ZBRK1 Acts as a Metastatic Suppressor by Directly Regulating MMP9 in Cervical Cancer

Cancer metastasis is the most common cause of death among cancer patients (1). It results from several highly organized sequential steps involving interactions between cancer cells and the host. Metastasis is a multistep process: invasion of tumor cells into adjacent tissues, entry of tumor cells into the systemic circulation (intravasation), survival in circulation, extravasation to distant organs, and finally growth of cancer cells to produce secondary tumors (2, 3). Recently, gene expression analyses of human breast carcinomas with known clinical outcomes revealed profiles associated with disease progression and identified groups of genes, including cell migration genes, whose characteristic expression patterns can predict the risk of metastatic recurrence (4–8). However, details of candidate genes involved in other aspects of cancer metastasis/invasion processes remain underinvestigated. Precisely how tumor cells become metastatic is still largely unknown, especially in terms of a transcriptional factor that serves as a repressor in metastasis/invasion.

Krupple-associated box (KRAB)–containing zinc finger (KZF) proteins comprise a group of the most widely distributed transcriptional repression proteins in mammals. They are composed of a KRAB domain at the NH₂ terminus and tandem C2H2 class zinc fingers at the COOH terminus. Several KZF proteins modulate cell growth and survival and are implicated in malignant disorders (9). Although a majority of KZF proteins function to repress RNA polymerase II–mediated transcription, their respective target genes, which underlie transcriptional regulation mechanisms, and their cellular activities remain largely unclear. Members of the really interesting new gene (RING) family are found throughout eukaryotic cells and play key functions in processes as diverse as development, oncogenesis, viral replication, and apoptosis (10, 11).

The KRAB domain is divided into A and B boxes and is required for repression of transcription by recruiting corepressors such as KRAB domain–associated protein 1 (KAP1; ref. 12). KAP1 was initially identified as a transcriptional corepressor and its NH₂-terminal RING-B boxed coiled-coil domain serves as a protein-protein interaction region for binding to the KRAB domain of KZF proteins (12–14). The tandem C2H2 class zinc fingers at the COOH terminus were used to bind specific DNA sequences. It has been suggested that some cofactors may cooperate with those zinc fingers to enhance the DNA-binding specificity (15, 16).

Zinc finger and BRCA1-interacting protein with KRAB domain-1 (ZBRK1), which was first identified in a yeast two-hybrid screening for proteins associated with BRCA1, is a typical KRAB-containing zinc finger protein that contains a highly conserved KRAB domain at the NH₂ terminus, eight consecutive C2H2 zinc finger motifs, and a CTRD domain for BRCA1 interactions at the COOH terminus (17). Its zinc finger repeats were suggested to recognize a consensus DNA-binding element, GGGxxxCAGxxxTTT, or to be involved in protein interactions (18). Two corepressors of RING members, BRCA1 and KAP1, were shown to interact with ZBRK1 in coordinating transcriptional regulation of diverse DNA damage response genes (19). Some of the ZBRK1-affected downstream targets, including GADD45 and p21, have been reported (20–22). Recently, ZBRK1 has been identified as cooperating with the CtIP/BRCA1 to repress angiopoietin-1 (ANG1) gene activation (22) and may play a role in tumor angiogenesis, implying that it may act as a potential tumor suppressor. However, whether ZBRK1 plays a direct role in tumor progression, especially in metastasis, has yet to be shown.

In this investigation, we found that reduction of ZBRK1 expression was observed in highly malignant cervical cancer cells compared with the counterpart normal tissue. Ectopic expression of ZBRK1 in HeLa cells significantly inhibits its neoplastic phenotypes. This article is the first to report on this significant discovery in cervical cancer cells; it seems that a reduction of ZBRK1 allows the growth of cancer cells, whereas an increase of ZBRK1 has been shown to inhibit cell growth both in vitro and in vivo, suggesting that ZBRK1 can act as a tumor suppressor. Interestingly, analyses of gene expression patterns of these cells revealed groups of genes not only critical for cell proliferation but also for cell motility being downregulated. Consistently, ectopic expression of ZBRK1 inhibits HeLa cell migration, in part by directly repressing transcription of the MMP9 metastatic gene. This study also presents the first demonstration of the direct negative repression of the transcriptional regulation of the MMP9 gene. This is further validated in cervical cancer specimens, in which loss of ZBRK1 expression is inversely correlated with the elevated expression of MMP9. All things considered, these results suggest that ZBRK1 plays a critical role in tumor progression, especially in metastasis, by directly modulating metastatic genes.

Mammalian cells, U2OS, and various cervical cancer cell lines were cultured in complete medium containing DMEM, 10% fetal bovine serum (FBS), 100 μg/mL streptomycin, and 100 units/mL penicillin. U2OS and HeLa cells stably expressing enhanced green fluorescent protein (EGFP) or EGFP-ZBRK1 were generated by hygromycin B selection.

For cell proliferation assay, daily cell culture samples were counted in a Neubauer chamber after trypsinization. Viability was assessed by trypan blue exclusion. Assays were performed in triplicates. For focus-forming assay, U2OS and HeLa cells stably expressing EGFP or EGFP-ZBRK1 were plated at low density (1,000 cells per plate) onto 100-mm dishes cultured with DMEM containing 5% FBS and the media were changed every 3 ± 4 d. After 2 wk, the number of colonies was counted with the Sigmascan software program by staining with 2% methylene blue in PBS. For anchorage-independent assay, 1 mL of 0.3% agarose in complete growth medium containing 2,000 HeLa cells stably expressing EGFP or EGFP-ZBRK1 were seeded onto a hard agar base composed of 1.5 mL 10% FBS-DMEM and 0.6% agarose in six-well plates as described (23). After 14 to 21 d, cells were stained with 0.05% crystal violet and the colonies were scored for statistical analysis. All the above experiments were performed at least twice in triplicate.

Total RNA was extracted by using TRIzol RNA extraction agent (Invitrogen). The isolated RNAs were subjected to reverse transcription reaction with SuperScript III (Invitrogen) for cDNA synthesis. PCR was performed with the pairs of specific primers as follows: ZBRK1, 5′-GACATATGGAAAGTTGATCATGTGCTG-3′ and 5′-ATTCACTGCACACATGATGCTTCTCTA-3′; GAPDH, 5′-CCATCACCATCTTCCAGGAG-3′ and 5′-CCTGCTTCACCACCTTCTTG-3′. The total RNA samples harvested from cells stably expressing EGFP or EGFP-ZBRK1 were subjected to microarray analysis at WEIGENE Biotech. Comp, using Agilent Whole Human Genome Oligo 4 × 44K array.

Athymic (nu/nu) nude mice (4–6 wk of age) were obtained from the National Laboratory Animal Center and fed with Laboratory Autoclavable Rodent Diet (LabDiet). All animal work was done in accordance with the protocol approved by the Institutional Animal Care. Aliquots of 3 × 10⁶ EGFP and EGFP-ZBRK1 HeLa cells were inoculated s.c. into the dorsal rear flanks of nude mice. After injection, tumor size was measured by external caliper and tumor volumes were calculated using a standard formula as follows: V = height × width × depth. Mice were sacrificed at 11 d after the cell injection.

Twelve female 4- to 6-wk-old nonobese diabetic-severe combined immunodeficient (NOD-SCID) mice were injected with 1 × 10⁶ EGFP or EGFP-ZBRK1 HeLa cells. Cells were washed and harvested in PBS and injected i.v., through the lateral tail vein, in a volume of 0.1 mL. At 14 d after injection, when mice had not died but are with significant morbidity, all mice were euthanized and their lungs were removed. The number of surface metastases per lung was counted.

After U2OS and HeLa cells were grown to confluence in 6-mm culture plates, an artificial “wound” was created using a P-200 pipette tip to scratch on the confluent cell monolayer. Cell migration into the scratched region was recorded using a Nikon microscopy system (Nikon Instrument) at 0 and 48 h, and the number of cells in the scratched region was quantified as the percentage of wound healing.

A migration assay kit (ECM 566 QCM 96-well Haptotaxis Cell Migration Assays-Collagen 1, Fluorimetric from Chemicon) and an invasion assay kit (ECM 555 QCM 96-well cell invasion assay from Chemicon) were used according to the manufacturer's instructions. Cells were seeded at a density of 1.5 × 10⁵ per well in 100 μL DMEM onto the upper chambers and DMEM with 1% FBS (150 μL) was added to the lower chambers of the 96-well plates. The plates were incubated at 37°C and 5% CO₂ for 6 h to allow migration or invasion through the membrane. Cells that had migrated through to the bottom of the insert membrane were dissociated from the membrane and detected by CyQUANT GR dye. The activity of cell migration and invasion was calculated as the percentage of the fluorescence relative to the controls.

The chromatin immunoprecipitation (ChIP) assay was carried out essentially as described by Furuta and colleagues (22), with a minor modification. Briefly, the sheared chromatin fragments were immunoprecipitated with antibodies specific to GFP, ZBRK1, or control mouse IgG at 4°C overnight. After dissociating the DNAs from immunoprecipitated chromatin, the DNAs underwent PCR amplification. For PCR amplification of specific regions (A and B) of the MMP9 genomic locus, the following sets of primers were used: A fragment of primers, forward 5′-TGCCCGTAATCCTAGCACTTTGGGA-3′ and reverse 5′-CCTCACTCCTTTCTTCCTAGCCAGC-3′; B fragment of primers, forward 5′-AGGCTGTCAGGGAGGGAAAAAGAGG-3′ and reverse 5′-AGAAAGGGCTTACACCACCTCCTCC-3′. The PCR conditions are as follows: 32 cycles of 30 s at 94°C, 25 s at 56°C, and 1 min at 72°C.

The reporters bearing the different fragments of the human MMP9 promoter were generated by PCR with genomic DNA as the template. The primers for the PCR reaction were as follows: MMP9/−1940, 5′-GGGGTACCAGTGACTTGCCCAAGGTCACATAGC-3′; MMP9/−940, 5′-GGGGTACCTACAGGAATGAGCCACCATACCTGG-3′; and MMP9/+113, 5′-CCCAAGCTTTGAGATTGGTTCTCAGGTCTCCAGG-3′. The PCR fragments were cloned into the multicloning sites of the promoter-less vector, pGL2-basic vector, and verified by sequencing. For the heterologous reporters, the various copies of repeated ZBRK1- or mutant ZBRK1-binding motifs on the MMP9 promoter were inserted into pGL2-promoter vector. These heterologous reporters were introduced into HeLa cells by Lipofectamine 2000 according to the manufacturer's instruction, and then the lysates of transfectants were harvested for luciferase assay.

The presence of MMP9 in the supernatants of EGFP and EGFP-ZBRK1 HeLa cells was analyzed with gelatin zymograms as described previously (24). Briefly, the cells were incubated in serum-free DMEM with or without 100 ng/mL of epidermal growth factor (EGF), and the supernatants were collected after stimulation for 24 h, clarified by centrifugation, normalized to cell number, mixed with nonreducing Laemmli sample buffer, and separated by electrophoresis in 10% SDS-PAGE containing 1 mg/mL gelatin (Sigma). After electrophoresis, the gels were renatured by washing in 2.5% Triton X-100 solution (Sigma) twice for 30 min to remove all SDS. The gels were then incubated in 50 mmol/L Tris-HCl (pH 7.4), 5 mmol/L CaCl₂, 1% Triton X-100, and 0.02% NaN₃ at 37°C overnight. After incubation, the gels were stained with 0.05% Coomassie brilliant blue R-250 for 1 h at room temperature and then destained in distilled water. The MMP9 activities were visible as clear bands on a blue background where the gelatin substrate had been hydrolyzed by enzyme activity.

Twelve patients with cervical cancer were surgically resected at the National Cheng Kung University Hospital. Total RNA samples were extracted from tumor tissues and adjacent unaffected cervical or colon tissues. All tumor specimens of patients were obtained from surgically resected tissues that had previously been pathologically assessed at the National Cheng Kung University Hospital. The fresh tissue samples were immediately cut into small pieces, snap-frozen in liquid nitrogen, and stored in a deep freezer. Total RNA was extracted from the tumorous and paired nontumorous tissues using the TRIzol RNA extraction reagent.

The lentiviral expression vectors pLKO.1-shLuc, containing 5′-CTTCGAAATGTCCGTTCGGTT-3′, and pLKO.1-shZBRK1, containing 5′-GCTAACCATGAACGACTTCAT-3′, were obtained from the National RNAi Core Facility located at the Genomic Research Center of Institute of Molecular Biology, Academia Sinica. Virus was produced as described using Lipofectamine 2000 cotransfection of Phoenix cells with the pLKO.1-shLuc or pLKO.1-shZBRK1 vector together with pMD2.G and psPAX2.

ZBRK1 is a transcriptional repressor and a potential tumor suppressor because it cooperates with BRCA1/CtIP in modulating ANG1 expression (22). It has not been known whether its expression level has any relationship with tumor malignancy. To test this possibility, we examined the mRNA expression level of ZBRK1 in a panel of cervical cancer cells and clinical cervical cancer specimens. It found that 7 of 7 cervical cancer cell lines and 9 of 12 cervical cancer specimens expressed significantly reduced amounts of ZBRK1 RNA when compared with the normal cervical tissues (Fig. 1A and B). This result suggests that reduction of ZBRK1 expression may be involved in tumorigenesis. To further corroborate this observation, stable cell lines bearing ectopically expressed EGFP-tagged ZBRK1 in HeLa and U2OS cells were established for comparison with the parental cells in growth rate, plating colony formation efficiency, soft-agar colony formation, and tumor formation in mice. As shown in Fig. 2, HeLa cells constitutively expressing ZBRK1, either clone 1 or 6, which expressed different levels of ZBRK1, grew slower (Fig. 2A) and formed less colonies in plating efficiency (Fig. 2B), less colonies in soft agar (Fig. 2C), and much smaller tumors in nude mice (Fig. 2D). These results suggest that ZBRK1 expression inversely correlates with the malignancy of tumor cells and ZBRK1 acts as a tumor suppressor to inhibit tumor progression.

To search the genes regulated by ZBRK1, two sets of microarray analyses using stable cell lines, EGFP-ZBRK1 and EGFP HeLa cells and EGFP-ZBRK1 and EGFP U2OS cells, were performed. About 304 and 196 upregulated genes, and 473 and 341 downregulated genes, were identified from these two sets of expression profiles of U2OS and HeLa cells, respectively (Supplementary Fig. S2A and B). Interestingly, 36 ZBRK1-downregulated genes and 23 ZBRK1-upregulated genes were commonly shared between these two ZBRK1 ectopically expressed cells (Supplementary Table S1). The expression of some of these commonly regulated genes was further validated by reverse transcription-PCR (RT-PCR; Supplementary Fig. S1D and E). The ZBRK1-regulated genes were divided into groups by the Ingenuity program according to their known biological functions (Supplementary Table S2). Based on this analysis, it seemed that the major functions of these genes affected by ZBRK1 could be categorized into cell proliferation, which is consistent with the observation as described in Fig. 2, and cell migration. Subsequently, to verify the potential role of ZBRK1 in regulating cell migration, we then performed a wound-healing assay and showed a slow-healing effect in EGFP-ZBRK1 HeLa cells in a dose-dependent manner when compared with the parental cells (Fig. 3A). Similar results were also observed in U2OS cells (30–40% reduction; Supplementary Fig. S3A). The migration and invasion assays were performed individually, which showed coincident results with the wound-healing assay (Supplementary Fig. S3B; Fig. 3B). To further validate this notion, a poorly metastatic/invasive cell line A431 with higher expression levels of ZBRK1 mRNA (Fig. 1) as well as protein (Fig. 3C) was used for depletion of ZBRK1 through lentiviral-mediated shRNA knockdown approach. As shown in Fig. 3C, knockdown of ZBRK1 expression in A431 cells promoted cell migration in the QCM Haptotaxis Cell Migration Assay, suggesting that the reduction of ZBRK1 has the advantage of promoting cancer cell migration. To further evaluate the effects of ZBRK1 on metastasis/invasion, EGFP-ZBRK1 and EGFP HeLa cells were injected into the lateral tail vein of scid mice, and their metastasis and growth in lung tissues were measured. After 4 weeks, the injected control tumor cells had formed 8 to 20 metastatic nodules per lung in all six mice. In contrast, mice injected with the same amount of EGFP-ZBRK1 HeLa cells formed 0 to 3 nodules (Fig. 3D). These results show that ectopic expression of ZBRK1 reduced or eliminated the metastasis/invasion ability of cancer cells.

Among the cell motility genes potentially regulated by ZBRK1 (Supplementary Table S2), MMP9 is of particular interest because MMP9 expression was the most significantly reduced in tumor specimens derived from HeLa clone 6 grown in mice when compared with HeLa-C tumors (Fig. 4A,, left). Similarly, using shRNA to knockdown ZBRK1 mRNA in HeLa cells, which ectopically expressed ZBRK1, expression of MMP9 mRNA was elevated in these cells when compared with the control shRNA (Fig. 4A,, right). These results strongly suggest that ZBRK1 downmodulates MMP9 expression. A putative ZBRK1-binding motif, −1504/−1518, was identified on the MMP9 promoter. Therefore, we performed an EMSA by using in vitro–translated ZBRK1 to incubate with the fragment −1524 to −1497 on the MMP9 promoter region. As shown in Supplementary Fig. S4, ZBRK1 can bind to the putative ZBRK1-binding motif. Next, we cloned the MMP9 promoter to verify whether ZBRK1 regulates MMP9 transcription through its promoter region. Ectopic expression of ZBRK1 specifically reduced the MMP9 reporter activity containing the putative ZBRK1-binding motif, MMP9−1940/+113, but not the MMP9−940/+113 reporter, with no putative ZBRK1-binding motif (Fig. 4B). To further confirm the repressive effect of the ZBRK1-binding motif on the MMP9 promoter, heterologous reporters generated by inserting a wild-type ZBRK1-binding motif, −1524/−1497 oligonucleotides, or a mutated ZBRK1-binding motif, −1524/−1497 oligonucleotides, into the pGL2-promoter vector were used for the reporter assay. The MMP9−1524/−1497 reporters pMMP9-Zx1 and pMMP9Zx3 showed repressive activity in dose-dependent manners, but the reporter bearing a mutant ZBRK1-binding motif, pMMP9-mZx2, showed no repressive activity (Fig. 4C). These results suggest that ZBRK1 can inhibit MMP9 promoter activity through the ZBRK1-binding motif, −1524/−1497. Importantly, ectopic expression of ZBRK1 overcame the transcriptional activation by tumor necrosis factor-α (TNF-α) or EGF, which is known to induce MMP9 reporter activity (Fig. 4D,, left; refs. 25, 26). Consequentially, MMP9 enzymatic activity induced by EGF was also reduced in cells ectopically expressing ZBRK1 (Fig. 4D,, right). To determine whether ZBRK1 indeed binds to the specific region containing the ZBRK1-binding motif of the MMP9 promoter in vivo, ChIP assay was performed, which showed that the A fragment, a PCR product of −1706/−1315, but not the B fragment, a PCR product of −526/−150, of the MMP9 promoter region was detected following immunoprecipitation by ZBRK1 antibodies or GFP antibodies in EGFP or EGFP-ZBRK1 cells (Fig. 5). These results strongly suggest that ZBRK1 directly downregulates MMP9 expression by specifically binding to the MMP9 promoter region containing the ZBRK1-binding motif.

Next, to test whether repression of MMP9 expression by ZBRK1 also occurs in clinical cancer specimens, we then collected 12 pairs of cervical cancer and noncancer tissues from the same patient for comparison of their expression pattern. Because we have previously reported that SUZ12, a polycomb protein, is overexpressed in specimens of cervical cancer (23), we then analyzed the three genes, ZBRK1, SUZ12, and MMP9, using the collected specimens. As shown in Supplementary Fig. S5 and Table 1, reduction of the ZBRK1 expression was observed in 9 of 12 pairs that inversely correlated with the upregulation of MMP9 with 100% coincidence. Although the expression of SUZ12 was also elevated in the majority of cervical cancer samples (9 of 12), the inverse correlation with ZBRK1 was not completely concordant (Table 1). These data further support the notion that ZBRK1 downregulates MMP9 in cervical cancer specimens.

ZBRK1 is a transcriptional repressor with potential activity in tumor suppression because of its association with BRCA1, a bona fide tumor suppressor. In this communication, our results indicate that ZBRK1 plays an important role in suppressing tumor progression, including metastasis. First, ZBRK1 expression was significantly lower in highly malignant cervical cancer cells and clinical cervical cancer specimens than in normal tissue counterpart, whereas ectopic expression of ZBRK1 in HeLa cells significantly inhibits its neoplastic phenotypes, including tumor growth in nude mice. Second, the metastasis-suppressing activity is shown by ectopic expression of ZBRK1 that inhibits HeLa cell metastasis shown in both cell mobility assay in culture and experimental metastatic assay in mice. Third, ZBRK1 directly represses transcription of the metastatic gene MMP9, and loss of ZBRK1 expression is inversely correlated with the elevated expression of MMP9 in cervical cancer specimens. These results indicate that ZBRK1 has a critical role in tumor progression, notably in metastasis.

Reduction of the expression of ZBRK1 in cervical tumors and cell lines provides a clue suggesting that ZBRK1 may behave as a tumor suppressor. Downregulation of ZBRK1 was observed in many cervical cancer cell lines and 75% of cervical cancer specimens (n = 12; Fig. 1; Table 1). Interestingly, our preliminary data indicate that the reduction of ZBRK1 expression is also observed in other cancers, including 50% of hepatocellular carcinomas (n = 8) and 37.5% of colorectal cancers (n = 8).⁹

Intriguingly, overexpression of ZBRK1 has the potential to inhibit cancer cell migration and suppress metastasis activity (Fig. 3), suggesting that ZBRK1 may behave as a metastasis suppressor. It was reported that metastasis was inhibited in an in vivo assay, whereas tumorigenicity was not significantly affected following reexpression of a metastasis-suppressor gene in a tumor cell line (27, 28). This is an essential criterion as a metastasis-suppressor gene does not affect the growth of the primary tumor. However, the proliferation assay (Fig. 2A) and downstream targets, including EGR1 and HMGA2, shown in profiling suggested that ZBRK1 could also be a proliferation suppressor. Therefore, we could not rule out the potential of the involvement of proliferation in these experimental migration assays. Even so, these current results suggested that ZBRK1 serves as an upstream regulator integrating the proliferation and cell migration in tumor progression. Therefore, ZBRK1 may be a useful therapeutic target by reactivation of ZBRK1 expression to inhibit metastasis of tumor cells through blocking of the metastatic cascade. Further pursuit of this possibility is warranted.

Further, expression of ZBRK1 inversely correlated with the expression of a group of genes participating in cell proliferation and mobility, particularly a well-characterized metastatic gene, MMP9 (Supplementary Table S2), in both cell lines as well as in clinical specimens. MMP9 is thought to play important roles in invasion and metastasis and has been shown to have multiple effects on cell motility (29). Although extrinsic stimulators in this activation include EGF, PMA, and TNF-α, other transcriptional activators, including AP-1 and CBP/p300, have been reported to be involved in the transcriptional activation of the MMP9 gene (30–32). Negative regulators of the MMP9 transcription, such as interferon-γ and interleukin-10, have also been reported to inactivate MMP9 transcription (33, 34). However, the precise mediator for MMP9 was not elucidated until this report revealed that ZBRK1 directly binds to the promoter and serves as a transcriptional repressor of MMP9. Interestingly, how these external stimulators and negative regulators integrate to modulate the derepression of ZBRK1 and subsequently activate the transcription of MMP9 will be a complicated but exciting subject to resolve.

Based on the global profiling regulated by ZBRK1, several other interesting candidates involved in the processes of migration and invasion were identified (Supplementary Fig. S2; Supplementary Table S1), which account for the function of ZBRK1 as a tumor suppressor, especially in repression of metastasis/invasion. For instance, ZBRK1 inhibits MMP3, LAMA1, FGF18, EGR1, and HMGA2 in addition to MMP9 but activates ICAM1, ANK1, NCAM1, and VIM. Therefore, the inactivation of ZBRK1, which results in the increase of MMP3 and MMP9, and the decrease of ICAM1 and NCAM1 may benefit the invading cancer cells migrating from the extracellular matrix of origin. ZBRK1 directly represses the MMP9 expression through binding to its promoter, similarly to other ZBRK1-regulated genes, including ANG1 (22). Although the detailed mechanism of transcriptional repression remains to be unraveled, one possibility is through its partner proteins such as BRCA1 and KAP1 to differentially regulate these genes. It was reported that ZBRK1 interacts with BRCA1 through its COOH terminus to negatively regulate the transcription of GADD45 and ANG1 (35), whereas p21 transcription was repressed by ZBRK1 through KAP1 recruitment (21). It is likely that either BRCA1 or KAP1 will work with ZBRK1 to repress MMP9. For those genes repressed by ZBRK1, it will be interesting to classify which one is going through BRCA1 or KAP1 or both. Furthermore, how ZBRK1 activates the expression of ICAM1 and other genes is another unknown process. ZBRK1 apparently can directly bind to the putative ZBRK1 motif on the ICAM1 promoter based on our EMSA. However, it is not known which coactivator will interact with ZBRK1 to turn on the expression of ICAM1. Elucidating each specific regulation mode of ZBRK1 will provide important information to explain its tumor-suppressive effects.

ZBRK1 mutation rarely occurs in tumor cells (35, 36). This current study clearly showed that reduction of ZBRK1 transcription in 75% of cervical cancer specimens is crucial for cancer progression. Therefore, elucidation of ZBRK inactivation in cancer cells would be an interesting issue. The ZBRK1 gene locates on chromosome 19q13.41. A recent publication showed a high frequency of 19q12 deletion (37), yet the frequently deleted locus is 19p13.11-q12 (15% cases), which does not include the ZBRK1 gene. This suggests that the transcriptional regulation or the loss of heterozygosity may explain the inactivation of ZBRK1 gene, which will be the focus of the next pursuit to fully elucidate the contribution of ZBRK1 during cancer progression.

No potential conflicts of interest were disclosed.

We thank Christine C. Hsieh and Dr. Kazi Ahmed for the review of the manuscript.

Grant Support: Grants NSC 98-2320-B-006-036-MY3 and the National Cheng Kung University landmark grant of C007 (J.-M. Wang) and NIH RO1 94170 (W.-H. Lee). Funding to pay the Open Access publication charges for this article was provided by grant NSC 98-2320-B-006-036-MY3 (J.-M. Wang).

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.