specified time points and the percentage of surviving cells was determined as the ratio of the number of surviving cells treated with TNF-α versus the number of surviving cells that were mock-treated. As evident in Fig. 6B, the percentage of cells surviving TNF-α is significantly greater (P<0.05) in HCT116 cells expressing control shRNA compared with cells expressing the FAT10 shRNA. Similarly, upon TNF-α treatment, more colonies on soft agar were observed in HCT116 cells carrying control shRNA compared with cells where the FAT10 gene expression was silenced (Fig. 6C). These results suggest that FAT10 plays a role in facilitating the survival of cells upon TNF-α insult. In summary, FAT10 plays an important role in protecting cells against TNF-α-induced cell death.
Discussion
Chronic inflammation has long been associated with cancer development (Schafer and Werner, 2008), and the proinflammatory cytokine, TNF-α, had been shown to play an important role in promoting inflammation-associated tumorigenesis (Balkwill, 2009) through the NF-κB pathway (Karin, 2006). More than 150 target genes (Pahl, 1999) of NF-κB have been identified, of which some might be involved in tumorigenesis by regulating the balance between apoptosis and cell survival upon TNF-α induction (Aggarwal, 2003). Here, we demonstrate that FAT10, the ubiquitin-like protein that is overexpressed in the tumors of various cancers (Heighway et al., 2002; Lee et al., 2003; Martinez et al., 2003), can be induced by TNF-α through the NF-κB pathway (Figs 1 and 2) mediated via TNFR1 (Fig. 2A), which is one of two receptors that TNF-α can act through (MacEwan, 2002). The action of TNF-α through the TNFR1 was implicated to play important roles in tumorigenesis because deletion of TNFR1 resulted in lower cancer incidence upon carcinogen treatment (Popivanova et al., 2008). The action of TNF-α through the TNFR2 was suggested to mediate tumor suppression (Zhao et al., 2007), probably by enhancing TNFR1-induced apoptosis (Fotin-Mleczek et al., 2002). Hence, the induction of FAT10 expression by TNF-α through the TNFR1 strongly implicates potential roles of FAT10 in promoting tumorigenesis.
TNF-α induces chromosomal instability through FAT10. (A) TNF-α induces CIN. HCT116 cells were grown with or without TNF-α treatment for ten passages before they were karyotyped. Inset: RT-PCR shows that that FAT10 expression is induced by TNF-α. Data are expressed as mean ± s.d. of results of three karyotype analyses, each examining 40–45 cells. (B) TNF-α induced less CIN in HCT116 cells in which FAT10 expression was inhibited. Different clones of HCT116 cells (CTRia, CTRib, FAT10ia and FAT10b cells) were cultured with (right) or without (left) TNF-α treatment for ten passages before karyotyping was performed. 45-55 cells were karyotyped. Inset: Real-time RT-PCR shows that that FAT10 expression is induced by TNF-α. *P<0.01 compared with clones expressing control shRNA. (C) TNF-α induces less CIN in SW620 cells where FAT10 expression is inhibited. SW620 cells stably expressing either control shRNA (SW620CTRi cells) or shRNA against FAT10 (SW620FAT10i cells), were cultured with (right) or without (left) TNF-α treatment for ten passages before karyotyping was performed. 40 cells were karyotyped. Inset: Western blot analyses show that that FAT10 expression is induced by TNF-α. (D) Overexpression of FAT10 induces chromosomal instability in HCT116 FATi cells. Stable HCT116 FATib cells were transfected with control vector or with a mutant FAT10 plasmid (mFAT10) that is resistant to the FAT10-specific siRNA. The cells were then treated with TNF-α for 24 hours and then grown in normal medium until the next subculture. The transfection and treatment procedures were repeated ten times before the cells were harvested and karyotyped. Data represent mean ± s.d. of results of three karyotype analyses, each examining 40–50 cells. Inset: western blot analyses showing FAT10 expression.
Notably, we observed that TNF-α stimulation resulted in the reduction of the mitotic duration in cells by ~20% through FAT10 (Fig. 3), suggesting that TNF-α might play some roles in the induction of premature anaphase through FAT10. This is consistent with previous reports implicating TNF-α in cell cycle regulation through the regulation of the expression of cell cycle proteins (Yu et al., 2000) and the promotion of cell proliferation (Son et al., 2004; Spaczynski et al., 1999), as well as our previous observations that overexpressed FAT10 leads to abbreviated mitosis (Ren et al., 2006). We have also demonstrated that, under TNF-α treatment, FAT10 can mediate the redistribution of the mitotic checkpoint protein MAD2 away from the kinetochores during mitosis, suggesting that TNF-α deregulates the cell-cycle through the effect of FAT10 on the distribution of MAD2 during mitosis (Fig. 4). Significantly, we provide evidence to demonstrate that TNF-α can induce CIN through FAT10 in both the HCT116 as well as SW620 cells (Fig. 5), which consistent with our current understanding that premature exit from mitosis causes numerical CIN (Schvartzman et al., 2010). Curiously, it was mentioned in one study as ‘data not shown’ that aneuploidy was not observed in DDC (diethyl1,4-dihydro-2,4,6-trimethyl-3,5-pyridinedi carboxylate)-fed mice in which FAT10 expression is induced (Oliva et al., 2008).
The role that aneuploidy plays in tumorigenesis remains under intense investigation. Recently, it has become clearer that aneuploidy is not merely a passive passenger phenotype but is implicated as a potential tumorigenic driving force (Schvartzman et al., 2010; Weaver and Cleveland, 2006) that precedes oncogenic transformation (Duesberg and Li, 2003) although it might suppress tumorigenesis in certain genetic contexts and cell types (Holland and Cleveland, 2009; Weaver et al., 2007). CIN has also been suggested to deregulate apoptosis surveillance, leading to the selection of cells with increased cell survival fitness and the potential to become tumorigenic (Castedo et al., 2006; Zhivotovsky and Kroemer, 2004). Because TNF-α, a molecule involved in chronic inflammation, is known to induce cells to undergo apoptosis (Van Antwerp et al., 1998) and because one of the genes that it induces, FAT10, is highly upregulated in various cancers and can cause CIN, it is thus worthwhile to investigate whether FAT10 can modulate the apoptotic effect of TNF-α to help clarify the role of FAT10 in tumorigenesis. As evident from Fig. 6, when the FAT10 gene was silenced using shRNA, increased apoptosis, decreased percentage of cell survival, and decreased potential of these cells to grow in soft agar was observed. These results suggest that FAT10 might play a role in protecting cells against TNF-α-induced cell death. This is consistent with previous observations that lymphocytes of FAT10 knockout mice were more prone to spontaneous apoptotic death (Canaan et al., 2006).
High levels of FAT10 protect cells from death caused by TNF-α. (A) TNF-α induces greater apoptosis in HCT116 FAT10i cells. Stable HCT116 CTRi and FAT10i cells were grown in the presence or absence of TNF-α for 24 hours before apoptosis assay was performed. Similar data were observed in three independent experiments. (B) Less HCT116 FAT10i cells survived TNF-α treatment. CTRi and FAT10i cells were grown for 24 hours (Day 0) before being treated with or without TNF-α for the specific period and live cells counted. Survival rate is expressed as the number of live TNF-α cells versus the number of mock-treated cells. Data are expressed as mean ± s.d. of three replicates. *P<0.05 compared with control shRNA-expressing clones. (C) Fewer FAT10i cells under TNF-α treatment can grow on soft agar. About 5,000 CTRib or FAT10ib cells were seeded in soft agar and grown with TNF-α of the specified concentration. Three weeks later, the colonies were stained for visualization. Similar data was obtained in three separate experiments.
Interestingly, the level of induction of the FAT10 gene expression by TNF-α alone seems to be a little low, raising the concern of whether such low induction of the FAT10 gene is able to significantly affect chromosome stability or tumorigenesis in vivo. Different cells seem to respond differently to different cytokines, with TNF-α playing a more important role in inducing FAT10 expression in HCT116 and SW620 (Fig. 1) cells and IFN-γ playing a greater role in inducing FAT10 expression in HepG2 (Fig. 1) and Hepa1-6 cells (Oliva et al., 2010a). Nonetheless, a variety of different cytokines, often working in synergy (Bartee et al., 2008), are usually concomitantly elevated during inflammation (Wang et al., 1999). As shown in Fig. 1, TNF-α can synergistically enhance the effect of IFN-γ on FAT10 expression, suggesting that perhaps FAT10 expression might be sufficiently induced by TNF-α and other cytokines like IFN-γ during inflammation to modulate chromosome stability or tumorigenesis in vivo, although this remains to be validated.
Another interesting observation made from this study is that continuous induction by TNF-α is necessary for the expression of FAT10 to be sustained and CIN to be observed (Fig. 5), suggesting that FAT10 perhaps plays a more significant role during chronic inflammation. This is consistent with the observation that chronic inflammation-associated cancers often take a long time to develop. For example, development of HCC usually takes more than 30 years after chronic infection with hepatitis B or C virus is first diagnosed (Thorgeirsson and Grisham, 2002).
Hypothetical model of how FAT10 functions in the pathway of chronic inflammation-induced tumorigenesis.
In summary, this study provides evidence that FAT10 mediates the effect of TNF-α in causing numerical CIN, and protects cells from TNF-α-induced cell death, implicating a potential role of FAT10 in chronic inflammation-induced tumorigenesis.
These observations might have potential applications. Various agents that block TNF-α (e.g. anti-TNF antibodies such as infliximab) have been shown to have clinical applications (Anderson et al., 2004; Balkwill, 2009). However, because of the adaptive protective purpose of inflammation, inhibiting this important proinflammatory cytokine might result in unwanted adverse side effects that could be potentially serious (Anderson et al., 2004). We hypothesize that targeting molecules further downstream in the TNF-α pathway that is more directly associated with tumorigenesis might be a better choice for a therapeutic target for inflammation-associated cancers (e.g. HCC) as this will result in less side effects. This is because TNF-α, being a pleiotropic molecule, plays many other important and diverse roles, in addition to being a tumor promoter. Fig. 7 summarizes the pathway from chronic inflammation to tumorigenesis. Chronic inflammation often induces the release of TNF-α, which activates the NF-κB pathway. Activation of the NF-κB pathway activates the expression of numerous genes that play diverse roles in inflammation, including a subset that plays roles in tumorigenesis. Because FAT10 is shown in this study to be one of the downstream molecules in the pathway of inflammation-induced tumorigenesis, FAT10 might represent an attractive therapeutic target to investigate. FAT10-deficient mice exhibit very minimal phenotype change, suggesting that if FAT10 is, in fact, a key molecule in chronic inflammation-associated HCC, targeting the FAT10 gene might be feasible and more desirable than targeting upstream genes like TNF-α or NF-κB as it is unlikely to result in extensive unintended side effects. Hence, an important future direction will be validation of the role of FAT10 in tumorigenesis in a mice model.
Materials and Methods
Cell culture, transfection, antibody, primers, siRNA and reagents
All cell lines in this study were purchased from American Type Culture Collection (ATCC, Manassas, VA) and cultured under the recommended conditions. Transfection of plasmids was performed using Lipofectamine 2000 (Invitrogen). siRNAs were transfected using siPORT transfection reagent (Applied Biosystems) or electroporated with the BTX ECM830 instrument (BTX Instrument, Holliston, MA). Rabbit anti-FAT10 polyclonal antibodies were generated as previously described (Lee et al., 2003). Rabbit anti NF-κB p65 and goat anti-actin antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Primers for reverse-transcription polymerase chain reaction (RT-PCR) analyses are listed in supplementary material Table S1. siRNAs that were used in this study are listed in supplementary material Table S2. Recombinant human TNF-α and IFN-γ were purchased from Invitrogen (Carlsbad, CA) and Roche Applied Science (Indianapolis, IN), respectively.
Constructs and stable cell lines
HCT116 cells stably expressing H2B–GFP (Kanda et al., 1998) were generated by transfecting these cells with pBOS-H2BGFP (BD Biosciences) and growing them in 5 ug/ml blasticidin (Invitrogen) for about 3 weeks. Colonies stably expressing the fluorescent protein were selected and characterized using a Zeiss LSM 510 confocal laser scanning microscope (Heidelberg, Germany) (supplementary material Fig. S3A) and a live cell imaging Zeiss Axiovert 200 M microscope (supplementary material Fig. S3C). TNF-α induction of FAT10 expression was confirmed by real-time RT-PCR (supplementary material Fig. S3B).
Double-stranded oligonucleotides encoding either FAT10 or control shRNAs were cloned downstream of the H1 promoter in a plasmid that also carried the Neomycin-resistance gene under the PGK (phosphoglycerate kinase) promoter (supplementary material Fig. S4A). These constructs were then transfected into HCT116 cells and the cells treated with 1 mg/ml G418 for 3 weeks to establish stable clones. Two clones stably expressing either the control (CTRia and CTRib cells) or FAT10 shRNAs (FAT10ia and FAT10ib cells) were selected and their identity and cellular characteristics were validated by RT-PCR (supplementary material Fig. S4B), karyotyping (supplementary material Fig. S4C) and CGH microarray (supplementary material Fig. S4D). SW620 cells stably expressing control or FAT10 shRNAs were similarly established.
Stable HCT116 CTRib and FAT10ib cells were stably transfected with pBOS-H2BGFP to generate H2BGFP-CTRib and H2BGFP-FAT10ib cells so as to facilitate the monitoring of their mitotic profiles.
To establish whether reintroducing the FAT10 protein into low FAT10-expressing HCT116 cells carrying shRNA against FAT10 would restore the CIN phenotype observed when FAT10 expression is high, we used PCR-guided site-directed mutagenesis to engineered a mutant FAT10 that did not alter the amino acid sequence but was resistant to FAT10 shRNA (supplementary material Fig. S5B). This mutant FAT10 was cloned downstream of the CMV promoter (supplementary material Fig. S5A,B). This construct was transfected into cells stably expressing FAT10 shRNA. Western blot analyses validated the expression of this mutant FAT10, despite the presence of FAT10-specific siRNA in the same cells (supplementary material Fig. S5C).
RT-PCR quantitation
Total RNA was extracted from cells using TRIzol reagent (Invitrogen) as per the manufacturer's protocol. The RNA concentration was measured using a NanoDrop1000 Spectrophotometer (Thermo Scientific, Pittsburgh, PA). Reverse transcription was performed using 1 μg of RNA, oligo (dT)17 primer and SuperScript II kit (Invitrogen) in 10 μl reaction. In the reverse transcription for TNFR2, a gene-specific primer was used (supplementary material Table S1). For quantitative PCR, 0.1 μl of the reverse transcription product was used with respective primers (supplementary material Table S1) in 10 μl reaction using either the Multiplex PCR kit (Qiagen) for agarose gel analyses or QuantiTect SYBR Green PCR kit (Qiagen) for analyses using the 7500 Fast Real-Time PCR system (Applied Biosystems).
Cytogenetic analysis
The analyses of chromosome numbers were performed as previously described (Ren et al., 2006).
Apoptosis and proliferation analysis
To determine the apoptosis profile, cells were stained using the Annexin V-PE apoptosis detection kit I (BD Biosciences) and analyzed with FACSCalibur flow cytometer (BD Biosciences). Cell numbers were determined using the Beckman Coulter AcT Diff hematology analyzer (Beckman Coulter).
Soft agar colony formation assay was performed in six-well culture plates: 1.5 ml of 0.7% low melting agarose solution was added into each well and left to solidify. Approximately 5,000 cells were then mixed with 0.4% agarose solution and added to the top of the 0.7% agarose gel. After the top layer had also solidified, 3 ml of medium was then added and the cells incubated for 3 weeks. The colonies formed were stained with 0.01% methylene blue dissolved in 40% methanol.
Statistics
Two-tailed Student's t-test was utilized to evaluate significance.
Acknowledgements
We thank Alexandra Pietersen and laboratory members for help with FACS sorting.
Footnotes
-
Funding
This work was supported by grants from the Academic Research Fund, National University of Singapore (NUS) [grant number R-183-000-163-112]; the National Medical Research Council (NMRC) of Singapore [grant number NMRC/1030/2006]; and the Singapore Millennium Foundation (SMF), as well as block fundings from the National Cancer Centre, Singapore to C.L.
-
Supplementary material available online at http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.087403/-/DC1
- Accepted June 13, 2011.
- © 2011.
References
Article navigation
Related articles
Cited by...
More in this TOC section
Similar articles
Other journals from The Company of Biologists
First Person interviews
Have you seen our First Person interviews with the early-career first authors of our papers? The authors talk about their work in and out of the lab, the journeys that led them to where they are now and the scientists who inspired them along the way. Recently, we caught up with first authors Maitreyi Rathod, Sushmita Chatterjee, Lotte Vanheer, Rachel Furlong, Pamela Adami, Ruiqi Wang and Chen Yu.
Travelling Fellowship – New imaging approach unveils a bigger picture
Find out how Pamela Imperadore’s Travelling Fellowship grant from The Company of Biologists took her to Germany, where she used new imaging techniques to investigate the cellular machinery underlying octopus arm regeneration. Don’t miss the next application deadline for 2020 travel, coming up on 29 November. Where will your research take you?
preLights – Meet the preLighters: an interview with Maiko Kitaoka
Maiko Kitaoka is a graduate student in the lab of Rebecca Heald at the University of California, Berkeley. Here she studies the cause of chromosome mis-segregation defects in Xenopus hybrids. We caught up with Maiko to discuss her research, science communication, ballet, preprints and more.
Journal Meeting – Cell Dynamics: Host-Pathogen Interface
Registration is now open for the third instalment of the highly successful Cellular Dynamics Meeting Series, and will focus on ‘Host-Pathogen Interface’. The meeting will take place 17-20 May 2020, and further information is available here.
Cell Science at a Glance – Adaptor protein complexes and disease at a glance
A new poster from the Robinson lab summarises what is known about the five adaptor protein complexes and discuss how this helps to explain the clinical features of different genetic disorders.
JCS joins the Review Commons initiative
Journal of Cell Science is pleased to be a part of the new and exciting Review Commons initiative, launched by EMBO and ASAPbio. Streamlining the publishing process, Review Commons enables high-quality peer review to take place before journal submission. Papers submitted to Review Commons will be assessed independently of any journal, focusing solely on the paper’s scientific rigor and merit.
Articles of interest in our sister journals
Casein kinase 1α decreases β-catenin levels at adherens junctions to facilitate wound closure in Drosophila larvae
Chang-Ru Tsai, Michael J. Galko
Development
Spherical spindle shape promotes perpendicular cortical orientation by preventing isometric cortical pulling on both spindle poles during C. elegans female meiosis
Elizabeth Vargas, Karen P. McNally, Daniel B. Cortes, Michelle T. Panzica, Brennan M. Danlasky, Qianyan Li, Amy Shaub Maddox, Francis J. McNally
Development