Activator protein 1 transcription factors are fundamental to v-rasHa-induced changes in gene expression in neoplastic keratinocytes.

The induction of mouse skin papillomas by initiation-promotion protocols is associated with aberrant expression of epithelial markers in the tumor mass. Similarly, initiation of mouse keratinocytes with a retrovirus encoding the v-rasHa gene (v-rasHa keratinocytes) causes characteristic alterations of epidermal gene expression (A. A. Dlugosz et at, Cancer Res., 54: 6413-6420, 1994). Because activator protein 1 (AP-1) proteins are likely targets of Ras activation, we have examined the role of AP-1 factors in v-rasHa keratinocytes. Introduction of v-rasHa into keratinocytes up-regulates c-Fos, deltaFos B, and Fra-1 transcripts and protein levels in nuclear extracts. The expression of Jun proteins is not significantly altered in v-rasHa keratinocytes. Transduction of cells with v-rasHa results in increased AP-1-dependent transcriptional activity, which is also simulated by transfection of keratinocytes with either c-Fos or deltaFos B but not Fra-1, suggesting that the up-regulation of c-Fos and deltaFos B contributes to this effect. To explore the role of AP-1 proteins in regulating keratinocyte markers in v-rasHa keratinocytes, we blocked the binding of AP-1 proteins to DNA by infecting keratinocytes with an adenovirus encoding a dominant-negative Fos mutant (A-FOS). A-FOS replaces endogenous Fos proteins in the formation of heterodimers with Jun family members and thus prevents the AP-1 transcription factor from binding to DNA. In v-rasHa keratinocytes, the A-FOS virus reversed the suppression of keratins 1 and 10 transcripts and protein, which is characteristically seen in tumors and v-rasHa keratinocytes. A-FOS also increased protein levels but reduced transcripts for the late marker, loricrin, a component of the cornified envelope. These findings indicate that AP-1 proteins are involved in the changes in gene expression that define the v-rasHa phenotype in mouse keratinocytes.

CRE DNA binding proteins bind to the AP-1 target sequence and suppress AP-1 transcriptional activity in mouse keratinocytes.

Previously, we have shown that nuclear extracts from cultured mouse keratinocytes induced to differentiate by increasing the levels of extra-cellular calcium contain Fra-1, Fra-2, Jun B, Jun D and c-Jun proteins that bind to the AP-1 DNA binding sequence. Despite this DNA binding activity, AP-1 reporter activity was suppressed in these cells. Here, we have detected the CREB family proteins CREB and CREMalpha as additional participants in the AP-1 DNA binding complex in differentiating keratinocytes. AP-1 and CRE DNA binding activity correlated with the induction of CREB, CREMalpha and ATF-1 and CREB phosphorylation at ser133 (ser133 phospho-CREB) in the transition from basal to differentiating keratinocytes, but the activity of a CRE reporter remained unchanged. In contrast, the CRE reporter was activated in the presence of the dominant-negative (DN) CREB mutants, KCREB and A-CREB, proteins that dimerize with CREB family members and block their ability to bind to DNA. The increase in CRE reporter activity in the presence of these mutants suggests that CRE-mediated transcriptional activity is suppressed in keratinocytes through protein-protein interactions involving a factor that dimerizes with the CREB leucine zipper. In experiments where the A-CREB mutant was co-transfected with an AP-1 reporter construct, transcriptional activity was also increased indicating that a CREB family member binds AP-1 sites and represses AP-1 transcriptional activity as well. Exogenous expression of the transcriptional repressor CREMalpha down-regulated both CRE and AP-1 reporters in keratinocytes suggesting that this factor may contribute to the suppression of AP-1 transcriptional activity observed in differentiating keratinocytes.

Opposing activities of c-Fos and Fra-2 on AP-1 regulated transcriptional activity in mouse keratinocytes induced to differentiate by calcium and phorbol esters.

The major differentiation products of maturing keratinocytes contain AP-1 regulatory motifs, and AP-1 DNA binding activity increases in cultured keratinocytes induced to differentiate by calcium. Here, we have analysed AP-1 transcriptional activity in mouse keratinocytes treated with calcium and 12-O-tetradecanoyl phorbol-13-acetate (TPA), two agents that induce terminal differentiation of keratinocytes with different phenotypic consequences. Reporter constructs representing multimers of AP-1 sequences found in keratinocyte marker genes demonstrated that the calcium-induced AP-1 DNA binding activity does not correlate with transcriptional activation. Moreover, expression from active subunits of the profilaggrin and spr 1 promoters increased in calcium-treated keratinocytes when the AP-1 sites were disrupted, indicating that AP-1 may negatively regulate certain promoters in these cells. In contrast, AP-1 reporter activity was increased in keratinocytes treated with TPA. This induction was dependent upon the expression of c-Fos since AP-1 transcriptional activity was not increased in TPA-treated keratinocytes derived from c-fos null mice. Analysis of AP-1 protein expression in calcium- and TPA-treated keratinocytes demonstrated that only TPA increased the expression of c-Jun, while Jun B and Jun D were induced by both of these agents. c-Fos was expressed only in TPA treated keratinocytes, Fra-2 was expressed only in calcium-treated cells, and Fra-1 was expressed in both. Exogenous expression of Fra-2 repressed AP-1 transcriptional activity in TPA-treated keratinocytes, while c-Fos expression activated the AP-1 sequence in calcium-treated keratinocytes. These data indicate that Fra-2 and c-Fos play opposing roles in regulating AP-1 activity in keratinocytes and that multiple inducer-dependent regulatory pathways may exist for the expression of keratinocyte differentiation markers.

Identification of differentially expressed genes in chemically induced skin tumors.

Previous studies have demonstrated a role for the fos gene in promoting malignant conversion of mouse skin tumors. In the study reported here, differential display was performed to identify fos- and jun-regulated genes that are differentially expressed during premalignant progression. Total RNA isolated from variants of the papilloma cell line SP-1 transduced with retroviral vectors expressing v-jun and v-fos alone or in tandem was analyzed for the presence of differentially expressed transcripts by using 35 different primer combinations. Differentially expressed clones were rescreened by dot-blot analysis by using cDNA from chemically induced tumors with a high or low risk of malignant conversion. Three differentially displayed fragments were isolated in this analysis. Homology searches indicated that these fragments shared significant homology with the apoptosis inhibitor bcl-2, human alternative splicing factor/splicing factor 2 (ASF/SF2), and a novel gene not present in the GenBank or EMBL databases. In situ hybridization indicated that the expression levels of the bcl-2 homolog increased with malignant potential in chemically derived mouse skin tumors. A similar analysis indicated that expression of the ASF/SF2 homolog was greater in papillomas than in normal skin or in squamous cell carcinomas. Transcripts for this gene were most abundant in the granular layer. The expression pattern of the third differential display fragment was consistent with that of a tumor suppressor gene. This gene was expressed at very high levels in normal skin and benign papillomas but was essentially undetectable in squamous cell carcinomas. Through this approach, we identified known and novel genes that may contribute to malignant progression in epidermal tumors.

Loss of retinoic acid receptors in mouse skin and skin tumors is associated with activation of the ras(Ha) oncogene and high risk for premalignant progression.

Retinoic acid receptor transcripts (RARalpha and RARgamma) are decreased in benign mouse epidermal tumors relative to normal skin and are almost absent in carcinomas. In this report, the expression of RARalpha and RARgamma proteins was analyzed by immunoblotting in benign skin tumors induced by two different promotion protocols designed to yield tumors at low or high risk for malignant conversion. RARalpha was slightly reduced in papillomas promoted with 12-O-tetradecanoylphorbol-13-acetate (low risk) and markedly decreased or absent in papillomas promoted by mezerein (high risk). However, mezerein also caused substantial reduction of RARalpha in nontumorous skin. RARgamma was not detected in tumors from either protocol and was greatly reduced in skin treated by either promoter. Both RARalpha and RARgamma proteins were decreased in keratinocytes overexpressing an oncogenic v-ras(Ha) gene, and RARalpha was underexpressed in a benign keratinocyte cell line carrying a mutated c-ras(Ha) gene. Introduction of a recombinant RARalpha expression vector into benign keratinocyte tumor cells reduced the S-phase population and inhibited [3H]thymidine incorporation in response to retinoic acid. Furthermore, transactivation of B-RARE-tk-LUC by retinoic acid was markedly decreased in keratinocytes transduced with the v-ras(Ha) oncogene (v-ras(Ha)-keratinocytes). Blocking protein kinase C function in v-ras(Ha)-keratinocytes with bryostatin restored RARalpha protein to near normal levels, reflecting the involvement of protein kinase C in RARalpha regulation. Both RARalpha and RARgamma are down-regulated in cultured keratinocytes by 12-O-tetradecanoylphorbol-13-acetate, further implicating PKC in the regulation of retinoid receptors. Our data suggest that modulation of RARs could contribute to the neoplastic phenotype in mouse skin carcinogenesis and may be involved in the differential promoting activity of mezerein and 12-O-tetradecanoylphorbol-13-acetate, particularly for selecting tumors at high risk for malignant conversion.

Differentiation of mouse keratinocytes is accompanied by PKC-dependent changes in AP-1 proteins.

The conversion of cultured basal keratinocytes to the spinous and granular cell phenotypes seen in the skin can be stimulated by raising the levels of extracellular calcium. Here we show that AP-1 DNA binding activity is very low in primary cultures of basal keratinocytes, but that this activity is induced 24-48 h after increasing the concentration of extracellular calcium from 0.05 to 0.12 mM. As such, the induction of AP-1 DNA binding activity correlates with events occurring during the terminal stages of keratinocyte differentiation. Calcium-induced AP-1 DNA binding complexes consist of Fra-1, Fra-2, c-Jun, JunB and JunD and are independent of c-Fos, since the induction of DNA binding activity and the composition of the AP-1 binding complexes are identical in differentiating keratinocytes derived from c-fos null and wild type mice. The formation of calcium-induced AP-1 binding complexes is regulated by protein kinase C (PKC) and requires a functional PKCalpha isozyme, as determined through pharmacological down-modulation of specific PKC isozymes in differentiating keratinocytes. Moreover, PKC activation is required for the increased expression of Fra-2, JunB and JunD in the nucleus of differentiating cells in vitro. This observation provides a link between the obligate activation of PKC during keratinocyte differentiation and the nuclear response required to alter gene expression. In vivo expression patterns suggest that the predominant AP-1 heterodimer in the granular layer consists of Fra-2 and JunB while a JunD and Fra-1 complex predominates the spinous layer of mouse epidermis. These findings suggest distinct functions for different AP-1 proteins in the regulation of events related to keratinocyte maturation.

Staurosporine induces a sequential program of mouse keratinocyte terminal differentiation through activation of PKC isozymes.

Staurosporine (stsp) induces assembly of cornified envelopes in mouse keratinocyte cultures. To clarify whether this effect is the consequence of a coordinated differentiation program similar to that observed in epidermis, we assessed the expression of multiple differentiation-specific markers in stsp-treated keratinocytes. In medium containing 0.05 mM Ca2+, in which the basal cell phenotype is normally maintained, stsp induced dose-dependent increases in keratin 1, epidermal and keratinocyte transglutaminases, SPR-1, loricrin, and profilaggrin mRNA. Based on nuclear run-on analysis, stsp-mediated marker expression was found to be due at least in part to increased transcription. Since protein kinase C (PKC) activation is required for keratinocyte differentiation, we tested whether stsp influenced this signaling pathway. Stsp induced the translocation of multiple PKC isoforms from the cytosol to membrane and/or cytoskeletal fractions, inducing isozyme downregulation within 24 h. Moreover, AP-1 DNA binding activity was elevated in stsp-treated keratinocytes, consistent with the notion that this agent influences keratinocyte-specific gene expression via the PKC pathway. Stsp-mediated marker expression was inhibited by the PKC inhibitor GF 109203X. In cells pre-treated with bryostatin 1 to selectively down-modulate specific PKC isoforms, stsp-induced loricrin, filaggrin, and SPR-1 expression was suppressed when PKC alpha, epsilon, and/or delta were downregulated, suggesting that these isozymes may be necessary for marker expression in response to this agent. Thus, in addition to its effects on cornified envelope assembly, stsp induces a coordinate program of differentiation-specific keratinocyte gene expression that is mediated at least in part by the PKC signaling pathway.

c-fos is required for malignant progression of skin tumors.

The proto-oncogene c-fos is a major nuclear target for signal transduction pathways involved in the regulation of cell growth, differentiation, and transformation. Using the multistep skin carcinogenesis model, we have directly tested the ability of c-fos-deficient mice to develop cancer. Upon treatment with a tumor promoter, c-fos knockout mice carrying a v-H-ras transgene were able to develop benign tumors with similar kinetics and relative incidence as wild-type animals. However, c-fos-deficient papillomas quickly became very dry and hyperkeratinized, taking on an elongated, horny appearance. While wild-type papillomas eventually progressed into malignant tumors, c-fos-deficient tumors failed to undergo malignant conversion. Experiments in which v-H-ras-expressing keratinocytes were grafted onto nude mice suggest that c-fos-deficient cells have an intrinsic defect that hinders tumorigenesis. These results demonstrate that a member of the AP-1 family of transcription factors is required for the development of a malignant tumor.

Ultraviolet irradiation and c-jun over-expression regulates replication of polyoma sequences in WOP cells through a PEBP2 binding site.

Mouse fibroblast cells (WOP) express permissive factors which support polyoma DNA replication. However, electroporation into WOP cells of a mammalian expression vector that encodes the c-jun cDNA results in repression of polyoma DNA replication in a dose-dependent manner. In previous studies we have shown that UV-irradiation is capable of mediating a similar effect on polyoma DNA replication. When c-jun over-expression was combined with ultraviolet (UV)-irradiation, polyoma DNA replication decreased further. The repression of replication mediated by c-jun appears to be mediated by factor(s) that bind to PEBP4/2 target sequences as oligomers bearing the PEBP2/4 target site were capable of restoring polyoma DNA replication when added to UV-treated or c-jun over-expressing cells. The binding to the PEBP2/4 is partially dependent on the availability of AP-1 proteins, since an AP-1 target sequence can efficiently compete one of the three complexes formed with the PEBP2 target site. PEPB2 sequences do not, however, affect binding to the AP1 site. The effect of PEBP2 on polyoma replication is not dependent on the adjacent AP-1 site since PEBP2 could restore replication of polyomavirus which is mutated at the AP-1 sequence. A similar replication pattern was noted in a deletion mutant of polyoma which lacks PEBP4, yet, contains an intact PEBP2 binding sequence, suggesting that PEBP2 is the principle target for mediating repression of polyoma DNA replication.

Regulation of c-jun by lung carcinogens in Clara cells of hamsters.

In vitro differentiated hamster Clara cells were used to study the effects of lung carcinogens on the regulation of the c-jun oncogene. Northern blot analysis revealed a decrease in the expression of jun transcripts 24 h following the exposure of Clara cells to the direct acting forms of benzo[a]pyrene (BPDE*) or 5-methylchrysene (5MeCDE). To determine whether this decrease was mediated at the transcriptional level, we have used CAT reporter constructs driven by nested deletions of the 5' non-coding regulatory region of the c-jun oncogene. While BPDE was capable of activating certain regulatory domains of the c-jun promoter, this activation was not observed with either 5MeCDE or the less active lung carcinogens BADE or 6MeCDE. Analysis of enhancer elements identified the SP1 target site as a strong silencer after BPDE treatment. While positive regulatory element(s) mediating activation of c-jun by BPDE were localized within the promoter region up to -1639, further upstream sequences reduced this transcriptional activation. Thus, when the complete promoter region, up to -4500, was tested, no transcriptional activation was noted following BPDE treatment. These observations suggest that the regulation of c-jun in Clara cells exposed to potent lung carcinogens is mediated at the post-transcriptional level, possibly by reducing the stability and, in turn, the half life of c-jun mRNA. Overall, in contrast to the response of c-jun to numerous carcinogens and stress inducing agents noted in various other cell systems, our findings suggest the existence of a tissue-specific regulatory response for c-jun.

PEBP2--a modulator of polyoma DNA replication.

Previously, we have shown that integrated copies of polyoma DNA can be induced to replicate in rat fibroblasts (H3 cells) exposed to a DNA-damaging agent. In the current study, we demonstrate that UV-irradiation of mouse fibroblasts (WOP cells), transiently transfected with polyoma DNA, results in repression of polyoma replication. Cotransfection of oligomers representing wild-type but not mutated forms of the PEBP2 target sequence restored levels of viral replication indicating a role of PEBP2 binding proteins in mediating this effect. DNA-binding assays revealed that a different subset of complexes was formed with the PEBP2 target sequence when nuclear proteins from sham and UV-irradiated WOP and H3 cells were compared, suggesting that the activities of PEBP2 binding proteins are differentially regulated upon UV-irradiation in these two cell types. The ability of PEBP2 to modulate polyoma replication following UV-irradiation in WOP cells suggests a potential role of PEBP2 proteins in the cellular response to DNA damage.

Expression and transcriptional activity of AP-1, CRE, and URE binding proteins in B16 mouse melanoma subclones.

The expression and DNA binding activity of members of the activating protein-1 (AP-1) and activating transcription factor (ATF) families of transcription factors were analyzed in sham and ultraviolet (UV)-irradiated subclones of the B16 mouse melanoma cell system. The four subclones we used represent sequential stages in the development and progression of malignant melanoma and exhibit differences in growth and metastatic potential. Western blot analysis revealed differential expression of some AP-1 (c-jun, jun-B, and jun-D) and ATF (43- and 47-kDa cyclic AMP-responsive element binding protein (CREB) family members) in the different subclones; while c-jun expression was noted in the subclones with the greater malignant potential, jun-D was expressed in those with the lesser malignant potential. Furthermore, a delicate balance between the two forms of CREB was noted; the 47-kDa CREB appeared, when expressed exclusively, in subclones that exhibit a greater malignant potential. Electrophoretic mobility shift assays using AP-1, CRE, and UV-responsive element (URE) consensus sequences indicated that distinct complexes were formed with extracts from each of the four subclones. The complexes were competitively inhibited by each of the target sequences used, suggesting that "cross-talk" occurs between some AP-1 and ATF family members in this cell system. Moreover, a multimer of the URE sequence, cloned upstream of a chloramphenicol acetyltransferase reporter gene, was transcriptionally active and responsive to UV irradiation in two of the four subclones. UV-related transcriptional activation was directly correlated with the expression of a 43-kDa CREB. Together, these observations identify members of AP-1 and CREB families whose expression and activities correlate with the malignant potential of subclones that represent different stages in melanoma development and progression.

UV-responsive element (TGACAACA) from rat fibroblasts to human melanomas.

When taken together, several lines of evidence suggest that the URE-bound proteins are associated with DNA replication. (1) A trans-acting factor of 60 kDa (which may include the 68-kDa URE-binding proteins) was found to be induced by UV and to mediate polyoma DNA replication; (2) the URE was able to compete for the binding of factors that promote polyoma replication in rodent cells; (3) URE-bound proteins are expressed to a higher extent at the S phase of the cell cycle; and (4) they are induced following treatment with aphidicholin. These observations may suggest that the URE may play a role in growth "release" (as opposed to growth "arrest") which would assist in restoring normal growth following DNA damage. It is clear that URE-bound proteins consist of multiple transcription factors, some of which are well characterized (i.e., jun, CREB, and fos families); however, it is likely that the growth release phenomenon we relate to is also mediated by (1) other members of these transcription factor families which have not been identified as yet and (2) a specific combination of known transcription factors which bind to this response element under certain circumstances. This hypothesis is outlined in the enclosed model (Fig. 3).

General purpose lamps induce polyoma DNA replication in H3 cells.

We have previously demonstrated the ability of UVC (254 nm) radiation to induce asynchronous polyoma replication in rat fibroblast cells (H3 line) that contain an integrated copy of polyoma virus. In the present study we show that general purpose lamps can induce polyoma replication in these cells as well. The amount of UV radiation emitted by three different light sources was determined and the effects of each source on the replication of polyoma DNA was assessed. Our findings indicate that a 100 W incandescent lamp had a minimal effect on replication, whereas a 90 s exposure to a halogen lamp or a 160 W mercury vapor lamp induced replication 1.5-fold and 2-fold, respectively, in comparison with nontreated controls. We have previously shown that asynchronous polyoma replication in H3 cells involves UV-inducible cellular protein factors. Our present results indicate that these factors are also activated by exposure to commonly used lamps that emit comparable doses of UV radiation.

A transcriptional inhibitor induced in human melanoma cells upon ultraviolet irradiation.

Nuclear proteins from human melanoma cells exhibit strong binding activity to the UV response element (TGACAACA); however, this binding is inhibited following UV-C irradiation. In contrast, the binding of nuclear proteins from rodent fibroblasts and human keratinocytes to the UV-responsive element is initially weak and increases significantly upon UV irradiation. The addition of nuclear proteins from UV-irradiated melanoma cells to those prepared from nonirradiated cells inhibited the binding to the UV-responsive element in a concentration-dependent manner. Fast protein liquid chromatographic analysis of nuclear proteins from UV-irradiated melanoma cells revealed 12 and 14 kilodalton proteins within a fraction which also contained the inhibitory activity. The inhibitor blocks the binding of proteins to three other target sequences, AP1, CREB, and PEBP2, as well as the in vitro transcription of SV40 promoter sequences. The inhibitor was also found in UV-irradiated melanocytes, suggesting that it is tissue specific. The induction of a transcriptional inhibitor in response to UV irradiation represents a regulatory event that may play an important role in the transcriptional response of both normal and malignant melanocytes to UV irradiation.

Expression pattern of proteins that bind to the ultraviolet-responsive element (TGACAACA) in human keratinocytes.

We previously identified an ultraviolet (UV)-responsive element (URE; TGACAACA) that plays a role in both transcription and replication of polyoma sequences. Mouse polyclonal antibodies were raised against affinity-purified URE-bound proteins to characterize their expression patterns. These antibodies specifically recognized two of four URE-bound proteins, of 40 and 68 kDa. The 68-kDa protein was constitutively expressed in human keratinocytes, while the expression of the 40-kDa protein was induced by UV irradiation. Of the two, the 68-kDa protein bound to the URE with greater affinity than the 40-kDa protein, as determined by southwestern analysis. The expression of the 40-kDa protein increased as early as 1 h after UV irradiation of both rat fibroblasts and human keratinocytes and correlated with increased binding to the URE in an electrophoretic mobility shift assay. Other types of damage, as well as heat shock and serum stimulation, also induced the expression of this protein, suggesting that it may play a role in cellular response to stress or damage. The 40-kDa protein was expressed at the highest levels in the S phase of the cell cycle and was induced by aphidicolin, suggesting that it has a role in DNA replication. All together, these results suggest that exposure of human keratinocytes to damage- and stress-inducing agents modulates the expression of proteins that may play a role in regulating cellular response to DNA damage.

Functional role of the ultraviolet light responsive element (URE; TGACAACA) in the transcription and replication of polyoma DNA.

We have previously identified a novel 8 bp sequence (UV-responsive element, URE: TGACAACA) present in the regulatory region of polyoma DNA that interacts with protein factors induced in rat fibroblast cells by exposure to UV light. In the present study, we demonstrate through competitive binding assays that this sequence is distinct from the partially homologous AP1 and CRE target sequences. The proteins that bind to the URE appear to have transcriptional activity in UV-exposed rat fibroblasts. In addition, the URE appears to play a role in promoting the replication of polyoma DNA as determined through two different experimental approaches. Together, these findings suggest that the URE is a novel DNA binding element that interacts with proteins involved in the transcription and replication of polyoma sequences.

Aldolase-DNA interactions in a SEWA cell system.

In this report we demonstrate the novel finding that aldolase A interacts with DNA sequences in mouse SEWA sarcoma cells. This interaction was initially observed through the identification of a 40 kDa protein which was eluted from a DNA affinity chromatography column consisting of the long terminal repeat (LTR) of the endogenous intracisternal A-type particle (IAP). Microsequencing analysis identified this 40 kDa protein as the glycolytic enzyme, aldolase A. The use of specific anti-aldolase antibodies enabled the identification and subsequent purification of aldolase from the nuclear protein fraction of two SEWA sublines, one that is adherent and one that grows in suspension. In order to confirm our initial finding that aldolase is capable of interacting with DNA, proteins from each subline were immunopurified with anti-aldolase antibodies, eluted and then tested for their ability to interact with IAP-LTR DNA sequences. Interestingly, only aldolase derived from the anchorage dependent SEWA cells was capable of interacting with the IAP-LTR, however, several cell lines derived from human tumors also exhibited this activity. Subsequent studies revealed the ability of aldolase to interact with some but not every DNA sequence tested, implying that there may be a minimal DNA conformation and/or sequence requirement for this activity. The presence of aldolase A in the nuclei and its ability to interact with certain DNA sequences suggest a novel role for this metabolic enzyme.