IJC International Journal of Cancer
Restoration of absent in melanoma 2 (AIM2) induces G2/M cell cycle arrest and promotes invasion of colorectal cancer cells Georgios Patsos1, Anja Germann2, Johannes Gebert1 and Susanne Dihlmann1,3 1
Department of Applied Tumor Biology, Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany Fraunhofer Institut fuer Biomedizinische Technik (IBMT), Department of Biohybrid Systems, Molecular Cell and Tissue Engineering, St. Ingbert, Germany 3 Department of General Pathology, Institute of Pathology, University Hospital Heidelberg, Heidelberg, Germany
Cancer Cell Biology
2
Absent in melanoma 2 (AIM2) is a member of the interferon-inducible HIN-200 protein family. Recent findings point to a role of AIM2 function in both inflammation and cancer. In response to foreign cytoplasmic DNA, AIM2 forms an inflammasome, resulting in caspase activation in inflammatory cells. Moreover, AIM2 reduces breast cancer cell proliferation and mammary tumor growth in a mouse model and shows a high frequency of frameshift mutations in microsatellite unstable (MSI-H) gastric, endometrial and colorectal cancers. However, the consequences of AIM2 restoration in AIM2-deficient colon cancer cells have not yet been examined. Using different constructs for expression of AIM2 fusion proteins, we found that AIM2 restoration clearly suppressed cell proliferation and viability in HCT116 cells as well as in cell lines derived from other entities. In contrast to previous reports from breast cancer cells, our cell cycle analyses of colon cancer cells revealed that AIM2-mediated inhibition of cell proliferation is associated with accumulation of cells at late S-phase, resulting in G2/M arrest. The latter correlated well with upregulation of cyclin D3 and p21Waf1/Cip1 as well as with inhibition of cdc2 activity through Tyr-15 phosphorylation. Furthermore, AIM2 restoration affected the adhesion of colorectal cancer cells to fibronectin and stimulated the invasion through extracellular matrix-coated membrane in transwell assays. Consistent with this phenotype, AIM2 induced the expression of invasion-associated genes such as VIM and MCAM, whereas ANXA10 and CDH1 were downregulated. Our data suggest that AIM2 mediates reduction of cell proliferation by cell cycle arrest, thereby conferring an invasive phenotype in colon cancer cells.
The absent in melanoma 2 (AIM2) gene belongs to the IFI200/HIN-200 family of IFN-inducible genes found in both human and mouse.1 Members of this gene family (IFI16, MNDA, IFIX and AIM2 in humans; Ifi202a, Ifi202b, Ifi203, Ifi204 and D3 in mouse) encode proteins that are characterized by a conserved sequence domain of 200 amino acids (HIN-200 domain). HIN-200 proteins were initially grouped based on their hematopoietic expression, IFN-inducibility and nuclear localization. In addition, there is growing evidence for a role of these proteins as regulators of cell proliferation and differentiation in other cell types.1,2
Key words: colorectal cancer, HIN-200, cell adhesion, cell cycle, invasion Additional Supporting Information may be found in the online version of this article. Grant sponsor: Deutsche Forschungsgemeinschaft; Grant number: GE-592/4-1 DOI: 10.1002/ijc.24905 History: Received 8 Apr 2009; Accepted 27 Aug 2009; Online 30 Sep 2009 Correspondence to: Susanne Dihlmann, Department of General Pathology, Institute of Pathology, University Hospital Heidelberg, INF220/221, D-69120 Heidelberg, Germany, E-mail:
[email protected]
C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Four recent reports point to a role of AIM2 in the innate immune system. The authors identified AIM2 as an important inflammasome component that senses potentially dangerous cytoplasmic DNA and regulates caspase activation.3–6 Although this activity places AIM2 function in the context of the cell’s defense against viral infection, its impact on cell growth of noninfected cells is less clear. Cytoplasmic overexpression of AIM2 was described to reduce proliferation and to increase the susceptibility to cell death in transfected murine fibroblasts.7 In contrast, AIM2 was localized within the nucleus in transfected or IFN-c-treated human melanoma cell lines without significantly affecting the growth or survival of these cells.8 Cumulative evidence supports the idea that AIM2 may also play a role in carcinogenesis. Exogenous AIM2 expression was shown to reduce human breast cancer cell proliferation by inhibition of NF-jB transcriptional activity and to suppress mammary tumor growth in a mouse model.9 Furthermore, silencing of the AIM2 gene by DNA methylation was shown to be associated with immortalization of cells.10 Finally, the coding 10-bp polyadenine (A10) tract in exon 6 of the AIM2 gene appears to be positively selected for frameshift mutations in human preneoplastic and neoplastic MSI lesions of the colon and endometrium (www.seltarbase.org/11). By sequencing of the entire coding region of AIM2 we recently demonstrated a high frequency of frameshift and
Patsos et al.
Material and Methods Cell lines and expression plasmids
Human colorectal cancer cell line HCT116 was obtained from ECACC (Salisbury, UK) (http://www.ecacc.org.uk); human embryonic kidney 293 cells and human cervical carcinoma cell line HeLa were received from the German Cancer Research Centre Tumorbank or CLS Cell Lines Services (Heidelberg, Germany). All cell lines were grown in RPMI 1640 (Invitrogen/Life Technologies, Karlsruhe, Germany) supplemented with 10% Fetal Calf Serum (FCS), 100 U/ml penicillin and 100 mg/ml streptomycin using standard conditions. The pTK-hyg and pEGFP vectors were obtained from Clontech Laboratories (Takara Bio Europe/Clontech, Saint Germain-en-Laye, France). The Enhanced Green Fluorescent Protein (EGFP)-AIM2 expression plasmid was generated as described previously.13 The bidirectional, tetracycline-responsive expression vector pBI-EGFP-Flag-AIM2 was kindly provided by M.C. Hung (University of Texas Health Science Center at Houston, Houston, TX9).
was determined by staining using standard procedures (Supporting Information Fig. 1a). HCT116-tet-AIM2 cells were established by cotransfection of HCT116tTA with pBI-EGFPFlag-AIM2 and pTK-hyg, followed by selection of resistant clones with 200 lg/ml hygromycin B (Merck Biosciences, Calbiochem, Darmstadt, Germany) that was replaced every 48 hr. Doxycycline (1 lg/ml) was added to suppress FlagAIM2 expression during selection. Hygromycin-resistant clones were isolated and doxycycline was withdrawn for 4 days to induce expression. All clones were analyzed for expression of EGFP and Flag-AIM2 by western blotting. In addition, EGFP expression was examined by fluorescence microscopy of individual clones.
Immunoblotting
Equal cell numbers of stably transfected cells were lysed in luciferase lysis buffer and applied to SDS-PAGE and blotting using standard procedures as described previously.17 Detection of specific proteins was done by incubation with anti-EGFP antibody (Clontech Laboratories), anti-Flag antibody (Sigma Aldrich, Taufkirchen, Germany) or anti-actin antibody (MP Biomedicals, Heidelberg, Germany), each diluted 1:1,000 in blocking solution. Upon incubation of the blots with a secondary antibody (rabbit-anti-mouse IgG peroxidase; Dako, Hamburg, Germany), Luminol Reagent (Luminol Reagent, Santa Cruz Biotechnology, Heidelberg, Germany) was added as a substrate for visualization by enhanced chemoluminescence. For cell cycle analysis, cells were synchronized by serum starvation overnight and subsequent growth for 24 hr in a growth medium containing 10% FCS. The following antibodies were used for detection of cell cycle proteins at dilutions recommended by the manufacturer: anti-phospho-cdc2 [Tyr15], anti-cyclin D3 and anti-p21Waf (all from NEB Cell Signaling Technology, Frankfurt, Germany).
Generation of HCT116tTA cells and HCT116-tet-AIM2 cells
All transfections were performed using Fugene HD transfection reagent (Roche Diagnostics, Mannheim, Germany) according to the manufacturer’s instructions. HCT116 cells stably expressing EGFP-AIM2 fusion constructs were generated by direct selection for G418 resistance. HCT116tTA cells were created by transfection of HCT116 cells with pUHD15.1 encoding the tetracycline-responsive transactivator tTA14 together with pSV2-neo.15 Clones resistant to G418 were assayed for transactivation by transient transfection with the tTA-responsive luciferase reporter plasmid pUHC-13.3.14 One of the clones stably expressing tTA and showing tight regulation of transcriptional activity in response to doxycycline was selected for secondary transfections. To further check the activity of the transactivator (tTA) in combination with the bidirectional promoter (pBI), HCT116-tTA cells were transiently transfected with pBI-lacZ-AIM2 (AIM2 cDNA cloned into pBI-3)16 and beta-galactosidase activity C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Colony forming assay and cell growth analysis
HCT116, 293 and HeLa cells were grown in 6-well plates and transfected with 2 lg of either pEGFP or pEGFP-AIM2 expression vector using Fugene HD transfection reagent. Equal transfection efficiency was verified at 24 hr after transfection by fluorescence microscopy for EGFP (data not shown). The cells were allowed to grow for 3 weeks in the presence of 600 lg/ml G418 with 3 medium changes, after which the cells were washed, fixed with methanol and stained with crystal violet. All experiments were repeated twice. Cell proliferation rate was determined upon seeding 104 cells of stably transfected HCT116 clones (as indicated in the figures) in 6-well plates in the absence of antibiotic selection. Cell numbers were determined at the indicated times. When reaching confluence, the cultures were split 1:10 and cell numbers were converted according to dilution. Each experiment was performed in duplicate.
Cancer Cell Biology
additional mutations in primary MSI-H colon cancers and cell lines, many of which were shown to be biallelic. In addition, AIM2 promoter hypermethylation conferred insensitivity to IFN-c-induced AIM2 expression of some MSI-H colon cancer cell lines.12 The functional consequences of these mutational and epigenetic inactivations, however, are still unknown. Although the findings in breast cancer cells point to growth-suppressive features of AIM2, it is still unclear whether AIM2 indeed acts as a tumor suppressor, counteracting the oncogenic survival mechanisms of cancer cells, such as temporary cell cycle arrest and metastasis. To further address AIM2 function in colorectal cancer we here generated different constructs for recombinant restoration of wildtype AIM2 expression in AIM2-negative HCT116 cells and investigated their effects on cell growth, cell cycle and adhesion/migration properties.
1839
AIM2 activity in colon cancers
1840
Cell viability assay
Cells were plated in microtiter plates at densities of 103 or 104 cells/well and 10 ll of WST-1 (water soluble tetrazolium; ROCHE Diagnostics, Mannheim, Germany) was added after 24, 48, 72 or 96 hr. The cells were further grown for 30 min and absorbance was measured using a microplate ELISA reader (absorbance at 450 versus 650 nm, according to the instructions of the manufacturer).
Cancer Cell Biology
Cell death analysis
Cell clones were harvested including nonadherent cells, washed in phosphate-buffered saline (PBS) and adjusted to 1 106 cells in 0.5 ml PI staining solution (0.1% Triton X100, 50 lg/ml propidium iodide in PBS). Percentage of cell death was determined by FACS analysis using CellQuest software. Activation of caspase 3 was analyzed by immunoblotting, using an anti-caspase-3 antibody that detects both procaspase 3 and its cleavage products (IMGENEX/Biomol, Hamburg, Germany). Antibodies for detection of caspase 1 (Calbiochem/Merck, Darmstadt, Germany) and apoptosisassociated speck-like protein containing a CARD (ASC; Alexis Biochemicals/Enzo Life Sciences, Loerrach, Germany) were used to analyze inflammasome activation. The detection was performed as described earlier. Cell cycle analysis
Parental HCT116, mock-transfected HCT116 (clone D3) and HCT116-tet-AIM2 (clones D1 and B8) were grown to a density of 80% and the culture medium was replaced by RPMI1640 supplemented with 1% FCS for synchronization of the cell cycle overnight. The cells were then transferred back to medium containing 10% FCS for 24 hr followed by harvesting and counting of both shed cells and adherent cells. After washing in PBS, the cells were adjusted to equal numbers of 2 106, fixed in ice cold 70% ethanol overnight and stored at 20 C. Twenty-four hours prior to FACS analysis the samples were resuspended in PBS, complemented by 25 lg/ ml DNase-free RNase and 10 ll of propidium iodide for staining in the dark. DNA content was determined by flow cytometry using BD CellQuest software for acquisition and MultiCycle AV DNA analysis software (Phoenix Flow Systems, San Diego, CA) for data analysis. Adhesion and invasion assays
For adhesion assay, 96-well plates were coated with 20 lg/ml of fibronectin in PBS overnight at 4 C. All wells were incubated with 0.1% bovine serum albumin (BSA) for 1 hr in cell culture medium to block unspecific binding and then washed twice with PBS. The cells were trypsinized and incubated in serum-free medium containing 0.02% BSA for 1 hr to allow for receptor recovery. Subsequently, the cells were plated at a density of 50,000 cells/well in the same medium for the indicated times and allowed to attach. Nonadherent cells were removed by washing with PBS and adherent cells were fixed
in methanol for 10 min, washed in PBS and stained with crystal violet for 15 min. After photography, the cells were lysed in 10% acetic acid to release the dye and absorbance was determined in an ELISA reader for quantification at 620 nm. Eight wells were analyzed for each cell clone. For quantification of enlarged, multinuclear cells, 2 randomly selected areas of each clone were selected from the photographs and the percentage of multinuclear cells was determined manually by counting the cells (300–500 cells were included in total for each clone). Cell invasion assays were done on a 96-well chamber system (QCM 96-Well Cell Invasion assay; Chemicon International/Millipore, Darmstadt, Germany) according to the instructions of the manufacturer. Briefly, 20,000 cells/well were resuspended in serum-free medium and allowed to attach to inserts coated with a layer of extracellular matrix (ECM; ECMatrix solution is a solid gel of basement membrane proteins prepared from the Engelbreth Holm-Swarm mouse tumor). The inserts were placed into a feeder tray containing cell culture medium with 10% FCS. After 24 hr, invasive cells which migrated through the ECM-coated membrane (8 lm pore size) were washed in PBS, detached from the bottom of the insert, lysed and stained with a fluorescent dye for fluorimetric quantification. Six wells were analyzed for each cell clone. RNA extraction and RT-PCR
RNA was extracted from different clones and parental HCT116 cells by means of RNeasy Mini-Kit (Qiagen, Hilden, Germany) following the instructions of the manufacturer. For cDNA synthesis, 1 lg of total RNA was transcribed using oligo-dT primers and SuperScript II reverse transcriptase (Invitrogen, Karlsruhe, Germany) following the manufacturer’s instructions. For analysis of full-length AIM2 expression, templates were amplified by PCR before and after reverse transcription into cDNA using the following primers: forward primer, AIM2-forII: cDNA: 50 -ctg atc cca aag ttg tca gat g-30 ; and reverse primer, AIM2-revIII: 50 -ctt ctc tat gtt ttt ttt ttg gcc-30 . As a control for mRNA integrity, RT-PCR analysis of ACTIN B was included. (Detailed PCR conditions are available on request.) Relative gene expression levels of VIM, MCAM, ANXA10, and CDH1 were assayed by real-time PCR using SYBR green Master Mix (Applied Biosystems, Darmstadt, Germany) for analysis with the 7300 Real-Time PCR System (Applied Biosystems). 18s RNA was used as internal control for standardization. Primer sequences are described in Supporting Information Table I. All assays were performed in triplicate and relative mRNA levels were assayed using the 2DDCt method provided by the 7300 System SDS Software.
Results Characterization of AIM2 constructs and cell lines stably expressing AIM2
Previous reports have demonstrated that overexpression of AIM2 reduces proliferation in fibroblasts and breast cancer C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Patsos et al.
Figure 1. Generation of HCT116 cells stably expressing AIM2. (a) Schematic representation of the pEGFP-AIM2 expression constructs used for the generation of EGFP-AIM2 fusion constructs and of the tTA-encoding plasmid pUHD15.1 and pBI-EGFP-Flag-AIM2 constructs used for the generation of HCT116-tet-AIM2 cells expressing Flag-AIM2 and EGFP from a bidirectional promoter. (b) Western blotting of cell lysates derived from HCT116 cells transiently transfected with pEGFP or pEGFP-AIM2 was performed using anti-EGFP antibody. Expression of b-actin was analyzed as a loading control. (c) RT-PCR analysis demonstrating expression of EGFP-AIM2 in HCT116 clones 1A3 and 9D1. Parental HCT116 cells and HCT116 cells stably transfected with pEGFP were used for comparison. Positive control: cDNA derived from the pEGFPAIM2 expression plasmid was included as a template for PCR. (d) Western blotting derived from HCT116 clones D3, A5, D1 and B8 was performed from whole cell lysates using anti-Flag or anti-EGFP antibodies. Clones D3 and A5 represent HCT116 cells or HCT116-tTA cells, respectively, stably transfected with pTK-hyg to induce hygromycin resistance.
cell lines. However, the effect of AIM2 restoration on cell proliferation, adhesion and migration of colon cancer cells has not yet been examined. We therefore generated different vectors for restoration of recombinant AIM2 expression in HCT116 cells lacking functionally active AIM2.12 Wild-type AIM2 cDNA (EGFP-AIM2) was cloned into the pEGFP-C3 expression vector and verified by sequencing (Fig. 1a). EGFP-AIM2 fusion protein was readily detected in lysates from HCT116 cells transiently transfected with the vector via western blot analysis using anti-EGFP antibody. The observed molecular weight of EGFP-AIM2 was 68 kDa which was consistent with the predicted size (Fig. 1b). Upon transfer of the EGFP-AIM2 fusion construct into HCT116 cells and selection for resistance toward G418, we isolated 22 clones, 6 of which showed stable and constitutive expression of an EGFP-AIM2 transcript (Supporting Information Fig. 1b). Two of these clones, 1A3 and 9D1, were selected for subsequent functional analysis (Fig. 1c). C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
As AIM2 was expected to repress the growth of transfected cells, we additionally established HCT116 transfectants that stably express AIM2 under control of a tetracycline-repressible promoter (Fig. 1a). HCT116-tTA cells were used for expression of Flag-AIM2 fusion protein and EGFP from a bidirectional promoter (pBI-EGFP-Flag-AIM2; Fig. 1a) to generate HCT116-tet-AIM2 cells. Only 2 out of 30 hygromycin-resistant clones showed expression of EGFP and FlagAIM2 in the absence of doxycycline. However, this expression was not suppressed by addition of doxycycline as demonstrated by western blot analysis of cell lysates using anti-Flag antibody (Fig. 1d, clones B8 and D1), fluorescence microscopy and FACS analysis (data not shown). As the 2 HCT116-tet-AIM2 clones were able to grow despite constitutive AIM2 expression at low levels, and the same system was successfully used by others despite some basal AIM2 expression,9 we decided to investigate AIM2 function by comparison of HCT116-tet-AIM2 cells versus mock-transfected cells
Cancer Cell Biology
1841
Cancer Cell Biology
1842
Figure 2. AIM2 inhibits colony formation in different cell lines. (a) Equal numbers of HCT116, HeLa and 293 cells were transfected with pEGFP-AIM2 or the empty vector pEGFP. Colonies were stained with crystal violet 20 days after transfection and selection with G418. (b) HCT116-tet-AIM2 clones D1 and B8 constitutively expressing Flag-AIM2 and the control clone D3 lacking AIM2 expression were seeded at a density of 104 cells per well in 6-well plates. Colonies were stained with crystal violet 14 days later and photographed. The picture shows 1 representative experiment that was performed in duplicate.
(HCT116 and HCT116tTA þ pTKhyg, clones D3 and A5, respectively).
AIM2 represses the growth of colon cancer cell lines without inducing apoptosis
We next investigated the effects of AIM2 restoration on cell growth in different cell types. Upon recombinant expression and selection with G418, EGFP-tagged AIM2 clearly suppressed the colony formation of HCT116, HeLa and 293 cells (Fig. 2a). The remaining colonies growing in AIM2-transfected cells did not show EGFP fluorescence, indicating that AIM2 expression was lost or silenced (data not shown). To investigate in more detail the cell growth/survival of colon cancer cells upon reconstitution of wild-type AIM2, we determined the cell number by directly counting the selected clones of stably transfected HCT116-EGFP-AIM2 and HCT116-tet-AIM2 cells over a 17-day period (Figs. 3a and
AIM2 activity in colon cancers
3b). All clones expressing either EGFP-AIM2 or Flag-AIM2 were repressed in cell proliferation. HCT116 clones 1A3 and 9D1, stably expressing EGFP-AIM2, completely stopped growth and most of the cells died when kept in culture for more than 3 weeks (Fig. 3a). The surviving cells restarted to grow but had lost AIM2 expression. Both clones still maintained the AIM2 gene in their genomic DNA as indicated by PCR (Supporting Information Figs. 1c and 1d); however, AIM2 protein expression was completely silenced (data not shown). This indicates that loss of AIM2 expression was not caused by outgrowth of cells lacking the AIM2 transgene but rather by suppression of its transcription or translation. HCT116-tet-AIM2 clones D1 and B8 stably expressing FlagAIM2 proliferated continuously, but at a much lower rate than mock-transfected HCT116-tTA cells (Fig. 3b), and produced smaller colonies within 3 weeks of growth (Fig. 2b). For indirect measurement of viable cell number we determined the overall metabolic activity of different HCT116 clones over a period of 5 days using the WST-1 proliferation assay for colorimetric quantification. Again the proliferation of HCT116-tet-AIM2 clones D1 and B8 stably expressing Flag-AIM2 was much lower than that of mock-transfected HCT116 clone D3 (Figs. 3c and 3d). When the cells reached confluence, the metabolic activity of AIM2-expressing cells declined, whereas the viability of mock-transfected HCT116 cells remained unaffected (after 48 hr; Fig. 3d). The mechanism by which AIM2 reduces cell growth in HCT116 cells seems to be independent from apoptosis. As determined by PI staining and cytometric flow analysis (Fig. 4a), the number of apoptotic cells in HCT116-tet-AIM2 cells did not differ from that in parental or mock-transfected cells. Furthermore, no activation of caspase-3 was observed in AIM2-expressing cells (Fig. 4b). The recently identified role of AIM2 in activation of inflammasome-mediated cell death in response to cytoplasmic DNA3–6 suggested the involvement of caspase-1 or ASC in AIM2-induced reduction of cell growth. However, neither caspase-1 nor ASC was expressed at detectable levels in HCT116 control or AIM2-expressing cells (Fig. 4b), indicating that AIM2 acts independently from these molecules in HCT116 cells. AIM2-mediated effects on cell cycle, cdc2 Tyr-15 phosphorylation, p21Waf1/Cip1 and cyclin D3 expression levels
To further define the mechanism of AIM2-mediated growth inhibition, we analyzed cell cycle phase distribution of synchronized cells by flow cytometry. As shown in Figure 5a, AIM2 expression results in a clear accumulation of cells in the late S-phase, thereby delaying transition to G2/M phase. The average percentage of cells in S-phase increased from 15.9% and 21.6% in parental and mock-transfected HCT116 cells to 27.9% in HCT116-tet-AIM2 clone D1 and to 37.7% in clone B8, respectively (Fig. 5b). As AIM2 expression is higher in clone B8 than in clone D1 (Fig. 5c), the extent of G2/M arrest also correlates with the relative level of AIM2. C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Patsos et al.
Cancer Cell Biology
1843
Figure 3. Reconstitution of AIM2 in HCT116 inhibits cell growth. (a, b) Cell growth assays of the AIM2 transfectants and control vector transfectants. The graph displays results of direct cell counting over a 3-week period and represents double measurements for 2 experiments. Bars indicate SD (note: some bars are too small to be visible). (a) Clones 1A3 and 9D1 represent HCT116 cells stably transfected with pEGFP-AIM2. Parental HCT116 cells and HCT116 cells stably transfected with pEGFP were used as control. (b) Clones D1 and B8 represent HCT116-tTA cells stably expressing Flag-AIM2. Mock-transfected HCT116 cells (HCT116-hyg, clone D3) was used as control. (c, d) Cell viability assay of clones D1 and B8 stably expressing Flag-AIM2 versus control clone D3. A total of 103 (c) or 104 (d) cells were grown over a 5-day period and cell viability was determined by addition of WST-1 and photometric analysis every 24 hr. Error bars indicate SD of 3 independent samples.
Based on the observation that AIM2 restoration could promote cell cycle arrest we sought to identify which cell cycle regulators could mediate AIM2’s inhibitory effects. We therefore examined the expression of cyclin D1, cyclin D3, Rb, phospho-Rb (Ser 795 and Ser 807/811), p27Kip1, p21Waf1/Cip1, p16INK4A, p15INK4B, p53, phospho-p53, phospho-cdc2 (Tyr15), CDK6 and phospho-GSK3b (Ser9) in HCT116-tet-AIM2 and control cells by western blotting. After synchronization of cells by serum deprivation and subsequent growth for 24 hr with serum, 3 of these proteins were clearly altered in HCT116-tet-AIM2 cells when compared with AIM2-deficient cells (Fig. 5c). Tyr15 phosphorylation of cdc2, a catalytic subunit of the M-phase promoting factor (MPF), was strongly increased in HCT116-tet-AIM2 C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
clone B8, indicating a decrease of its kinase activity. The effect was slightly weaker in clone D1, according to the much lower AIM2 expression in these cells. In addition, expression of the universal cdk inhibitor p21Waf1/Cip1 was upregulated in both HCT116-tet-AIM2 clones. As neither p53 expression nor p53 phosphorylation differed between AIM2-proficient and AIM2-deficient cells (data not shown), the upregulation of p21Waf1/Cip1 seems to occur through a p53-independent mechanism. Finally, expression of cyclin D3, a regulator of the G2/M DNA damage checkpoint,18 was likewise elevated in HCT116-tet-AIM2 cell clones (Fig. 5c). In summary, our data indicate that AIM2 expression results in delayed transition from G2 to M-phase of the cell cycle, thereby decelerating proliferation.
AIM2 activity in colon cancers
1844
Cancer Cell Biology
Before invading the surrounding tissue, cancer cells of epithelial origin must degrade the underlying basement membrane. Thus, to investigate whether AIM2 expression alters the cells’ invasiveness, parental and HCT116-tet-AIM2 cells were seeded onto a filter that was coated with a complex mixture of ECM proteins and exposed to 10% fetal calf serum. The number of cells that digested the ECM and migrated through a microporous membrane was determined by staining with a fluorescence dye after 24 hr. Interestingly, the invasiveness of HCT116-tet-AIM2 clones D1 and B8 was significantly increased compared with their parental counterpart (Fig. 7a). To gain further insight into the molecular mechanisms underlying this effect, we assessed expression of a number of invasion/metastasis-associated genes by realtime PCR. In agreement with the cells invasion behavior, transcription of VIM (vimentin), and MCAM (melanoma cell adhesion molecule) was induced up to 10-fold, whereas ANXA10 (Annexin A10) and CDH1 (E-cadherin) were downregulated by 6- and 25-fold, respectively, in HCT116-tetAIM2 cells (Fig. 7b).
Discussion Figure 4. AIM2 expression does not induce apoptosis in HCT116 cells. (a) Percentage of cell death as determined by PI staining and flow cytometric analysis. (b) Immunoblot for detection of uncleaved/inactive and cleaved/active forms of caspases 1 and 3. Lysates derived from IFNc-treated and untreated SW480 cells were used for comparison (IFNc þ/). Contr*: for detection of ASC, cell lysate of 293T cells transfected with ASC (Alexis Biochemicals/ Axxora, Gruenberg, Germany) was used as positive control. For detection of caspase-1, cell lysate derived from a Morbus Crohn mucosa was used as positive control.
AIM2-mediated effects on cell matrix adhesion and invasion of colon cancer cells
Reduced cell proliferation in combination with cell cycle arrest is a hallmark of colorectal metastasis, when disseminated tumor cells migrate and invade through the surrounding tissue.19–21 Invasion is initiated by adherence to and spreading along the blood vessel endothelium which is covered with ECM proteins. Accordingly, we addressed the effect of AIM2 expression on matrix-dependent adhesion and invasion. Fibronectin-induced cell adhesion was observed in both the control and HCT116-tet-AIM2 cells (Figs. 6a and 6b). However, HCT116-tet-AIM2 cells spread more rapidly, were more flattened, and formed large cell bodies containing multiple nuclei. After 2 hr on fibronectin, control cells were still rounded in part, whereas HCT116-tet-AIM2 cells were tightly attached to the surface and the number of enlarged cells was strongly increased (Fig. 6c). The total amount of HCT116tet-AIM2 cells attached to fibronectin did not differ from control cells (Fig. 6b). Moreover, no significant difference in adhesion between HCT116-tet-AIM2 and control cells was observed on laminin-coated surface (data not shown).
Recent findings in breast cancer cells suggested that AIM2 might exert tumor-suppressive activity by reducing cell proliferation and tumor growth in a mouse model. In addition, its high frequency of mutations in a subpopulation of colon and gastric cancers pointed to a role in gastrointestinal tumor progression.11,12 However, the mechanism of AIM2 action in carcinogenesis remains unclear. In principle, 2 different mechanisms are conceivable: AIM2 might act in a cell autonomous fashion, such as affecting cell growth by direct interaction with internal signaling pathways. Alternatively, AIM2 expression might interact with the tumor microenvironment for example by altering presentation of antigens or secretion of cytokines. We here investigated cell-autonomous effects of wild-type AIM2 restoration in a colorectal cancer cell line, lacking endogenous and IFN-c-induced AIM2 expression.12 Like some other HIN-200 proteins, AIM2 is thought to mediate IFN-directed effects by regulating cell proliferation.1 However, previous studies regarding cell growth upon induction or overexpression of AIM2 were inconsistent. Although AIM2 did not affect cell growth when reintroduced into melanoma cells,8 it was shown to repress proliferation of murine fibroblasts7 and human breast cancer cells.9 According to our data, AIM2 significantly inhibited cell growth as shown by reduced colony formation in HCT116, HeLa and 293 cells. Moreover, restoration of AIM2 in HCT116, lacking endogenous AIM2 expression,12 clearly repressed cell proliferation and cell viability. However, and in contrast to a recent report suggesting induction of apoptosis by AIM2 expression in breast cancer cells,9 there was no evidence for AIM2-induced apoptosis in HCT116 cells. As assessed by FACS analysis, AIM2 expression only marginally increased the percentage of apoptotic cells and no cleavage of caspase 3 could be identified in AIM2-transfected cells. C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
1845
Cancer Cell Biology
Patsos et al.
Figure 5. AIM2 expression affects cell cycle phase distribution. HCT116 cells, mock-transfected HCT116 cells (HCT116-hyg, clone D3) and AIM2-expressing cells (clones B8 and D1) were synchronized by serum starvation overnight and grown for 24 hr in the presence of 10% FCS. For flow cytometric analysis, cells were harvested and stained with propidium iodide, and minimum of 10,000 cells per sample were analyzed. (a) One representative out of 3 cell cycle analyses is shown. (b) Cell cycle phase ratios of 3 independent analyses (mean and SD are shown). (c) AIM2 expression increases protein levels of cyclin D3 p21Waf1/Cip1 and induces Tyr15-phosphorylation of cdc2. Cell lysates were obtained from synchronized cells shown in (b) and analyzed by western blotting for the presence of Tyr15-cdc2, p21Waf1/Cip1, cyclin D3 and actin. One representative out of 2 independent experiments is shown.
Instead, our data demonstrate that AIM2-mediated inhibition of cell proliferation is associated with accumulation of cells during late S-phase, thereby reducing transition to G2/ M phase. In accordance with these findings, 3 proteins known to be involved in the regulation of G2/M transition were altered upon AIM2 reconstitution: Cdc2, a catalytic subunit of the MPF, displayed increased phosphorylation at Tyr15 in HCT116-tet-AIM2 cells, indicating its inactivation. The progression of eukaryotic cells from G2 to M-phase of C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
the cell cycle depends on the activity of a functional complex formed by cdc2 and cyclin B22 and mitotic entry was shown to be delayed by phosphorylation of cdc2 at Tyr15.23 Thus, the observed G2/M arrest induced by AIM2 in HCT116 cells may be mediated through inhibition of cdc2. In addition, induction of p21Waf1/Cip1 in HCT116-tet-AIM2 cells further supports a role of AIM2 in inhibition of G2/M transition. Our findings are in good agreement with data on breast cancer cells delayed in the G2-phase independent from a G1
AIM2 activity in colon cancers
Cancer Cell Biology
1846
Figure 6. AIM2 expression affects cell adhesion and increases cell invasiveness. (a) Samples of parental, mock-transfected (D3) and AIM2expressing HCT116 cells (clones B8 and D1) were seeded at 50,000 cells per well on uncoated (upper panel) or fibronectin-coated tissue culture plates (lower panel, 2 hr). Cells were fixed in methanol and stained with crystal violet for microphotographs. (b) Cells grown on fibronectin for 1 or 2 hr were lysed in 10% acetic acid, to assess the incorporated dye released by adherent cells. Data are expressed as the mean of 8 experiments 6 SD. (c) The percentage of enlarged, multinuclear cells was determined by counting the cells from 2 equal areas of each cell clone.
arrest,24 where increased contents of p21Waf1/Cip1 were found in parallel with cdc2 inhibition. Moreover, a similar delay in G2/M transition was recently reported from another member of the HIN-200 family, the mouse homolog p205.25 Like AIM2, p205 contains only one of the 200 amino acid domains characterizing this protein family. The authors demonstrated that p205 expression results in elevated levels of p21Waf1/Cip1 and can induce growth arrest independent from p53 and Rb by delaying G2/M. Finally, we demonstrated that HCT116-tet-AIM2 cells show a marked increase in cyclin D3 level. High abundance of cyclin D3 is often found in quiescent cells and its elevated expression was reported to be associated with induction of differentiation in a number of cell lines and mammalian tissues in situ.26 Moreover, cyclin D3 appears to regulate the
G2/M DNA damage checkpoint.18 However, as p21Waf1/Cip1 and cyclin D3 expression are also often elevated during G1 phase, we cannot rule out an additional impact of AIM2 on G1/S transition. The observation that AIM2 inhibits proliferation in several cell types and suppressed tumor growth in a breast cancer model in mice9 suggested AIM2 to be tumor suppressive. However, our present findings disagree with this hypothesis. Although reconstitution of AIM2 in HCT116 cells reduced cell proliferation, HCT116-tet-AIM2 cells displayed increased in vitro invasiveness through complex ECM. Furthermore, expression of the invasion-associated genes VIM, MCAM, ANXA10 and CHD1 was altered in AIM2-reconstituted cells toward the expression pattern of a metastatic phenotype. Aberrant expression of vimentin (VIM) and melanoma cell C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Figure 7. AIM2 enhances invasion of colorectal cancer cells. (a) Invasion assay of different HCT116 cell clones through extracellular matrix-coated cell culture chambers toward 10% FCS as chemoattractant. Twenty-four hours after attachment, cells that invaded the membrane were detached and quantified by fluorimetric analysis. Data show the mean of 6 experiments 6 SD (*p ¼ 0.04; **p ¼ 0.009). (b) AIM2 expression affects transcription of invasion-associated genes in colorectal cancer cells. Relative expression of VIM, MCAM, ANXA10 and CDH1 was assessed by real-time PCR from HCT116-tet-AIM2 cells (clones B8 and D1) versus mock-transfected HCT116 cells (D3). Data are expressed as the mean of triplicates; bars: minimum and maximum fold expression of triplicates.
adhesion molecule (MCAM) were repeatedly reported to induce cell migration and invasion of epithelial cancer cells.27–31 Interestingly, Stat3, which is induced by IFN-c just as AIM2, was likewise described to enhance vimentin expression,32 implicating a role of AIM2 in the IFN-Stat3-signaling cascade. Downregulation of Annexin A10 (ANXA10) and Ecadherin (CDH1) in AIM2-reconstituted HCT116 cells further supports the involvement of AIM2 in acquisition of metastatic features, because reduced expression of these genes is associated with vascular invasion and poor prognosis in epithelial cancers.33–35 Thus, in contrast to previous findings in C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
1847
breast cancer cells, AIM2 does not appear to induce a classical tumor suppressive phenotype in colon cancer cells. The reason for this discrepancy is unknown, so far. However, it may result from different methods used for functional analysis of AIM2. We focused our study on cell cycle analysis and migration/invasion properties, whereas Chen et al. studied tumor growth of primary tumors without exploring metastasis.9 Some additional clues to understanding AIM2-mediated changes in cellular behavior might come from our observations of cells grown on fibronectin. The glycoprotein fibronectin is a major component of ECM that is known to stimulate intracellular signaling cascades and is implicated in the pathobiology of carcinogenesis.36 Upon binding to specific integrin receptors expressed on the surface of tumor cells, fibronectin was, for example, shown to turn nontumorigenic colon adenocarcinoma-derived cells into tumorigenic cells,37 to induce epithelial–mesenchymal transition in cervical cancer progression,38 and to be associated with more aggressive growth of breast and lung cancers.39,40 We cannot conclude from our data that whether fibronectin– integrin signaling directly induces multinuclear cells in HCT116-tet-AIM2 cells grown on fibronectin. However, it should be noted that formation of multinucleated giant cells was likewise observed in macrophages upon stimulation by fibronectin–integrin signaling in a former study.41 Alternatively, the formation of giant, multinuclear cells might be caused by incomplete mitosis in consequence of the G2/M cell cycle arrest in AIM2-expressing cells. Finally, fibronectin–integrin signaling is also involved in the downregulation of E-cadherin-dependent cell–cell contacts.42 It remains to be investigated whether downregulation of the E-cadherinencoding gene CDH1, as observed in HCT116-tet-AIM2 cells, is involved in the development of multinuclear cells or in increasing invasiveness. In summary, our results point to a more complex role of AIM2 in cancer cells. Although we could confirm previous reports describing the growth suppressive activities of AIM2 in tumor cells, HCT116-tet-AIM2 cells did not undergo apoptosis. Instead, the cell cycle arrest observed upon AIM2 reconstitution may reflect mechanisms associated with transition to a different state, where cancer cells reduce proliferation in favor of migration and invasion during metastasis. Although we cannot exclude the possibility that our data are restricted to cells tolerating a low constitutive expression level of AIM2, our findings imply that AIM2 is involved in the regulation of cellular migration/invasion. Bearing in mind that microsatellite unstable colon cancers are often infiltrated by INF-c–secreting lymphocytes,43 a similar selection for constitutive INF-c–induced AIM2 expression in single tumor cells is conceivable. Future analysis using animal models will help to determine the physiologic relevance of this phenomenon and to clarify whether colon cancer cells increase their metastatic potential upon AIM2 reconstitution in vivo.
Cancer Cell Biology
Patsos et al.
AIM2 activity in colon cancers
1848
Acknowledgements The authors thank Mrs. Susanne Eiermann and Mrs. Sigrun Himmelsbach for technical assistance in cloning of AIM2 cDNA and for performing some
of the western blots. They also thank Dr. Volker Ehemann for help with cell cycle analysis. This study was supported by a grant from the Deutsche Forschungsgemeinschaft (to J.G. and S.D.).
References
Cancer Cell Biology
1.
Ludlow LE, Johnstone RW, Clarke CJ. The HIN-200 family: more than interferoninducible genes? Exp Cell Res 2005;308: 1–17. 2. Ouchi M, Ouchi T. Role of IFI16 in DNA damage and checkpoint. Front Biosci 2008; 13:236–9. 3. Roberts TL, Idris A, Dunn JA, Kelly GM, Burnton CM, Hodgson S, Hardy LL, Garceau V, Sweet MJ, Ross IL, Hume DA, Stacey KJ. HIN-200 proteins regulate caspase activation in response to foreign cytoplasmic DNA. Science 2009;323: 1057–60. 4. Fernandes-Alnemri T, Yu JW, Datta P, Wu J, Alnemri ES. AIM2 activates the inflammasome and cell death in response to cytoplasmic DNA. Nature 2009;458: 509–13. 5. Hornung V, Ablasser A, Charrel-Dennis M, Bauernfeind F, Horvath G, Caffrey DR, Latz E, Fitzgerald KA. AIM2 recognizes cytosolic dsDNA and forms a caspase-1activating inflammasome with ASC. Nature 2009;458:514–8. 6. Burckstummer T, Baumann C, Bluml S, Dixit E, Durnberger G, Jahn H, Planyavsky M, Bilban M, Colinge J, Bennett KL, Superti-Furga G. An orthogonal proteomic-genomic screen identifies AIM2 as a cytoplasmic DNA sensor for the inflammasome. Nat Immunol 2009;10: 266–73. 7. Choubey D, Walter S, Geng Y, Xin H. Cytoplasmic localization of the interferoninducible protein that is encoded by the AIM2 (absent in melanoma) gene from the 200-gene family. FEBS Lett 2000;474:38–42. 8. Cresswell KS, Clarke CJ, Jackson JT, Darcy PK, Trapani JA, Johnstone RW. Biochemical and growth regulatory activities of the HIN-200 family member and putative tumor suppressor protein AIM2. Biochem Biophys Res Commun 2005;326:417–24. 9. Chen IF, Ou-Yang F, Hung JY, Liu JC, Wang H, Wang SC, Hou MF, Hortobagyi GN, Hung MC. AIM2 suppresses human breast cancer cell proliferation in vitro and mammary tumor growth in a mouse model. Mol Cancer Ther 2006;5:1–7. 10. Kulaeva OI, Draghici S, Tang L, Kraniak JM, Land SJ, Tainsky MA. Epigenetic silencing of multiple interferon pathway genes after cellular immortalization. Oncogene 2003;22:4118–27. 11. Woerner SM, Kloor M, Mueller A, Rueschoff J, Friedrichs N, Buettner R, Buzello M, Kienle P, Knaebel HP,
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Kunstmann E, Pagenstecher C, Schackert HK, et al. Microsatellite instability of selective target genes in HNPCC-associated colon adenomas. Oncogene 2005;24: 2525–35. Woerner SM, Kloor M, Schwitalle Y, Youmans H, Doeberitz MV, Gebert J, Dihlmann S. The putative tumor suppressor AIM2 is frequently affected by different genetic alterations in microsatellite unstable colon cancers. Genes Chromosomes Cancer 2007;46:1080–9. Patsos G, Andre S, Roeckel N, Gromes R, Gebert J, Kopitz J, Gabius HJ. Compensation of loss of protein function in microsatellite-unstable colon cancer cells (HCT116): a gene-dependent effect on cell surface glycan profile. Glycobiology 2009;19: 726–34. Gossen M, Bujard H. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc Natl Acad Sci USA 1992;89:5547–51. Southern PJ, Berg P. Transformation of mammalian cells to antibiotic resistance with a bacterial gene under control of the SV40 early region promoter. J Mol Appl Genet 1982;1:327–41. Baron U, Freundlieb S, Gossen M, Bujard H. Co-regulation of two gene activities by tetracycline via a bidirectional promoter. Nucl Acids Res 1995;23:3605–6. Dihlmann S, Kloor M, Fallsehr C, von Knebel DM. Regulation of AKT1 expression by beta-catenin/Tcf/Lef signaling in colorectal cancer cells. Carcinogenesis 2005;26:1503–12. Mullany LK, White P, Hanse EA, Nelsen CJ, Goggin MM, Mullany JE, Anttila CK, Greenbaum LE, Kaestner KH, Albrecht JH. Distinct proliferative and transcriptional effects of the D-type cyclins in vivo. Cell cycle (Georgetown, TX) 2008;7:2215–24. Palmqvist R, Rutegard JN, Bozoky B, Landberg G, Stenling R. Human colorectal cancers with an intact p16/cyclin D1/pRb pathway have up-regulated p16 expression and decreased proliferation in small invasive tumor clusters. Am J Pathol 2000; 157:1947–53. Palmqvist R, Sellberg P, Oberg A, Tavelin B, Rutegard JN, Stenling R. Low tumour cell proliferation at the invasive margin is associated with a poor prognosis in Dukes’ stage B colorectal cancers. Br J Cancer 1999;79:577–81. Jung A, Schrauder M, Oswald U, Knoll C, Sellberg P, Palmqvist R, Niedobitek G, Brabletz T, Kirchner T. The invasion front
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
of human colorectal adenocarcinomas shows co-localization of nuclear betacatenin, cyclin D1, and p16INK4A and is a region of low proliferation. Am J Pathol 2001;159:1613–17. Solomon MJ, Glotzer M, Lee TH, Philippe M, Kirschner MW. Cyclin activation of p34cdc2. Cell 1990;63:1013–24. McGowan CH, Russell P. Human Wee1 kinase inhibits cell division by phosphorylating p34cdc2 exclusively on Tyr15. EMBO J 1993;12:75–85. Barboule N, Lafon C, Chadebech P, Vidal S, Valette A. Involvement of p21 in the PKC-induced regulation of the G2/M cell cycle transition. FEBS Lett 1999;444:32–7. Asefa B, Dermott JM, Kaldis P, Stefanisko K, Garfinkel DJ, Keller JR. p205, a potential tumor suppressor, inhibits cell proliferation via multiple pathways of cell cycle regulation. FEBS Lett 2006;580: 1205–14. Bartkova J, Lukas J, Strauss M, Bartek J. Cyclin D3: requirement for G1/S transition and high abundance in quiescent tissues suggest a dual role in proliferation and differentiation. Oncogene 1998;17:1027–37. Sommers CL, Heckford SE, Skerker JM, Worland P, Torri JA, Thompson EW, Byers SW, Gelmann EP. Loss of epithelial markers and acquisition of vimentin expression in adriamycinand vinblastine-resistant human breast cancer cell lines. Cancer Res 1992;52: 5190–7. Gilles C, Polette M, Piette J, Delvigne AC, Thompson EW, Foidart JM, Birembaut P. Vimentin expression in cervical carcinomas: association with invasive and migratory potential. J Pathol 1996;180: 175–80. Zabouo G, Imbert AM, Jacquemier J, Finetti P, Moreau T, Esterni B, Birnbaum D, Bertucci F, Chabannon C. CD146 expression is associated with a poor prognosis in human breast tumors and with enhanced motility in breast cancer cell lines. Breast Cancer Res 2009;11:R1. Wu GJ, Fu P, Wang SW, Wu MW. Enforced expression of MCAM/MUC18 increases in vitro motility and invasiveness and in vivo metastasis of two mouse melanoma K1735 sublines in a syngeneic mouse model. Mol Cancer Res 2008;6: 1666–77. Wu GJ, Peng Q, Fu P, Wang SW, Chiang CF, Dillehay DL, Wu MW. Ectopical expression of human MUC18 increases
C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
Patsos et al.
35. Guarino M, Rubino B, Ballabio G. The role of epithelial-mesenchymal transition in cancer pathology. Pathology 2007;39: 305–18. 36. Larsen M, Artym VV, Green JA, Yamada KM. The matrix reorganized: extracellular matrix remodeling and integrin signaling. Curr Opin Cell Biol 2006;18:463–71. 37. Lopez-Conejo MT, Olmo N, Turnay J, Lopez De Silanes I, Lizarbe MA. Interaction of fibronectin with human colon adenocarcinoma cells: effect on the in vivo tumorigenic capacity. Oncology 2002;62:371–80. 38. Lee MY, Chou CY, Tang MJ, Shen MR. Epithelial-mesenchymal transition in cervical cancer: correlation with tumor progression, epidermal growth factor receptor overexpression, and snail upregulation. Clin Cancer Res 2008;14: 4743–50. 39. Yao ES, Zhang H, Chen YY, Lee B, Chew K, Moore D, Park C. Increased beta1 integrin is associated with decreased
C 2009 UICC Int. J. Cancer: 126, 1838–1849 (2010) V
40.
41.
42.
43.
survival in invasive breast cancer. Cancer Res 2007;67:659–64. Ritzenthaler JD, Han S, Roman J. Stimulation of lung carcinoma cell growth by fibronectin-integrin signalling. Mol Biosyst 2008;4:1160–9. Kao WJ, Lee D, Schense JC, Hubbell JA. Fibronectin modulates macrophage adhesion and FBGC formation: the role of RGD, PHSRN, and PRRARV domains. J Biomed Mater Res 2001;55: 79–88. Azare J, Leslie K, Al-Ahmadie H, Gerald W, Weinreb PH, Violette SM, Bromberg J. Constitutively activated Stat3 induces tumorigenesis and enhances cell motility of prostate epithelial cells through integrin beta 6. Mol Cell Biol 2007;27: 4444–53. Smyrk TC, Watson P, Kaul K, Lynch HT. Tumor-infiltrating lymphocytes are a marker for microsatellite instability in colorectal carcinoma. Cancer 2001;91: 2417–22.
Cancer Cell Biology
metastasis of human prostate cancer cells. Gene 2004;327:201–13. 32. Wu Y, Diab I, Zhang X, Izmailova ES, Zehner ZE. Stat3 enhances vimentin gene expression by binding to the antisilencer element and interacting with the repressor protein ZBP-89. Oncogene 2004; 23:168–78. 33. Liu SH, Lin CY, Peng SY, Jeng YM, Pan HW, Lai PL, Liu CL, Hsu HC. Downregulation of annexin A10 in hepatocellular carcinoma isassociated with vascular invasion, early recurrence, and poorprognosis in synergy with p53 mutation. Am J Pathol 2002;160:1831–7. 34. Peng SY, Ou YH, Chen WJ, Li HY, Liu SH, Pan HW, Lai PL, Jeng YM, Chen DC, Hsu HC. Aberrant expressions of annexin A10 short isoform, osteopontin and alphafetoprotein at chromosome 4q cooperatively contribute to progression and poor prognosis of hepatocellular carcinoma. Int J Oncol 2005;26:1053–61.
1849
