The evolutionarily conserved TREX (Transcription/Export) complex physically couples transcription, messenger ribonucleoprotein particle biogenesis, RNA processing, and RNA export for a subset of genes. HPR1 encodes an essential component of the S. cerevisiae TREX complex. HPR1 loss compromises transcriptional elongation, nuclear RNA export, and genome stability. Yet, HPR1 is not required for yeast viability. Thoc1 is the recently discovered human functional orthologue of HPR1. Thoc1 is expressed at higher levels in breast cancer than in normal epithelia, and expression levels correlate with tumor size and metastatic potential. Depletion of Thoc1 protein (pThoc1) in human cancer cell lines compromises cell proliferation. It is currently unclear whether Thoc1 is essential for all mammalian cells or whether cancer cells may differ from normal cells in their dependence on Thoc1. To address this issue, we have compared the requirements for Thoc1 in the proliferation and survival of isogenic normal and oncogene-transformed cells. Neoplastic cells rapidly lose viability via apoptotic cell death on depletion of pThoc1. Induction of apoptotic cell death is coincident with increased DNA damage as indicated by the appearance of phosphorylated histone H2AX. In contrast, the viability of normal cells is largely unaffected by pThoc1 loss. Normal cells lacking Thoc1 cannot be transformed by forced expression of E1A and Ha-ras, suggesting that Thoc1 may be important for neoplastic transformation. In sum, our data are consistent with the hypothesis that cancer cells require higher levels of pThoc1 for survival than normal cells. If true, pThoc1 may provide a novel molecular target for cancer therapy. [Cancer Res 2007;67(14):6657–64]
- RNA processing
The accumulation of genetic and epigenetic alterations in cancer cells endows them with unwanted proliferative and metastatic potential. However, these alterations can also handicap cancer cells with unique vulnerabilities such that it is possible to identify particular genes whose function is more critical for the viability of cancer cells than for normal cells ( 1). Mutations in such genes are formally synthetic lethal with the genetic and epigenetic alterations present in cancer cells. Proteins encoded by such synthetic lethal genes identify molecular targets for therapy because antagonizing their function will be more toxic to cancer cells than to normal cells, thus yielding superior therapeutic index. The inability to carry out synthetic lethal screens in cultured human cells has limited the number of genes identified whose inactivation is uniquely toxic to human cancer cells, although this is likely to change with the advent of high-throughput gene silencing technologies.
The human Thoc1 gene, also known as hHpr1 or p84, encodes a protein that was originally identified as a nuclear matrix component that binds the retinoblastoma tumor suppressor protein ( 2). Alterations in the nuclear matrix and resulting changes in nuclear structure have long been recognized to correlate with tumor progression, prompting their use as biomarkers for the diagnosis of cancer ( 3). Indeed, overexpression of Thoc1 has recently been documented in human breast cancer with pThoc1 levels correlating with tumor size and metastases ( 4). We and others have recently identified pThoc1 as a functional orthologue of the S. cerevisiae HPR1 gene ( 4– 6). The HPR1-encoded protein (Hpr1p) is a component of the TREX (Transcription/Export) complex that physically couples the elongating RNA polymerase II with factors important for messenger ribonucleoprotein particle (mRNP) formation, RNA processing, and nRNA export ( 7– 9). Yeast TREX is composed of the salt-resistant THO subcomplex containing the four proteins Hpr1p, Tho2p, Mft1p, and Thp2p, all of which are essential for efficient transcriptional elongation of a subset of yeast genes ( 10). THO associates with two proteins involved in nuclear RNA export, Sub2p and Yra1p, to form the larger TREX complex ( 9). Sub2p is the yeast orthologue of human UAP56 that has been implicated in RNA splicing ( 11). Hpr1p genetically and physically interacts with RNA polymerase II ( 12, 13) and is essential for recruitment of Sub2p to genes regulated by TREX ( 14). Loss of Hpr1p impairs both transcriptional elongation and nuclear RNA export ( 4, 8, 9, 15– 18). Whereas Hpr1p is an essential component of the TREX complex, it is not essential for yeast viability ( 19).
The human TREX complex is composed of proteins encoded by Thoc1 (yeast HPR1), Thoc2 (yeast THO2; ref. 20), hTex1 (yeast TEX1), UAP56 (yeast SUB2; ref. 21), Aly (yeast YRA1; ref. 22), and other genes that are not evolutionarily conserved with yeast ( 4, 6). Depletion of pThoc1 compromises both transcriptional elongation and cell proliferation in the cancer cell lines that have been tested ( 4, 5). Thoc1 is also required for early embryonic development in the mouse ( 23). Cells of the blastocyst inner cell mass that include embryonic stem cells rapidly die on loss of maternally supplied pThoc1. Interestingly, conditional ablation of Thoc1 in adult mammary epithelia has no detectable effect on viability or gland function. 1 These observations suggest the hypothesis that Thoc1 may be required for the viability of indefinitely self-renewing cells, like cancer cells, but not for normal differentiated cells. We have tested this possibility by comparing the sensitivity of isogenic normal and oncogene-transformed cells to loss of pThoc1. We have also tested whether Thoc1 is required for oncogene-mediated neoplastic transformation in vitro.
Materials and Methods
Cell culture and viruses. HeLa, HEK293, WI38, IMR90 (passage 12), and HCT116 cell lines have been obtained from American Type Culture Collection. The p53−/− HCT116 subline has previously been described ( 24). The IMR5C, IMR90, and oncogene-transduced IMR90 derivative cell lines have also been previously described ( 25). Murine embryonic fibroblast (MEF) cultures were isolated from F/− and F/+ littermate embryos at 13.5 days of gestation by standard procedures. The construction and genotyping of the floxed Thoc1 allele has previously been described ( 26). Cells were cultured in DMEM supplemented with 10% fetal bovine serum, 2 mmol/L l-glutamine, and 0.5% penicillin and streptomycin and grown at 37°C in a humidified atmosphere of 5% CO2.
The Ad-lacZ adenoviral stock is from the M.D. Anderson Cancer Center Vector Core Facility. The AdCre (a recombinant adenovirus expressing the Cre recombinase) adenoviral stock is purchased from the Gene Transfer Vector Core of the University of Iowa ( 27). Expression of the transgene in each recombinant adenovirus is driven by the early promoter of cytomegalovirus. Infections are typically done at a multiplicity of infection (MOI) of 25 to 50 infectious units per cell.
Recombinant retroviruses designed to express E1A (E1A.12S.Lpc) or Ha-ras (c-H-RasV12.pWzl) have previously been described ( 25). Early-passage MEFs (passage ≤5) were infected with freshly prepared retroviral stocks and subsequently selected for puromycin and hygromycin to enrich for E1A- and Ha-ras–expressing cells.
Small interfering RNA. The control (SC and 2M) and Thoc1 (N52 and N54) small interfering RNAs (siRNA) have previously been described ( 5). For siRNA transfections, cells were plated at 1 × 105 per well in DMEM supplemented with 10% fetal bovine serum (without antibiotics) in six-well plates. Twenty-four hours later, the siRNAs (synthesized by Xeragon, Inc.) were transfected at 10 nmol/L by using Lipofectamine 2000 as described by the manufacturer (Invitrogen). Typically, cells were collected or treated as indicated 72 h after transfection.
Cell viability, cell cycle, apoptosis, and β-galactosidase assays. Cell viability was measured by direct counting of trypan blue–excluding cells using a hemacytometer or by 23-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) assay done as recommended (Roche Applied Science) with similar results. Absorbance readings for the XTT assays were read at 490 nm using an Elx808 microplate reader (Bio-Tek Instruments, Inc.).
For cell cycle distribution analysis, single-cell suspensions at 1 × 106/mL in 1-mL PBS were prepared. Cells were fixed by adding 3 mL of ice-cold 70% ethanol and kept at −20°C at least overnight. Cells were then washed with PBS and incubated for 15 min at 37°C in 1-mL propidium iodide solution (50 μg/mL propidium iodide, 20 μg/mL RNase in PBS). The signals were collected on a FACScan flow cytometer (Becton Dickinson) and analyzed with ModFit LT software (Verity Software House, Inc.).
The fraction of apoptotic cells was measured by using the Annexin V-phycoerythrin apoptosis detection kit (BD Biosciences PharMingen) according to the manufacturer's recommendations. Briefly, the treated cells were collected by trypsinization, washed, and resuspended in 1× Annexin V binding buffer at a concentration of 1 × 106/mL. Annexin V-phycoerythrin and 7-amino-actinomycin D were added and cells incubated for 15 min at 25°C in the dark. Stained cells were diluted in 1× binding buffer and analyzed on a FACScan flow cytometer (Becton Dickinson) within 1 h. WinList software was used to calculate the fraction of apoptotic cells (Verity Software House).
An ortho-nitrophenyl-β-d-galactopyranoside (ONPG) assay was used according to the manufacturer's recommendations (Promega) to measure β-galactosidase activity. Cells were trypsinized, washed, and lysed in 1× Reporter Lysis Buffer for 15 min. The extract was clarified by centrifugation at 16,000 × g for 2 min at 4°C and protein concentration of the extracts quantitated using the Bradford assay (Bio-Rad). Thirty micrograms of protein per sample were diluted in 1× Reporter Lysis Buffer and 2× Assay Buffer [200 mmol/L sodium phosphate (pH 7.3), 2 mmol/L MgCl2, 100 mmol/L β-mercaptoethanol, 1.33 mg/mL ONPG]. The mixture was incubated at 37°C for 30 to 60 min. The absorbance of samples was measured at 405 nm using an Elx808 microplate reader (Bio-Tek Instruments).
Western blotting. Cells were harvested, washed in ice-cold PBS, and extracted in Lysis250 buffer [50 mmol/L Tris (pH 7.4), 250 mmol/L NaCl, 5 mmol/L EDTA, 0.1% NP40], supplemented with a cocktail of protease and phosphatase inhibitors (Sigma). Extracts were clarified by centrifugation and protein concentration was measured by Bradford assay (Bio-Rad). Twenty micrograms of total protein for each sample were resolved by 7.5% SDS-PAGE, transferred to polyvinylidene difluoride membrane, and the blot was blocked with 5% nonfat dry milk in 20 mmol/L Tris (pH 7.5), 150 mmol/L NaCl, 0.1% Tween 20 (TBST) for 1 h at room temperature. The blot was incubated with primary antibody diluted in fresh TBST-5% milk and the bound antibody detected using horseradish peroxidase–conjugated secondary antibody and chemiluminescence (Amersham Biosciences Corp.). The following antibodies were used: mouse monoclonal anti-pThoc1 (GeneTex); mouse monoclonal anti–β-actin (Calbiochem); mouse monoclonal anti–heat shock protein 70 (Stressgen); mouse monoclonal anti–poly(ADP-ribose) polymerase (PARP; Santa Cruz Biotechnology); and mouse monoclonal anti–phospho-histone H2AX (Upstate).
Loss of Thoc1 compromises the viability of oncogene-transformed fibroblasts but not normal fibroblasts. We have previously generated oncogene-transformed derivatives of normal diploid IMR90 human fibroblasts ( 25). An E1A/Ha-ras–transduced IMR90-derived line is immortal but not tumorigenic. An E1A/Ha-ras/c-myc–transduced IMR90 line is immortalized and tumorigenic in vivo. We have used this panel of human fibroblast cells to compare the effects of pThoc1 depletion on normal and oncogene-transformed human cells. Examination of pThoc1 levels in the cell panel indicates that the oncogene-transformed lines express significantly higher levels of pThoc1 than the normal parental cells ( Fig. 1A, left ), similar to previously published data comparing breast cancer cells and normal mammary epithelia ( 4).
We transfected siRNA directed against Thoc1 into each cell line to deplete pThoc1 and assessed the effects on accumulation of viable cells over time. The targeted siRNA oligonucleotides used (N52 and N54) were previously verified to be specific for Thoc1 ( 5). Thoc1 siRNA, but not a mismatch control siRNA, was able to deplete pThoc1 from each of the cell lines ( Fig. 1A, right). To verify that siRNA-mediated pThoc1 depletion inhibited TREX activity, we assayed the expression of a bacterial β-galactosidase reporter gene. Thoc1 was previously noted to be required for efficient transcriptional elongation of the G/C-rich bacterial β-galactosidase gene in both yeast and mammalian cells ( 5, 16). Cells transfected with Thoc1 or control siRNA were infected with a recombinant, β-galactosidase–expressing adenovirus and then assayed for β-galactosidase activity. Consistent with Western blot analysis, Thoc1 siRNA treatment reduced β-galactosidase activity in all three cell lines relative to cells transfected with control siRNA ( Fig. 1B). Depletion of pThoc1 activity was relatively greater in normal IMR90 cells than in the transformed derivatives (see below). Loss of pThoc1 in parental IMR90 cells had no detectable effect on cell accumulation compared with control siRNA–treated cells ( Fig. 1C and D). In contrast, Thoc1 siRNA caused a significant decrease in the accumulation of E1A/Ha-ras– or EIA/Ha-ras/c-myc–transformed IMR90 cells.
We investigated whether the decrease in the accumulation of transformed cells might be due to induction of apoptotic cell death because a significant number of pThoc1-depleted cells stained positively for trypan blue. Thoc1 siRNA treatment of EIA/Ha-ras/c-myc–transformed cells caused an ∼2-fold increase in the fraction of cells staining positively for the apoptosis marker Annexin V, relative to control siRNA–treated cells ( Fig. 2A ). Thoc1 siRNA treatment had no detectable effect on apoptosis in IMR90 cells. Similarly, Thoc1 siRNA caused a significant increase in the fraction of EIA/Ha-ras/c-myc–transformed cells containing fragmented DNA, as assessed by terminal transferase dUTP nick end labeling (TUNEL) assay, relative to control siRNA–treated cells (Supplementary Fig. S1A). We also examined whether there were changes in cell cycle distribution on pThoc1 loss. No significant difference in cell cycle distribution was observed in either IMR90 or EIA/Ha-ras/c-myc–transformed cells on pThoc1 depletion ( Fig. 2B). These data suggested that loss of cell accumulation in transformed cells was due to apoptotic cell death rather than changes in the cell cycle, and that oncogene-transformed human fibroblasts were significantly more sensitive to pThoc1 depletion than normal human fibroblasts.
We have investigated two additional normal diploid human fibroblast cell lines to ensure that insensitivity to pThoc1 depletion was not unique to IMR90 cells. Thoc1 siRNA efficiently depleted pThoc1 in both WI38 and IMR5C normal diploid human fibroblasts ( Fig. 3A ), yet loss of pThoc1 had no detectable effect on the growth rate of these cells in culture ( Fig. 3B and C). Coincident with the decrease in Thoc1 protein levels, Thoc1 siRNA treatment also decreased the level of β-galactosidase activity by >2-fold compared with control siRNA–transfected cells ( Fig. 3D). Thus, efficient depletion of pThoc1 protein and TREX activity had no detectable effect on the growth of normal human fibroblasts in vitro. Because the oncogene-transformed cell lines lose viability on pThoc1 loss, the modest siRNA-mediated depletion of pThoc1 observed in these cells was probably due to selection for viable cells that escaped siRNA-mediated Thoc1 gene silencing.
We have recently generated a conditionally excisable allele of the murine Thoc1 gene that facilitates stable depletion of pThoc1 ( 26). The modified Thoc1 allele (Thoc1F) contains lox P sites within introns 5 and 7. Mice homozygous or hemizygous for this allele are normal. Cre-mediated recombination between the sites excises intervening exons 6 and 7, deleting evolutionary conserved coding sequence and creating a premature stop codon by reading frame shift. Mice homozygous for the Cre-excised allele phenocopy mice homozygous for a previously characterized Thoc1 null allele ( 23). To test the effects of complete genetic loss of Thoc1, relative to intermediate levels of depletion achievable by siRNA, and to ensure that the results described above were not caused by off-target effects of siRNA-mediated gene silencing, we have used MEFs containing the conditional Thoc1 allele to compare the effects of Thoc1 loss on the viability of normal and transformed cells.
Early-passage (<5) MEFs from littermate embryos heterozygous (Thoc1F/+) or hemizygous (Thoc1F/−) for the conditional allele were transformed into neoplastic cells by retroviral mediated transfer of the E1A and Ha-ras oncogenes ( 28). Both Thoc1F/+ and Thoc1F/− MEFs were efficiently transformed by E1A/Ha-ras as indicated by the large number of drug selection–resistant cell colonies that exhibited classic transformed morphology (data not shown). To test whether Thoc1 is required for the viability of these oncogene-transformed cells, we infected them with AdCre to convert the Thoc1F allele to a null allele. AdCre caused a significant reduction in the accumulation of Thoc1F/− transformed cells but not of Thoc1F/+ transformed cells ( Fig. 4A ). We have thus far failed to recover viable E1A/Ha-ras–transformed cell clones that lack pThoc1. We also investigated whether normal primary MEFs required Thoc1 for viability. Early-passage MEFs isolated from Thoc1F/− and Thoc1F/+ mice were treated with AdCre and monitored for accumulation of viable cells. In contrast to E1A/Ha-ras–transformed cells, there was no statistically significant difference in the relative accumulation of Thoc1F/− or Thoc1F/+ cells subsequent to AdCre infection. These observations suggested that, similar to human fibroblasts, Thoc1 was required to support the viability of oncogene-transformed MEFs but not of normal primary MEFs.
To test this, we treated cells with AdCre and assayed for induction of apoptosis by Annexin V staining. AdCre caused a significant increase in the percentage of apoptotic cells in transformed Thoc1F/− cell cultures but not in transformed Thoc1F/+ cell cultures that retain a wild-type Thoc1 allele ( Fig. 4B). We did not observe significant apoptosis in primary MEFs of either genotype subsequent to AdCre treatment. There was a small increase in apoptosis observed in E1A/Ha-ras–transformed Thoc1F/+ cells as well as in primary MEFs of both genotypes, probably caused by nonspecific toxicity of adenovirus infection or Cre-mediated DNA damage.
AdCre infection efficiently depleted pThoc1 from both primary and E1A/Ha-ras–transformed Thoc1F/− cells ( Fig. 4C). As expected, AdCre treatment had no detectable effect on pThoc1 levels in either primary or E1A/Ha-ras–transformed Thoc1F/+ cells. Thus, despite equally efficient pThoc1 depletion, induction of apoptosis and loss of viability only occurred at detectable levels in oncogene-transformed fibroblasts, not normal MEFs. We also noted that E1A/Ha-ras–transformed MEFs expressed significantly higher levels of pThoc1 than primary MEFs of the same genotype (compare E/R and MEF samples in Fig. 4C relative to the loading control). This observation was consistent with the increased pThoc1 expression observed in oncogene-transformed human fibroblasts. In sum, these data suggested that cancer cells may be dependent on higher levels of pThoc1 for survival than normal cells.
If true, Thoc1 may be required for neoplastic transformation. To test this possibility, we have attempted to transform early-passage primary Thoc1F/− and Thoc1F/+ MEFs with E1A/Ha-ras subsequent to AdCre-mediated excision of the Thoc1 floxed allele. We have verified excision of the floxed allele by Western blot analysis of MEF protein extracts before E1A/Ha-ras transduction ( Fig. 5A ). As expected, Thoc1F/− MEFs treated with AdCre exhibit significant depletion of pThoc1 relative to mock-treated controls. Thoc1F/+ MEFs, both Cre or mock pretreated, generate numerous colonies of morphologically transformed cells on retroviral mediated E1A/Ha-ras gene transfer and drug selection ( Fig. 5B). Mock-treated Thoc1F/− MEFs from two independent MEF isolates also generate abundant transformed colonies on transduction of E1A/Ha-ras. However, very few transformed cell colonies are recovered from Thoc1F/− primary MEFs lacking pThoc1. The few transformed colonies that are recovered have escaped AdCre-mediated excision of Thoc1 because all of those cell clones tested still express pThoc1 ( Fig. 5C). Similarly, Thoc1F/− primary MEFs treated with Cre and spontaneously immortalized by continuous in vitro culture under the 3T3 protocol also retain pThoc1. We have been unable to isolate any E1A/Ha-ras–expressing or spontaneously immortalized cell clones that lack pThoc1. These observations suggest that Thoc1 is required for spontaneous immortalization and E1A/Ha-ras–mediated transformation of primary MEFs.
Depletion of pThoc1 compromises the viability of human epithelial cancer cell lines. The data presented above showed that neoplastic fibroblasts are more dependent on pThoc1 for survival than normal fibroblasts, and that neoplastic fibroblasts deprived of pThoc1 die via apoptotic cell death. We wished to determine whether human epithelial cancer cell lines also undergo apoptotic cell death on pThoc1 loss and to examine possible mechanisms that might trigger this apoptotic response. Thoc1 siRNA efficiently mediated depletion of pThoc1 in a number of different human cancer cell lines including HeLa, 293, HCT116, a p53-null derivative of HCT116, and MCF-7 cells ( Fig. 6A ). Depletion of pThoc1 reduced the accumulation of cells in each of these lines ( Fig. 6B), consistent with the data presented above and previously published observations ( 4, 5).
To ascertain if pThoc1 loss also induced apoptotic cell death in human epithelial cancer cells, we assayed pThoc1-depleted HeLa cells for apoptosis. HeLa cells depleted of pThoc1 and exhibiting the typical 2-fold reduction in cell accumulation showed a >2-fold increase in the percentage of cells stained for the apoptosis marker Annexin V, relative to control siRNA–transfected cells ( Fig. 6C). Increased apoptosis was verified by assaying cleavage of PARP, a known substrate of caspase proteases, in siRNA-transfected HeLa cells. Specific depletion of pThoc1 caused an increase in PARP cleavage ( Fig. 6D). Thoc1 siRNA–treated HeLa, MCF-7, and HCT116 cells also showed an increase in the fraction of cells positive in the TUNEL assay relative to control siRNA–treated cells (Supplementary Fig. S1A). Finally, treatment of pThoc1-depleted HeLa cells with the pan-caspase ZVAD peptide inhibitor diminished the fraction of Annexin V–positive cells (Supplementary Fig. S1B). In sum, these data indicated that depletion of pThoc1 in human epithelial cancer cell lines inhibited proliferation by induction of apoptosis and subsequent loss of cell viability. It should be noted that sensitivity to loss of pThoc1 was independent of p53 and caspase-3 status. HCT116 cells that retain wild-type p53 and a p53-null HCT116 derivative cell line were both sensitive to Thoc1 siRNA. We also failed to observe consistent increases in p53 and p21 levels in pThoc1-depleted HeLa cells, suggesting that the p53 pathway was not activated (data not shown). MCF-7 cells that lack caspase-3 were also sensitive to Thoc1 siRNA, and we did not detect caspase-3 activation on pThoc1 depletion.
How pThoc1 deprivation triggers apoptotic cell death in cancer cells is currently unknown, but a number of potential mechanisms are conceivable. Loss of pThoc1 compromises transcriptional elongation due to improper incorporation of nascent RNA into mRNPs. Nascent RNA can then hybridize to the ssDNA template emerging from the RNA polymerase complex, thus generating stable RNA:DNA R-loop structures that are known to be recombinogenic and may compromise DNA integrity ( 17). Thoc1 loss, therefore, may directly lead to DNA damage. To test this possibility, we have examined the accumulation of phosphorylated histone H2AX, a known marker for DNA damage, in pThoc1-depleted normal and cancer cells. Phosphorylated histone H2AX is undetectable in Thoc1 or control siRNA–treated normal IMR90 cells ( Fig. 7 ). In contrast, Thoc1 siRNA treatment increases the levels of phosphorylated histone H2AX in EIA/Ha-ras/c-myc–transformed IMR90 cells as well as in HeLa cells, compared with control siRNA–treated cells. These results are consistent with the possibility that Thoc1 loss may lead to accumulation of DNA damage.
We describe a number of observations that suggest that cancer cells may differ from normal cells in their requirements for Thoc1. Thoc1 expression is higher in oncogene-transformed human and mouse fibroblasts compared with normal fibroblasts. This mimics the increased Thoc1 expression observed in primary human breast cancer compared with normal mammary epithelium ( 4) and the abundant levels of pThoc1 typically observed in human cancer cell lines that have been tested ( 5). However, higher pThoc1 levels are not merely a reflection of the faster proliferation rate typical of cancer cells. Loss of pThoc1 in cancer cell lines and oncogene-transformed human or mouse fibroblasts inhibits cell accumulation through induction of apoptosis and subsequent loss in cell viability. In contrast, loss of pThoc1 in normal human or mouse fibroblasts has no detectable effect on viability. Thus, the requirements for Thoc1 to support viability are different in the tested normal and cancer cells. We have been unable to recover viable E1A/Ha-ras–transformed or spontaneously immortalized MEFs that lack pThoc1. Thus, cells undergoing neoplastic transformation or immortalization are unable to adapt to the absence of Thoc1, suggesting that Thoc1 may be required.
There are potential caveats in interpreting these experiments. Because we cannot recover viable E1A/Ha-ras–expressing cells that lack pThoc1, we cannot exclude the possibility that pThoc1 loss inhibits neoplastic transformation by compromising expression of E1A/Ha-ras or the drug resistance genes. However, experiments have been done in the absence of drug selection with similar results. E1A/Ha-ras transduction of Thoc1F/− MEFs, but not Cre-pretreated Thoc1F/− MEFs lacking pThoc1, generates cultures of immortal, transformed cells without drug selection. 2 Thus, the inability to recover transformed cells in the absence of pThoc1 is independent of drug selection. In addition, MEFs lacking pThoc1 are viable, yet we do not recover viable cells on transduction of these cells with E1A/Ha-ras. This suggests that E1A/Ha-ras must be expressed sufficiently in the absence of pThoc1 to trigger a biological response. We note that pThoc1 loss also inhibits spontaneous immortalization of MEFs and the viability of many different human cancer cell lines. The effects of Thoc1, therefore, are not specific to E1A/Ha-ras transduction and likely reflect a general requirement for Thoc1 in cancer cells.
We have previously observed that Thoc1 is required for the viability of blastocyst-stage mouse embryos ( 23). Blastocyst-stage embryos are composed largely of stem cells that, like cancer cells, have an extended potential for replication and self-renewal. Thus, Thoc1 may be required more generally for the maintenance of extended replicative potential. Consistent with this hypothesis, we have observed that primary MEFs genetically ablated for Thoc1, but viable and able to proliferate in culture, undergo premature conversion to a senescence-like state. 3 Similarly, loss of the yeast orthologue HPR1 is not lethal but compromises life span ( 29). It is conceivable that lack of sufficient pThoc1 limits replicative potential in normal cells by induction of a cellular senescence program, thereby inhibiting immortalization and neoplastic transformation. Up-regulation of Thoc1 may be required, therefore, to facilitate immortalization and neoplastic transformation. Because cancer cells are unable to sustain cell viability on acute loss of pThoc1 activity, the effects of pThoc1 loss must be dominant to the effects of deregulated E1A/Ha-ras expression studied here as well as to other alterations that facilitate neoplastic transformation in a variety of human cancer cell lines tested here.
How Thoc1 loss triggers apoptosis in cancer cells is unknown. Because all detectable pThoc1 is apparently within the TREX complex ( 6), the simplest explanation is that loss of pThoc1 compromises TREX function. Loss of TREX function could have several conceivable effects on cells. Loss of TREX activity may adversely affect the generation of translatable mRNA from a subset of genes required to maintain viability. Loss of TREX activity may also compromise normal telomere maintenance. In yeast, loss of the Thoc1 orthologue HPR1 is associated with defects in telomere maintenance ( 30). Defects in telomere maintenance would be expected to influence replicative potential and viability. Alternatively, deficient mRNP biogenesis in the absence of Thoc1 may trigger R loop formation and DNA strand breaks ( 31). Such DNA lesions could trigger apoptotic cell death in cancer cells if they are unable to efficiently repair them. The accumulation of phosphorylated histone H2AX on pThoc1 depletion observed here is consistent with this possible mechanism, but additional studies will be required to identify and verify the mechanism underlying loss of cancer cell viability.
Irrespective of the precise mechanism, mutational inactivation of Thoc1 is synthetic lethal with the genetic and epigenetic alterations associated with a number of cancer cell lines of different type and origin. Thus, Thoc1 may represent a novel molecular target for cancer therapy. Therapy that blocks pThoc1 activity is expected to preferentially compromise the viability of cancer cells, potentially yielding superior therapeutic index. Because the mechanism of pThoc1 action is novel, utilization of pThoc1 as a therapeutic target may yield unique clinical responses and opportunities for novel combination therapy. For example, yeast deficient in the Thoc1 orthologue HPR1 are synthetic lethal with topoisomerase mutations ( 32) and are more sensitive to DNA damage ( 33). Depletion of pThoc1 in human cancer cell lines renders them more sensitive to camptothecin and cisplatin ( 5). These observations suggest that therapeutic inhibition of pThoc1 in human cancer will increase sensitivity to topoisomerase poisons and possibly other forms of genotoxic therapy.
Thoc1 protein functions in the newly discovered TREX complex, a representative of a class of complexes that regulate gene expression subsequent to transcriptional initiation. This class of protein complexes may specify posttranscriptional “operons” that facilitate protein expression from coordinately regulated genes of diverse size and structure ( 34). Although there is increasing appreciation for the importance of such complexes, they are understudied relative to the transcription factors that govern the initiation of transcription. As such, their relevance to carcinogenesis is largely undocumented. However, the von Hippel-Lindau tumor suppressor protein is a known inhibitor of the elongin transcription elongation factor, although it is unclear if this function is critical for tumor suppression ( 35). The interaction of the retinoblastoma tumor suppressor protein and pThoc1 may reflect another interaction between a tumor suppressor gene product and a transcription elongation/RNA processing factor. Such interactions suggest that complexes that regulate gene expression at the level of mRNP formation and RNA processing may provide a largely untapped source of novel molecular targets for cancer therapy.
Grant support: American Cancer Society-Institutional Research Grant no. 02-197-01 (A.W. Lin) and National Cancer Institute grants CA70292 and CA125665 (D.W. Goodrich).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Terry Beerman for kindly providing the p53−/− HCT116 cell line; members of the Lin and Goodrich labs for helpful discussions; and Aimee Stablewski of the RPCI Gene Targeting Core, supported by NIH Cancer Center Support Grant CA016056, for helpful advice and assistance in the construction of Thoc1 mutant mice.
Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).
Current address for A.W. Lin: Department of Basic Sciences, College of Osteopathic Medicine, Touro University, Vellejo, CA 94592.
↵1 Wang and Goodrich, unpublished.
↵2 Y. Li and A. Lin, unpublished observation.
↵3 Y. Li and D.W. Goodrich, unpublished observation.
- Received August 31, 2006.
- Revision received April 27, 2007.
- Accepted May 8, 2007.
- ©2007 American Association for Cancer Research.