Cyclin-dependent kinase 2 (cdk2) is a small serine/threonine kinase that regulates cell cycle progression. Cdk2 activity is tightly controlled by several mechanisms, including phosphorylation and dephosphorylation events. Cables is a recently described novel cdk-interacting protein. In proliferating cells, Cables was predominantly localized in the nucleus by cell fractionation and immunostaining. Expression of Cables in HeLa cells inhibited cell growth and colony formation. Cables enhanced cdk2 tyrosine 15 phosphorylation by the Wee1 protein kinase, an inhibitory phosphorylation, which led to decreased cdk2 kinase activity. The gene encoding Cables is located on human chromosome 18q11-12, a site that is frequently lost in squamous, colon, and pancreas cancers. We found that Cables was strongly expressed in normal human epithelial cells including squamous and glandular mucosa. Breast and pancreatic cancers show strong Cables expression; however, loss of Cables expression was found in approximately 50–60% of primary colon and head and neck cancer specimens. Lack of Cables expression was associated with loss of heterozygosity on chromosome 18q11. The data provide evidence for a Cables-mediated interplay between cdk2 and Wee1 that leads to inhibition of cell growth. Conversely, loss of Cables may cause uncontrolled cell growth and enhance tumor formation.

Cdks2 comprise a family of serine/threonine protein kinases that have been shown to be key regulators of cell cycle progression. The cdks require association with regulatory subunits known as cyclins for activation. In addition to binding cyclins, post-translational phosphorylation and dephosphorylation events regulate cdk activity (1). Phosphorylation of the threonine residue in the T loop (T160 on cdk2 or T161 on cdc2) by cdk-activating kinase is an obligatory step in kinase activation, and a threonine to alanine mutation of this residue renders the cdk inactive. On the other hand, phosphorylation of the threonine 14 and tyrosine 15 (Y15) residues by the Wee1 family of dual specificity kinases is inhibitory for the cdks, and dephosphorylation of these residues by the Cdc25 family of phosphatases coincides with cdk activation.

There is a clear connection between cdk regulation and cancer. Overexpression of certain cyclins contributes to cell transformation. A mutant cyclin-A protein that lacked the cyclin destruction box as a result of hepatitis B virus insertion was implicated in the tumorigenesis of a hepatoma (2). Cyclin D1 has been identified as the Bcl-1 proto-oncogene, which is overexpressed by translocation or amplification in parathyroid adenomas, lymphomas, squamous cell carcinomas of the head and neck, and breast tumors (3, 4, 5). Some of the cdk inhibitory molecules are tumor suppressor genes, such as p16, which has alternatively been termed multiple tumor suppressor 1 (6). Another cdk interactor, p21, is activated by p53 and blocks cdk activity and cell cycle progression (7, 8). In cells lacking functional p53, the failure to induce p21 after DNA damage may contribute to the increased incidence of chromosomal abnormalities and genetic instability in transformed cells. Similarly, many tumor cell lines proceed through the cell cycle with damaged DNA suggesting there is a defect in the regulation of cdk2 T14/Y15 phosphorylation. The defect could lie in the Wee1-mediated cdk2 T14/Y15 phosphorylation or Cdc25 phosphatase, which removes the cdk inhibitory phosphorylations. Cdc25 has been implicated as an oncogenic protein in colorectal, stomach, breast, and lung cancers (9).

Cables is a novel cdk-interacting protein that has been found to interact with cdk2, cdk3, and cdk5 (10, 11). In neurons, Cables acts as a link between cdk5 and c-Abl. Enhanced cdk tyrosine phosphorylation occurs in the presence of Cables and modulates the activity of the cdk5, which is important in neurite outgrowth. However, Cables is also expressed in most cell lines, and high levels of Cables are found in embryonic and adult tissues, suggesting that it has a role in cell proliferation and the development of multiple organ systems.

We found that the Cables is predominantly located in the nucleus of proliferating cells, and expression of Cables in actively growing cells inhibited cell growth. Cables enhanced cdk2 Y15 phosphorylation by Wee1, an inhibitory phosphorylation, and caused decreased cdk2 kinase activity. The Cables gene is located on chromosome 18q11-12, a site of allelic loss in many human cancers. We found lack of Cables protein expression in 50–60% of primary human colon and head and neck squamous cell carcinomas. The data suggest that Cables and cdk2 tyrosine phosphorylation are involved in cell growth regulation, and loss of Cables may be involved in the pathogenesis of human cancers.

Cell Culture and Generation of Stable Cell Lines.

HeLa and COS7 cells were propagated in DMEM with 4.5 g/liter of glucose, 10% FCS, and penicillin/streptomycin. Transient transfections in COS7 cells were performed using the calcium phosphate method with 10–20 μg of total DNA. For generation of stable cell lines, full-length Cables was cloned into the pCIN4 vector (12) and the resulting plasmid was transfected into HeLa cells using Lipofectamine (Life Technologies, Inc.). Stable cell lines were selected by growth in 50 μg/ml G418. Control cells were made by transfection of vector alone. The growth rates of both the Cables and control cells were determined by seeding 1 × 104 cells/well and counting the cells in triplicate for up to 13 days. Colony-formation efficiency was determined by plating single cells at 1 × 103/100-mm dish and incubating plates undisturbed for 10 days to allow for colony formation. Cells were fixed and Giemsa stained, and colonies of 25 or more cells were counted as positive. Seeding efficiency was determined by seeding 2 × 106 cells/100-mm dish and counting the number of adherent cells after 6 h. To examine anchorage-independent growth, single cells (1 × 103 cells/60-mm dish) were dispersed in 0.36% Nobel agar in DMEM-20% fetal bovine serum and layered in a 0.6% agar base. Media was added every 3 days, and colonies were counted after 5 weeks.

DNA Constructs.

Full-length Cables was obtained from a mouse neonatal brain cDNA library and a human fetal brain cDNA library (Clontech) and sequenced. For production of bacterial fusion proteins, the Cables cDNA was subcloned in-frame into pGEX-4T-2 (Pharmacia) at the BamHI site. The cdk2, Wee1, and cdk2Y15F constructs were a gift of Li-Huei Tsai (Harvard Medical School, Boston, MA).

Protein Analysis.

Cell lysate was produced in E1A lysis buffer [50 mm Tris-HCl (pH 7.5), 250 mm NaCl, 5 mm EDTA (pH 8.0), 0.1% NP40, 5 mm DTT, 10 mm NaF, 1 mm phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml Na3VO4]. Proteins were analyzed by direct Western blotting (50 μg/lane) or blotting after immunoprecipitation. Cell extracts were immunoprecipitated with anti-cdk2 (pAb M2; Santa Cruz Biotechnology) and anti-Cables (pAb 64; Ref. 10). Immunoprecipitates were collected by binding to protein A-Sepharose. Western blots were probed with anti-cdk2 (mAb D12; Santa Cruz Biotechnology), anti-cdc2 (pAb H-297; Santa Cruz Biotechnology), anti-Cables (pAb 64), anti-phosphotyrosine (mAb 4G10; Upstate Biotechnology), anti-cyclin E (mAb HE12; gift of Ed Harlow, Harvard Medical School), anti-cyclin A (mAb C160; gift of E. Harlow), anti-p21 (mAb CP74; gift of E. Harlow), anti-p27 (mAb HBB6; gift of E. Harlow), and anti-phospho-Cdc2 (Tyr15; pAb; New England Biolabs).

Cell Fractionation.

Cells were resuspended in STM buffer [10 mm Tris-HCl (pH 8.0), 25 m sucrose, 10 mm MgCl2, 0.1 mm DTT, 10 mm NaF, 1 mm phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml Na3VO4] and homogenized. Lysates were centrifuged at 600 × g for 15 min. Supernatant was centrifuged at 100,000 × g, and the recentrifuged supernatant with 0.05% NP40 was saved as the cytoplasmic fraction. The pellet from the initial centrifugation was resuspended in STM buffer plus 0.5% NP40 and centrifuged at 600 × g for 15 min. The supernatant was saved as the membrane fraction, and the pellet was saved as the nuclear fraction, which is resuspended in radioimmunoprecipitation assay (RIPA) buffer.

In Vitro Kinase Assay.

Kinase assays were performed by washing immunoprecipitates three times with lysis buffer and once with kinase buffer [50 mm HEPES (pH 7.0), 10 mm MgCl2, and 1 mm DTT]. Cdk2 levels were equalized by Western blotting before the kinase assay. Subsequently, the beads were incubated with kinase buffer containing 0.5 μg of histone H1 and 5 μCi of [γ-32P]ATP in a final volume of 50 μl at room temperature for 30 min.

Cloning of Human Cables cDNA and FISH Mapping of Cables.

The human Cables cDNA was isolated from a human fetal brain cDNA library using the mouse Cables cDNA and was used as a probe for chromosomal in situ hybridization. Chromosomal slides were prepared from lymphocytes isolated from human blood and cultured in MEM supplemented with 10% FCS and phytohemagglutinin. The lymphocyte cultures were treated with bromodeoxyuridine (0.18mg/ml) to synchronize the cell population. Synchronized cells were released from the block and recultured for 6 h. Slides were made using hypotonic treatment, fixation, and air-drying. The human Cables cDNA was biotinylated with dATP using the Life Technologies, Inc. BioNick labeling kit and used for FISH. The procedure for FISH detection was performed according to standard procedures (13, 14). FISH signals and DAPI banding pattern were recorded separately by photographs, and the assignment of the FISH mapping data with chromosomal bands was achieved by superimposing FISH signals with DAPI-banded chromosomes.

Reverse Transcription-PCR of Cables in Human Tumors.

The Cables mRNA of human tumor xenografts was studied using a panel of eight normalized, first-strand cDNA preparations from human tumor xenografts (Human Tumor MTC Panel; Clontech) according to directions. Cables PCR primers were 5′-GCAGGAGGACTGTGGCCTTGAGGAG-3′, 5′-CTGTGTGCTGGGGCATGTGTGCTGT-3′, and 3′-GGCCCTTGGCTGTCCTCGGGGCCAGTG-5′, which amplified the very COOH-terminal 350-bp and 250-bp of the open reading frame. G3PDH PCR primers (included in kit) were used as a cDNA normalization control.

Immunohistochemistry.

GST fusion proteins were expressed in Escherichiacoli and purified with reduced glutathione beads. A specific antibody was raised against a GST-tagged human Cables in rabbits and affinity purified. The antibody recognized a protein of about Mr 70,000 in cell lysates on SDS-PAGE that comigrated with Cables synthesized in rabbit reticulocyte lysates in vitro and that was recognized by the antimouse Cables antisera described previously (10). Formalin-fixed, paraffin-embedded sections of normal and tumor tissue were stained with affinity-purified anti-Cables antisera at a 1:200 dilution using a microwave-enhanced avidin-biotin-staining method (15, 16). Negative control sections were immunostained under the same conditions substituting preabsorbed antisera and preimmune rabbit antisera for primary antibodies. The specificity of the affinity-purified antisera was demonstrated by (a) lack of staining with preabsorbed antisera and (b) strong staining of COS7 cells transfected with Cables with little to no staining of nontransfected cells, which contained little Cables by Western blot.

LOH.

Archival DNA was extracted from formalin-fixed specimens by using standard methods (17). Briefly, sections 4-μm thick were mounted on slides, and non-neoplastic tissue was removed with a clean blade by superimposing the unstained section with the corresponding stained section. The tissue was deparaffinized, rehydrated through xylene and alcohol, and scraped into a microfuge tube, and DNA was extracted with phenol and chloroform mixture. DNA was also extracted from paired normal tissue in all of the cases. LOH in the region of the Cables gene on chromosome 18q was examined by using the highly polymorphic markers D18S44 and D18s1107, which map to 18q11. Loss was scored as described previously (17).

Cables Is Predominantly a Nuclear Protein.

Cables cDNA encodes a protein of 568 amino acid residues with a predicted molecular size of Mr 63,000. Cables displays little sequence homology to other known proteins in the databases. It does, however, show weak homology to cyclin A and weaker homology to cyclin C over an ∼200 amino acid stretch in the COOH-terminal third of the protein that may be the cdk-interacting region. Cables contains two tyrosine-based sorting motifs (YXXLE), which have been implicated in axonal growth cone sorting and explain its presence in the axonal growth cone in mature neurons (10). There are three classic nuclear localization signals (18) composed of three basic amino acids and either histidine or proline at amino acid positions 48, 54, and 406. These nuclear signals exhibit both pattern 4 (4 residue) and pattern 7 (starting with P and followed within three residues by a basic segment). Another bipartite nuclear localization signal is located at position 264 (19). Moreover, a cytoplasmic/nuclear discrimination score based on amino acid composition predicts that Cables is a nuclear protein with a 94.1% reliability score (20).

To examine Cables localization in proliferating cells, Cables immunohistochemistry was performed on cell cultures and tissue sections using affinity-purified antihuman Cables antisera (Figs. 1 and 6). Preabsorbed antisera were used as a control to show specificity. Diffuse nuclear staining was seen in both COS7 cells transfected with Cables and human fibroblasts (Fig. 1). COS7 cells contain little to no endogenous Cables by Western blot (data not shown), and untransfected cells showed no Cables staining with the affinity-purified anti-Cables antisera (Fig. 1,A). Similarly, diffuse strong nuclear staining was seen in normal squamous and glandular epithelium using paraffin-embedded tissue sections and a microwave-enhanced staining method (Fig. 6). Cell fractionation studies confirmed the immunohistochemical staining results seen in proliferating cells and brain tissue (data not shown). In proliferating cells (HeLa and CEM) Cables was predominantly (90%) localized to the nuclear fraction with only small amount in the cytoplasmic and membrane fractions. In contrast, fractionation studies of adult mouse brain showed that 60–70% of Cables was present in the nuclear fraction and approximately 30% was in the membrane fraction (data not shown). Thus, Cables is likely to be predominantly a nuclear protein in proliferating cells but also is located at axonal growth cones where it appears to have a specific function in neurite outgrowth.

Cables Inhibits Cell Growth.

In proliferating cells, Cables is a nuclear protein that interacts with multiple cdks, including cdk2 and cdk3 (10, 11). To examine the role of Cables in cell growth, HeLa cell lines stably expressing Cables were generated. The pCIN4 vector was used to ensure that all of the antibiotic-resistant cells expressed the recombinant protein (12). Two stable cell lines were generated that showed an increased amount of Cables protein by both Western blot (Fig. 2,A) and immunostaining. Three control lines (vector alone) were also generated and compared with the Cables cell lines. Both Cables cell lines showed approximately 10-fold more Cables than endogenous levels found in the control HeLa cells, and immunohistochemical staining showed diffuse nuclear staining in both the control and Cables cell lines. Growth characteristics of the three control and two Cables cell lines were studied (Table 1). Each experiment was performed three times using all of the five cell lines, and the range of values for all of the experiments is given. The growth rate of the cell lines was determined by seeding 1 × 104 cells/well and counting the cells for up to 13 days. Increased numbers of control cells were present after 24 h and throughout the experiment (Fig. 3). The doubling time in the exponential growth phase for the control cells was calculated to be between 9 and 12 h (American Type Culture Collection reports similar doubling time for HeLa cells) and between 24 and 28 h for the cells overexpressing Cables. Differences in growth rates should be accompanied by differences in the rate of DNA synthesis. [3H]thymidine was added to identical numbers of cells, and uptake was measured after 30 min, 1 h, and 2 h. The Cables cell lines showed a 35% (range, 31–38) decrease in [3H]thymidine uptake compared with control cells at each time point. In addition, the Cables cell lines showed reduced colony formation (a measure of both seeding efficiency and growth) but equal seeding efficiency. Approximately 55% (range, 47–58) of the control cells formed colonies compared with 18% (range, 16–20) of the Cables cells, despite approximately 80–85% seeding efficiency for both groups. The cell cycle profile of the control and Cables cell lines was analyzed by flow cytometry and found to be similar. No consistent difference in the percentage of cells in the G1 phase, S phase, or G2-M phase of the cell cycle was noted between Cables and control cell lines. Similarly no change in the cell cycle profile was seen in Saos2 and HeLa cells transiently transfected with either Cables or a control vector. Terminal deoxynucleotidyl transferase-mediated nick end labeling assays, as well as careful observation of cell cultures, showed no increase in apoptosis or dead cells in the Cables versus control cell lines. These data suggest that changes in doubling time and growth rate are attributable to lengthening of multiple phases of the cell cycle rather than to a cell cycle block or cell death.

Cables Enhances Y15 Phosphorylation of Cdk2 by the Wee1 Kinase.

Cables associates with cdk2 (Fig. 2,B), and the COOH-terminal 200 amino acids show weak homology to cyclin A and cyclin C. To examine whether Cables growth inhibition was a result of binding to cdk2 and blocking cyclin binding, Cables-associated proteins were studied. Upon cotransfection, association of Cables and cdk2, cyclin A, cyclin E, and p21 could be readily demonstrated. No association was seen with cdc2 (cdk1) and p27. Equivalent levels of cdk2, immunoprecipitated directly or through Cables from HeLa cell lysates, showed approximately equivalent levels of cyclin A, cyclin E, and p21 (Fig. 2 B), suggesting that Cables is present in a multimolecular complex with cdk2, cyclin A/E, and p21.

The observation that Cables linked cdk5 and c-Abl led us to look at cdk2 tyrosine phosphorylation. Interestingly, cdk2 became tyrosine phosphorylated when Cables or Wee1 was overexpressed, and the level of tyrosine phosphorylation increased exponentially (approximately 30 times basal level) when both Cables and Wee1 were coexpressed (Fig. 2,C). Phosphorylation of the Y15 residue is a known regulatory event for the cdks (1). This residue is conserved in cdk2. A Y15 to phenylalanine cdk2 mutant (F15) was no longer phosphorylated by Cables and Wee1 (data not shown), demonstrating that Cables-enhanced tyrosine phosphorylation of cdk2 occurs on the highly conserved Y15 residue. To further verify that Cables enhanced cdk2Y15 phosphorylation, cdk2 tyrosine phosphorylation was examined in the stable Cables and control cell lines, using a phosphotyrosine cdc2 (Tyr15) antibody that is specific for Y15-phosphorylated cdc2 and cdk2 (Fig. 2,D). The Cables cell lines showed a 5-fold increase in Y15-phosphorylated cdk2 compared with the control cell lines, whereas the total amount of cdk2 immunoprecipitated from the cells was the same (Fig. 2,D). Both Cables-overexpressing cell lines showed similar increased levels of Y15-phosphorylated cdk2 (and decreased kinase activity). Y15-phosphorylation of cdc2 and cdk2 by the Wee1 family kinases is inhibitory and must be relieved by the cdc25 family of phosphatases for kinase activation. As expected, we found that cdk2 immunoprecipitated from the Cables cell lines had significantly decreased levels of histone H1 kinase activity than from the control cell lines (Fig. 2 E). Similarly, in cells transfected with cdk2, Cables, and Wee1, comparable levels of Cables-associated cdk2 had much less histone H1 kinase activity than cdk2 that was directly immunoprecipitated from the cell extracts.

Cables Is Not Expressed in One-half of Head and Neck Squamous Cell and Colon Cancers.

The human Cables cDNA was isolated from a human fetal brain cDNA library using the mouse Cables cDNA (GenBank accession no. AF348525) and was used as a probe for FISH analysis. Cables was found to lie on human chromosome 18, region q11.2-q12.1 (Fig. 4). A search of the genome databases revealed one BAC clone of human chromosome 18 (AC021244) that matched the human Cables sequence, confirming Cables location on human chromosome 18.

Chromosome 18q abnormalities are found in many human tumors, especially colorectal and pancreatic adenocarcinomas and head and neck squamous cell carcinomas (21, 22, 23). Expression of Cables in actively growing cells inhibited cell growth, so we questioned whether Cables might be affected in chromosome 18q loss. First-strand cDNA from eight human tumor xenografts (Clontech), including lung, colon, breast, prostate, pancreas, and ovary tumors, were screened with two sets of Cables-specific primers from the very COOH-terminal portion of human Cables and G3PDH primers as a control (Fig. 5). Seven tumors showed approximately equal levels of Cables cDNA; however, the eighth tumor (one of two colon cancers) showed significantly decreased Cables cDNA with both primer sets. G3PDH cDNA levels were approximately the same in all of the tumors. These results suggested that Cables may be involved in chromosome 18q loss.

To further study Cables expression in primary human tumors, a specific antibody was raised against a GST-tagged human Cables and affinity purified. Cables immunohistochemistry was performed on cell cultures and tissue sections using affinity-purified Cables antisera. Adherent cultured cell preparations of COS7 cells, which contain little endogenous Cables by Western blot, were transfected with a mammalian Cables expression construct. The untransfected COS7 cells showed little nuclear staining, whereas the transfected cells showed strong nuclear positivity (Fig. 1,A), suggesting the antisera was specific. Paraffin sections of tumors and normal tissue were then stained with the anti-Cables antisera. Preabsorbed antiserum was also used as a negative control. Colon, pancreas, and squamous tumors, which show chromosome 18q loss, as well as breast cancers, which do not often have chromosome 18q abnormalities, were studied (Table 2). All of the tumor slides contained adjacent normal tissue on the same slide to serve as an internal control. Normal colon, breast, pancreas, and squamous tissue (80 different specimens) showed strong nuclear staining of the squamous and glandular epithelium in all of the cases (Fig. 6). There was also nuclear staining of the scattered lymphocytes and plasma cells within the lamina propria. Similarly, all of the cases of breast and pancreas cancer showed strong nuclear staining of the tumor cells. In contrast, 11 of 20 colonic adenocarcinomas (55%) showed lack of nuclear Cables staining, whereas the other nine cases showed strong nuclear staining for Cables (Table 2 and Fig. 6). Similarly, 12 of 20 squamous cell carcinomas of the head and neck (60%) showed no Cables staining. All of the sections contained normal squamous and glandular mucosa, which served as a positive control, and the normal epithelial cells showed strong nuclear staining in all of the cases, including those in which the tumor was negative. Furthermore, positive staining of lymphocytes and plasma cells present adjacent to negative tumor cells served as an additional internal positive control. Faint cytoplasmic staining was seen in both normal and tumor cells, including many tumors without nuclear staining. The significance of this finding is not clear at this time. These results suggest that loss of Cables is involved in the pathogenesis of some human cancers with chromosome 18q loss. Furthermore, lack of Cables expression is specific to colon and squamous cancers because all cases of breast cancer, which do not often have 18q loss, and pancreas cancer, which has been found to have 18q loss in most cases, showed strong nuclear staining for Cables.

To examine the correlation between Cables expression and loss of chromosome 18q11-12, 20 cases of colon and squamous cancer and normal tissue away from the tumor were studied for LOH using highly polymorphic markers that map to 18q11 (D18S44 and D18S1107) in a “blind” manner. These cases were selected on the basis of having areas composed predominantly of tumor with less desmoplastic stroma and inflammation. LOH studies were performed three times on each sample. Sixteen cases were heterozygous with both markers, whereas two cases were homozygous with D18S44 and heterozygous with D18S1107 and two cases were homozygous with D18S1107 and heterozygous with D18S44. Ten cases (five colon and five squamous) with strong Cables staining showed no LOH on 18q11. In contrast, 8 of 10 cases of (four colon and four squamous) that lacked Cables staining showed definite LOH of 18q11 compared with the paired normal tissue (Fig. 7). Two cases were indeterminate; tumor DNA did not amplify well in one colon cancer and one squamous cancer showed inconsistent results between trials.

Cancer develops from the transformation of normal epithelium, to a dysplastic epithelial lesion, and ultimately to invasive carcinoma (24). In the colon, this progression is accompanied by a number of recently characterized genetic alterations (25). Inactivation of the adenomatous polyposis gene marks one of the earliest events in colorectal carcinoma followed by oncogenic K-ras mutations. Later events include inactivation of the tumor-suppressor gene p53 on chromosome 17p and LOH on the long arm of chromosome 18 (18q). Chromosome 18q abnormalities are also common in pancreas cancer and squamous cell cancer of the head and neck. FISH mapping of pancreatic cancer showed that all of the cases had lost at least one copy of chromosome 18q and that most breakpoints mapped to 18q11 (22). Similarly, loss or deletion of chromosome 18q is one of the most common chromosome abnormalities in squamous cell cancers, occurring in 55% to 65% of tumors (26). More sensitive studies with LOH showed that 75% of squamous cancers showed 18q loss at one or more loci and in up to 66% this involved 18q11 (23). Furthermore, 25% of the tumors showed loss of 18q11.1-q12.3 without distal loss of chromosome 18.

Three candidate tumor-suppressor genes, deleted in colon cancer (DCC), Smad4 (DPC4), and Smad2, have been identified in the distal portion of chromosome 18q. DCC was recently shown to be the netrin-1 receptor and bind directly to netrin-1 (27, 28). Furthermore, DCC is expressed in both normal colonic mucosa and both primary and metastatic colon cancer (29), and DCC-null mice do not develop tumors (30). The Smad proteins mediate transforming growth factor-β effects and regulate genes involved in cell cycle control. Biallelic inactivation of Smad4 occurs in greater than 60% of pancreas tumors with few mutations identified in the Smad genes in squamous and colon cancers (31, 32, 33, 34, 35). Thus, it is likely that chromosome 18q harbors at least one more proximally located gene that is involved in human cancer. This is supported by the proximal loss of 18q11-12 without distal 18q loss in some tumors.

In this study, we cloned human Cables cDNA and mapped its genomic location to chromosome 18, region q11.2-12.1. Previous studies (21) of colon cancer specimens found that 38% had lost an entire chromosome 18, and approximately 65% that lost alleles of a subset of markers on chromosome 18q showed LOH in the region that included Cables. Similarly, up to 66% of squamous cancers showed loss of the region of chromosome 18 that included Cables (23). Thus, we generated antibodies against human Cables to assess Cables expression in tumors. We were able to detect nuclear Cables protein in routinely processed, formalin-fixed paraffin-embedded tissue sections using affinity-purified polyclonal antisera. Nuclear staining for Cables was detected in normal squamous and glandular epithelium from the breast, pancreas, colon, and head and neck. However, 55% of colon cancers and 60% of squamous cancers showed loss of nuclear Cables staining. In each of these cases, normal mucosa was present on the same section and showed strong nuclear staining. In addition, inflammatory cells, which also show nuclear Cables staining, were present in and around the negative carcinoma cells. To examine the specificity of the Cables loss in these cancers, we also studied breast cancers, which do not commonly show chromosome 18q abnormalities (36), for loss of Cables. All of the cases showed strong staining without Cables loss. Furthermore, lack of Cables staining was associated with LOH of chromosome 18q11-12 in colon and squamous cancers. Eight of 10 cases with loss of Cables expression showed LOH of chromosome 18q11-12. No LOH was detected in 10 cases of colon and squamous cancer with Cables staining. Only cases that had areas of predominant tumor with less desmoplastic stroma and inflammation were studied for LOH. Even so, in one case, tumor DNA did not amplify and one case was inconclusive, possibly because of admixed non-neoplastic cells. In addition, pancreatic carcinomas are known to lose one copy of chromosome 18q in the majority of cases (32) and all showed strong Cables staining. Therefore, loss of Cables expression is not likely to be a general consequence of 18q LOH and appears to be specific to colon and squamous cancers.

Cables was discovered recently (10, 11) as a cdk-interacting protein. It acts as a link or cable between the cdks and nonreceptor tyrosine kinases, such as c-Abl (10). To examine whether loss of Cables could affect cell growth, we created stable cell lines with mildly elevated ectopic expression of Cables. Growth of these cells was significantly inhibited compared to the vector control cells, with a decreased doubling time and colony formation rate but identical plating efficiency. The cell cycle profile and degree of apoptosis was similar in both Cables and control cell lines. Inhibition of cell growth by Cables may be related to decreased cdk2 activity. Cdk2 is involved at multiple points in the cell cycle, including the G1-S transition, initiation, and maintenance of DNA replication (S-phase) and entry and progression through mitosis (37, 38), which could explain a normal cell cycle profile despite slower growth. Cables exists in a multiprotein complex with at least cdk2, cyclin A or cyclin E, and p21, so it does not inhibit cdk2 activity by displacing the cyclin molecule. Recently, an association between PKCeta and cyclin E/cdk2/p21 complex was shown to inhibit cdk2 activity by causing dephosphorylation of cdk2 Thr160 during keratinocyte differentiation (39). Cables appears to inhibit cdk2 activity by enhancing Wee1 tyrosine 15 phosphorylation of cdk2, which is an inhibitory phosphorylation for the kinase. In transfected cells, Cables and Wee1 act together to dramatically increase cdk2 tyrosine phosphorylation. Despite the functional interaction between Cables and Wee1, we could not demonstrate a stable interaction between these two proteins in transfected cell or endogenous cell lysates, suggesting that the interaction is more functional than physical. Cables expressing cell lines showed significantly increased levels of Y15 phosphorylated cdk2 and decreased kinase activity.

Loss of Cables could provide a growth advantage to neoplastic cells by allowing faster progression through the cell cycle. Lack of nuclear Cables expression in primary human cancers is frequent in colon and squamous cancers and is likely to be related to 18q LOH plus gene mutations in the remaining gene, leading to an unstable or truncated protein, or hypermethylation of CpG islands in its promoter. The latter hypothesis is attractive, because widespread genomic hypomethylation, which occurs in the setting of localized hypermethylation, has been reported in these cancers (40, 41). Hypermethylation of CpG islands within promoter sequences of specific tumor-suppressor genes can have a potent silencing effect and has been reported for INK4p16 gene in 28–55% of colon cancers (42, 43).

When taken together, our data suggest that Cables is involved in both cell growth and differentiation and may act as a crossover signal between the two processes (Fig. 8). Cables serves as an adapter molecule that facilitates both cdk5 and cdk2 Y15 phosphorylation with different consequences. Cdk5Y15 phosphorylation by Cables and c-Abl increases kinase activity and contributes to its role in neuronal migration and neurite outgrowth. In primary neuronal cultures, expression of antisense Cables caused neurite shortening similar to that seen for dominant-negative cdk5. In contrast, cdk2Y15 phosphorylation by Cables and Wee1 is an inhibitory phosphorylation that inhibits cell growth. Furthermore, strong expression of Cables in both normal squamous and glandular epithelium suggests a role for Cables in epithelial development. Loss of Cables may impair growth inhibition and cell differentiation, which can lead to or enhance uncontrolled cell growth and eventually cancer. Lack of expression of Cables in up to 60% of colon and squamous cancers, along with its chromosomal location and role in growth control, suggests it has a role in the pathogenesis of these tumors. However, mutational analysis of the Cables gene and mutant Cables mice will be important to determine what role, if any, Cables plays in human cancer.

Fig. 1.

Cables is present in the nucleus of proliferating cells. COS7 cells transfected with Cables (A) and human (Wi38) fibroblasts (B) were stained with affinity-purified anti-Cables antibody. Cables is predominantly detected in the nucleus with less membrane and cytoplasmic staining.

Fig. 1.

Cables is present in the nucleus of proliferating cells. COS7 cells transfected with Cables (A) and human (Wi38) fibroblasts (B) were stained with affinity-purified anti-Cables antibody. Cables is predominantly detected in the nucleus with less membrane and cytoplasmic staining.

Close modal
Fig. 2.

Cables enhances cdk2 Y15 phosphorylation by Wee1. A, a Western blot of HeLa Cables and control cell lysates (50 μg) was probed with anti-Cables antisera. B, immunoprecipitations with antibodies to cdk2 and Cables from HeLa cell lysates followed by Western blot for cyclin A, cdk2, and p21. Cyclin E blot (data not shown) also showed equivalent levels in both cdk2 and Cables lanes. C, COS7 cells were transfected with cdk2 and Cables, Wee1, or both, and cell lysates were subjected to cdk2 immunoprecipitation followed by an antiphosphotyrosine Western blot. In the lower panel, the lysates were reprobed with antibody to cdk2. Numbers indicate fold induction over the basal level. D, immunoprecipitations of cdk2 from Cables and control cell lysates were probed with an antibody specific for Y15-phosphorylated cdk2 and cdk2. E, Cables and control cell lysates were subjected to cdk2 immunoprecipitation followed by in vitro [γ-32P]ATP kinase assays using histone H1 as a substrate. Equivalent levels of cdk2 were immunoprecipitated (D).

Fig. 2.

Cables enhances cdk2 Y15 phosphorylation by Wee1. A, a Western blot of HeLa Cables and control cell lysates (50 μg) was probed with anti-Cables antisera. B, immunoprecipitations with antibodies to cdk2 and Cables from HeLa cell lysates followed by Western blot for cyclin A, cdk2, and p21. Cyclin E blot (data not shown) also showed equivalent levels in both cdk2 and Cables lanes. C, COS7 cells were transfected with cdk2 and Cables, Wee1, or both, and cell lysates were subjected to cdk2 immunoprecipitation followed by an antiphosphotyrosine Western blot. In the lower panel, the lysates were reprobed with antibody to cdk2. Numbers indicate fold induction over the basal level. D, immunoprecipitations of cdk2 from Cables and control cell lysates were probed with an antibody specific for Y15-phosphorylated cdk2 and cdk2. E, Cables and control cell lysates were subjected to cdk2 immunoprecipitation followed by in vitro [γ-32P]ATP kinase assays using histone H1 as a substrate. Equivalent levels of cdk2 were immunoprecipitated (D).

Close modal
Fig. 3.

Growth of stable HeLa cell lines with Cables or control vector. The doubling time of Cables cell lines (at the exponential growth phase) was twice that of the control cell lines.

Fig. 3.

Growth of stable HeLa cell lines with Cables or control vector. The doubling time of Cables cell lines (at the exponential growth phase) was twice that of the control cell lines.

Close modal
Fig. 4.

FISH mapping of Cables gene. Top left panel, the FISH signals on human chromosome 18; top right panel, the same mitotic figure stained with DAPI to identify human chromosome 18. Bottom panel illustrates the detailed mapping results from 10 high quality photographs.

Fig. 4.

FISH mapping of Cables gene. Top left panel, the FISH signals on human chromosome 18; top right panel, the same mitotic figure stained with DAPI to identify human chromosome 18. Bottom panel illustrates the detailed mapping results from 10 high quality photographs.

Close modal
Fig. 5.

Comparison of Cables and G3PDH mRNA expression in human tumors. cDNA from multiple human tumors (Clontech) was used as a PCR template with Cables (A and B) and G3PDH (C) primers. Cables cDNA (Lane 1) was used as a positive control followed by a negative control (Lane 2). Cables mRNA was decreased in one of two colon cancers but not in the other tumors (breast, lung, and pancreas shown).

Fig. 5.

Comparison of Cables and G3PDH mRNA expression in human tumors. cDNA from multiple human tumors (Clontech) was used as a PCR template with Cables (A and B) and G3PDH (C) primers. Cables cDNA (Lane 1) was used as a positive control followed by a negative control (Lane 2). Cables mRNA was decreased in one of two colon cancers but not in the other tumors (breast, lung, and pancreas shown).

Close modal
Fig. 6.

Cables is present in normal tissue but is lost in some colon and squamous cancers. Immunohistochemical staining of normal tissue and tumors with affinity-purified Cables antisera. Normal breast, colon, pancreas, and squamous epithelium show strong nuclear Cables staining. Invasive breast and pancreas carcinomas also show strong nuclear staining. In contrast, nuclear Cables is not detected in approximately one-half of invasive colon and squamous cell carcinomas.

Fig. 6.

Cables is present in normal tissue but is lost in some colon and squamous cancers. Immunohistochemical staining of normal tissue and tumors with affinity-purified Cables antisera. Normal breast, colon, pancreas, and squamous epithelium show strong nuclear Cables staining. Invasive breast and pancreas carcinomas also show strong nuclear staining. In contrast, nuclear Cables is not detected in approximately one-half of invasive colon and squamous cell carcinomas.

Close modal
Fig. 7.

Representative examples of tumor-specific allelic loss on chromosome 18q. Both cases are informative. Colon cancers with LOH (A) and without LOH (B) of 18q11-12, using microsatellite marker D18S44. T, tumor sample; N, paired normal control. Arrow, the lost allele in the tumor sample.

Fig. 7.

Representative examples of tumor-specific allelic loss on chromosome 18q. Both cases are informative. Colon cancers with LOH (A) and without LOH (B) of 18q11-12, using microsatellite marker D18S44. T, tumor sample; N, paired normal control. Arrow, the lost allele in the tumor sample.

Close modal
Fig. 8.

Model illustrating possible role of Cables in growth inhibition and development. In response to growth inhibitory and developmental signals, Cables enhances cdk2Y15 phosphorylation by Wee1, an inhibitory phosphorylation, which leads to decreased cdk2 activity and growth inhibition. Cables also enhances cdk5Y15 phosphorylation by activated c-Abl, which leads to increased cdk5 activity and is critical for proper neuronal development. Cables expression in normal epithelium suggests that Cables may also play a role in epithelial cell development. Loss of Cables in some colon and squamous cancers suggests that Cables may be involved in the pathogenesis of these tumors.

Fig. 8.

Model illustrating possible role of Cables in growth inhibition and development. In response to growth inhibitory and developmental signals, Cables enhances cdk2Y15 phosphorylation by Wee1, an inhibitory phosphorylation, which leads to decreased cdk2 activity and growth inhibition. Cables also enhances cdk5Y15 phosphorylation by activated c-Abl, which leads to increased cdk5 activity and is critical for proper neuronal development. Cables expression in normal epithelium suggests that Cables may also play a role in epithelial cell development. Loss of Cables in some colon and squamous cancers suggests that Cables may be involved in the pathogenesis of these tumors.

Close modal

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.

2

The abbreviations used are: cdk, cyclin-dependent kinase; FISH, fluorescent in situ hybridization; DAPI, 4′,6-diamidino-2-phenylindole; GST, glutathione S-transferase; LOH, loss of heterozygosity; DCC, deleted in colon cancer.

Table 1

Growth characteristics of cables and control cell lines

ControlCables
Doubling time 9–12 h 24–28 h 
Thymidine uptake 100% 62–68% 
Colony formation 47–58% 16–20% 
Seeding efficiency 81–85% 82–86% 
Soft agar growth 0.4–0.6 
ControlCables
Doubling time 9–12 h 24–28 h 
Thymidine uptake 100% 62–68% 
Colony formation 47–58% 16–20% 
Seeding efficiency 81–85% 82–86% 
Soft agar growth 0.4–0.6 
Table 2

Cables expression in human tumors

Nuclear Cables staining
PositiveNegative
Normal breast (20) 20 
Breast cancer (20) 20 
Normal colon (20) 20 
Colon cancer (20) 11 
Normal pancreas (20) 20 
Pancreas cancer (20) 20 
Normal squamous mucosa (20) 20 
Squamous cell carcinoma (20) 12 
Nuclear Cables staining
PositiveNegative
Normal breast (20) 20 
Breast cancer (20) 20 
Normal colon (20) 20 
Colon cancer (20) 11 
Normal pancreas (20) 20 
Pancreas cancer (20) 20 
Normal squamous mucosa (20) 20 
Squamous cell carcinoma (20) 12 
1
Morgan D. O. Principles of cdk regulation.
Nature (Lond.)
,
374
:
131
-134,  
1995
.
2
Wang J., Chenivesse X., Henglein B., Brechot C. Hepatitis B-virus integration in a cyclin A gene in a hepatocellular carcinoma.
Nature (Lond.)
,
343
:
555
-557,  
1990
.
3
Motokura T., Bloom T., Kim H. G., Juppner H., Ruderman J. V., Kronenberg H. M., Arnold A. A novel cyclin encoded by a bcl-linked candidate oncogene.
Nature (Lond.)
,
350
:
512
-515,  
1991
.
4
Rosenberg C. L., Wong E., Petty E. M., Bale A. E., Tsujimoto Y., Harris N. L., Arnold A. PRAD1, a candidate BCL1 oncogene: mapping and expression in centrocytic lymphoma.
Proc. Natl. Acad. Sci. USA
,
88
:
9638
-9642,  
1991
.
5
Wang T. C., Cardiff R. D., Zukerberg L., Lees E., Arnold A., Schmidt E. V. Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice.
Nature (Lond.)
,
369
:
669
-671,  
1994
.
6
Bonetta L. Open questions on p16.
Nature (Lond.)
,
370
:
180
1994
.
7
El-Deiry W. S., Tokino T., Velculescu V. E., Levy D. B., Parsons R., Trent J. M., Lin D., Mercer W. E., Kinzler K. W., Vogelstein B. WAF1, a potential mediator or p53 tumor suppression.
Cell
,
75
:
817
-825,  
1993
.
8
Xiong Y., Hannon G. J., Zhang H., Casso D., Kobayashi R., Beach D. p21 is a universal inhibitor of cyclin kinases.
Nature (Lond.)
,
366
:
701
-704,  
1993
.
9
Parsons R. Phosphatase and tumorigenesis.
Curr. Opin. Oncol.
,
10
:
88
-91,  
1998
.
10
Zukerberg L. R., Patrick G. N., Nikolic M., Humbert S., Wu C. L., Lanier L. M., Gertler F. B., Vidal M., Van Etten R. A., Tsai L. H. Cables links cdk5 and c-Abl and facilitates cdk5 tyrosine phosphorylation, kinase upregulation, and neurite outgrowth.
Neuron
,
26
:
633
-646,  
2000
.
11
Matsuoka M., Matsuura Y., Semba K., Nishimoto I. Molecular cloning of a cyclin-like protein associated with cyclin dependent kinase 3 (cdk3) in vivo.
Biochem. Biophys. Res. Commun.
,
273
:
442
-447,  
2000
.
12
Rees S., Coote J., Stables J., Goodson S., Harris S., Lee M. G. Bicistronic vector for the creation of stable mammalian cell lines that predisposes all antibiotic-resistant cells to express recombinant protein.
Biotechniques
,
20
:
102
-110,  
1996
.
13
Heng H., Squire J., Tsui L. C. High resolution mapping of mammalian genes by in situ hybridization to free chromatin.
Proc. Natl. Acad. Sci. USA
,
89
:
9509
-9513,  
1992
.
14
Heng H., Tsui L. C. Modes of DAPI banding and simultaneous in situ hybridization.
Chromosoma (Berl.)
,
102
:
325
-332,  
1993
.
15
Hsu S. M., Raine L., Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures.
J. Histochem. Cytochem.
,
29
:
577
-580,  
1981
.
16
Yang W. I., Zukerberg L. R., Motokura T., Arnold A., Harris N. L. Cyclin D1 (bcl1, Prad1) protein expression in low-grade B-cell lymphoma and reactive hyperplasia.
Am. J. Pathol.
,
145
:
86
-96,  
1994
.
17
Young J., Leggett B., Ward M., Thomas L., Buttenshaw R., Searle J., Chenevix-Trench G. Frequent loss of heterozygosity on chromosome 14 occurs in advanced colorectal carcinomas.
Oncogene
,
8
:
671
-675,  
1993
.
18
Hicks G. R., Raikhel N. V. Protein import into the nucleus: an integrated view.
Annu. Rev. Cell Dev. Biol.
,
11
:
155
-188,  
1995
.
19
Robbins J., Dilworth S. M., Laskey R. A., Dingwall C. Two interdependent basic domains I nucleoplasmin nuclear targeting sequence: identification of a class of bipartite nuclear targeting sequence.
Cell
,
64
:
615
-623,  
1991
.
20
Reinhardt A., Hubbard T. Using neural networks for prediction of the subcellular location of proteins.
Nucleic Acids Res.
,
26
:
2230
-2236,  
1998
.
21
Thiagalingam S., Lengauer C., Leach F. S., Schutte M., Hahn S., Overhauser J., Willson J., Markowitz S., Hamilton S. R., Kern S. E., Kinzler K. W., Vogelstein B. Evaluation of candidate tumour suppressor genes on chromosome 18 in colorectal cancers.
Nat. Genet.
,
13
:
343
-346,  
1996
.
22
Hoglund M., Gornova L., Jonson T., Dawiskiba S., Andren-Sandberg A., Stenman G., Johansson B. Cytogenetic and FISH analysis of pancreatic carcinoma reveal breaks in 18q11 with consistent loss of 18q12-qter and frequent gain of 18p.
Br. J. Cancer
,
77
:
1893
-1899,  
1998
.
23
Jones J. W., Raval J. R., Beals T. F., Worsham M. J., van Dyke D. L., Esclamado R. M., Wolf G. T., Bradford C. R., Miller T., Carey T. F. Frequent loss of heterozygosity on chromosome arm 18q in squamous cell carcinomas.
Arch. Otolaryngol. Head Neck Surg.
,
123
:
610
-614,  
1997
.
24
Fearon E. R., Vogelstein B. A genetic model for colorectal tumorigenesis.
Cell
,
61
:
759
-767,  
1990
.
25
Chung D. The genetic basis of colorectal cancer: insights into critical pathways of tumorigenesis.
Gastroenterology
,
119
:
854
-865,  
2000
.
26
Van Dyke D. L., Worsham M. J., Benninger M. S., Krause C. J., Baker S. R., Wolf G. T., Drumheller T., Tilley B. C., Carey T. E. Recurrent cytogenetic abnormalities in squamous cell carcinomas of the head and neck region.
Genes Chromosomes Cancer
,
9
:
192
-206,  
1994
.
27
Keino-Masu K., Masu M., Hinck L., Leonardo E. D., Chan S. S., Culotti J. G., Tessier-Lavigne M. Deleted in colorectal cancer (DCC) encodes a netrin receptor.
Cell
,
87
:
175
-185,  
1996
.
28
Stein E., Zou Y., Poo M., Tessier-Lavigne M. Binding of DCC by netrin-1 to mediate axon guidance independent of adenosine A2B receptor activation.
Science (Wash. DC)
,
291
:
1976
-1982,  
2001
.
29
Gotley D. C., Reeder J. A., Fawcett J., Walsh M. D., Bates P., Simmons D. L., Antalis T. M. The deleted in colon cancer (DCC) gene is consistently expressed in colorectal cancers and metastases.
Oncogene
,
13
:
787
-795,  
1996
.
30
Fazeli A., Dickinson S. L., Hermiston M. L., Tighe R. V., Steen R. G., Small C. G., Stoeckli E. T., Keino-Masu K., Masu M., Rayburn H., Simmons J., Bronson R. T., Gordon J. I., Tessier-Lavigne M., Weinberg R. A. Phenotype of mice lacking functional deleted in colorectal cancer (DCC) gene.
Nature (Lond.)
,
386
:
796
-804,  
1997
.
31
Eppert K., Scherer S. W., Ozcelik H., Pirone R., Hoodless P., Kim H., Tsui L. C., Bapat B., Gallinger S., Andrulis I. L., Thompsen G. H., Wrana J. L., Attisano L. MADR2 maps to 18q21 and encodes a TGF-β-regulated MAD-related protein that is functionally mutated in colorectal carcinoma.
Cell
,
86
:
543
-552,  
1996
.
32
Hahn S. A., Hoque A., Moskaluk C. A., da Costa L., Schutte M., Rozenblum E., Seymour A. B., Weinstein C. L., Yeo C. J., Hruban R. H., Kern S. E. Homozygous deletion map at 18Q21.1 in pancreatic cancer.
Cancer Res.
,
56
:
490
-494,  
1996
.
33
Schutte M., Hruban R. H., Hedrick L., Cho K. R., Nadasdy G., Weinstein C., Bova G. S., Issacs W. B., Cairns P., Nawroz H., Sidransky D., Casero R., Meltzer P. S., Hahn S. A., Kern S. E. DPC4 gene in various tumor types.
Cancer Res.
,
56
:
2527
-2530,  
1996
.
34
Takagi Y., Kohmura H., Futamura M., Kida H., Tanemura H., Shimokawa K., Saji S. Somatic alterations of the DPC4 gene in human colorectal cancers in vivo.
Gastroenterology
,
111
:
1369
-1372,  
1996
.
35
Takagi Y., Kohmura H., Futamura M., Aoki S., Ymaguchi K., Kida H., Tanemura H., Shimokawa K., Saji S. Somatic alterations of the SMAD-2 gene in human colorectal cancers.
Br. J. Cancer
,
78
:
1152
-1155,  
1998
.
36
Schenk M., Leib-Mosch C., Schenck I. U., Jaenicke M., Indraccolo S., Saeger H. D., Dallenbach-Hellweg G., Hehlmann R. Lower frequency of allele loss on chromosome 18q in human breast cancer than in colorectal tumors.
J. Mol. Med.
,
74
:
155
-159,  
1996
.
37
Hu B., Mitra J., van den Heuvel S., Enders G. H. S and G2 phase roles for Cdk2 revealed by inducible expression of a dominant-negative mutant in human cells.
Mol. Cell. Biol.
,
21
:
2755
-2766,  
2001
.
38
Furuno N., den Elzen N., Pines J. Human cyclin A is required for mitosis until mid-prophase.
J. Cell Biol.
,
147
:
295
-306,  
1999
.
39
Kashiwagi M., Ohba M., Watanabe H., Ishino K., Kasahara K., Sanai Y., Taya Y., Kuroki T. PKCeta associates with cyclin E/cdk2/p21 complex, phosphorylates p21 and inhibits cdk2 kinase in keratinocytes.
Oncogene
,
19
:
6334
-6341,  
2000
.
40
Goelz S. E., Vogelstein B., Hamilton S. R., Feinberg A. P. Hypomethylation of DNA from benign and malignant human colon neoplasms.
Science (Wash. DC)
,
228
:
187
-190,  
1985
.
41
Jones P. A., Laird P. W. Cancer epigenetics comes of age.
Nat. Genet.
,
21
:
163
-167,  
1999
.
42
Guan R. J., Fu Y., Holt P. R., Pardee A. B. Association of K-ras mutations with p16 methylation in human colon cancer.
Gastroenterology
,
116
:
1063
-1071,  
1999
.
43
Liang J. T., Chang K. J., Chen J. C., Lee C. C., Cheng Y. M., Hsu H. C., Wu M. S., Wang S. M., Lin J. T., Cheng A. L. Hypermethylation of the p16 gene in sporadic T3N0M0 stage colorectal cancers: association with DNA replication error and shorter survival.
Oncology (Basel)
,
57
:
149
-156,  
1999
.