1514 Altered Expression of bcl-2 Family Member Proteins in Nonmelanoma Skin Cancer Maryse Delehedde, Ph.D. Song H. Cho, B.S. Mona Sarkiss, M.D., Ph.D. Shawn Brisbay, B.S. Michael Davies, B.S. Adel K. El-Naggar, M.D., Ph.D. Timothy J. McDonnell, M.D., Ph.D. Department of Molecular Pathology, The University of Texas, M. D. Anderson Cancer Center, Houston, Texas. BACKGROUND. Differentiation, proliferation, and cell death are coordinated tightly within the epidermis. Alterations within keratinocytes that disrupt these processes are believed to contribute to the development of nonmelanoma skin cancers (NMSC). In the current study the authors examined the expression of selected members of the bcl-2 gene family in the skin and in case-matched samples of NMSC. METHODS. Immunohistochemistry was performed on tissue sections using antibodies against bcl-2, bcl-x, bax, and bak. Case-matched frozen nonneoplastic skin samples and tumor tissues were used for Western blot analysis. RESULTS. In normal epidermis, bcl-2 oncoprotein is expressed in keratinocytes of the basal layer but is down-regulated in suprabasal layers. The proapoptotic bax protein is expressed at low levels in basal keratinocytes and is up-regulated in suprabasal layers. The bcl-x and bak proteins both are expressed in the basal and spinous strata but are down-regulated in the granular cell layer. Both bcl-2 and bax were diffusely cytosolic whereas bcl-x and bak exhibited a distinct perinuclear distribution. Squamous cell carcinomas (SCC) were negative for bcl-2 whereas bcl-2 increased 5.5-fold in basal cell carcinomas (BCC). The distribution of bcl-x and bax proteins within BCC and SCC overlapped and were associated with squamous differentiation. Bax protein was increased twofold to threefold in NMSC. An increase in bak protein also was observed in SCC. However, bak was diffusely cytosolic within BCC in contrast to the perinuclear distribution in nonneoplastic keratinocytes. CONCLUSIONS. These findings suggest that altered expression of bcl-2 family members may play a role in the pathogenesis of NMSC. Cancer 1999;85:1514 –22. © 1999 American Cancer Society. KEYWORDS: bcl-2, skin cancer, apoptosis, differentiation. Supported by Grant NCI P01 CA68233. Dr. Maryse Delehedde is supported by a postdoctoral fellowship from the Fondation pour la Recherche Médicale. Song Cho is supported by a Cancer Biology Training Grant NIH T32 CA60440. Address for reprints: Timothy J. McDonnell, M.D., Ph.D., Department of Molecular Pathology, Box 89, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, TX 77030. Received May 11, 1998; revision received December 2, 1998; accepted December 2, 1998. © 1999 American Cancer Society A balance between cell proliferation and cell death is necessary for the normal development and maintenance of cellular homeostasis in the epidermis. Considerable evidence indicates that the development and progression of several types of cancer result from the disruption of normal cell death and proliferation.1 In fact, molecular alterations in genes known to regulate these two fundamental processes occur frequently in nonmelanoma skin cancer (NMSC). In this context, activation of the ras gene family2,3 and inactivation of the p53 tumor suppressor gene4,5 are known to contribute to the pathogenesis of NMSC. We and others previously have reported that the antiapoptotic bcl-2 protein is expressed at high levels in basal cell carcinomas (BCCs) but is not expressed at detectable levels in squamous cell carcinomas (SCCs) of the skin.6-10 The contribution of bcl-2 expression to multistep skin carcinogenesis recently was established in a bcl-2 Family Member Proteins in Skin Cancer/Delehedde et al. transgenic model that utilized the human keratin 1 promoter to target expression of a human bcl-2 transgene specifically to the epidermis.11 The HK1.bcl-2 transgenic mice had a reduced rate of apoptosis in response to ultraviolet (UV) irradiation and were more susceptible to tumor formation in response to UV and chemical carcinogens. Bcl-2 is now recognized to be a member of an expanding multigene family of cell death regulators.12 Members of the bcl-2 gene family share significant sequence homology and can be divided functionally into two categories: the antiapoptotic members (bcl-2, bcl-xL, mcl-1, and bcl-w) and the proapoptotic members (bax, bak, bad, bcl-xS, bid, and hrk). Current evidence suggests that the differential expression of antiapoptotic and proapoptotic bcl-2 family proteins is an important determinant of the susceptibility of a cell to undergo apoptosis.13 The potential contribution of these bcl-2 family members to the pathogenesis of NMSC essentially is unknown. In this report, we examined the expression of selected members of the bcl-2 gene family in the normal skin and in NMSC using immunohistochemical and Western blot analysis. MATERIALS AND METHODS Tumor Samples Normal and corresponding NMSC specimens were obtained from the pathology files of the Department of Pathology at the University of Texas, M. D. Anderson Cancer Center in Houston. This material was comprised of 17 cases of BCC, 4 cases of basosquamous carcinoma, and 14 cases of SCC varying from well differentiated to poorly differentiated. Tissues were excised prospectively after frozen-section evaluation to include tumor and adjacent normal skin in certain blocks. When available, separate normal skin and tumor samples corresponding to the tissue blocks were snap frozen in liquid nitrogen and stored at -80 °C until used. The histopathologic diagnosis was established on routine sections stained with hematoxylin and eosin. Immunohistochemistry Immunohistochemical staining was performed on formalin fixed, paraffin embedded tissue samples. Serial sections (5 mm) were deparaffinized in xylene and rehydrated in descending concentrations of alcohols, then placed in phosphate-buffered saline (PBS) for 5 minutes. Each specimen was evaluated by two pathologists (T.J.M. and A.K.E.N.). 1515 bcl-2 Staining After rehydration, sections were incubated with the primary antibody, a mouse antihuman bcl-2 antibody (Clone 124; Dako Co., Carpinteria, CA), at a dilution of 1:80 for 1 hour in a moist chamber. After washing in Tris-buffered saline (TBS), the sections were incubated with rabbit antimouse antibody diluted 1:25 in TBS for 30 minutes at room temperature. The slides were rinsed in TBS and then incubated with alkaline phosphatase antialkaline phosphatase (APAAP) at a dilution of 1:50 for 20 minutes. To increase the intensity of the staining, the antimouse antibody and APAAP steps were repeated for 3 cycles at 10 minutes of incubation each. An alkaline phosphatase substrate, Fuchsin (Dako Co.) was added and incubated for 10 minutes at room temperature in the dark and then rinsed in water. Slides finally were counterstained with Mayer hematoxylin (Sigma Chemical Co., St Louis, MO) and mounted. Infiltrating lymphocytes served as internal positive controls. Negative controls were comprised of consecutive tissue sections of each case in which the primary antibody was omitted. bax, bcl-x, and bak Staining To block nonspecific binding, the slides first were covered in PBS containing 30% methanol and 0.3% of hydrogen peroxide for 10 minutes. The anti-bax N-19, anti-bcl-x S-18, and anti-bak N-20 antibodies were obtained from Santa Cruz, CA. For bax and bcl-x staining, the sections then were incubated with 1% normal goat serum in PBS for 20 minutes. After washing in PBS, these slides were incubated for 1 hour at room temperature with a 1:100 dilution of primary antibody in PBS containing 1% normal goat serum. The slides were rinsed with PBS and covered with a 1:100 dilution of peroxidase-conjugated goat antirabbit immunoglobulin (Ig) G antibody in PBS containing 1% goat serum for 45 minutes at room temperature. Finally, after an additional 30 minutes with Vectastain ABC horseradish peroxidase (Vector Laboratories, Burlingame, CA), the staining was revealed with the chromogen diaminobenzidine and counterstained with hematoxylin. The same procedure was used for bak protein staining except that the normal goat serum was replaced by 1% normal rabbit serum, the working dilution of the primary antibody was 1:50, and the secondary antibody was a rabbit antigoat IgG coupled with horseradish peroxidase. Proliferating Cell Nuclear Antigen Staining Endogenous peroxidases first were quenched with 0.3% hydrogen peroxide in methanol for 10 minutes 1516 CANCER April 1, 1999 / Volume 85 / Number 7 FIGURE 1. Distribution of bcl-2 family members in normal epidermis. (A) In the nonneoplastic epidermis, bcl-2 expression was confined strictly to basal keratinocytes. The suprabasal layers were uniformly bcl-2 negative. (B) bcl-x staining was diffusely cytosolic within the basal cells, but more intense, punctate, and perinuclear in the keratinocytes of the spinous layer. (C) The proapoptotic bax protein was expressed at low levels in basal keratinocytes and was up-regulated in the spinous and granular cell layers. (D) bak staining was homogenous and mostly diffuse in the cytoplasm of the keratinocytes in the basal layer, but exhibited a perinuclear and punctate pattern in the differentiated cells of the spinous layer (3 400). and washed in PBS. Nonspecific binding was blocked using 1% normal goat serum in PBS for 30 minutes. The sections then were incubated with the monoclonal antiproliferating cell nuclear antigen (PCNA) antibody (PC 10; Dako Co.), diluted at 1:100 in PBS containing 0.5% Tween 20 and 0.5% bovine serum albumin for 60 minutes.14 Negative controls were incubated with PBS without antibody. After washing in PBS, the slides were incubated with a biotinylated antimouse IgG secondary antibody diluted 1:100 with 1% goat serum in PBS for 30 minutes at room temperature. After an additional 30 minutes with Vectastain ABC horseradish peroxidase (Vector Laboratories), the substrate was added for the appropriate time period (range, 5-15 minutes). This resulted in PCNA positive cells being labeled brown. Slides were counterstained with hematoxylin and mounted. Gel Electrophoresis and Immunoblotting Case-matched, frozen, nonneoplastic skin samples and tumor tissues from ten patients were available for Western blot analysis. Tumor samples were comprised of at least 75% neoplastic tissue as verified by frozen-section evaluation at the time of acquisition. Aliquots of the total crude protein extract were electrophoresed on 12.5% sodium dodecyl sulfate-polyacrylamide gels. After transfer of the proteins onto nitrocellulose membranes, filters were blocked overnight at 4 °C in blotting solution (PBS, 5% Carnation nonfat dry milk [Carnation,], and 0.05% Tween 20). Incubation with the primary antibody (antihuman bcl-2 6C8, anti-bax N-19, anti-bcl-x S-18, and anti-bak N-20 [all from Santa-Cruz, CA]) diluted at 1:500 was performed for 1 hour at room temperature in the blotting solution. After 5 washes with 0.05% Tween 20 in PBS, the filters were incubated with peroxidase-labeled anti-IgG antibodies (goat antihamster, goat antirabbit, goat antirabbit, and rabbit antigoat, respectively) diluted 1:1000 in the blotting solution. After several washes with 0.05% Tween 20 in PBS, immunoreactive proteins were detected with enhanced chemiluminescence-Western chemiluminescence detection on Hyperfilm (Amersham Life Science, Arlington Heights, IL). RESULTS In the normal epidermis, bcl-2 expression was confined to basal keratinocytes that exhibited cytoplasmic staining with perinuclear enhancement. The suprabasal keratinocytes uniformly were bcl-2 negative (Fig. 1A). The distribution of bcl-x protein in non-neoplastic epidermis varied from a high level of expression in basal keratinocytes and that progressively was downregulated in the stratum spinosum and stratum granulosum (Fig. 1B). Bcl-x was diffusely cytosolic within the basal keratinocytes, but distinctly punctate and mostly perinuclear in keratinocytes of the spinous layer. Keratinocytes of the granular cell layer showed low or undetectable bcl-x immunoreactivity and the cornified layer was bcl-x negative. The localization of the proapoptotic bax protein in the normal epidermis showed essentially no overlap with the distribution of bcl-2 protein (Fig. 1C). Basal keratinocytes rarely exhibited bax immunoreactivity. The bax protein was diffusely cytosolic in keratino- bcl-2 Family Member Proteins in Skin Cancer/Delehedde et al. 1517 FIGURE 2. Distribution of bcl-2 family members in basal cell carcinoma (BCC). (A) bcl-2 protein was expressed strongly in basal cell carcinomas Frequently, the bcl-2 staining exhibited by the BCC was increased compared with adjacent nonneoplastic basal keratinocytes. (B) bcl-x and (C) bax staining appeared diffusely cytosolic in BCC. (D) bak staining appeared diffusely cytosolic in marked contrast to the punctate perinuclear pattern exhibited by the majority of nonneoplastic keratinocytes (3200). cytes of the spinous and granular cell layers. Bax protein was undetectable in the cornified layer. The distribution of the proapoptotic bak (bcl-2 homologous antagonist/killer) protein in the normal epidermis is shown in Figure 1D. Bak protein was present in keratinocytes from the basal layer to the granular layer. Bak protein was homogenous and diffusely cytosolic in the keratinocytes of the basal layer, but mostly was perinuclear in the differentiated cells of the spinous layer. Keratinocytes of the granular cell layer rarely exhibited bak immunoreactivity. Bak protein was not observed in the cornified layer. Bcl-2 protein was expressed at high levels in all BCCs examined in this study (Fig. 2A). The level of bcl-2 protein present in the BCC frequently was more intense than that observed in basal keratinocytes in adjacent nonneoplastic epidermis. In contrast, neoplastic keratinocytes within BCCs expressed relatively low levels of bcl-x protein. Bcl-x staining appeared weak and diffusely cytosolic in the neoplastic cells (Fig. 2B). Bax staining was heterogeneous in individual cases of BCC and the subcellular staining pattern mostly was cytosolic (Fig. 2C). Comparison of adjacent tissue sections strongly suggests that expression of bcl-x and Bax virtually is completely overlapping in BCC. Bak staining within malignant keratinocytes in BCC was diffusely cytosolic, which is in marked contrast to the punctate perinuclear pattern exhibited by suprabasal epidermal keratinocytes (Fig. 2D). SCC showed no immunohistochemically detectable bcl-2 protein (Fig. 3A). In these cases infiltrating lymphocytes present in the dermis served as internal positive controls for bcl-2 immunoreactivity. The dis- tribution of bcl-x protein in SCC in general correlated with areas exhibiting histologic evidence of squamous differentiation (Fig. 3B). Moreover, in the poorly differentiated SCCs, we also observed enhanced immunoreactivity for bcl-x protein in individual tumor cells that exhibited cytologic evidence of keratinization. bax protein was expressed at relatively high levels in SCC compared with nonneoplastic epidermis (Fig. 3C). The distribution of bax protein within individual SCCs overlapped with bcl-x in areas of squamous differentiation. Similar to bax, the distribution of bak protein correlated with histologic evidence of squamous differentiation in SCC (Fig. 3D). In contrast to nonneoplastic epidermal keratinocytes, bak protein appeared to be diffusely cytosolic, not perinuclear, in the majority of neoplastic cells of SCC and basosquamous carcinomas. In those cases of basosquamous carcinoma exhibiting foci of both well differentiated SCC and BCC, bcl-2 protein was limited to the basal cell component of the tumor (Fig. 4A). Although the expression of bcl-x protein also was found in the basaloid cells, it was expressed more strongly in foci of squamous differentiation (Fig. 4B). As noted in BCC, the bax staining pattern in the basal part of the tumor was heterogeneous. However, bax immunoreactivity was relatively higher in the foci of squamous differentiation (Fig. 4C). The bak staining was increased in areas of squamous differentiation in basosquamous carcinoma (Fig. 4D). In general, the staining intensity of bcl-x, bax, and bak proteins was increased with the squamous differentiation in malignant keratinocytes of both SCC and basosquamous carcinomas. 1518 CANCER April 1, 1999 / Volume 85 / Number 7 FIGURE 3. Distribution of bcl-2 family members in squamous cell carcinoma (SCC). (A) All SCCs exhibited undetectable levels of bcl-2. Lymphocytes served as internal positive controls. (B) The distribution of bcl-x protein correlated with areas of squamous differentiation. bcl-x protein was characteristically perinuclear and punctate in cells exhibiting squamous differentiation. As shown in panel C, bax was expressed uniformly in SCC samples; the staining generally was homogenous and diffusely cytosolic. (D) The distribution of bak protein within malignant keratinocytes was associated with the squamous differentiation of the tumors (3200). FIGURE 4. Distribution of bcl-2 family members in basosquamous carcinoma. (A) In basosquamous carcinoma, bcl-2 protein was detected within the neoplastic cells of the basaloid component. (B) In contrast, the bcl-x protein mainly was associated with areas of squamous differentiation. Bcl-x staining primarily was punctate and perinuclear in the neoplastic cells surrounding keratin pearls. (C) The level of bax protein appeared enhanced in areas of squamous differentiation. (D) The distribution of the bak protein was very diffuse in the cytoplasm of the cells in basal part of the tumor, but exhibited a characteristic pattern of perinuclear and punctate staining in areas of keratinization (3200). The stromal components of the dermis and within NMSCs exhibited negligible or undetectable levels of the bcl-2 family proteins examined in this study with the exception of infiltrating lymphocytes. The immunohistochemical observations for nonneoplastic epidermis and NMSC are summarized in Table 1. Immunohistochemical evaluation of PCNA revealed that the majority of proliferating cells, as anticipated, were localized to the basal stratum in nonneoplastic epidermis (Fig. 5A). A loss of PCNA positivity was associated with areas of keratinization and squamous differentiation in both SCC (Fig. 5B) and basosquamous carcinomas (Fig. 5C). In contrast, in BCC, PCNA positive cells in general were scattered uniformly throughout the neoplasm (Fig. 5D). Tumor cells exhibiting cytologic features consistent with apoptotic cell death (such as cell shrinkage, loss of junctional continuity, chromatin condensation, and formation of apoptotic bodies) generally comprised # 1% of BCCs. Apoptotic cells present within SCC typically were associated with areas exhibiting histologic evidence of squamous differentiation and keratinization (data not shown). Immunohistochemical analysis of the expression of selected bcl-2 family member proteins in NMSC was confirmed and extended using Western blot analysis. Total protein extracts were derived from casematched, adjacent nonneoplastic skin and NMSC specimens (Figs. 6A and 6B). The level of bcl-2 protein was increased approximately 5.5-fold in BCC when bcl-2 Family Member Proteins in Skin Cancer/Delehedde et al. TABLE 1 Summary of Immunostaining Results for Nonneoplastic Epidermis and NMSC Nonneoplastic epidermis Basal layer Suprabasal layer Basal cell carcinoma (17 cases) Squamous cell carcinoma (14 cases) Basosquamous carcinoma Basal component Squamous component (4 cases) bcl-2a bcl-x bax bak 111 2 111 111 11 1 11 1 1 2 11 1/2 111 11 Heterogeneous 111 Homogeneous 111 2 1 11 11 111 1/2 11 11 NMSC: nonmelanoma skin cancer. a Staining intensity was normalized to infiltrating lymphocytes. Data are representative of all tissue samples examined. normalized to matched nonneoplastic epidermis (P , 0.05). The level of bcl-2 protein in SCC was not demonstrably different from normal epidermis. Immunoblotting of nonneoplastic epidermis as well as NMSC samples indicated that the bcl-x protein demonstrable by immunohistochemical techniques was likely the antiapoptotic, 31-kilodalton (kD) bcl-xL form of the protein because the proapoptotic, 19-kD bcl-xS form of the bcl-x protein was undetectable in all nonneoplastic epidermis and NMSC samples examined. However, levels of bcl-xS expression below the sensitivity of the immunoblot procedure cannot be excluded completely. The total amount of bcl-xL protein in tumor samples was not significantly different from the nonneoplastic epidermis. A 2-fold and a 3-fold increase in the level of bax protein was observed in BCC and SCC, respectively, compared with nonneoplastic epidermis (P , 0.05). A 2-fold increase in the amount of bak protein was observed in SCC (P , 0.05). However, no significant modulation of bak protein was observed in BCC compared with nonneoplastic epidermis. DISCUSSION The control of the individual processes of cell death, cell division, and cell differentiation is inherently complex. Furthermore, the molecular regulation of each of these processes must necessarily be integrated, temporally and cytoarchitecturally, within self-renewing complex epithelia such as the epidermis. In this regard, an initially descriptive assessment of potentially critical regulatory molecules within the epidermis may provide insight into the mechanisms by which these processes are regulated. 1519 In this study, the expression and distribution of specific apoptosis-regulating members of the bcl-2 gene family were assessed in the epidermis and in NMSC using immunohistochemical and immunoblotting techniques. The keratinocytes of the basal layer of the epidermis expressed high levels of the bcl-2 and bcl-x proteins. Immunoblotting of nonneoplastic epidermis using anti-bcl-x antibodies demonstrated the presence of the antiapoptotic, 31-kD bcl-x long form of the protein. The proapoptotic, 19-kD bcl-x short form of the protein either was not expressed or was below the level of sensitivity of the immunoblot. Although to our knowledge antibodies that discriminate bcl-xL from bcl-xS using immunohistochemistry currently are unavailable, our findings are consistent with the interpretation that the predominant, if not exclusive, form of bcl-x expressed in nonneoplastic epidermis is bcl-xL. To our knowledge immunohistochemical detection of bcl-x protein in basal keratinocytes has not been demonstrated previously.15,16 The basis of these discrepant findings potentially may be attributed to tissue source (surgical specimens vs. autopsy), fixation techniques (frozen-section, formalin, or Buin fixative), or variations in the polyclonal anti-bcl-x antibodies used in these studies. The inability to detect bcl-xS in epidermal keratinocytes using Western blot analysis is consistent with previous observations.16 In contrast to the antiapoptotic bcl-2 and bcl-xL proteins, levels of the proapoptotic bax and bak proteins are comparatively modest in basal keratinocytes. This may be related to the self-renewal capacity of these cells, which would necessitate relative resistance to stress-induced cell death. However, the expression and the cytoarchitectural distribution of the numerous other bcl-2 family members and other cell death regulatory proteins within the epidermis largely is unknown. Suprabasal keratinocytes preferentially expressed the proapoptotic bax and bak proteins. This commitment to terminal differentiation is associated with the immediate down-regulation of bcl-2 in suprabasal keratinocytes and the loss of bcl-x expression in keratinocytes of the stratum granulosum. The relative distribution of the proapoptotic and antiapoptotic members of the bcl-2 family supports the contention that terminal differentiation in the epidermis represents a specialized and tightly controlled physiologic form of apoptosis.17-19 It was recently reported that bak protein is able to interact preferentially with bcl-xL relative to bcl-2.20 It is interesting to note that the distribution of bcl-x and bak proteins within the epidermis is completely overlapping. A correlation between the expression of the bcl-2-related proteins and keratinocyte differentiation 1520 CANCER April 1, 1999 / Volume 85 / Number 7 FIGURE 5. Proliferative activity in nonneoplastic epidermis and nonmelanoma skin cancer. (A) In nonneoplastic epidermis, proliferating cell nuclear antigen (PCNA) was expressed predominantly in basal keratinocytes as anticipated. (B) In squamous cell carcinoma, PCNA positive cells were not found in areas of keratinization (3200). (C) In basal cell carcinoma the PCNA staining was heterogeneous and a loss of PCNA positivity was observed in the cells involved in the foci of squamous differentiation (3200). (D) In basosquamous carcinoma, only few PCNA positive cells were found throughout the tumor (3400). FIGURE 6. Immunoblot analysis of bcl-2 family proteins in normal epidermis and in nonmelanoma skin cancer (NMSC). (A) A representative Western blot analysis of total protein extracts for bcl-2 family proteins is presented. (B) The results of scanning densitometry after normalization of expression to actin loading and levels of expression in nonneoplastic epidermis show that the relative amount of bcl-x protein was not substantially different from the nonneoplastic epidermis. In contrast, the relative amount of bax protein was modified strongly in NMSC. Compared with bax expression in normal tissue, a threefold and twofold increase, respectively, in the amount of the bax protein was observed in squamous cell carcinoma (SCC) and in basosquamous carcinoma (BCC). Moreover, a 5.5-fold increase in the relative amount of bcl-2 protein was observed in BCC. Finally, a twofold increase in the relative amount of bak protein was observed in SCC but not BCC. These findings confirm the relative expression levels observed with immunohistochemical techniques. bcl-2 Family Member Proteins in Skin Cancer/Delehedde et al. in vitro recently has been reported.19 These authors showed that the terminal differentiation of gingival keratinocytes, induced by high calcium concentrations, is accompanied by a decrease in bcl-x and increase in bax. Together these findings indicate that cell death regulatory proteins of the bcl-2 family are expressed differentially within the epidermis and suggest that they may contribute to the normal cellular growth, differentiation, and homeostasis. It may be anticipated that altered expression of these proteins could contribute to the pathogenesis of NMSC. Our findings and those of others provide consistent evidence that BCCs express high levels of bcl-2 protein.6-9 In this report we demonstrated a fivefold increase of bcl-2 protein in BCC compared with nonneoplastic epidermis. A corresponding increase in bcl-2 protein was not observed in SCC. A comparatively modest, but significant, increase in the level of bax protein was observed in SCC and BCC. It is interesting to note that an approximately twofold increase in the level of bak protein was observed in SCC but not BCC. The level of bcl-xL protein was not altered substantially in either SCC or BCC compared with nonneoplastic epidermis. The process of apoptosis is believed to be regulated by the ratio of antiapoptotic proteins and proapoptotic proteins present in the cell.13 This rheostat model may even apply to the role bcl-2 family member proteins play in the process of keratinocyte differentiation and their contribution to skin carcinogenesis. Direct evidence for the contribution of bcl-2 family proteins to multistep skin carcinogenesis recently has been provided. Targeting of bcl-2 expression specifically to the epidermis was accomplished using a human keratin 1 promoter construct.11 The bcl-2 transgenic protein was expressed transmurally in the epidermis and resulted in focal areas of hyperplasia. Keratinocytes from the bcl-2 transgenic mice were resistant to apoptosis induction by UV irradiation and chemical carcinogens. In addition, HK1.bcl-2 transgenic mice exhibited a shorter latency and higher frequency of tumor formation compared with control littermates. A similar transgenic model used the human keratin 14 promoter to target expression of bcl-xL specifically to the epidermis.21 The HK14.bcl-xL epidermis was resistant to cell death induction by etoposide and UV irradiation. Moreover, the observed increase in the proapoptotic members of the bcl-2 family within NMSCs was unexpected because recent evidence suggests that these proteins may act as tumor suppressors. A reduction in the level of bax-a in breast carcinoma relative to normal breast tissue has been observed.22 In metastatic breast carcinoma, reduced bax expression may 1521 be predictive of a poor response to chemotherapy.23 In addition frequent frame shift mutations of bax were found in the mutator phenotype of colon adenocarcinomas, suggesting that bax inactivation may contribute to colorectal carcinogenesis.24 Direct evidence of bax tumor suppressor activity has been provided from experiments using genetically engineered strains of mice.25 To our knowledge somatic mutations involving the bax gene product have not yet been reported in NMSC nor have previously documented examples of bax mutations been associated with an increase in the steady-state levels of bax protein. Alternatively, the elevated levels of death effector proteins observed in SCC suggest that a cell death suppressor of the bcl-2 family other than bcl-2 or bcl-xL may be involved in these neoplasms. Candidate proteins would include MCL-1,26 A1,26 bcl-w,27 bfl-1,28 or an as yet unidentified antiapoptotic bcl-2 family member. It is interesting to note that regions exhibiting histologic evidence of squamous differentiation or keratinization in SCC and basosquamous carcinomas showed immunohistochemical evidence of a reduction in bcl-2 protein and enhanced expression of bcl-x, bax, and bak. bax protein generally was diffusely cytosolic whereas bcl-x and bak showed a distinctly punctate and perinuclear distribution. The significance of the subcellular distribution of these proteins with respect to their ability to function in the regulation of cell death is an active area of investigation. In this regard, recent observations suggest that the subcellular localization of bcl-2 may alter apoptosis sensitivity.29,30 In addition, the significance of the correlation between the modulation of these proteins with loss of PCNA positivity, apoptosis induction, and acquisition of differentiated cytologic features remains to be elucidated. The results of the current study indicate that the modulation of bcl-2 family member proteins is coordinated with differentiation in nonneoplastic epidermal keratinocytes and in NMSC. Ongoing studies will determine whether variations in these proteins correlate with rates of apoptosis and clinical response to nonsurgical therapeutic interventions. Studies also are currently ongoing to evaluate directly whether inactivating mutations of death effector proteins contribute to multistep skin carcinogenesis in vivo. REFERENCES 1. 2. McDonnell TJ. Cell division versus cell death: a functional model of multistep neoplasia. Mol Carcinog 1993;8:209-13. Ananthaswamy HN, Prince JE, Goldberg LH, Bales ES. Detection and identification of activated oncogenes in human skin cancers occurring on sun-exposed body sites. Cancer Res 1988;48:3341-8. 1522 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. CANCER April 1, 1999 / Volume 85 / Number 7 Pelling JC, Sheng X, Betz NA. Role of Ha-ras oncogene in skin cancer. In: Mukktar H, editor. Skin cancer: mechanisms and human relevance. New York: CRC Press, Inc., 1995:28391. Ziegler A, Johnason AS, Leffell DJ, Simon JA, Sharma HW, Kimmelmam J, et al. Sunburn and p53 on the onset of skin cancer. Nature 1994;372:773-6. Quinn AG. Molecular genetics of human non melanoma skin cancer. Cancer Surv 1996;26:89-113. Rodriguez-Villanueva J, Colome MI, Brisbay S, McDonnell TJ. The expression and localization of bcl-2 protein in normal skin and in non-melanoma skin cancers. Pathol Res Pract 1995;191:391-8. Morales-Ducret CR, Van de Rijn M, Lebrun DP, Smoller BR. Bcl-2 expression in primary malignancies of the skin. Arch Dermatol 1995;131:909-12. Cerroni L, Kerl H. Aberrant bcl-2 protein expression provides a possible mechanism of neoplastic cell growth in cutaneous basal cell carcinoma. J Cutan Pathol 1994;21:398403. Nakagawa K, Yamamura K, Maeda S, Ichihashi M. Bcl-2 expression in epidermal keratinocytic diseases. Cancer 1994; 74:1720-4. Verhaegh ME, Sanders CJG, Arends JW, Neuman H. Expression of the apoptosis-suppressing protein bcl-2 in non-melanoma skin cancer. Br J Dermatol 1995:132:740-4. Rodriguez-Villanueva J, Greenhalgh DA, Wang XJ, Bundmann DS, Cho SH, Delehedde M, et al. Human keratin1.bcl-2 transgenic mice aberrantly express keratin 6, exhibit reduced sensitivity to keratinocyte cell death induction and are susceptible to skin tumor formation. Oncogene 1998;16: 853-63. McDonnell TJ, Beham A, Sarkiss M, Andersen MM, Lo P. Importance of the bcl-2 family in cell death regulation. Experientia 1996;52:1008-17. Korsmeyer SJ, Shutter JR, Veis DJ, Merry DE, Oltvai ZN. Bcl-2/Bax: a rheostat that regulates an anti-oxidant pathway and cell death. Semin Cancer Biol 1993;4:327-32. Delehedde M, Boilly B, Hondermarck H. Differential responsiveness of human breast cancer cells to basic fibroblast growth factor. A cell kinetics study. Oncol Res 1995;7: 399-405. Wrone-Smith T, Johnson T, Nelson B, Boise LH, Thompson CB, Nunez G, et al. Discordant expression of bcl-x and bcl-2 by keratinocytes in vitro and psoriatic keratinocytes in vivo. Am J Pathol 1995;146:1079-88. Krajewski S, Krajewska M, Shabaik A, Wang HG, Irie S, Reed JC. Immunohistochemical analysis of in vivo patterns of bcl-X expression. Cancer Res 1994;54:5501-7. McCall CA, Cohen JJ. Programmed cell death in terminally differentiating keratinocytes: role of endogenous endonuclease. J Invest Dermatol 1991;97:111-4. 18. Polakowska R, Piacentini M, Bartlett R, Goldsmith LA, Haake A. Apoptosis in human skin development: morphogenesis, periderm and stem cells. Dev Dyn 1994;199:176-88. 19. Maruoka Y, Harada H, Mitsuyasu T, Yuji S, Kurokawa H, Kajiyama M, et al. Keratinocytes become terminally differentiated in a process involving programmed cell death. Biochem Biophys Res Commun 1997;238:886-90. 20. Farrow SN, White JHM, Martinou I, Raven T, Pun KT, Grinham CJ, et al. Cloning of a bcl-2 homologue by interaction with adenovirus EIB. Nature 1995;374:731-3. 21. Pena JC, Fuchs E, Thompson CB. Bcl-x expression influences keratinocyte survival but not terminal differentiation. Cell Growth Differ 1997;8:619-29. 22. Bargou RC, Daniel PT, Mapara MY, Bommer K, Wagner C, Kallinich B, et al. Expression of the bcl-2 gene family in normal and malignant breast tissue: low bax-alpha expression in tumor cells correlates with resistance towards apoptosis. Int J Cancer 1995;60:854-9. 23. Krajewski S, Blomqvist C, Franssila K, Krajewska M, Wasenius VM, Niskanen E, et al. Reduced expression of proapoptotic gene BAX is associated with poor response rates to combination chemotherapy and shorter survival in women with metastatic breast adenocarcinoma. Cancer Res 1995;55: 4471-8. 24. Rampino N, Yamamoto H, Ionov Y, Li Y, Sawai H, Reed JC, et al. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 1997;275(5302):967-9. 25. Yin C, Knudson CM, Korsmeyer SJ, VanDyke T. Bax suppresses tumorigenesis and stimulates apoptosis in vivo. Nature 1997;385:637-40. 26. Sedlak TW, Oltvai ZN, Yang E, Wang K, Boise LH, Thompson CB, et al. Multiple bcl-2 family members demonstrate selective dimerizations with bax. Proc Natl Acad Sci USA 1995; 92:7834-8. 27. Gibson L, Holmgreen SP, Huang DCS, Bernard O, Copeland NG, Jenkins NA, et al. Bcl-w, a novel member of the bcl-2 family, promotes cell survival. Oncogene 1996;13:665-75. 28. Choi SS, Park IC, Yun JW, Sung YC, Hong SI, Shin HS. A novel bcl-2 related gene, bfl-1 is overexpressed in stomach cancer and preferentially expressed in bone marrow. Oncogene 1995;11:1690-8. 29. Zhu W, Cowie A, Wasfy GW, Penn LZ, Leber B, Andrews DW. Bcl-2 mutants with restricted subcellular location reveal spatially distinct pathways for apoptosis in different cell types. EMBO J 1996;15:4130-41. 30. Bruel A, Karsenty E, Schmid M, McDonnell TJ, Lanotte M. Altered sensitivity to retinoid-induced apoptosis associated with changes in the subcellular distribution of Bcl-2. Exp Cell Res 1997;233:281-7.