12-O-Tetradecanoylphorbol-13-acetate and UV Radiation-induced Nucleoside Diphosphate Protein Kinase B Mediates Neoplastic Transformation of Epidermal Cells*

The molecular changes associated with early skin carcinogenesis are largely unknown. We have previously identified 11 genes whose expression was up- or down-regulated by 12-O-tetradecanoylphorbol-13-acetate (TPA) in mouse skin keratinocyte progenitor cells (Wei, S.-J., Trempus, C. S., Cannon, R. E., Bortner, C. D., and Tennant, R. W. (2003) J. Biol. Chem. 278, 1758-1768). Here, we show an induction of a nucleoside diphosphate protein kinase B (NDPK-B) gene in response to TPA or UV radiation (UVR). TPA or UVR significantly induced the expression of NDPK-B both in vivo hyperplastic mouse skin and in vitro mouse JB6 Cl 41-5a epidermal cells. Indeed, this gene was also up-regulated in TPA or UVR-mediated skin tumors including papillomas, spindle cell tumors, and squamous cell carcinomas, relative to adjacent normal skins. Functional studies by constitutive expression of nm23-M2/NDPK-B in TPA susceptible JB6 Cl 41-5a and TPA-resistant JB6 Cl 30-7b preneoplastic epidermal cell lines showed a remarkable gene dosage-dependent increase in foci-forming activity, as well as an enhancement in the efficiency of neoplastic transformation of these cells in soft agar but no effect on proliferation in monolayer cultures. Interestingly, stable transfection of the nm23-M2/NDPK-B del-RGD or G106A mutant gene in JB6 Cl 41-5a cells selectively abrogated NDPK-B-induced cellular transformation, implicating a possible Arg105-Gly106-Asp107 regulatory role in early skin carcinogenesis.

Mouse skin carcinogenesis is a complex multistage process that progresses through distinct stages of initiation, promotion, and progression to malignancy (1,2). The molecular changes associated with the early stages of skin tumor formation have yet to be determined. Tg⅐AC mice, which carry the coding region of the v-Ha-ras oncogene fused to a fetal -globin gene promoter (3), are considered to be genetically initiated and have a higher sensitivity to promotional stimuli including TPA 1 (3) and full thickness wounding (4), or carcinogens such as UV radiation (UVR) (5) and 7,12-dimethylbenz[a]anthra-cene (6). These features establish the in vivo Tg⅐AC mouse model as a valuable tool to study the early stages of skin carcinogenesis.
In an earlier study with a combination of fluorescence-activated cell sorting, switching mechanism at the 5Ј-end of RNA templates cDNA amplification, and mouse cDNA array technology, we identified 11 genes whose expression changed significantly in ␣6 ϩ CD34 ϩ keratinocytes harvested from TPAtreated mice relative to cells from untreated mice. Nine genes, including galectin-7, nm23-M2/NDPK-B, cytoskeletal epidermal keratin 14, deleted in split hand/split foot gene 1 (Dss1), DNA double strand break repair RAD21 homolog, transcription termination factor 1, thymosin ␤ 4 , calpactin I light chain, and 40 S ribosomal protein SA, were up-regulated, and two genes, apolipoprotein E precursor and acidic keratin complex 1 gene 15, were down-regulated by TPA (7). Dss1, a gene associated with a heterogeneous limb developmental disorder called split hand/split foot malformation (8), has recently been identified as a novel TPA-inducible gene expressed in keratinocyte progenitor cells, with possible involvement in early skin tumorigenesis (7). This novel approach was highly effective in the in vivo identification of TPA-inducible effector genes that might lead to neoplastic transformation. The protein kinase nm23-M2/ NDPK-B was another one of nine TPA-up-regulated genes and was selected for further characterization.
Nm23-H2, known as Nm23-M2 or NDPK-B (38), is a basic protein recently identified as the human PuF factor, a transcriptional activator of the c-myc proto-oncogene (39,40). Mutational analysis has identified residues and domains of Nm23-H2 that are involved in DNA binding, implicating a role in the regulation of genes important to cell proliferation, differentiation, and cancer development (40,41). In colorectal carcinomas, nm23-H2 is among the most abundant overexpressed transcript, suggesting that nm23-H2 helps maintain the malignant phenotype in these tumors (35). Recently, nm23-M2/NDPK-B was also identified as a novel potential disease locus that was involved in mouse leukemic transformation (42).
The biological functions of Nm23-M2/NDPK-B in cellular transformation still remain unknown. To gain insights into the contribution of this protein to early skin carcinogenesis, we characterized and studied the biological properties of Nm23-M2/NDPK-B. In this study, our data showed that Nm23-M2/ NDPK-B was significantly induced either in vivo animal models or in vitro cell cultures by TPA or UVR and appeared to have a critical role in mediating neoplastic transformation of epidermal cells in the early stages of skin carcinogenesis. Using site-directed mutagenesis analysis, we further identified an RGD consensus sequence domain site in Nm23-M2/NDPK-B that was involved in the potentiation of cellular transformation activity. Nm23-M2/NDPK-B represents an attractive candidate mediator of TPA-or UVR-induced tumor promotion.

Animals and Treatments
8 -10-week-old female homozygous Tg⅐AC mice were obtained from the Taconic Laboratory of Animals and Services (Germantown, NY). Animal studies were carried out in compliance with the National Institutes of Health Guidelines for Humane Care and Use of Laboratory Animals. The dorsal skin surface of groups of four homozygous female Tg⅐AC mice were dosed twice weekly for 2 weeks with 5 g of TPA in 200 l of acetone. Untreated control mice were sacrificed on day 1 (designated as NS). Four dosing protocols were used as follows. Mice were dosed on day 1 and sacrificed on day 5 (designated as 1TPA); mice were dosed on days 1 and 5 and sacrificed on day 8 (designated as 2TPA); mice were dosed on days 1, 5, and 8 and sacrificed on day 12 (designated as 3TPA); and mice were dosed on days 1, 5, 8, and 12 and sacrificed respectively at 48 h (designated as 4TPA), 7 days (designated as 4TPA ϩ 7 days), 14 days (designated as 4TPA ϩ 14 days), and 21 days (designated as 4TPA ϩ 21 days) after the last dose. TPA-induced papillomas and malignant tumors (spindle cell tumors and squamous cell carcinomas) were identified, removed, and characterized as described previously (6). Tg⅐AC mice were irradiated with combination of 30 -40% UVA and 60 -70% UVB, as described previously (5). After three exposures to UVA/UVB with 8.67 kJ/m 2 per exposure, total cumulative dose of 26 kJ/m 2 , skin tissues (designated as UV1-UV4) were collected 24 h after the last exposure. Some animals were held for papilloma development (designated as UVP1-UVP6), and tumors were collected and stored at Ϫ80°C.

RT-PCR
Single-stranded cDNA was prepared from total RNA using the Moloney murine leukemia virus reverse transcriptase SuperScript II (Invitrogen) with oligo(dT) primer and used as a template for PCR. PCR primers for mouse nm23-M2/NDPK-B were as described above. The forward and reverse primers of ␤ 2 -microglobulin gene (217 bp in size), used as an internal control, are 5Ј-GAC TGG TCT TTC TAT ATC CTG G-3Ј and 5Ј-CTT TCT GCG TGC ATA AAT TG-3Ј, respectively. PCR cycling was as follows: denaturation (94°C, 45 s), annealing (58°C, 45 s), and extension (72°C, 2 min) for 30 cycles. The reaction was carried out in a PerkinElmer-9600 thermal cycler, and PCR products were analyzed using 2% agarose gels. DNA was quantified using Quantity One software version 4.0 (Bio-Rad).

In Situ Hybridization
An in situ hybridization assay was performed as previously described (45). Briefly, skin tissues were removed from Tg⅐AC mice treated or untreated with multiple doses of TPA and fixed overnight in 10% neutral buffered formalin. Tissues were paraffin-embedded, and sections (6 m) were cut onto SuperFrost Plus microscope slides (Daigger, Vernon Hills, IL). The sections were deparaffinized and rehydrated by successive washes in xylene and graded alcohols to 2ϫ SSC, and then 2 ϫ 10 6 cpm of [␣-35 S]UTP-labeled mouse nm23-M2/NDPK-B sense or antisense riboprobes was applied to slides. Riboprobes were prepared from T7/T3-U19/nm23-M2/NDPK-B (1053321) plasmid linearized with EcoRI (antisense) or HindIII (sense) using an in vitro T7 or T3 Riboprobe kit (Promega, Madison, WI). Following 40°C overnight hybridization, tissues were washed in 2ϫ SSC plus 50% formamide at 40°C and then in 2ϫ SSC, 1ϫ SSC, 0.5ϫ SSC, and 0.5ϫ SSC for 30 min each wash at room temperature. To remove unbound probe, tissues were incubated with 20 l of RNase (10 mg/ml). After several washes, the slides were dehydrated in graded alcohols and completely air-dried. The slides were then dipped into NTB-3 autoradiographic emulsion (Eastman Kodak Co.), exposed for 10 days at room temperature in the dark, dried in a light-tight container, and developed in Kodak D19 developer and fixer. The sections were counterstained with hematoxylin, covered with coverslips, and photographed under dark-field illumination (model BX51, Olympus Optical Co., Tokyo, Japan).

Immunoblot Analysis
Cells were washed with ice-cold phosphate-buffered saline and lysed in ice-cold modified radioimmunoprecipitation (RIPA) buffer consisting of 50 mM Tris-HCl (pH 7.4), 1% Nonidet P-40, 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml each aprotinin, leupeptin, and pepstatin, 1 mM Na 3 VO 4 , and 1 mM NaF. Cell suspensions were gently rocked on an orbital shaker in a cold room for 15 min to lyse cells. Lysates were centrifuged at 14,000 ϫ g for 15 min at 4°C. The skin tissues or tumors were homogenized and sonicated in ice-cold RIPA buffer and ultracentrifuged at 100,000 ϫ g for 1 h at 4°C. Protein concentration was determined by Bradford assay (Bio-Rad). Proteins were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride membranes (Amersham Biosciences). Membranes were stained with primary antibodies, detected using horseradish peroxidase-conjugated secondary antibodies (1:3000) (Amersham Biosciences) and enhanced chemiluminescence (Amersham Biosciences). Membranes were stripped and rehybridized with anti-␣-tubulin mouse mAb (1:1000) or anti-actin rabbit pAb (1:1000) as a control to confirm equal loading. Protein quantitation was determined by ImageQuant software version 5.1 (Amersham Biosciences), and relative quantity is shown below the panels of Figs. 3B, 4 -6, and 8.
Characterization of Cell Growth-Growth curves were generated as described previously (46). In brief, 2 ϫ 10 3 cells were grown as described above. The medium was changed every 3-4 days. Cell number was counted in triplicate on a hemocytometer every other day for 14 days.
Anchorage-independent Growth Assay-Colony formation in soft agar was assayed as described previously (46). In a 60-mm tissue culture dish, 1 ϫ 10 4 cells were resuspended in 0.33% Noble agar in Eagle's MEM with 10% FBS and layered over 5 ml of 0.5% agar in Eagle's MEM with 10% FBS. Cells were grown at 37°C in 95% air plus 5% CO 2 , and colonies with more than eight cells were counted and photographed 18 days postseeding.

nm23-M2/NDPK-B Expression Is Induced following TPA
Treatment-Our previous microarray data indicated that mouse nm23-M2/NDPK-B was increased 5.4-fold in integrin ␣ 6 ϩ CD34 ϩ keratinocytes isolated from TPA-treated Tg⅐AC mice skins (7). We have reported previously that keratinocytes expressing ␣ 6 and CD34 surface markers represent a subpopulation of follicle-derived cells exhibiting properties of progenitor cells (47). This up-regulation was confirmed by relative RT-PCR analysis that revealed a 3.5-fold increase in integrin ␣ 6 ϩ CD34 ϩ keratinocytes isolated from TPA-treated Tg⅐AC mice, relative to integrin ␣ 6 ϩ CD34 ϩ keratinocytes from untreated mice (Fig. 1). The trend of nm23-M2/NDPK-B overexpression was in agreement with our previous microarray experiment (7). To define the specificity of tissue distribution of nm23-M2/ ؉ CD34 ؉ keratinocytes. TPA-treated Tg⅐AC mouse skin keratinocytes carrying the cell surface markers integrin ␣ 6 and CD34 were isolated by fluorescence-activated cell sorting (47). Control cells were harvested from animals not treated with TPA. RNase-free DNase Itreated total RNAs from TPA-treated (TPA(ϩ)␣6 ϩ CD34 ϩ ) or -untreated integrin ␣ 6 ϩ CD34 ϩ keratinocytes (TPA(Ϫ)␣6 ϩ CD34 ϩ ) were assayed by RT-PCR, as described under "Experimental Procedures." The products were subjected to 2% agarose gels. Mouse nm23-M2/NDPK-B plasmid DNA was used as a positive control (PC) and showed a band of 456 bp in size. The ␤ 2 -microglobulin gene (217 bp in size) was used as an internal control. DNA was quantified using Quantity One software version 4.0 (Bio-Rad). V ANDPK-B, a panel of tissues from normal Tg⅐AC mice was examined by immunoblot analysis using a rat monoclonal antibody specific to mouse Nm23-M2/NDPK-B, with no crossreactivity with mouse Nm23-M1/NDPK-A molecule (27). As shown in Fig. 2, the specific signal was detected under reducing conditions as a band of 17.5-kDa in size, which was seen in varying intensities in heart, brain, kidney, liver, skin, stomach, spleen, and small intestine. Nm23-M2/NDPK-B was expressed at higher levels in heart and kidney. In liver, spleen, stomach, small intestine, brain, and skin the expression levels of nm23-M2/NDPK-B were very low, whereas Nm23-M2/NDPK-B was barely detectable in lung and ovary. To determine the localization of nm23-M2/NDPK-B message in mouse skin tissues, in situ hybridization was employed. Although some nonspecific background was apparent in the outermost cornified cell envelope, the signal primarily localized to the stratified squamous epithelial regions (indicated by the arrows in Fig. 3) with antisense but not sense [␣-35 S]UTP-labeled mouse nm23-M2/ NDPK-B mRNA. Some expression was evident in hair follicles, especially in the third dose of TPA-treated skins in Fig. 3A (XI). Signal was not detectable in the dermis, adipose, and muscle tissues (Fig. 3A). However, there was no detectable nm23-M2/ NDPK-B mRNA message in normal skins (Fig. 3A, II), relative to the sense probe control (Fig. 3A, III). Total protein lysates of skins harvested from Tg⅐AC mice treated with various doses of TPA were used for immunoblot analysis. These doses induce extensive hyperplasia (see Fig. 3A; IV, VII, and X). Fig. 3B revealed an induction of Nm23-M2/NDPK-B in mouse skin homogenates following TPA treatment. Untreated mouse skin was used as a negative control and expressed very low levels of Nm23-M2/NDPK-B (Fig. 3B, lane 1). Mouse nm23-M2/ NDPK-B expression was induced 2.5-fold following one dose of TPA (5 g), with a maximal induction of about 8 -10-fold between the second and the fourth dose of 5-g TPA treatment (Fig. 3B). Nm23-M2/NDPK-B increased with hyperplasia and was maintained at high levels (3.8-fold) even 21 days after the last of four doses of TPA (Fig. 3B). In contrast, Nm23-M1/ NDPK-A was not significantly up-regulated in TPA-induced hyperplastic skin tissues (data not shown).
Overexpression of nm23-M2/NDPK-B in Skin Tumor Cell Lines and Skin Tumors-In vivo and in vitro studies showed that TPA was able to induce an increase in nm23-M2/NDPK-B gene expression in Tg⅐AC mice keratinocytes, in hyperplastic Tg⅐AC mice skins, and in TPA-susceptible mouse JB6 Cl 41-5a epidermal cells. The gene expression of nm23-M2/NDPK-B was also examined by immunoblot analysis in the mouse skin tumor cell lines as well as in TPA-induced skin tumors. The protein level of Nm23-M2/NDPK-B was found to respectively increase 10-and 2.8-fold in two mouse skin tumor cell lines, Tg⅐AC43 and FVB/N217 (relative to TPA-untreated normal keratinocytes isolated from Tg⅐AC mice) (TPA(Ϫ)KCs) (Fig.  5A). In addition, NDPK-B expression was also increased in one human epidermoid carcinoma cell line A431 (5.0-fold), when compared with human HaCaT keratinocytes (Fig. 5A). The keratinocytes isolated from the skins of Tg⅐AC mice treated with four doses of TPA (TPA(ϩ)KCs) were used as a positive control and showed a 3.5-fold induction (Fig. 5A, lane 2). Nm23-M2/NDPK-B protein was higher in TPA-mediated Tg⅐AC mice skin tumors, including 15 papillomas (3.1 Ϯ 1.0-fold) and three malignant tumors (3.7 Ϯ 1.8-fold) (one spindle cell tumor (5.8fold) and two squamous cell carcinomas (2.7-and 2.6-fold)), than in adjacent normal skins (1.0 Ϯ 0.3-fold) (Fig. 5B).
The nm23-M2/NDPK-B Gene Is Induced by UVR-To examine whether nm23-M2/NDPK-B is induced in response to UVR in vivo, we irradiated Tg⅐AC mice with a combination of UVA and UVB (30 -40% UVA and 60 -70% UVB). After three UVA/ UVB exposures, the skin was found to be extensively hyperplastic, keratinized, and inflamed (data not shown). The UVexposed tissues were collected 24 h after the last exposure and snap frozen in liquid nitrogen for preparation of total protein lysates. Immunoblot analysis revealed that Nm23-M2/ NDPK-B was induced by 7.2-10-fold in these four mouse skin tissues following UV exposure (designated as UV1-UV4), relative to the low level expression found in nonexposed skin (designated as NS) (Fig. 6A). To further investigate if this induction persistently occurred in UVR-mediated skin tumors, six papillomas were examined. Our result indicated that Nm23-M2/NDPK-B was increased 7-10-fold in UVR-mediated skin papillomas (designated as UVP1-UVP6), relative to untreated normal Tg⅐AC mice skins (NS) (Fig. 6B).
Constitutive expression of Ras family proteins and other oncogenic proteins increase focus-forming capability and decrease growth contact inhibition of normal untransformed cells (46). Our results showed that constitutive expression of mouse nm23-M2/NDPK-B increases foci formation 3.8-(2 g of DNA) to 4.5-fold (4 g of DNA) and 4.0-(2 g of DNA) to 5.3-fold (4 g of DNA) in mouse JB6 Cl 30-7b and JB6 Cl 41-5a epidermal cell lines, respectively (Fig. 7B). In addition, the increase in fociforming activity appears to have gene dosage-dependent effects (Fig. 7B). However, expression of mouse nm23-M2/NDPK-B did not change the foci-forming properties in fibroblasts such as NIH/3T3 and Rat-1 cells. These results demonstrate that nm23-M2/NDPK-B alters normal contact inhibition in mouse epidermal cell lines, suggesting that nm23-M2/NDPK-B may have some oncogenic properties.
Transformed cells have a growth advantage in monolayer culture and acquire capacity for anchorage-independent growth (46). The effects of mouse nm23-M2/NDPK-B expression on these growth characteristics were measured in NIH/ 3T3, Rat-1, JB6 Cl 30-7b, and JB6 Cl 41-5a cells (Fig. 7C). The cells transfected with vector only or mouse nm23-M2/NDPK-B were grown for 48 h and then selected for at least 10 days with medium containing 400 g/ml of antibiotic neomycin analog G418. Stable cell clones expressing vector only or mouse nm23-M2/NDPK-B were seeded at a cell density of 2 ϫ 10 3 in a 6-well tissue culture plate to assay for cell growth. Constitutive expression of mouse nm23-M2/NDPK-B was not correlated with increased growth rate in epidermal cells such as JB6 Cl 30-7b and JB6 Cl 41-5a. On the contrary, our results showed that clones overexpressing mouse nm23-M2/NDPK-B in NIH/3T3 and Rat-1 fibroblasts grew at a significantly slower rate than control vector-only cells (Fig. 7C). Furthermore, stable clones expressing vector only or mouse nm23-M2/NDPK-B were seeded at 1 ϫ 10 4 in a 60-mm soft agar tissue culture plate to assay for anchorage-independent growth. Colony formation efficiency increased ϳ3.3and 4.0-fold when mouse nm23-M2/ NDPK-B was expressed in JB6 Cl 41-5a and JB6 Cl 30-7b cells, respectively (Fig. 7D). However, colony forming efficiency did not increase in mouse nm23-M2/NDPK-B-transfected fibroblasts such as NIH/3T3 and Rat-1 cells (Fig. 7D). Background colony formation was higher in JB6 Cl 41-5a cells than in JB6 Cl 30-7b cells.
An Increased Release of nm23-M2/NDPK-B from Cells following Exposure to TPA or UVR-Next we examined whether Nm23-M2/NDPK-B is released from the cells and is affected by stresses such as TPA and UVR. For these studies, TPA-susceptible mouse epidermal cell line JB6 Cl 41-5a and human keratinocyte cell line HaCaT, which is routinely used in UVR studies, were employed. Prior to TPA or UV exposure, subconfluent JB6 Cl 41-5a and HaCaT cell monolayers were starved in serum-free medium for at least 12 h. As seen in Fig. 8, Nm23-M2/NDPK-B was detected in serum-free conditioned media collected from untreated 18-h cultures of JB6 Cl 41-5a and HaCaT cells. At this time, cells were observed to still be viable, as ascertained by examination of trypan blue dye exclusion using an inverted microscope (data not shown). The result indicates that Nm23-M2/NDPK-B is released from the cells. More interestingly, the release of Nm23-M2/NDPK-B into the medium was enhanced about 2.5-3-fold in JB6 Cl 41-5a (Fig.   8A) and HaCaT cells (Fig. 8B) 18 h following exposure to TPA (10 ng/ml), UVA (10 J/cm 2 ) or UVB (10 mJ/cm 2 ). The increased release of Nm23-M2/NDPK-B may explain why the expression of NDPK-B in JB6 Cl 41-5a cells was significantly down-regulated 12-18 h after treatment with 10 ng/ml of TPA (Fig. 4B).

DISCUSSION
In this study, we further characterize the biological functions of the NDPK-B gene product using the in vitro JB6 mouse epidermal clonal genetic variant cell system that has proven to be valuable in studying tumor promoter-dependent biological events occurring during preneoplastic progression (48). In this way, we first demonstrate that NDPK-B is a potentially important TPA-or UVR-responsive gene required for neoplastic transformation in epidermal cells.
Our RT-PCR data showed a significant up-regulation in mouse nm23-M2/NDPK-B gene expression in integrin ␣ 6 ϩ CD34 ϩ keratinocytes following TPA treatment (Fig. 1). This induction is closely associated with the TPA-and UVR-induced promotion stage of skin carcinogenesis in mice (Figs. 3 and 6A). Immunoblot analysis also showed an in vitro TPA-induced increase in Nm23-M2/NDPK-B protein level in JB6 Cl 41-5a preneoplastic epidermal cells (Fig. 4, A and B). In addition, a dramatic increase in gene expression of NDPK-B occurs consistently in skin tumor cell lines (Fig. 5A) and in TPA-or UVR-mediated mouse skin tumors as well (Figs. 5B and 6B). These results clearly suggest that nm23-M2/NDPK-B is a novel TPA-or UVR-responsive gene that may be a useful marker for early skin tumorigenesis.
Although our results indicate a close correlation between nm23-M2/NDPK-B gene expression and the tumor promotion stage (induced by TPA or UVR), the specific functions of nm23-M2/NDPK-B in this process, such as proliferation, differentiation, or neoplastic transformation, remains unknown. Earlier reports identified NDPK-B as a differentiation-inhibitory factor, which is able to inhibit the differentiation of several hematopoietic cell lines in vitro, where the inhibition was independent of the phosphotranferase activity, as demonstrated with NDPK-B mutants lacking the enzymatic activity (14,49). Moreover, several hematopoietic tumor cell lines stained positive for NDPK-B in flow cytometric analysis (50), and upregulation of the human NDPK-B gene was also observed in normal lymphocytes induced to proliferate with phytohemagglutinin (51). Down-regulation of the NDPK-B and c-myc genes was also reported in 1,25-dihydroxyvitamin D 3 -induced differentiation (52). These results suggest that NDPK-B may play a critical role in the inhibition of differentiation or the promotion of proliferation of these cells. As seen in Fig. 7, we found that nm23-M2/NDPK-B was efficiently expressed in NIH/3T3, Rat-1, JB6 Cl 30-7b, and JB6 Cl 41-5a (Fig. 7A). The elevated expression of mouse nm23-M2/NDPK-B in genetically modified NIH/3T3, Rat-1, JB6 Cl 30-7b, and JB6 Cl 41-5a stable clones, which were transfected with pcDNA3.1D/nm23-M2/NDPK-B mammalian expression vector and selected with G418 for at least 10 days, does not significantly promote cell proliferation under standard in vitro tissue culture conditions (Fig. 7C) but appears to markedly increase foci-forming activity in epidermal cell lines (e.g. JB6 Cl 30-7b and JB6 Cl 41-5a) in a gene dosage-dependent manner but not in fibroblast cell lines (e.g. NIH/3T3 and Rat-1) (Fig. 7B). Notably, we found the rate of cell growth to be selectively inhibited in nm23-M2/NDPK-B-overexpressing mouse fibroblasts, NIH/3T3 and Rat-1. Our data consistently showed a decrease of proliferation rate from 100% down to about 1% (NIH/3T3) or 50% (Rat-1) and implicated that more than 50% of the growth inhibition observed in fibroblasts was mediated by activation of Nm23-M2/NDPK-B-initiated signaling pathways (Fig. 7C), presumably suggesting that NDPK-B plays an important role in the maintenance of integrity of skin tissues. More importantly, overexpression of mouse nm23-M2/NDPK-B increased colony-forming efficiency in JB6 Cl 41-5a and JB6 Cl 30-7b cells but not in NIH/3T3 and Rat-1 cells using anchorage-independent growth assay (Fig. 7D). Thus, the ability of nm23-M2/NDPK-B to regulate cellular transformation may be specific for epithelial cells.
NDPKs have previously been reported to localize at different subcellular localizations, such as nucleus (53), cytoplasm (54), and mitochondria (37), with low levels in the plasma membrane (50). Previous studies have already revealed that inhibitory factor was purified from the conditioned medium of the mouse myeloid cell line M1, although NDPKs have no signal peptide for their secretion (49,55). Anzinger et al. (56) recently found secretion of NDPK-B by cells isolated from human breast, colon, pancreas, and lung tumors. In addition, Willems et al. (57,58) also demonstrated that extracellular NDPK modulated normal hematopoietic cell differentiation. Our data have shown NDPK-B release into the serum-free media of mouse JB6 Cl 41-5a epidermal cells and an about 2.5-3.0-fold increase in release levels 18 h following TPA treatment (Fig. 8). All previously identified NDPKs, except that in bacteria, share a specific tripeptide Arg-Gly-Asp (RGD) domain (27). To further determine whether the RGD domain of NDPK-B protein is required for the enhancement of neoplastic transformation in epidermal cells, site-directed mutagenesis was employed to create mutants for deletion or point substitution. Immunoblot analysis showed that Nm23-M2/NDPK-B proteins were released into the serum-free conditioned media collected from JB6 Cl 41-5a individual stable clones, which were transfected with vector only, wild type (WT), delRGD, G106A, S122P, or H118F nm23-M2/NDPK-B gene (Fig. 9A). The elevated expression of mouse nm23-M2/NDPK-B (WT) into JB6 Cl 41-5a cells acquired the susceptibility to transformation in an in vitro soft agar assay. Indeed, constitutive expression of mouse nm23-M2/ NDPK-B (delRGD) or nm23-M2/NDPK-B (G106A) mutant gene in JB6 Cl 41-5a cells selectively abolished NDPK-B-induced anchorage-independent growth (Fig. 9, B and C). Thus, the enhancement of cellular transformation requires the presence of the RGD consensus sequence domain. Previous reports indicated that the point substitutions of nm23-H2/NDPK-B at His 118 with Phe (H118F) and at Ser 122 instead of Pro (S122P) result in a defective kinase activity and phosphoryl transfer activity, respectively (59,60). Interestingly, overexpression of mouse nm23-M2/NDPK-B (S122P) (lack of phosphoryl transfer activity) and nm23-M2/NDPK-B (H118F) (lack of kinase activity) did not significantly reduce the efficiency of cellular transformation and suggests that NDPK-B kinase activity is not required for transformation (Fig. 9, B and C). Taken together, these results point to a crucial regulatory role of the RGD consensus sequence domain, but not the catalytic domain, in NDPK-B-mediated cellular transformation.
Although NDPK-B protein can be released into the medium in our cell model systems, we cannot exclude the possible effects of NDPK-B in the nucleus. Nm23-H2/NDPK-B has recently been identified as the human PuF factor (40) and can transactivate a human c-myc proto-oncogene via a functional nuclease hypersensitive element (39). Activation of the c-myc proto-oncogene contributes to cellular transformation, mitogenesis, differentiation, and apoptosis (61)(62)(63). Recent reports also indicated that c-Myc promotes differentiation of human epidermal stem cells (64), and c-myc activation in transgenic mouse epidermis results in mobilization of stem cells and differentiation of their progeny (65). Moreover, the constitutive expression of human c-myc has been demonstrated to deplete epidermal stem cells by reducing ␤ 1 integrin expression (66). Mutational analysis of human Nm23-H2/NDPK-B has identified amino acid residues and structural domains that are involved in DNA binding, implicating a multifunctional role in the regulation of genes (e.g. c-myc) important to cell proliferation, differentiation, and cancer development (41,67). Research is in progress to further identify how NDPK-B, integrin, and c-Myc coordinately regulate cell growth, differentiation, and transformation in epidermal stem cells following treatment with TPA.
Due to inefficiency of the antibody for immunohistochemical staining in mouse skins, we used in situ hybridization to localize the nm23-M2/NDPK-B mRNA message. Our findings on the localization of nm23-M2/NDPK-B mRNA in homozygous Tg⅐AC mice skin tissues following treatment with variable doses of TPA showed a dramatically increased expression in the stratified squamous epithelial regions, in particular in basal layer and spinosum, and with some expression in hair follicles (Fig.  3A). Indeed, the induction of nm23-M2/NDPK-B mRNA seems to occur in a TPA dose-dependent manner, consistent with immunoblot analysis (Fig. 3B). Unlike immunoblot analysis, in situ hybridization did not detect the nm23-M2/NDPK-B signal in normal mouse skin tissues, probably in part due to the relative insensitivity of in situ hybridization. Interestingly, our previous study indicates that the expression patterns of Dss1 and Nm23-M2/NDPK-B were similar in respect to tissue distribution and localization (7). Moreover, recently reported data indicate that human Nm23-H2/NDPK-B binds to singlestranded oligonucleotides in a non-sequence-specific manner but that it exhibits strong specificity for single-stranded DNA (68). One Dss1 functional model, as shown by its 37% acidic residue (aspartic and glutamic acid-rich domains, Ϫ21 charge at pH 7.0) and its 13% aromatic residue content, suggests that it mimics oligonucleotides, possibly regulating the accessibility of a subset of the putative DNA binding sites on the helical and OB1 domains of BRCA2 (69). Recently, a Dss1/BRCA2 interaction was shown to be required for proficiency in DNA repair, recombination, and genome stability in the fungi Ustilago maydis (70). Postel et al. (18) also revealed that catalysis of DNA cleavage and nucleoside triphosphate synthesis by human FIG. 9. Deletion of Arg 105 -Gly 106 -Asp 107 consensus sequence domain or point substitution Gly 106 with Ala in mouse Nm23-M2/ NDPK-B selectively abrogates cellular transformation. A, immunoblot analysis. The deletion or point substitution was introduced into mouse nm23-M2/NDPK-B cDNA by the ExSite TM PCR-Based site-directed mutagenesis kit, as described under "Experimental Procedures." These mutants include delRGD, G106A, S122P, and H118F. JB6 Cl 41-5a cells were transfected with pcDNA3.1 vector only (Vector), wild type nm23-M2/NDPK-B (WT), or mutant types nm23-M2/NDPK-B including delRGD, G106A, S122P, and H118F plasmid DNAs. To obtain stable clones, transfected cells were selected under 400 g/ml G418 for at least 10 days. Serum-free conditioned media were collected and concentrated from stable clones and analyzed by immunoblotting to confirm the release of mouse NDPK-B proteins. B, anchorage-indepen-dent growth assay. JB6 Cl 41-5a-stable cell clones were seeded at a density of 1 ϫ 10 4 into 0.33% soft agar over a 0.5% agar bottom layer. Cells were grown at 37°C in 95% air plus 5% CO 2 , and a colony with more than eight cells was counted 18 days after seeding. The data represent an average of two experiments. Nm23-H2/NDPK-B share an active site, thus implicating a DNA repair function for the NDPK-B molecule. It raises a possibility that Dss1, in addition to binding the DNA repair protein BRCA2 (71), could interact with the DNA repair protein Nm23-M2/NDPK-B in mouse skins following exposure with TPA. Although we have found that Dss1 binds directly with Nm23-M2/NDPK-B in in vitro studies using TNT Quick Coupled Transcription/Translation Systems, 2 further studies will be necessary before the significance of this unique proteinprotein interaction can be fully understood.
In summary, our studies provide a new insight into the biological function of mouse Nm23-M2/NDPK-B and demonstrate that it is likely to play a unique role in mediating TPAor UVR-induced skin carcinogenesis. NDPK-B was capable of being released outside of cells in culture, and an increase in release was observed in response to TPA and UVR. This protein kinase nm23-M2/NDPK-B was also induced in hyperplastic mouse skins and in TPA-mediated mouse skin tumors. In addition, constitutive expression of mouse nm23-M2/NDPK-B in preneoplastic epidermal cells not only promoted foci-forming activity but also enhanced the process of neoplastic transformation in these cells. More interestingly, stable transfection of the mouse nm23-M2/NDPK-B (delRGD) or nm23-M2/NDPK-B (G106A) mutant gene in JB6 Cl 41-5a cells selectively abrogated NDPK-B-induced cellular transformation, implicating a possible RGD regulatory role in early skin carcinogenesis.