Expression patterns of programmed cell death 4 protein in normal human skin and some representative skin lesions


Dr Sachiko Matsuhashi, Division of Hepatology and Metabolism, Department of Internal Medicine, Saga Medical School, Saga University, 5-1-1 Nabeshima, Saga 849-8501, Japan, Tel.: +81 952 34 2362, Fax: +81 952 34 2017, e-mail:


Abstract:  Expression of a tumor suppressor gene, programmed cell death 4 (PDCD4), was investigated at the protein level in the human skin. Immunohistochemically, PDCD4 protein expressed mainly in suprabasal layers, while PDCD4-positive and -negative areas were observed discontinuously in the basal cell layer of the epidermis. In hair follicles, the suprabulbar area including the hair and inner root sheath was immunoreactive, while the bulbar area, containing germinative cells which were strongly proliferating cell nuclear antigen (PCNA)-positive, was not or less. PDCD4 therefore appears to be important in the differentiation of hair follicles. PDCD4-positive cells were localized in the inside layers while PCNA-positive cells were located in the basal layer in the outer root sheath of hair follicles. The cells of sebaceous glands and sweat glands also were PDCD4-positive. The PDCD4 protein was localized mostly in nuclei of cutaneous cells. PDCD4 expression was found to be suppressed in the epidermis overlying an adult T-cell lymphoma (ATL), possibly reflecting a paracrine effect of factors produced by ATL cells. PDCD4 expression was suppressed in the keratinocyte cell line HaCaT by exposure of cultures to epidermal growth factor, transforming growth factor-β1 or hepatocyte growth factor. Immunohistochemically, various skin cancers tended to show less PDCD4 expression than normal skin. Promotion of expression might prove useful in preventing or treating certain skin cancers.


First isolated during screening of a glioma cDNA library with the monoclonal antibody Pr-28 recognized an antigen involved in the cell cycle (1,2), the human programmed cell death 4 (PDCD4) gene (H731) mapped at 10q24 (3). The mouse homologue, Pdcd4, was isolated as a gene that was upregulated in association with apoptosis (4) and downregulated by treatment of cells with topoisomerase inhibitors (5). PDCD4/Pdcd4 expression was shown in many systems to be upregulated when apoptosis was induced (4,6,7).

Carcinogenesis involves initiation and promotion of excessive cell growth. Potent promoters in epidermal cell carcinogenesis include 12-O-tetradecanoylphorbol 13-acetate (TPA), epidermal growth factor (EGF) and reactive oxygen species (8,9). Pdcd4 was shown to suppress the transformation of JB6 mouse epidermal cells exposed to promoters (10,11). Recently the PDCD4 protein was shown to be suppressed in some lung cancers relative to expression in normal lung tissues; relatively high PDCD4 expression in tumors was associated with tumor slower progression and better prognosis (12). The protein levels also decreased in hepatoma tissues compared with the corresponding normal liver tissues (7). Jansen et al. examined PDCD4 expression levels in the NCI60 cells (National Cancer Institute drug-screening panel of 60 human cancer cells) and reported that expression was reduced in may cell lines and most frequent reduction was found in the cell lines derived from lung, kidney and central nervous system tumors (13).

At the molecular level, Pdcd4-protein was found to inhibit activator protein-1 (AP-1) transactivation activities (11,14). The transcription factors AP-1 and nuclear factor-kB play important roles in the promotion and progression phases of epidermal carcinogenesis (15). PDCD4/Pdcd4 also was shown to possess MA-3 domains homologous to the M1 domain of the translation initiation factor elF4G (16). The Pdcd4 protein associates with elF4A, which binds to elF4G in the initiation complex; this may inhibit the RNA helicase activity of elF4A, thereby inhibiting cap-dependent translations (17,18). On the other hand, human PDCD4 was shown to bind to the protein component of 40S ribosome S13 (RPS13) and the initiation factor elF4G (19).

PDCD4 protein is expressed ubiquitously, and has been localized to both cytoplasm and nucleus depending on cell type and conditions. In most cancer cell lines PDCD4 protein is detected in the cytoplasm, but PDCD4 also has been localized in tumor cell nuclei of some breast adenocarcinomas as well as in nuclei of cells making up normal small ducts of the breast (2). Expression patterns of PDCD4 in human skin have not yet been described. We immunohistochemically surveyed expression of the protein in normal human skin and in some cutaneous lesions.



Normal hair-bearing skin samples including some from the scalp were taken from the margins of excisional skin biopsy specimens obtained during routine surgical procedures. Several kinds of benign and malignant cutaneous tumors also were analysed. All procedures were performed with informed consent at Saga University Hospital.

Cell lines and cultures

HaCaT cells were provided by Dr N. E. Fusenig of the German Cancer Research Center, Heidelberg Germany, who established the cell line (20). Cells were cultured with Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (FBS) according to the protocol provided by Dr Fusenig. A431 and DJM-1 epidermal cancer cell lines from our laboratory stock also were cultured in DMEM with 10% FBS.


Anti-PDCD4 antibody was prepared by immunizing rabbits with recombinant H731 (PDCD4 isoform)-protein and partially purified as described previously (2). Anti-β-actin antibody was purchased from Biomedical Technologies (Stoughton, MA, USA). Anti-proliferating cell nuclear antigen (PCNA) monoclonal antibody was the product of Santa Cruz Biotechnology (Santa Cruz, CA, USA).

Immunostaining of tissues and cells

Deparaffinized sections cut from routinely fixed and embedded specimens were stained immunohistochemically with anti-PDCD4 antibody and anti-PCNA antibody using either an avidin–biotin complex (ABC) kit (DakoCytomation, Glostrup, Denmark) by the ABC peroxidase staining technique, or the Dako Envision system.

Cells were cultured on a Lab-Tek II chamber slide (Nalge Nunc International, Naperville, IL, USA), fixed for 10 min at room temperature with 4% paraformaldehyde and then treated for 1 min with cold acetone. The slide with fixed cells was stained with anti-PDCD4 antibody using the Dako ABC kit as described above.

Western blotting

Protein was extracted from tissues and cells in a lysis-buffer containing 50 mm Tris–HCl, pH 7.5, 150 mm NaCl, 1% sodium dodecyl sulphate (SDS), 1 mm phenylmethylsulfonyl fluoride, 10 μg/ml soybean trypsin inhibitor and 50 mm iodoacetamide. Amounts of total protein were determined with a protein assay kit (Bio-Rad, Hercules, CA, USA) using bovine serum albumin as a standard. NuPAGE sample buffer (Novex, San Diego, CA, USA) was mixed with 40 μg protein of each sample. Electrophoretic separation was carried out on an SDS-10% polyacrylamide gel. Bands in the gels were transferred electrophoretically to a polyvinylidene difluoride membrane (Bio-Rad). The membrane then was blocked by incubation with 5% skim milk in phosphate-buffered saline containing 0.1% Tween 20 for 1–2 h at room temperature, followed by incubation with primary antibody for 30 min at room temperature. Specific bands were visualized by further incubation with horseradish peroxidase-conjugated secondary antibody, followed by enhanced chemiluminescence detection (ECL system; Amersham, Buckinghamshire, UK) according to the manufacturer's instructions. The rabbit polyclonal anti-β-actin antibody (Biomedical Technologies) was used as an internal control.


PDCD4 expression in keratinocytes and epidermal carcinoma cell lines

To investigate PDCD4 expression in cells we developed an anti-PDCD4 antibody that recognized a 65-kDa protein in epidermal tissue by Western blotting (Fig. 1a, lane 1). The 65-kDa protein band disappeared when antigen-absorbed anti-PDCD4 antibody was used as described previously (2).

Figure 1.

 PDCD4-protein expression in epidermal tissues and cell lines. (a) Western blot analysis. Twenty micrograms of protein from each sample was analysed using anti-PDCD4 antibody after electrophoresis and membrane transfer. Extracts from the human epidermal tissue (lane 1), HaCaT cells (lane 2), A431 cells (lane 3) and DJM-1 cells (lane 4) were studied. The anti-PDCD4 antibody recognized a p65 band in the extracts. (b) Immunostaining of cell lines. HaCaT (i), A431 (ii) and DJM-1 (iii) cells were subjected to immunostaining with anti-PDCD4 antibody.

Intensity of PDCD4 expression according to Western blotting in a normal human keratinocyte cell line, HaCaT, was essentially the same as in epidermal tissue samples (Fig. 1a, lane 2), while expression was very slight in the epidermal carcinoma cell lines A431 and DJM-1 (Fig. 1a, lanes 3 and 4). As shown in Fig. 1b, HaCaT cells (Fig. 1b-i) were stained while A431 (Fig. 1b-ii) and DJM-1 (Fig. 1b-iii) cells stained rarely, in agreement with the results of Western blot analyses. Staining patterns of HaCaT cells were heterogeneous, mixing strongly stained cells and no or less stained cells. Staining of HaCaT cells was evident no longer when the antibody was absorbed with antigen protein.

PDCD4 expression in human skin

PDCD4 expression in human skin was localized immunochemically using anti-PDCD4 antibody. PDCD4 protein was detected particularly in the suprabasal cell layer, with less staining observed in the upper cell layer of the epidermis (Figs 2a and 3a). The protein was localized mostly in nuclei. The cornified layer was not stained, while basal cells were stained discontinuously.

Figure 2.

 Staining patterns obtained with anti-PDCD4 antibody in skin. a, Facial skin. PDCD4-positive and negative areas are present in the basal cell layer, while cells of suprabasal layers are stained diffusely. b, Scalp skin. The bulbar area, with germinative cells including proliferating and undifferentiated cell layers is less or unstained, while the suprabulbar area, with differentiating cells, shows staining. c, Scalp skin. Nuclei of outer root sheath (designated O) are PDCD4-positive. d, Sweat gland. Nuclei are well stained. SG, sebaceous gland; I, inner root sheath; H, hair; G, germinative cell layer; P, dermal papilla.

Figure 3.

 Staining patterns of serial sections obtained by simultaneous staining with either anti-PDCD4 antibody (a, c and e) or anti-PCNA antibody (b, d and f) in skin. a and b, Scalp skin was stained with anti-PDCD4 antibody (a) or with anti-PCNA antibody (b). PCNA-positive cells are localized in the basal and lower suprabasal layers of epidermis. Arrows in a show melanocytes. c and d, Hair follicle of scalp skin stained with anti-PDCD4 antibody (c) or with anti-PCNA antibody (d). Matrix cells are strongly stained with anti-PCNA antibody but rarely stained with anti-PDCD4 antibody. e and f, Hair follicle of scalp skin stained with anti-PDCD4 antibody (e) or with anti-PCNA antibody (f). PCNA-positive cells are localized in the outside layers and PDCD4-positive cells in the inside layers of the outer root sheath (ORS). P, dermal papilla, G, germinal cells, SG, sebaceous gland, O, ORS, H, hair.

Most of the outer root sheath (ORS) of the hair follicle were entirely stained (Fig. 2c) but in some follicles, the inner portion of ORS was weakly or negatively stained (Figs 2c and 3e). In the hair bulb, the basal cells surrounding the dermal papilla were not stained, while suprabasal cells including proliferating and undifferentiated cells were weakly or moderately stained (Fig. 2b). The cells in the hair shaft before cornification were PDCD4-positive (Fig. 2b). The dermal papilla was not stained. The cells of inner root sheath (IRS) surrounding the hair shaft was stained (Fig. 2b) although the fused IRS was PDCD4-negative (Fig. 2c). In sebaceous glands (Fig. 2a) the mature cells containing lipid droplets were weakly or negatively stained and peripherally located immature cells without lipid droplets were strongly stained (Fig. 2a). Ductal portion and secretary portion of the sweat glands were strongly stained (Fig. 2d).

As shown in Fig. 3b, PCNA-positive cells were located in the basal and lower suprabasal layers of epidermis. In the hair follicle, the matrix cells of hair bulb were strongly stained and the suprabulbar cells of differentiating cell area were also partly stained with anti-PCNA antibody (Fig. 3d), In the ORS, PCNA-positive cells were localized in the basal layer and PDCD4-positive cells were mostly found in the inside of PCNA-positive cell layers (Fig. 3e,f). These staining patterns with anti-PCNA antibody were in accordance with the finding by Soma et al. (21).

PDCD4 expression in cutaneous lesions

Seborrheic keratosis showed the coexistence of PDCD4-positive and -negative areas. The keratinized area including terminal differentiation was not stained (Fig. 4a). Basal cell carcinoma cells mostly were negative, but some cases were partly stained (Fig. 4b). Strong positive cells were crowded. Squamous cell carcinomas, including eccrine porocarcinomas, were unstained (Fig. 4c). Malignant melanomas rarely showed staining (not illustrated). Three samples of each cancer were observed and a representative of them was illustrated in each case.

Figure 4.

 Staining patterns in skin lesion. Seborrheic keratosis (a), basal cell carcinoma (b) and eccrine porocarcinoma (c) were stained using anti-PDCD4 antibody. Positive and negative areas are present in (a) and (b), but (c) shows no staining. PDCD4 expression is suppressed in the epidermis overlying an adult T-cell lymphoma tissue (designated A) compared with expression in surrounding epidermis (d, e and f). e and f show higher magnification images of areas between arrows and arrowheads in d, respectively.

As shown in Fig. 4d–f, PDCD4 expression in the epidermis overlying an adult T-cell lymphoma (ATL) (Fig. 4f) was less than expression in surrounding epidermis (Fig. 4e). Two samples of this case were observed, and one case was affected as shown in the figure but the other was not.

Effects of growth factors on PDCD4 expression in keratinocytes

Suppressing of PDCD4 expression in the epidermis overlying the ATL (Fig. 4d), suggested that growth factors produced by the ATL could have affected PDCD4 expression by keratinocytes. To examine the effect of growth factors on PDCD4 expression in keratinocytes, HaCaT cells were cultured in the presence of EGF, transforming growth factor (TGF)-β1 and hepatocyte growth factor (HGF); protein extracts then were analysed by Western blotting. All of these growth factors downregulated PDCD4 expression by HaCaT cells (Fig. 5). TPA showed no effect on the amount of PDCD4 in HaCaT cells (data not shown). These results indicated that PDCD4 expression was easily modulated by the environmental conditions of cells.

Figure 5.

 Effects of growth factors on PDCD4 expression by HaCaT cells according to Western blotting using anti-PDCD4 antibody. (a) HaCaT cells were cultured without growth factor (lane 1), with 10 ng/ml epidermal growth factor (EGF; lane 2) and 5 ng/ml transforming growth factor (TGF)-β 1 (lane 3) for 24 h. (b) HaCaT cells were cultured without hepatocyte growth factor (HGF; lane 4), with 1 ng/ml HGF (lane 5) or 10 ng/ml HGF (lane 6) for 48 h.


We presently localized PDCD4 expression in differentiating cell layers such as the suprabasal layer of the epidermis and the suprabulbar area of the hair follicle; expression was absent or less extensive in the epidermal basal cell layer and in the hair bulb, which is composed of proliferating and undifferentiated cells. The results indicate that PDCD4 may have an important role in differentiation of keratinocytes. Expression patterns may vary in the differentiation stages of epidermis and its appendages. Especially, in the hair follicle, expression may be changed in the anagen and catagen phases. The problems should be investigated in the future. Overall, a normal keratinocyte cell line, HaCaT cells, was found to abundantly express PDCD4 but staining in these cultures was heterogeneous with individual cells showing high to negative expression. This mixed staining pattern in HaCaT cell culture suggests that PDCD4 expression may vary according to stage in the cell cycle. In HaCaT cells, EGF and HGF stimulated proliferation (22,23), but TGF-β1 inhibited cell cycle progression by upregulating the p21(WAF1/Cip1) cell cycle inhibitor (24). The results that all of these growth factors suppressed PDCD4 expression of HaCaT cells also may suggest that PDCD4 expression is controlled in the cell cycle. The protein PDCD4 was expressed only to a very low degree in carcinoma cell lines (A431 and DJM-1).

In the epidermis and its appendages, we mostly localized PDCD4-protein to the cell nuclei. The protein also was reported to be found in the cell nuclei of small ducts of the breast and in differentiated adenocarcinomas of the breast (2). The protein accumulated in nuclei at the G0 phase of the cell cycle in a normal human fibroblast cell line, MRC5, while PDCD4 protein was localized mainly in the cytoplasm in most cancer cells, including the A431 and DJM-1 cell lines.

Cmarik et al. reported that Pdcd4 expression was low in promotion-sensitive (P+) cells and high in promotion-resistant (P-) cells in a JB6 mouse epidermal model of neoplastic transformation (10). They also reported that the expression decreased as the tumor progressed. Transgenic mice overexpressing Pdcd4 in the epidermis showed a neonatal phenotype with shorter hair than wild-type siblings reflecting early catagen entry; the transgenic mice also were resistant to cutaneous carcinogenesis induced by TPA (25). These results as well as ours indicate that proliferating cells were more likely to show low than high PDCD4 expression.

PDCD4 might contribute to differentiation of keratinocytes by inhibiting AP-1 activity that otherwise would induce cells to proliferate because PDCD4 was shown to inhibit AP-1 transactivation activities (11,14). Alternatively, we have reported previously that TGF-β1-induced apoptosis propagated through upregulation of PDCD4 expression and PDCD4 activated caspase 9 and 3 via mitochondria events in the hepatoma cell line Huh7 (7). Recently, Chaturvedi et al. reported that the caspases were activated on cornification in human epidermal equivalents (26). Therefore, PDCD4 might partly contribute to activate the apoptotic machinery caspase cascade on cornification of keratinocytes, considering that PDCD4-protein abundantly expressed in the differentiating keracinocyte cell layers. The activation of caspase cascade also was reported in the ultraviolet B (UVB)-induced apoptosis of HPV-immortalized human keratinocytes but the induction was death receptor independent (27). PDCD4 also might contribute to the UVB-induced apoptosis of keratinocytes, because we have obtained preliminary results that UVB modulated PDCD4 expression of HaCaT cells and that the Fas system could not induce PDCD4 expression of the hepatoma Huh7 cells (unpublished data).

PDCD4 expression was noted to be inhibited in the epidermis overlying an ATL, which could reflect paracrine effects of factors produced by ATL cells. We found that the growth factors EGF, TGF-β1 and HGF inhibited PDCD4 expression in HaCaT cells. TGF-β and interleukin (IL)-2 have been reported to be produced by ATL cells (28,29) while IL-2 has been found to inhibit PDCD4 expression by T cells and natural killer cells (30). Considering that growth and differentiation of epidermal cells is controlled by dermal factors (31), PDCD4 expression by basal cells may be regulated by these dermal factors while expression by germinal cells in the hair bulb may be modulated similarly by factors produced by dermal papilla cells. HGF was found to be expressed in dermal papilla cells and to stimulate hair growth (32). We suspect that PDCD4 expression is essential for keratinocyte differentiation and is controlled by various factors in the surrounding environment.

Allelic losses in the 10q24 region have been reported, where the PDCD4 gene have been mapped (33–36), but loss of the PDCD4 gene locus has not been reported in any cancer despite downregulation of this gene's expression. Proliferating cells may not require loss of the gene to escape the control of its gene product, as PDCD4 gene expression may be easily controlled by paracrine and/or autocrine mechanisms. The PDCD4 gene may prove to have potential as a molecular target in cancer prevention and therapy (15,37) by pharmacologic or dietary control of its gene expression.


We thank Mrs Yumiko Tsugitomi of Department of Internal Medicine, Saga Medical School, Saga University, for her technical assistance to stain tissue sections.