Phorbol Ester Up-regulates Phospholipase D1 but Not Phospholipase D2 Expression through a PKC/Ras/ERK/ NF (cid:1) B-dependent Pathway and Enhances Matrix Metalloproteinase-9 Secretion in Colon Cancer Cells

Despite its importance in cell proliferation and tumorigenesis, very little is known about the molecular mechanism under-lying the regulation of phospholipase D (PLD) expression. PLD isozymes are significantly co-overexpressed with cancer marker genes in colorectal carcinoma. Phorbol 12-myristate 13-acetate (PMA) treatment, as a mitogenic signal in colon cancer cells, selectively increases PLD1 expression in transcription and post-transcription. Moreover, experiments using intraperitoneal injection of PMA into mice showed selective PLD1 induction in the intestine and lung tissues, which suggests its physiological relevance in vivo . Therefore, we have undertaken a detailed analysis of the effects of PMA on the promoter activity of PLD genes. Protein kinase C inhibitors, but not a protein kinase A inhibitor, were found to suppress the up-regulation of PLD1 but not PLD2. Dominant-negative mutants of Ras, Raf, and MEK suppressed the induction and activity of PLD1. Moreover, metallo- protease-9; RT, reverse transcription; GAPDH, glyceraldehyde-3-phos-phate dehydrogenase; PI3K, phosphatidylinositol 3-kinase; MEK, mito- gen-activated protein kinase/extracellular signal-regulated kinase kinase; ERK, extracellular signal-regulated kinase.

Colorectal cancer (CRC) 2 is one of the leading causes of cancer death in much of the developed world, despite the facts that colonic malignancies can be effectively managed when detected early and that chemoprevention has shown some success in reducing the severity of the disease (1). Dysregulation of proliferation and apoptosis frequently occurs in the development of various human cancers. The unraveling of growth-regulatory pathways and the process of tumorigenesis has revealed new diagnostic and prognostic markers of disease as well as potential targets for therapeutic intervention. Therefore, the discovery of new cancer-associated genes and their mechanisms of action is important for overcoming CRC.
Phospholipase D (PLD) has emerged as a critical regulator of cell proliferation and survival (2). Abnormalities in PLD expression and activity have been observed in many human cancers (2). PLD hydrolyzes phosphatidylcholine, the predominant membrane phospholipid, and produces phosphatidic acid and free choline. Two mammalian isozymes of phosphatidylcholine-specific PLD (PLD1 and PLD2) have been identified and are differentially regulated. PLD1 is activated by protein kinase C (PKC) and small G proteins of the Rho and ADP-ribosylation factor families, whereas PLD2 is not directly activated by these regulators (3). We have reported that overexpression of the PLD isozymes causes anchorage-independent growth and neoplastic transformation of mouse fibroblasts (4). PLD1 has been reported as a critical downstream mediator of H-Ras-induced transformation (5). Mutation in the Ras oncogene is found in many human cancers (6) and can lead to the deregulated activation of PLD and important downstream signaling pathways, including Raf/MEK/ERK (7) and PI3K/Akt (8). Moreover, PLD has been shown to be involved in the resistance of cancer cells to chemotherapeutic drugs (9,10). Despite gathering evidence regarding the regulation of PLD activity in cell function, little is known about the functional role and regulatory mechanisms of PLD expression.
Phorbol esters, such as phorbol-12-myristate-13-acetate (PMA), can substitute for diacylglycerol (DAG; the endogenous PKC activator) and function as tumor promoters. Phorbol esters have been reported to modulate diverse cellular responses, such as gene transcription, mitogenesis, and differentiation, tumor promotion, and apoptosis, through the PKC signaling pathway (11,12). PMA protects cells from stimuliinduced apoptosis through a mechanism that is dependent on the activation of PKC (13). However, the signaling pathways involved in PMA-induced PLD expression have not been defined. HCT116 human colon cancer cells contain the most common K-ras mutation type found in colorectal cancers, the glycine to aspartate mutation at codon 12 (14). Because many growth factors utilize Ras-dependent signaling pathways, this cell line serves as a model system for studying alterations in gene expression that occur during the progression of CRC. Secretion of matrix metalloproteinases (MMPs) from cancer cells is an important stage of metastatic spread. MMPs hydrolyze components of the extracellular matrix and allow the invasion of cancer cells from their primary site to the circulation and secondary sites (15)(16)(17). Various stimuli, including PMA, have been shown to induce the secretion of MMP-9 from cancer cell lines (18), including human fibrosarcoma line HT 1080 (19,20). Thus, we investigated the involvement of PLD1 induction in PMA-stimulated MMP-9 secretion.
The present study was conducted to define the cellular mechanisms that regulate the expression of PLD isozyme in human colon cancer cells. Here, we demonstrate for the first time that induction of PLD1 through the PKC/Ras/ERK/NFBdependent pathway is involved in the secretion of MMP-9 in colorectal cancer cells.

EXPERIMENTAL PROCEDURES
Materials-PMA and DAG were purchased from Sigma. Various inhibitors were obtained from Calbiochem. The oligonucleotides of NFB-binding sites were purchased from Genotech (Daejeon, Republic of Korea). The siRNA of human PLD1 (nucleotides 1486 -1506, AAGGUGGGAC-GACAAUGAGCA) and PLD2 (nucleotides 2665-2685, AAUGGGGAUGCGGAUAAAGGC) were purchased from Dharmacon Research Inc. (Lafayette, CO). The siRNAs of PKC␣, Ras, and ERK were obtained from Cell Signaling Inc. (Beverly, MA). Dual luciferase assay kits were purchased from Promega (Madison, WI). Polyfect transfection reagent was purchased from Qiagen.
Animals and Treatments-ICR female mice (6 weeks of age, weighing ϳ18 -21 g) were obtained from Oriental Laboratories (Seoul, Republic of Korea). During this study, the mice received standard rodent chow and acidified tap water ad libitum. After acclimatization for 1 week, female mice were injected intraperitoneally with PMA three times at doses of 0.015 or 0.15 mg/kg over the course of 1 week (n ϭ 5 samples). Controls received an equal volume of corn oil in the same manner.
Preparation of Human CRCs Samples and Immunohistochemistry-Formalin-fixed and paraffin-embedded samples were obtained from surgical resections of 122 colorectal cancers. The tumor stage was classified according to Dukes' criteria (21). There were 12, 47, 55, and eight cases with stage A, B, C, and D, respectively. Two pathologists screened histological sections and selected areas of the representative tumor cells. Three tissue cores (0.6 mm in diameter) were taken from each tumor sample and placed in a new recipient paraffin block using a commercially available microarray instrument (Beecher Instruments, Micro-Array Technologies, Silver Spring, MD) according to established methods (22). One cylinder of normal colonic mucosa adjacent to each tumor was also transferred to the recipient block. Immunohistochemical analysis was performed as described previously (23).
Cloning of the Promoter Regions of Human PLD1 and PLD2-Approximately 1.9-and 2.6-kb genomic DNA fragments covering the 5Ј upstream regulatory sequences of human PLD1 or PLD2, respectively, were amplified from human genomic DNA and cloned into a pGL4.14b luciferase reporter plasmid (Promega, Madison, WI) and were then used for promoter analysis. For further detail, see supplemental data.
RNA Isolation and Reverse Transcription PCR (RT-PCR)-Total RNA was isolated from cells using TRIzol reagent (Sigma). First strand cDNA was synthesized using 5 g of total RNA with Moloney murine leukemia virus reverse transcriptase (Promega, Mannheim, Germany). The reaction was incubated at 80°C for 5 min and 42°C for 90 min, and Moloney murine leukemia virus reverse transcriptase was inactivated at 95°C for 5 min. The synthesized cDNA was amplified using PLD1, PLD2, and GAPDH primers with an Eppendorf thermocycler (Eppendorf Scientific, Westbury, NY). PCR products were analyzed by 1.5% agarose gel electrophoresis with ethidium bromide staining. For PCR conditions and primer sequences, see supplemental data.
Immunoprecipitation and Western Blotting-The cells or tissues were analyzed by immunoblotting or immunoprecipitation as described previously (24). ECL was used to detect the signal. The following antibodies were used: anti-␤-tubulin (Sigma); anti-NFB (p65) antibody (Santa Cruz Biotechnology); anti-PKC␣ antibody (Santa Cruz Biotechnology); and anti-ERK antibody (Cell Signaling Inc., Beverly, MA). Rabbit polyclonal anti-PLD antibody that recognizes both PLD1 and PLD2 was generated as described previously (4).
Chromatin Immunoprecipitation (ChIP) Assay-ChIP experiments were essentially performed as described previously (25), with minor modifications. For further details, see supplemental data.
PLD Activity Assay-PLD activity was assessed by measuring the formation of [ 3 H]phosphatidylbutanol, the product of PLDmediated transphosphatidylation, in the presence of 1-butanol as described previously (24).
Zymography-HCT116 cells were seeded in 12-well plates at 1 ϫ 10 5 cells/ml and were washed and refreshed with serumfree Dulbecco's modified Eagle's medium for 24 h. The conditioned media of the cells were determined by 0.1% casein zymography. Zones of caseinolytic activity were detected as clear bands against a blue background.
Measurement of Cell Viability and Apoptosis-For the cell viability assay, a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was performed. Absorbance was measured using a spectrophotometer at 540 nm, and viability was expressed relative to the control. Cell apoptosis was measured using an Annexin V binding assay (Molecular Probes), which was performed according to the protocol of the manufacturer. Annexin V-positive cells were quantitated. More than 200 cells were counted for each variable per experiment using fluorescence microscopy (Axiovert 200M, Zwiss, Germany).

Statistics-
The results are expressed as means Ϯ S.E. of the number of determinations indicated. Statistical significance of differences was determined by analysis of variance. The correlation between PLD expression and various clinicopathologic parameters and Ki-67, ␤-catenin, and phospho-Akt expression were assessed using the 2 and Bartholomew tests. A p value less than 0.05 was considered to indicate statistical significance.

PLD Isozymes Are Significantly Co-overexpressed with Cancer Marker Genes in Human Colorectal
Cancer-The expression and activity of PLD isozymes have been reported to be highly up-regulated in many cancer tissues (2), but the number of cases in each study was small. Thus, using a tissue microarray containing 122 CRCs obtained from surgery, we attempted to investigate the expression of PLD isozymes and their association with clinicopathological parameters involving tumor stage, lymph node metastasis, and other cancer marker genes, which are important factors in prognosis. The expression level of PLD isozymes was assessed by immunohistochemistry. As shown in Fig. 1A, using the antibody to PLD that recognizes both PLD1 and PLD2, PLD expression was detected as faint or negative in normal colonic mucosa. Occasional weak staining of PLD was observed in inflammatory cells in lamina propria. No immunoreactivity was observed in connective tissues; thus, these tissues served as internal negative controls. Immunopositivity for PLD protein was clearly marked in colorectal tumor cells (Fig. 1B). The tumors were interpreted as negative when immunostaining was weak, as in the corresponding normal colonic mucosa, or when immunopositive cells represented Ͻ5% of the cancer cells. Table 1 summarizes the relationship between clin-FIGURE 1. Expression of PLD, Ki-67, ␤-catenin, and p-Akt in normal colonic mucosa and cancers. A tissue microarray containing 122 human colorectal cancer tissues was subjected to immunohistochemical analyses with primary antibodies. A positive immunohistochemical reaction is indicated by diaminobenzidine (brown staining), and hematoxylin was used as a nuclear counterstain. Representative images from normal colonic mucosa and cancer tissue are shown. A, normal colonic mucosa. Note that immunoreactivity for PLD is not present in the normal colonic mucosal epithelial cells. Inflammatory cells in lamina propria demonstrate negative to weak staining. B, colon cancer. Note that tumor cells demonstrate strong labeling for PLD, whereas connective tissues in the stroma of cancer specimens are not stained for PLD. C, normal colonic mucosa showed immunonegativity for Ki-67. D, colon cancer cells showing strong nuclear staining for Ki-67. E, in normal colonic mucosa, ␤-catenin was lightly stained in the membranes of the colonic epithelial cells. F, colon cancers showed strong nuclear and/or cytoplasmic staining for ␤-catenin. G, normal colonic mucosa was immunonegative for p-Akt. H, colon cancer cells demonstrated strong cytoplasmic staining for p-Akt protein (original magnification was ϫ200).

TABLE 1 Relationship of PLD expression detected by immunohistochemistry and clinicpathological parameter in colorectal carcimoma
The correlation between PLD overexpression and clinicopathologic parameters involving tumor stage and lymph node metastasis was analyzed by immunohistochemistry using a tissue microarray containing 122 human colorectal carcinomas. In addition, the correlation between PLD expression and the expression of Ki-67, ␤-catenin, or p-AKT in these cancer samples was also analyzed for examination of proliferative capacity in CRCs with overexpression of PLD. A positive or negative sign means strong immunostained tissues and faint or negative immunostained tissues, respectively. L/N, lymph node.
icopathological features and PLD expression in colorectal carcinoma. Overexpression of PLD isozymes was detected in 74 (60.1%) of 122 colorectal cancers (Table 1). However, no significant association was found between the PLD expression level and tumor stage or lymph node metastasis (Table 1). Additionally, to determine the proliferative capacity in CRCs with overexpression of PLD, we also examined the correlation between the expression of PLD and the expression of Ki-67, ␤-catenin, or p-AKT in cancer samples. We compared the expression pat-tern of the proteins in tumor cells and corresponding normal cells and interpreted results as being positive or showing overexpression when immunostaining was stronger than that of corresponding normal cells.
In immunohistochemical analyses, corresponding normal colonic mucosa showed negative or weak staining for ␤-catenin, Ki-67, and p-Akt ( Fig. 1, C, E, and G). PLD isozymes were significantly cooverexpressed with early stage cancer marker genes, such as Ki-67, ␤-catenin, or p-Akt in colorectal carcinoma (Fig. 1, B, D, F, and H, and Table 1). These findings suggest that PLD plays an important role in the progression of early stages of development of colon carcinogenesis. PMA Up-regulates the Expression of PLD1, but Not PLD2, in Human Colon Cancer Cells and Mice-To investigate the signaling pathways responsible for PLD expression in CRC, we examined the effects of PMA on PLD expression in HCT116 cells. Treatment with PMA markedly increased PLD1 mRNAexpressioninatime-anddosedependent manner ( Fig. 2A). In contrast, PLD2 isozyme was not significantly affected by PMA. In addition, the induction of PLD1 mRNA levels by PMA resulted in the corresponding induction of PLD1 protein (Fig. 2B). Moreover, DAG increased the expression of PLD1 protein, but not PLD2, in a dose-dependent manner (Fig. 2C). To assess the physiological relevance in vivo, we examined the PLD expression after intraperitoneal injection of PMA into five different mice. Strikingly, PMA treatment significantly induced the expression of PLD1 in the intestine and lung tissues, compared with vehicle (Fig. 2D). However, PLD2 expression in these tissues was not detected by anti-PLD antibody, which recognizes both PLD1 and PLD2. Taken together, these results demonstrate that PMA, which is known to activate PKC and then PLD, selectively up-regulates the expression of PLD1 in HCT116 cells as well as mice.
Cloning of the Promoter Region of Human PLD1 and PLD2 Genes-To determine how PMA is able to regulate PLD gene expression and what its transcriptional mechanism might be, we have cloned the promoter regions of human PLD1 (pGL4- PLD1; 1.9 kb) transcribed from exon 2 among two alternate transcripts of PLD1 to be transcribed at two different transcription sites (exons 1 and 2) and human PLD2 (pGL4-PLD2; 2.6 kb) transcribed from exon 1, on the basis of human genome data from the NCBI data base. Supplemental Fig. S1 shows the schematic diagrams of genomic DNA structure and the promoter regions of the human PLD1 and PLD2 genes.
PMA Increases the Transcriptional Activity of PLD1 Promoter-To investigate whether PMA increases the transcriptional activities of PLD genes, HCT116 cells were transiently transfected with PLD-luc and treated with PMA. PMA increased the promoter activity of PLD1 in a time-and dosedependent manner but did not increase the promoter activity of PLD2 (Fig. 3, A and B). The PLD2 promoter showed higher basal activity compared with that of PLD1. These results demonstrate that PMA selectively up-regulates the transcriptional activity of PLD1 but not that of PLD2.
PMA-stimulated PLD1 Induction Is Mediated via the PKC Pathway-Therefore, to examine the signaling pathways leading to PMA-mediated PLD1 induction, HCT116 cells were pretreated with various PKC inhibitors (GF109203x, Go6976, Ro-31-8220, Rottlerin) and PI3K inhibitor (LY294001) or a protein kinase A inhibitor (H89) prior to the addition of PMA. As shown in Fig. 4A, PMA-induced PLD1 mRNA expression was blocked by all of the PKC inhibitors, inhibitors of PI3K, and Src tyrosine kinase but not by a protein kinase A inhibitor. Furthermore, the effects of these inhibitors on mRNA expression of PLD1 were correlated with the results obtained from the PLD1 promoter assay, which suggests that the effects are mediated through an increase in transcription (Fig. 4A). To further confirm this observation, the cells were transfected with various constructs with PLD1-luc. Wild-type PKC-␣, -␤II, -␦, orstimulated PLD1 promoter activity, but its dominant negative mutants suppressed both basal and PMA-stimulated promoter activities (Fig. 4B). Furthermore, dominant negative PI3K (p85) and Src, but not protein kinase A, inhibited PMA-stimulated PLD1 promoter activity (Fig. 4C). However, dominant active PI3K (p110) and Src enhanced its activity. Taken together, these results, which utilized complementary approaches, dem-  onstrate a contributory role for these PKCs, PI3K, or Src in PMA-mediated PLD1 induction.

PMA-induced PLD1 Expression Occurs in Other Human
Cancer Cell Lines-To examine whether PMA-induced expression of PLD1 is specific to the HCT116 cells or also occurs in other human cancer cells, the colon cancer cells (HCT116, CaCo2, LS174T, RKO, DLD1, and SNU-C5) and breast cancer cells (SK-BR3, MCF7) were treated with PMA, and PLD1 expression was analyzed by RT-PCR and Western blot analysis (Fig. 5A). The induction of PLD1 mRNA and protein was observed in the HCT116, CaCo2, SNU-C5, SK-BR3, and MCF7 cancer cells and was blocked by Ro31-8220. In contrast, PMAinduced PLD1 expression was not detected in LS174T, RKO, and DLD1 cells. Using antibody to pan-phospho-PKC, we found that the cells that were responsive to PMA (except for MCF7 cells) showed relatively low basal PKC activation compared with that of the cells that were unresponsive to PMA (Fig.  5B). For DLD1 cells, the PKC inhibitor did not greatly affect PLD1 expression. Although responsiveness to PMA may be partly related to the extent of basal PKC activation in different cancer cells, it is possible that these cell-specific differences in PMA-induced PLD1 expression might be related to differences in the genetic profiles of these tumors.
Ras/Raf/ERK Pathways Are Required for the Induction of PLD1 by PMA-We next investigated whether Ras, a downstream target of PKC (26), is also involved in the signaling cas-cade. As shown in Fig. 6A, expression of Ras-N17 (inactive mutant) significantly suppressed both basal and PMA-stimulated PLD1 promoter activity. Conversely, expression with Ras-V12 (active mutant) significantly stimulated its promoter activity. Ras activates multiple parallel pathways that involve Raf, MEK, and ERK kinases (26). Like Ras, Raf enhanced PLD1 promoter activity in a similar pattern (Fig. 6A). These results indicate that Ras and Raf are involved in PMA-induced PLD1 upregulation. Moreover, we investigated the role of MEK/ERK, acting downstream of Ras, in the regulation of PLD1 expression. Pretreatment with MEK inhibitors (PD98059, U0126) and overexpression of the dominant negative mutant of MEK2, but not p38 inhibitor (SB203580), significantly blocked PMA-mediated PLD1 expression, as measured by RT-PCR and a promoter assay (Fig. 6B). The JNK inhibitor had no effect on the induction of PLD1 (data not shown). PKC ␤II-stimulated PLD1 induction was suppressed by the expression of dominant negative mutants of Ras, Raf, MEK2, and ERK2 (Fig. 6C). Furthermore, we examined the relationship between enhanced PLD1 expression and PLD activity. As shown in Fig. 6D, transfection of PKC-␤II, Ras-V12, catalytically active Raf, and MEK, but not its dominant negative mutant, enhanced PLD activity, which suggests that PKC and these proteins are responsible for an increase in PLD activity in the colon cancer cells. In addition, PMA increased the phosphorylation of ERK, Akt, and Src kinase in a time-dependent manner (data not shown). Moreover, we investigated whether depletion of the supposedly involved protein affects PLD1 expression. Using siRNA, the depletion of PKC␣, Ras, and ERK2 reduced the endogenous PLD1 protein level (Fig. 6E). Taken together, these results suggest that PMA stimulated the induction of PLD1, and thus, the activity might be mediated via PKC/Ras/Raf/MEK/ERK signaling pathways in HCT116 colon cancer cells.
NFB Induces PMA-mediated PLD1 Expression, but Not PLD2 Expression, and Increases PLD Activity-PMA is known to activate NFB, which has been shown to be a downstream target gene of MEK/ERK (27). Thus, we investigated whether NFB plays a role in PMA-stimulated PLD1 induction. As shown in Fig. 7A, pretreatment with PDTC, an NFB inhibitor, and overexpression of the dominant negative mutant of IB␣ (S32A/S36A) significantly blocked PMA-mediated PLD1 transcriptional expression. However, ectopic expression of NFB (p65) enhanced the promoter activity, protein expression, and enzymatic activity of PLD1 in a dose-dependent manner but did not significantly affect those of PLD2 (Fig. 7, B-D). Furthermore, transfection of siRNA for NFB (p65) decreased PMAstimulated PLD1 protein expression (Fig. 7E). Taken together, these results demonstrate that NFB is required for the PMAinduced expression and activation of PLD1.
Identification of NFB-binding Sites Responsible for PMAinduced PLD1 Expression-Next, we attempted to examine the regions that are responsible for PMA-induced PLD1 expression. Treatment with PMA and ectopic expression of NFB enhanced the promoter activity of PLD1 in all of the deletion constructs (Fig. 8A). We found four putative NFB-binding sites in the PLD1 promoter using bioinformatics (Fig. 8B). To determine if NFB (p65) could directly bind to the PLD1 promoter in vivo and to further define the NFB response elements FIGURE 5. Cell-specific differences in PMA-mediated PLD1 induction. A, various cancer cells were pretreated with Ro31-8220 (5 M) and stimulated with PMA (100 nM) for 24 h. The proteins were extracted and analyzed by Western blotting using antibody to PLD1 or ␤-tubulin. Total RNA was extracted, and expression for PLD1 and GAPDH was determined by RT-PCR. B, to examine the endogenous PKC activation states, lysates from these cells were analyzed by immunoblotting using anti-pan-phospho-PKC antibody. These blots are representative of results obtained from three experiments. FEBRUARY 15, 2008 • VOLUME 283 • NUMBER 7

PMA-induced PLD1 Expression via NFB
to the promoter, a chromatin immunoprecipitation assay was carried out after treatment with PMA. PCR was performed using primers designed to amplify the four putative NFBbinding regions, as shown in Fig. 8B. PMA enhanced the binding of p65 to the two NFB-binding sites of the PLD1 promoter (NFB-1 and NFB-4), showing especially strong binding of NFB to the NFB-1 site. These results are comparable with those of the deletion promoter assay. As a positive control, the binding of NFB to the promoter of c-IAP2 was enhanced by PMA (28). These data suggest that NFB may play a role as an activator of the transcription of PLD1. Furthermore, mutagenesis of NFB-1 and NFB-4 sites attenuated PMA-induced PLD1 promoter activity (Fig. 8C). These findings demonstrate the functional importance of NFB in the PMA-induced induction of PLD1 in HCT116 cells.

PMA-induced PLD1 Induction Is Involved in the Secretion of MMP-9 via Ras/ERK/NFB in Colon Cancer
Cells-MMPs are well known to be involved in the invasion of tumor cells. PLD has been shown to be essential for the induction of MMP-9 expression by several stimuli (29 -31). Thus, we examined whether PMA-induced PLD1 induction is involved in the secretion of MMP-9 in HCT116 colon cancer cells. As shown in Fig. 9A, inhibitors that were specific for Ras (L744832), ERK (U0126), and NFB (PDTC), which are involved in the induction of PLD1 by PMA, suppressed PMA-induced MMP-9 secretion and PLD1 induction. Moreover, depletion of PLD1 inhibited PMA-induced MMP-9 secretion, as analyzed by zymography. Taken together, these results suggest that PMA-induced expression of PLD1 via a Ras/ERK/ NFB pathway might mediate invasiveness associated with the secretion of MMP-9 in HCT116 cells.
Depletion of PLD Decreases Cell Viability and Increases Apoptosis-Furthermore, we attempted to investigate the effect of the depletion of PLD isozyme on cell viability and apoptosis in cancer cells. As shown in Fig.  10, the depletion of PLD1 or PLD2 using siRNA decreased cell viability and increased apoptosis compared with those of control siRNA, as measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay and Annexin V binding assay. Depletion of both PLD1 and PLD2 significantly suppressed cell viability and enhanced apoptosis compared with that of either PLD isozyme. Taken together, these results suggest that

DISCUSSION
Abnormalities in the expression and activity of PLD have been reported in various human cancers using small numbers of tumor samples, but its clinicopathological significance remains unclear. In this study, using 122 colorectal carcinomas obtained from surgery, we found that PLD overexpression was detected in 74 (60.1%) of 122 colorectal cancers. However, no significant association was found between the PLD expression level and tumor stage or lymph node metastasis. Our observations indicated that PLD isozymes are significantly co-overexpressed in human CRCs with Ki-67, ␤-catenin, or p-Akt, which are involved in the early stages of the development of colon cancer. Thus, it is suggested that PLD may contribute to the progression of colorectal carcinoma in the clinical setting.
Despite extensive information regarding the regulation of PLD activity in cell function, the signaling mechanisms that regulate PLD expression in cancer remain undefined. In the present study, we demonstrate for the first time that activation of PKC by PMA enhances the expression of PLD1, but not that of PLD2, via an NFB-dependent pathway in colon cancer cells and suppresses apoptosis. Selective induction of PLD1 by PMA was also confirmed in the tissues of mice injected with PMA, thus suggesting its physiological relevance. Moreover, PMA-induced expression of PLD1 was not noted in all of the cancer cell lines examined. The specificity of PLD1 induction in certain cancer cells might be partly associated with endogenous PKC activation states, depending on cell types, but there are the other possible explanations for this phenomenon, including different genetic profiles of cancer cells or the sensitivity of the assays.
In this study, we found for the first time that PMA selectively upregulates the induction of PLD1, but not PLD2, via a Ras/Raf/MEK/ERK/ NFB signaling pathway in HCT116 colorectal cancer cells. Elevated expression of these induced proteins also enhanced PLD activity. Moreover, depletion of these supposedly involved proteins reduced the PLD1 protein level. Recently, Kikuchi et al. (32) reported that Ewing's sarcoma fusion protein induces the expression of the PLD2 gene but not that of the PLD1 gene. . The expression of PLD isozymes was examined using RT-PCR. The same methods were performed after transfection of pGL4-PLD1, and the luciferase activity was measured (bottom). B, the cells were cotransfected with NFB (p65) at various doses (0, 50, 100, 200, 400, and 600 ng) and with pGL4-PLD, and the luciferase activity was then measured. C, HCT116 cells were transfected with NFB (p65) at various doses (0, 0.5, 1, and 2 g), and the lysates were immunoprecipitated or immunoblotted with antibodies to the indicated antibodies. D, the cells were transfected with NFB (p65) at various doses for 24 h, labeled with [ 3 H]myristate for 12 h, and treated with or without PMA (100 nM) for 1 h, and the PLD activity was then measured as described under "Experimental Procedures." Results shown are means Ϯ S.E. of three independent experiments. #, p Ͻ 0.05 versus vector; *, p Ͻ 0.05 versus PMA treatment. E, the cells were transfected with siRNA specific for NFB or control siRNA (NS, nonspecific) and were treated with or without 100 nM PMA for 24 h. The lysates were immunoprecipitated and immunoblotted with anti-PLD antibody. The levels of PLD expression were determined by densitometer analysis. The expression levels of NFB (p65) and ␤-tubulin proteins were determined by Western blot analysis. These blots are representative of results obtained from three experiments.
NFB has been reported to regulate transcription of a number of target genes related to tumorigenesis and cell survival, including TRAF1, TRAF2, cIAP-1, cIAP-2, and FLIP (28,33,34). Analysis of the PLD1 promoter shows the presence of two functional NFB sites; the binding of these sites to the PLD1 promoter is enhanced by PMA, and the mutation of these sites suppresses PMA-induced PLD1 expression. NFB plays a role in the enhanced survival of colon cancer cells, and inhibition of NFB sensitizes colon cancers that are resistant to chemotherapy-induced apoptosis (35). These results suggest that NFB activation is a critical step in the tumor-promoting effect of PMA. Despite these important findings, the signaling pathways involved in PMA regulation of genes that are directly involved in cell death or death resistance remain largely unknown. Our current study shows that expression of PLD1 is regulated through the PKC-stimulated activation of NFB.
The secretion of MMP-9 from cancer cells (36,37) leads to hydrolysis of the extracellular matrix, thus enabling cells to  break out of their primary site into the circulation and from there to secondary sites. MMP-9 secretion is induced by various agonists that differ among cell lines (36). MMP-9 is secreted in response to PMA. PLD is known to be essential for the induction of MMP-9 secretion by several stimuli, including PMA. We found that inhibition of the proteins (Ras/ERK/NFB) involved in the PMA-stimulated induction of PLD1 suppresses PMAinduced MMP-9 secretion and PLD1 protein level. Reduced expression of PLD1 using siRNA also inhibited PMA-stimulated secretion of MMP-9. Furthermore, depletion of PLD decreased cell viability and increased apoptosis, thus suggesting the involvement of PLD in survival and anti-apoptosis.
In conclusion, we demonstrate selective induction of PLD1 in human colon cancer cells and mice by PMA. Taken together, our current study is, to our knowledge, the first to demonstrate that the induction of PLD1 through a PKC/Ras/ERK/NFB pathway is involved in the secretion of MMP-9 associated with invasiveness in colon cancer cells. In addition, our finding identifies PLD1 as a novel downstream target molecule of NFB, which contributes to tumorigenesis. Since PLD has emerged as a critical regulator of cell proliferation and invasion, the selective modulation of PLD expression levels in certain cancers might be useful in sensitizing resistant cancers to the effects of anti-cancer drugs. . Depletion of PLD increases apoptosis and reduces cell viability. HCT116 cells were transfected with control siRNA or siRNA specific for PLD1 or PLD2. A, cell viability was measured by a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay, as described under "Experimental Procedures." *, p Ͻ 0.05 versus either small interfering PLD transfection. B, apoptosis was measured using an Annexin V staining kit. Results shown are means Ϯ S.E. of four independent experiments. *, p Ͻ 0.05 versus either siPLD transfection.