Expression of Phosphorylated Ser⁷⁰ of Bcl-2 Correlates with Malignancy in Human Colorectal Neoplasms

Bcl-2 has been identified as an important suppressor of apoptosis. Expression of this molecule is believed to enhance tumor progression by countering apoptosis triggers, such as hypoxia, viral infections, cytokine deprivation, and the proapoptotic action of Bcl-2 family members (1–4). Recent studies have used in vitro assays to detect Bcl-2 phosphorylation activated by agents, including cytokines (e.g., interleukin-3), chemotherapy drugs [e.g., the microtubule-active drug paclitaxel (Taxol)], and serine/threonine kinases Cdc2, extracellular signal-regulated kinase, c-Jun NH₂-terminal kinase, and Cdk6/v-cyclin complex; all of these agents are activated during the mitotic phase (5–9). These data collectively revealed intriguing functional aspects of Bcl-2 in relation to apoptosis.

It is known that Bcl-2 undergoes multisite phosphorylation centered at the loop region that includes Thr⁵⁶, Thr⁶⁹, Ser⁷⁰, Thr⁷⁴, and Ser⁸⁷ (8, 10–12). Phosphorylation of Ser⁷⁰, Ser⁸⁷, and Thr⁶⁹ in this loop was observed by in vitro assay using cell lines treated with the microtubule-active agent Taxol. c-Jun NH₂-terminal kinase has been shown to phosphorylate Bcl-2 at Thr⁵⁶, Ser⁷⁰, Thr⁷⁴, and Ser⁸⁷ in bcl-2-transfected cell lines (8, 10–12). Interestingly, among these loop residues, Ser⁷⁰ is the most frequently observed, pointing toward its functional significance. In addition to these molecular studies, several investigations suggest that the phosphorylated form of Bcl-2 may contribute to advancement of human neoplasms, to include colorectal cancer (13, 14). Despite this accumulating data, the mechanistic role of Bcl-2 phosphorylation in the tumor biology remains unclear (15). Results from Taxol-treated cell lines imply that phosphorylation inhibits Bcl-2 antiapoptotic function, although interleukin-3-induced phosphorylation of Bcl-2 increased this function in interleukin-3-dependent cell lines (7, 10). Other studies propose that Bcl-2 phosphorylation may be involved in mitotic or meiotic events that are independent of apoptosis (15–19). In view of these sometimes conflicting results, an in vivo study of phospho-Bcl-2 expression would be valuable to the improved understanding of Bcl-2 in tumor progression. Here, we investigated the in vivo expression of phospho-Bcl-2, focusing on the phosphorylated Ser⁷⁰ (pSer⁷⁰) residue in colorectal neoplasms, including tubular adenomas and adenocarcinomas, to elucidate the molecular mechanism of Bcl-2 in the context of the actual disease setting.

Tissues and cell lines. Paraffin-embedded tissues of colorectal tumors were selected based on the diagnostic records of the Department of Clinical Pathology at Okayama University Hospital and Division of Pathology of Iwate Medical University (Morioka, Japan). In total, 209 colorectal tumors were examined, which included 75 cases of tubular adenomas, 119 cases of adenocarcinomas, and 15 cases of cancer in adenomas; all tissues were surgically excised between 1993 and 2002, representing clinically untreated primary tumors. Histologic typing of the adenocarcinomas was done according to the WHO International Histological Classifications (20) and the corresponding clinical data were retrospectively complied. Other clinicopathologic diagnosis for cases of the adenocarcinomas was based on tumor-node-metastasis system (21). A human colorectal adenocarcinoma cell line, SW620 (American Type Culture Collection, Manassas, VA), was cultured in RPMI 1640 with 10% fetal bovine serum (FBS). Cells were seeded on the tissue culture slides and treated with 200 nmol/L Taxol (Bristol-Myers Squibb K.K., Tokyo, Japan) for 8 hours. Samples were sequentially fixed with 4% paraformaldehyde in PBS for immunofluorescence. Baby hamster kidney (BHK) cells used for DNA transfection assay were also cultured in DMEM supplemented with 10% FBS.

All the patient-derived materials in the present study were used following the guideline of Japanese Society of Pathology.

Antibodies. Rabbit serum against human pSer⁷⁰ Bcl-2 was raised by 3-fold immunization with a synthetic phosphorylated oligopeptide encoding the 15 amino acids (RDPVARTpSPLQTPAA) at the loop region of human Bcl-2. After ELISA titer measurement, serum was collected and purified through an IgG affinity column.

For the preabsorption assay, phosphopeptide (10 μg; sc-16647p; Santa Cruz Biotechnology, Santa Cruz, CA) was incubated with 1 μg anti-pSer⁷⁰ Bcl-2 polyclonal antibody for 24 hours at 4°C, and the mixture was applied as primary antibody solution at the immunohistochemistry step to establish the specificity of the anti-pSer⁷⁰ Bcl-2 antibody. A mixture of 1 μg anti-pSer⁷⁰ Bcl-2 antibody and 10 μg recombinant truncated Bcl-2 protein (sc-4096; Santa Cruz Biotechnology) was also used in parallel to confirm the specific reactivity of the pSer⁷⁰ antibody. The following antibodies were used for immunohistochemistry: anti-Bcl-2 polyclonal antibody (sc-783; Santa Cruz Biotechnology) and anti-Ki-67 monoclonal antibody (mAb; MIB-1; DakoCytomation A/S, Glostrup, Denmark).

Double-labeled immunofluorescence. Immunofluorescence stainings used anti-pSer⁷⁰ Bcl-2 polyclonal antibody along with anti–heat shock protein 60 (Hsp60) mAb (BD Biosciences, San Jose, CA), anti-Ki-67 mAb, or anti-Myc-tag mAb (9E10; American Type Culture Collection) on paraffin-embedded tissue sections or cytosmeared sections of human bcl-2-transfected BHK cells or SW620 cells. Primary antibodies were employed in combination with mAb (anti-Hsp60 and anti-Ki-67) and a rabbit polyclonal antibody (anti-pSer⁷⁰ Bcl-2). For secondary antibodies, Cy3-labeled goat anti-mouse IgG was used to label anti-Hsp60 and anti-Myc-tag mAb, and FITC-labeled goat anti-rabbit IgG was employed for anti-pSer⁷⁰ Bcl-2 polyclonal antibody. For double immunofluorescence detecting Myc-tagged Bcl-2 and Ki-67 antigen, specimens were first stained with anti-Ki-67 mAb (second antibody: FITC-goat anti-mouse IgG), blocked with normal mouse serum (DakoCytomation), and sequentially moved to staining with Cy3-conjugated 9E10 (Sigma-Aldrich, St. Louis, MO). Immunofluorescence was detected with a confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany).

Immunohistochemistry. Enzymatic immunohistochemistry was done by standard indirect staining. For heat-induced antigen retrieval to detect endogenous pSer⁷⁰ Bcl-2 or Ki-67, deparaffinized slides were boiled in 10 mmol/L citrate buffer (pH 6.4) in a microwave oven for 10 minutes at 600 W. Endogenous peroxidase activity was blocked by incubation in 3% H₂O₂ for 20 minutes followed by incubation with goat serum for 10 minutes at room temperature. Samples were incubated with primary antibody overnight at 4°C, and signals were visualized using the Dako EnVision kit at the secondary immunostaining step (EnVision+, DakoCytomation) according to the manufacturer's instruction. Counterstaining was in Meyer's hematoxylin.

Expression of phosphorylated Ser⁷⁰ Bcl-2 in colorectal tumors. For colorectal adenocarcinomas, the cells expressing pSer⁷⁰ Bcl-2 were distributed among three major categories. The group in which the pSer⁷⁰-positive segment covered >50% of the cancerous area was classified as “diffuse positive,” whereas the group characterized by <20% coverage was termed “very localized.” An intermediate group was termed “localized.” These classifications are shown in Table 1. Expression of total Bcl-2 was evaluated as either positive or negative, because it was observed diffusely and uniformly with almost the same intensity in the positive cases. Results from immunostainings of both pSer⁷⁰ Bcl-2 and total Bcl-2 were independently evaluated by five pathologists before being categorized.

MIB-1 index. The MIB-1 index, a measure of Ki-67 staining, was assessed as the percentage of Ki-67-positive tumor cell nuclei among all tumor cell nuclei. More than 1,000 nuclei were evaluated from each section.

Apoptotic cell assay. Triplicate samples of 1 × 10⁵ SW620 cells transfected by 1 μg phosphomimetic mutants of Ser⁷⁰-Bcl-2 were assayed for the number of apoptotic cells based on morphologic shrinkage, nuclear fragmentation, or chromatin condensation with Hoechst dye staining 24 hours after transfection.

DNA transfections. BHK cells (1 × 10⁵) were transfected with 1 μg pcDNA3-Myc epitope–tagged human bcl-2 lacking the transmembrane domain (ΔTM-Bcl-2) or with pcDNA3-Myc-Ser⁷⁰Ala-ΔTM-bcl-2 in which Ser⁷⁰ was substituted with alanine by site-directed mutagenesis (Ser⁷⁰Ala-ΔTM-Bcl-2); transfection was by the calcium phosphate method (22). Sixteen hours after transfection, the cells were treated with 200 nmol/L Taxol and collected for further assays as shown in Fig. 1. To examine effect of phosphomimicking Ser⁷⁰ Bcl-2 expression on colon adenocarcinoma cells, 1 × 10⁵ SW620 cells were transfected with either 1 μg pcDNA3-Myc-bcl-2 (wild-type), Myc-Ser⁷⁰Ala-bcl-2, or Myc-Ser⁷⁰Asp-bcl-2 (Ser⁷⁰ substituted with aspartate by site-directed mutagenesis) by lipofection (LipofectAMINE 2000, Invitrogen, Carlsbad, CA). Twenty-four hours after transfection, the cells were stained by the indirect immunofluorescence method described above using 9E10 mAb and Cy3-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) for apoptosis assay.

Immunoblotting. Immunoblotting was used to determine the specificity of affinity-purified rabbit anti-pSer⁷⁰ Bcl-2 polyclonal antibody. Briefly, lysates of BHK cells expressing ΔTM-Bcl-2 or Ser⁷⁰-ΔTM-Bcl-2 were resolved by 12% SDS-PAGE gel. Proteins were transferred to nitrocellulose membrane and probed with the antibodies anti-Myc tag, anti-human Bcl-2, anti-pSer⁷⁰ Bcl-2, and anti-actin. Signals were developed by chemiluminescence detection using SuperSignal (Pierce Chemical Co., Rockford, IL).

Phosphorylation assay. Phosphorylation assay of Bcl-2 was done as described in previous studies (19, 23). Briefly, BHK cells expressing truncated Bcl-2 (Myc epitope–tagged ΔTM-Bcl-2) were first treated with 1 μmol/L A23187 (Calbiochem, San Diego, CA) or 250 ng/mL Taxol for 1 hour and then metabolically labeled with 0.5 mCi/mL (³²P) orthophosphate for 4 hours. The labeled cells were lysed in a 0.5% Triton X-100 lysis buffer containing protease inhibitors and phosphatase inhibitors. At this step, the same volume of the lysates from these three samples were aliquoted for immunoblotting probed with anti-actin mAb (Chemicon, Temecula, CA) to show the amount of total protein from each sample used for immunoprecipitation. Cellular lysates were then incubated with anti-Myc mAb-coated protein A beads to precipitate Myc-ΔTM-Bcl-2. The eluted proteins were fractionated by 12.5% SDS-PAGE; sequentially, the gel was analyzed by a phosphorimager (Bioimage Analyzer BAS2000-II, Fuji Photo Film Co., Honshu, Japan). Signal from the imager analysis was normalized against the actin levels obtained in the parallel immunoblot probed with anti-actin mAb.

Statistical analyses. Probabilities of survivals were calculated using the Kaplan-Meier method (24), and statistical significance of the differences between pSer⁷⁰ Bcl-2-positive and Bcl-2-negative groups was evaluated by the log-rank test. Comparisons between pSer⁷⁰ Bcl-2 expression and each clinical variable was calculated by nonparametric Mann-Whitney U test (age, sex, and differentiation; ref. 25) and by Spearman's rank-correlation test (wall penetration, lymph node metastasis, and stage) using ranks adjusted for multiple comparisons because the data did not show a normal distribution. Statistics were done by StatView J4.5 statistical analysis software (Hulinks, Tokyo, Japan). Two-sided P < 0.05 was the threshold for significance.

Characterization of the newly generated antibody against phosphorylated Ser⁷⁰ Bcl-2. We first examined the specificity of the pSer⁷⁰ Bcl-2 antibody by immunoblotting and ³²P phosphorylation assay using Taxol or ionomycin (A23187)–treated BHK cells expressing ΔTM-Bcl-2 based on the finding that Taxol, an inhibitor of microtubule assembly, induces hyperphosphorylation of Bcl-2 (5, 7, 11, 26, 27), whereas ionomycin was shown to dephosphorylate Bcl-2 (22). Immunoblots probed with anti-Myc-tag mAb revealed that BHK cells transfected with Myc epitope–tagged ΔTM-Bcl-2 (Myc-ΔTM-Bcl-2) showed a ΔTM-Bcl-2 mobility up-shift in response to Taxol treatment, whereas this mobility was shifted downward for cells treated with ionomycin (Fig. 1A, left). The same mobility shifts were detected in immunoblots probed with rabbit anti-Bcl-2 polyclonal antibody. The phosphorylation assay employing [³²P]dATPγ labeling showed that this mobility shift was caused by Taxol-induced phosphorylation of Myc-ΔTM-Bcl-2 (Fig. 1A, right, Taxol+). In immunoblots probed with anti-pSer⁷⁰ Bcl-2 antibody, the single band corresponding to up-shifted Myc-ΔTM-Bcl-2 was distinctly observed only in the sample treated with Taxol, and no band was observed in any sample derived from the cells expressing Myc-Ser⁷⁰Ala-ΔTM-Bcl-2, which substituted Ser⁷⁰ with alanine (Fig. 1A, left).

Specificity of the anti-pSer⁷⁰ Bcl-2 antibody was further examined by immunofluorescence of the Taxol-treated BHK cells transfected with Myc-ΔTM-Bcl-2. Double immunofluorescence using both anti-pSer⁷⁰ Bcl-2 polyclonal antibody and anti-Myc-tag mAb (9E10) revealed that these cells were strongly positive for anti-pSer⁷⁰ Bcl-2 antibody only in the presence of Taxol (Fig. 1B, top, Taxol− and Taxol+). In the preabsorption assay, the cells lacked significant signal only when the anti-pSer⁷⁰ antibody was absorbed with an excess amount of the pSer⁷⁰ oligopeptide before staining (Fig. 1C, bottom). We noted that, with respect to background activity, neither anti-total Bcl-2 nor anti-pSer⁷⁰ Bcl-2 reacted to endogenous Bcl-2 in BHK (Fig. 1A and B).

Taken together, the results in Fig. 1A and B show that anti-pSer⁷⁰ Bcl-2 antibody specifically recognizes the form of Bcl-2 phosphorylated at Ser⁷⁰. We then asked whether this antibody could detect endogenous pSer⁷⁰ Bcl-2, particularly as the intracellularly localized molecule, in the tubular adenoma cases using double-labeling immunofluorescence with combined anti-pSer⁷⁰ antibody and Hsp60. Hsp60 is a mitochondrial chaperone protein that localizes to mitochondrial outer membrane (28). The double immunofluorescence detected by confocal laser scanning microscopy showed that endogenous pSer⁷⁰ Bcl-2 mainly colocalized with Hsp60 at mitochondria (Fig. 1C) and, to a lesser extent, in the cytoplasm and Golgi apparatus (detailed data not shown).

Expression of phosphorylated Ser⁷⁰ Bcl-2 in colorectal adenomas and adenocarcinomas. The anti-pSer⁷⁰ Bcl-2-specific antibody was used to examine expression of the form of Bcl-2 phosphorylated at Ser⁷⁰ in colorectal neoplasms. Standard immunohistochemistry was employed to examine tumors, including 75 cases of tubular adenomas and 119 cases of adenocarcinomas at various stages (I-IV); all tumors were primary, lacking any pretreatment, and surgically removed between 1993 and 2002 (Table 1A). Expression of total Bcl-2 was first determined in these tumors by employing anti-Bcl-2 polyclonal antibody, which can detect both dephosphorylated and phosphorylated forms of Bcl-2 as shown in the immunoblot of Fig. 1A. As indicated in Table 1, all cases of tubular adenomas (75 of 75) were strongly positive for total Bcl-2. Similarly, >90% of the cases of adenocarcinomas (114 of 119) were positive for total Bcl-2, even the cases showing weak staining (Table 1A). Based on the immunohistochemical results from the adenocarcinomas, total Bcl-2 expression seems to be independent of the stage of differentiation (Table 1A). The expression of pSer⁷⁰ Bcl-2 was further investigated among the total Bcl-2-positive tumors by using the anti-pSer⁷⁰ Bcl-2 antibody. The expression pattern of pSer⁷⁰ Bcl-2 as detected by immunohistochemistry varied with each case of adenocarcinoma, whereas that of total Bcl-2 expression was uniform among the cancer cells in each case. Therefore, the patterns of pSer⁷⁰ expression were classified into four groups, including Bcl-2-negative cases, according to the proportion of pSer⁷⁰-positive segment in the total Bcl-2-positive area. With respect to frequency of positives, ∼76% of the cases of tubular adenomas diffusely and uniformly expressed pSer⁷⁰ Bcl-2 (Table 1B; Fig. 2A). In the adenocarcinomas, frequency of pSer⁷⁰ Bcl-2 expression seemed to depend on the stage of differentiation. Specifically, 63% of the well-differentiated cases were positive for the anti-pSer⁷⁰ antibody, whereas <12% of the poorly differentiated adenocarcinomas cases expressed pSer⁷⁰ Bcl-2 (Table 1B; Fig. 2B and D). The moderately differentiated adenocarcinomas showed an intermediate frequency (52%; Table 1B; Fig. 2C). The difference between tubular adenoma and adenocarcinoma with respect to pSer⁷⁰ Bcl-2 expression was found to be statistical significant (P < 0.0001). The difference between well-differentiated and poorly differentiated types was also found to be statistically significant (P < 0.0001). Distribution analysis revealed that over half the cases of well-differentiated adenocarcinomas showed diffuse or localized expression of pSer⁷⁰ Bcl-2. By contrast, only a small portion of cancer cells expressed pSer⁷⁰ Bcl-2 throughout the Bcl-2-positive cases of poorly differentiated adenocarcinomas (3 of 26 cases; Table 1B). These results indicate that well-differentiated adenocarcinomas express pSer⁷⁰ Bcl-2 with a higher frequency and over a broader area in comparison with poorly differentiated adenocarcinomas. To equalize the effects of the passage of time, we selected all of the cases of adenomas and adenocarcinomas derived from the same year (1995) and reexamined pSer⁷⁰ expression within this subset (data not shown); these results were in excellent agreement with the findings for the 1993 to 2002 range. Nonneoplastic epithelial cells at crypts of colorectal mucosa were also broadly positive for pSer⁷⁰ Bcl-2 and it seemed mainly accumulated in the mitochondria and Golgi apparatus, much as the tubular adenoma cells (Fig. 2E).

Correlation between expression of phosphorylated Ser⁷⁰ Bcl-2 and Ki-67 in cancer in adenoma. Because a distinct difference in the expression of pSer⁷⁰ Bcl-2 was recognized between tubular adenomas and adenocarcinomas, the cases of cancer in adenomas were subjected to further investigation to determine whether the phenomenon is also observed between a benign region and a malignant region originating from the same background. Among 15 cases of cancer in adenoma obtained from polypectomies, all showed strong pSer⁷⁰ Bcl-2 expression in the areas of the adenomas. Interestingly however, the areas of adenocarcinomas arising from the same cases lost substantial expression in approximately half of these instances. This result was statistically significant (P < 0.01; Table 1C). Although 7 of 15 (47%) cases were positive for the anti-pSer⁷⁰ antibody, pSer⁷⁰ Bcl-2 expression was not diffused over the entire cancerous area but was restricted to a localized region of 20% to 40% of total cancerous area (Table 1C). Based on this finding, we sought to determine if any biological markers correlated to pSer⁷⁰ Bcl-2 expression by employing these cancer in adenoma cases. Consequently, the expression of the Ki-67 antigen detected by MIB-1 mAb showed a definite inverse correlation with pSer⁷⁰ Bcl-2 expression. The cancerous pSer⁷⁰ Bcl-2-negative areas were predominantly positive for Ki-67 (MIB-1 index = 54.9), whereas both adenomatous areas and pSer⁷⁰-positive cancerous areas showed dramatically reduced Ki-67 expression (MIB-1 indices = 24.0 and 22.8, respectively; Fig. 3A and B). This observation was confirmed by double immunofluorescence using both pSer⁷⁰ Bcl-2 antibody (FITC-labeled) and MIB-1 mAb (Cy3-labeled; Fig. 3C) and was supported by statistical analysis as shown in Table 1 (Fig. 3A). We also noted that the inverse correlation between expression of pSer⁷⁰ Bcl-2 and that of Ki-67 antigen was seen at sites of vascular invasion in several cases of pSer⁷⁰-positive adenocarcinomas. In these cases, the foci of vascular invasion showed reduced or abrogated phospho-Bcl-2 expression and substantial Ki-67 antigen levels, although pSer⁷⁰ Bcl-2 was strongly expressed at the primary cancerous site (Fig. 3D).

Phosphomimetic Ser⁷⁰ Bcl-2 mutant affected cell cycle control and induced apoptosis to colon adenocarcinoma cells. For attempt to disclose the potential role of altered phosphorylation of Ser⁷⁰ Bcl-2 in colorectal cancers, we did transfection assay on the SW620 colon adenocarcinoma cell line using phosphomimetic mutants of Ser⁷⁰ Bcl-2. Interestingly, significant differences in cell cycle regulation and induction of apoptosis were observed between cells overexpressing phosphomimetic mutant Ser⁷⁰Asp-Bcl-2 and those overexpressing wild-type Bcl-2 or dephosphomimetic mutant Ser⁷⁰Ala-Bcl-2. The Ser⁷⁰Ala-overexpressing SW620 cells were highly proliferative; MIB-1 index was ∼88% at 12 hours after transfection. In contrast, MIB-1 index of the Ser⁷⁰Asp-overexpressing cells was 55%, which was even lower than that of the wild-type transfected cells (Fig. 4A, left, and B, left). Noticeably, in the Ser⁷⁰Asp-overexpressing cells, over half of them entered apoptosis at 24 hours after transfection, whereas majority of the cells expressing Ser⁷⁰Ala Bcl-2 or wild-type Bcl-2 were nonapoptotic (Fig. 4A, right, and B, right).

Specific correlation between phosphorylated Ser⁷⁰ Bcl-2 expression and clinicopathologic factors and influence of its expression on survival of patients with colorectal cancers. Phospho-Bcl-2 expression was statistically analyzed among patients subgroups defined by clinical and pathologic variables (age, sex, histologic differentiation of adenocarcinoma, wall penetration, lymph node metastasis, and clinical stage). First, a positive correlation with pSer⁷⁰ expression was observed in cancer cell differentiation by Mann-Whitney U test (P < 0.001; Table 2). Moreover, its expression was significantly correlated with both lymph node metastasis and clinical stages from Spearman's rank correlation test, whereas no specific correlation was found between pSer⁷⁰ expression and age, sex, or wall penetration (Table 2). Following on these results, we next investigated whether the pSer⁷⁰ Bcl-2 expression would show prognostic value in patients with colorectal adenocarcinomas. Univariate analysis (log-rank test) revealed that the survival of the pSer⁷⁰-negative patients (total of 61 cases) was significantly shorter than that of the pSer⁷⁰-positive patients (total of 58 cases; Fig. 4C). Two- and 5-year overall survival rates were 50.5% and 39.3% in the pSer⁷⁰-negative subgroup, respectively, whereas those rates were 65.9% and 52.4% in the pSer⁷⁰-positive subgroup.

The expression pattern of phosphorylated Bcl-2 in vivo had been unexplored previously, and its biomedical significance has accordingly been controversial (15). As one of the reasons, it is conceivable that the opposing results are strongly influenced by the different backgrounds of Bcl-2 family expression among the various cell lines. Our main purpose in the present study is to investigate in vivo expression of phospho-Bcl-2 in colorectal neoplastic tissues to reveal clues about its biological significance from the pathologic and clinical standpoints.

To this end, we investigated the expression of pSer⁷⁰ in the loop region of Bcl-2, because that residue is the most frequently observed Bcl-2 phosphorylation site in human colorectal adenomas and adenocarcinomas. Consequently, differential expression of pSer⁷⁰ Bcl-2 was observed between adenomas and adenocarcinomas. For tubular adenoma cases, >80% abundantly expressed pSer⁷⁰ in a diffuse and uniform pattern. In contrast to adenomas, the frequency of pSer⁷⁰ Bcl-2 expression in adenocarcinomas decreased in a differentiation-dependent manner. Detailed analysis by immunohistochemistry indicated that <12% of the poorly differentiated adenocarcinoma cases were positive for pSer⁷⁰ Bcl-2, whereas ∼60% of the well-differentiated adenocarcinoma cases were positive. Interestingly, loss of pSer⁷⁰ expression was also observed at sites of vascular invasion in some cases of pSer⁷⁰-positive adenocarcinomas.

From the cases of cancer in adenoma for detailed investigation, it was found that loss of pSer⁷⁰ Bcl-2 expression was observed at mitotically active cancer area, whereas adenomatous area maintained the expression. Thus, Ki-67 antigen showed a distinct inverse correlation with pSer⁷⁰ Bcl-2 in colorectal tumors in vivo. These immunohistochemical and statistical analyses supported the correlational significance of the absence of pSer⁷⁰ Bcl-2, combined with the presence of Ki-67-positive neoplastic cells of high mitotic activity, whereas Ki-67-negative cells tended strongly to express pSer⁷⁰ Bcl-2. In colon adenocarcinoma cell line as an in vitro model, transfection assays using phosphomimetic or dephosphomimetic Ser⁷⁰ bcl-2 mutants revealed that enhanced expression of phosphorylated form of Ser⁷⁰ Bcl-2 (Ser⁷⁰Asp-Bcl-2) seemed to promote cell cycle arrest that was proven by increasing MIB-1 index on Ser⁷⁰Asp-expressing cells, and induction of apoptosis was also observed in these cells. At this moment, detailed molecular mechanism is still necessary to be investigated because bcl-2 has been reported to be multifunctional and biological effect of its phosphorylation still remains to be complexity (15, 19, 22, 26, 29). However, it seemed, at least, that change in numbers of the Ki-67-expressing cells and apoptotic cells by the mutative Ser⁷⁰-Bcl-2 overexpression showed parallel behavior. Taking these findings together, it implies that phosphorylation at Ser⁷⁰ of Bcl-2 might converge to suppression of colorectal tumor progression.

In the clinicopathologic study, clinical factors, such as degree of lymph nodal metastasis and tumor stage, were revealed to be associated with pSer⁷⁰ Bcl-2 expression in addition to cellular differentiation of adenocarcinoma. Specifically, loss of the pSer⁷⁰ expression was more frequently recognized in the cases with advanced stage of lymph nodal metastasis and clinical stages, which suggests its utility as a biological marker for the extent of malignancy in colorectal cancer. Moreover, the result of nonparametric analysis indicated the shorter survival of the pSer⁷⁰ Bcl-2-negative patients in comparison with that of the pSer⁷⁰-positive patients. Thus, from clinicopathologic viewpoint, expression of pSer⁷⁰ Bcl-2 seems to be useful as a novel prognostic indicator for colorectal cancer patients. In the present study, we attempted to show the phosphorylation-regulation of Bcl-2 in colorectal tumors. Clearly, more investigation is needed to provide further insight on phospho-Bcl-2 molecular actions during cancer progression. However, we believe that our present work offers a logical framework for evaluating the specific molecular enhancement of malignancy and for integrating the increasing clinical knowledge with definitive in vivo and molecular studies of Bcl-2-contributed tumorigenesis.

We thank K.V. Myrick (Department of Molecular and Cellular Biology, Harvard University) for useful advice and critical reading of the article.