Register for Mechanobiology 2016!


Keratinocyte differentiation requires integrating signaling among intracellular ionic changes, kinase cascades, sequential gene expression, cell cycle arrest, and programmed cell death. We now show that Cl intracellular channel 4 (CLIC4) expression is increased in both mouse and human keratinocytes undergoing differentiation induced by Ca2+, serum and the protein kinase C (PKC)-activator, 12-O-tetradecanoyl-phorbol-13-acetate (TPA). Elevation of CLIC4 is associated with signaling by PKCδ, and knockdown of CLIC4 protein by antisense or shRNA prevents Ca2+-induced keratin 1, keratin 10 and filaggrin expression and cell cycle arrest in differentiating keratinocytes. CLIC4 is cytoplasmic in actively proliferating keratinocytes in vitro, but the cytoplasmic CLIC4 translocates to the nucleus in keratinocytes undergoing growth arrest by differentiation, senescence or transforming growth factor β (TGFβ) treatment. Targeting CLIC4 to the nucleus of keratinocytes via adenoviral transduction increases nuclear Cl content and enhances expression of differentiation markers in the absence of elevated Ca2+. In vivo, CLIC4 is localized to the epidermis in mouse and human skin, where it is predominantly nuclear in quiescent cells. These results suggest that CLIC4 participates in epidermal homeostasis through both alterations in the level of expression and subcellular localization. Nuclear CLIC4, possibly by altering the Cl and pH of the nucleus, contributes to cell cycle arrest and the specific gene expression program associated with keratinocyte terminal differentiation.


The CLIC family of chloride intracellular channels is composed of seven differentially expressed proteins that localize to the cytoplasm and intracellular organelles in many cell types (Suh and Yuspa, 2005). Five members of the CLIC family are highly homologous in size and sequence (CLIC1-5), whereas p64 (CLIC5B) and parchorin (CLIC6) have extended amino-terminal domains and are considerably larger. Crystallographic analysis of CLIC1 and CLIC4 indicates that the protein exists in both soluble and membrane bound forms and is structurally related to the omega class of glutathione S-transferases (Harrop et al., 2001; Littler et al., 2005). Several CLIC proteins have Cl selective channel activity (Ashley, 2003), but it is not yet clear if they form pores by themselves or participate in ion flux indirectly (Jentsch et al., 2002).

Recent data from several laboratories have suggested that CLIC family members also participate in cell signaling. CLIC3 interacts with ERK7 in the nucleus of mammalian cells and stimulates chloride conductance (Qian et al., 1999). p64 is a substrate for Fyn kinase, and the combination enhances chloride channel activity when co-expressed in HeLa cells (Edwards and Kapadia, 2000). CLIC1 is associated with TNFα release in amyloid-β-treated rat microglial cells (Novarino et al., 2004). CLIC4 is induced by TGFβ in mammary fibroblasts as they transdifferentiate into myofibroblasts (Ronnov-Jessen et al., 2002), and CLIC4 binds to dynamin 1 in a complex involving actin, tubulin and 14-3-3ζ in rat brain extracts (Suginta et al., 2001). CLIC4 also interacts with nuclear transport factor 2 (NTF2), Ran and importin-α, active components of the nuclear import machinery, suggesting that the nuclear localization signal (NLS)-recognition protein complex may be involved in CLIC4 nuclear trafficking (Suh et al., 2004). The physiological consequences of these interactions is under study, but CLIC1 and CLIC4 has been implicated in regulation of the G2-M phase of the cell cycle in CHO cells (Berryman and Goldenring, 2003), and CLIC4 is associated with adipocyte differentiation in 3T3-L1 cells (Kitamura et al., 2001).

In keratinocytes, an increase in CLIC4 expression is associated with Ca2+-induced differentiation, DNA damage, p53 upregulation and exposure to TNFα (Fernandez-Salas et al., 1999; Fernandez-Salas et al., 2002; Suh et al., 2004). In addition to cytoplasmic residence, CLIC4 is a component of the inner mitochondrial membrane, and maintenance of CLIC4 levels is essential for viability of keratinocytes (Fernandez-Salas et al., 2002). Thus, those conditions that elevate CLIC4 are associated with cell cycle arrest or cell death. CLIC4 elevation in differentiating keratinocyte could participate in the terminal phase of the differentiation process when keratinocytes undergo cell cycle arrest, karyorrhexis and die. Cytoplasmic CLIC4 translocates to the nucleus in stressed cells (Suh et al., 2004), and nuclear CLIC4 could also participate in other aspects of the differentiation response such as the gene expression changes that are associated with differentiation. Previous studies have suggested that changes in intracellular chloride and pH are associated with keratinocyte differentiation (Mauro et al., 1990; Mauro et al., 1993). This study was undertaken to document the changes in CLIC4 protein that may be associated with keratinocyte differentiation and to explore the potential function of these changes in the differentiation response of keratinocytes to increasing extracellular Ca2+.


CLIC4 is Ca2+ responsive and PKC dependent in mouse and human keratinocytes

Previous results had indicated that CLIC4 transcripts and protein increased in mouse keratinocytes induced to differentiate by Ca2+ (Fernandez-Salas et al., 1999). In mouse keratinocytes, CLIC4 protein and transcript levels increase after 24 hours relative to the level of extracellular Ca2+ (Fig. 1A,B). Both Ca2+ and serum also elevate CLIC4 in human foreskin keratinocytes (Fig. 1A,C). Ca2+ is known to activate several protein kinase C (PKC) isoforms in keratinocytes (Denning et al., 1995). To determine if PKC signaling is involved in CLIC4 expression, mouse and human keratinocytes were treated with the PKC activator TPA. As shown in Fig. 1D, this treatment increased CLIC4 levels within 24 hours also, suggesting a PKC-regulated process. To explore the connection among Ca2+ signaling, PKC activation, keratinocyte differentiation and CLIC4 induction, mouse keratinocytes induced to differentiate by 0.5 mM Ca2+ were treated with selective PKC inhibitors and probed for expression of CLIC4, keratin 1 (K1) and keratin 10 (K10; Fig. 2A). In control cells, both CLIC4 and the differentiation markers increase by 24 hours after Ca2+ induction. The PKCα selective inhibitor GO6976 did not inhibit either CLIC4 or the differentiation marker response to Ca2+. In fact, both responses increased with GO6976, consistent with the negative effect of PKCα activation on markers of spinous differentiation reported earlier (Lee et al., 1997). By contrast, the general PKC inhibitor bisindoylmaleimide-I at 10 μM and the PKCδ selective inhibitor rottlerin reduced expression of both CLIC4 and K1, and K10, implying that PKCδ regulates CLIC4 expression. Fig. 2B supports this conclusion since both differentiation markers and CLIC4 expression in Ca2+-induced keratinocytes were depressed by infection with an adenovirus encoding a PKCδ dominant negative construct previously shown to effectively and selectively inhibit PKCδ activity in keratinocytes (Li et al., 1999). In the presence of the dominant negative construct or the pharmacological inhibitor rottlerin, CLIC4 expression in high Ca2+ culture conditions was reduced toward or below the level in low Ca2+ culture. To test if CLIC4 is a specific and direct substrate for PKC, we incubated recombinant CLIC4 and CLIC1 with recombinant PKCα, β, γ, δ and ϵ in an in vitro kinase reaction (Fig. 2C). In all cases, phosphorylation was detected on both substrates, indicating a lack of specificity. Furthermore, we were unable to detect phosphorylation of CLIC4 after treatment of keratinocytes with TPA (data not shown). Taken together, the results suggest that CLIC4 is regulated by activation of PKCδ but not by a specific direct phosphorylation of CLIC4 by PKCδ. Instead, PKCδ may regulate the expression or stability of CLIC4 indirectly and the upregulation of CLIC serves as a downstream mediator of keratinocyte differentiation.

Fig. 1.

CLIC4 is upregulated in differentiating keratinocytes. (A) Newborn mouse and human foreskin keratinocytes were cultured in 0.05 mM Ca2+ medium for 4 days, then treated for 24 hours with the indicated concentrations of Ca2+ in the culture medium and lysates were examined by immunoblotting. (B) RNA isolated from mouse keratinocytes from A was amplified by RT-PCR, and CLIC4 and actin transcripts were detected by ethidium bromide staining of agarose gels. (C) Human foreskin keratinocytes were cultured in serum-free medium for 4 days and exposed to 5% FBS deleted of Ca2+ serum for the indicated times. (D) Mouse and human keratinocyte were treated with indicated doses of TPA for 24 hours, and lysates were analyzed by immunoblotting. The fold change for each band was determined by densitometry as described in Materials and Methods.

Upregulation of CLIC4 is required for Ca2+-induced keratinocyte differentiation

To determine if upregulation of CLIC4 is essential for Ca2+-induced differentiation in keratinocytes, we reduced CLIC4 level by expressing CLIC4 antisense in an adenoviral vector (AS-CLIC4 Ad) in keratinocytes exposed to 0.5 mM (Hi) Ca2+ medium. We have previously shown that this vector reduces CLIC4 expression in a dose-dependent manner (Suh et al., 2005). Under conditions in which AS-CLIC4 adenovirus (Ad) reduced CLIC4 and prevented the Ca2+-induced increase in CLIC4 expression, K1 and K10 proteins were not elevated at 24 hours after Ca2+ addition, and the expression of the late differentiation marker filaggrin was inhibited after 48 hours (Fig. 3A). Expression of AS-CLIC4 did not alter the endogenous level of PKCδ (Fig. 3A), suggesting that CLIC4 upregulation is an important component of the differentiation response to elevated Ca2+. To verify this observation by an independent differentiation marker, cell cycle parameters were monitored to determine if the rapid G1 cell cycle arrest associated with Ca2+-induced differentiation, was altered (Dotto, 1999). Within 24 hours of 0.5 mM Ca2+ exposure, BrdU pulse-labeled cells were absent when assayed by flow cytometry whereas the number in G0-G1 increased substantially (Fig. 3B,C). However, these changes were prevented by expression of CLIC4 antisense, and the proliferating population was maintained up to 48 hours as in the 0.05 mM Ca2+ medium (Fig. 3C). To verify CLIC4 is a specific CLIC family member that mediates the keratinocyte differentiation response, non-specific (NS) and specific shRNA constructs were developed for CLIC1 (sR-1), CLIC4 (sR-4) and CLIC5 (sR-5) and used for a transient knockdown of CLIC members (Fig. 3D,E). When transfected cells are induced to differentiate by 0.5 mM Ca2+, CLIC4 shRNA prevented upregulation of K10 but the other shRNAs did not. This result also distinguishes CLIC4 from other CLIC proteins that might have been influenced by AS-CLIC4 expression (Suh et al., 2005).

Fig. 2.

CLIC4 is a downstream target of PKCδ in keratinocytes induced to differentiate by Ca2+. (A) Mouse keratinocytes were pre-treated with the indicated doses of PKC inhibitors for 1 hour, and then 0.5 mM Ca2+ was added in the medium for 24 hours prior to lysis and immunoblotting. (B) Mouse keratinocytes were infected with PKCδ DN Ad for 17 hours before the treatment with 0.5 mM Ca2+ for 24 hours. Actin was used as a loading control. Bis, bisindolylmaleimide I; Ro, rottlerin; Go, Go 6976. Similar data were obtained from two independent experiments. The numbers over the CLIC4 bands in A and B represent fold change determined by densitometry. (C) Purified recombinant CLIC1 or CLIC4 proteins were incubated with purified, kinase active recombinant PKC isoforms in appropriate reaction buffers with [γ-32P]ATP, the reaction mixture was resolved by SDS-PAGE and then the phosphorylated proteins were visualized by phospho-imaging methods. Un, no recombinant PKC.

Fig. 3.

CLIC4 upregulation is required for Ca2+-induced differentiation in keratinocytes. K1, K10, Filaggrin, CLIC1, CLIC4, PKCδ and actin protein levels in mouse keratinocytes were determined by immunoblot analysis. (A) Mouse keratinocytes were co-infected with 5 MOI of Tet-Off Ad (to induce expression of the conditional AS) and increasing MOI (5 and 10, shaded triangle) of AS-CLIC4 Ad or Null Ad (N) for 17 hours before treatment with 0.5 mM Ca2+ for 24 hours. (B) Duplicate sets of mouse keratinocytes were treated with BrdU, and anti-BrdU-FITC stained cells of each sample were analyzed with flow cytometry. Cell populations: null Ad-infected without BrdU (upper left, N/(–)/Lo Ca2+), with BrdU (upper right, N/(+)/Lo Ca2+) both in 0.05 mM Ca2+; with 0.5 mM Ca2+ and BrdU (lower left, N/(+)/Hi Ca2+); co-infected with AS-CLIC4 Ad and Tet-Off Ad followed by Ca2+ and BrdU treatment (lower right, AS/(+)/Hi Ca2+). (C) The key indicates the gated population in each cell cycle phase. The percentage of cells in each cell cycle phase from the total cell population in N/(+)/Lo Cal2+, N/(+)/Hi Ca2+ and AS/(+)/Hi Ca2+ at 24 hours and 48 hours after Ca2+ induction is represented in a bar graph format. (D) Human 293 cells were transfected with constructs expressing nonspecific (NS) or CLIC4 (sR-CLIC4) shRNA for 48 hours, and specificity of CLIC4 knockdown was analyzed by immunoblots of cell lysates. (E) Mouse keratinocytes were transfected with constructs expressing shRNA for nonspecific (NS), CLIC1 (sR-1), CLIC4 (sR-4) or CLIC5 (sR-5) for 24 hours, and then treated with 0.05 mM (Lo) or 0.5 mM (Hi) Ca2+. Actin was used as a loading control. Similar data were obtained from at least two independent experiments. The numbers over the CLIC4 bands in A and B represent fold changes determined by densitometry.

CLIC4 translocates to the nucleus in differentiating keratinocytes in vitro and is localized to the nucleus in differentiating keratinocytes in vivo

Previous data have shown that cytoplasmic CLIC4 translocates to the nucleus of cells undergoing a variety of stress responses (Suh et al., 2004). Similarly, CLIC4 translocates to the nucleus in both mouse and human keratinocytes that are induced to differentiate by Ca2+ (Fig. 4A). Interestingly, CLIC4 also translocates to the nucleus in keratinocytes undergoing G1 arrest because of senescence (Fig. 4B) or serum starvation (Fig. 4C) or when arrested by TGFβ (Fig. 4D) or induced to differentiate by TPA (Fig. 4E) treatment. This intracellular trafficking of CLIC4 is dynamic since the nuclear CLIC4 in the serum-starved cells translocates back to the cytoplasm within 30 minutes upon addition of serum to the medium, suggesting that CLIC4 has a physiological function during cell cycle arrest and release (Fig. 4C). Translocation of CLIC4 to the nucleus in keratinocytes undergoing differentiation and senescence in vitro prompted us to investigate the localization of CLIC4 in vivo. In contrast to the cytoplasmic pattern of CLIC4 seen in the actively proliferating keratinocytes in culture, immunohistochemical studies demonstrate that CLIC4 is predominantly nuclear in the epidermis of adult human and mouse skin (Fig. 4F). CLIC4 is often nuclear in basal cells, but is excluded from the nucleus in a subpopulation of basal cells and some spinous cells. Overall CLIC4 cellular staining generally increases in the suprabasal region where differentiation is occurring. Almost all granular layer keratinocytes express CLIC4 in the nucleus. This staining pattern suggests that cycling keratinocytes are regulating the subcellular localization of CLIC4. It should be noted that nuclear CLIC4 can be detected in certain regions of the stromal compartment, most likely in endothelial cells where CLIC4 is known to be important for blood vessel tubulogenesis (Bohman et al., 2005).

Nuclear CLIC4 enhances the differentiation response

To address whether upregulation and nuclear translocation of CLIC4 is sufficient to induce expression of differentiation markers, keratinocytes in 0.05 mM Ca2+ medium were infected with adenoviruses encoding intact CLIC4 (Cyt-CLIC4) or CLIC4 with a nuclear targeting signal (Nuc-CLIC4). These adenoviral vectors infect keratinocytes with high efficiency and target appropriate organelles as described previously (Suh et al., 2004). At relatively low levels of expression over endogenous CLIC4 levels, both constructs upregulated CLIC4, K1 and K10 in the absence of a Ca2+ switch, but expression of the markers was greater when CLIC4 was targeted to the nucleus (Fig. 5A). Thus, a small increase in the exogenous Nuc-CLIC4 expression, above the endogenous level of CLIC4, caused a significant increase in the keratinocyte differentiation response. Primary keratinocytes cultured in basal medium (0.05 mM Ca2+) actively proliferate but individual cells spontaneously differentiate and detach from the culture dish. Confocal studies (Fig. 5B) show that the increased and nuclear expression of CLIC4 coincides (merged image) with the upregulation of K1 in keratinocytes that are spontaneously undergoing differentiation in the absence of extracellular Ca2+. In this setting, the nuclear CLIC4 is detected only in the subpopulation of differentiating keratinocytes while the cytoplasmic CLIC4 is detected in the neighboring cells. Taken together, the results shown in Fig. 5A and B suggest that nuclear translocation of CLIC4 coincides with the expression of early differentiation markers during keratinocyte differentiation.

Fig. 4.

CLIC4 translocates to the nucleus in differentiating and senescing keratinocytes in vitro and is primarily nuclear in epidermis in vivo. (A) Mouse primary keratinocytes and human foreskin keratinocytes were treated with indicated Ca2+ for 24 hours and immunostained with CLIC4 antibody. (B) Mouse keratinocytes were maintained in 0.05 mM Ca2+ medium and either fixed on the first day of culture (1 Day) or after 7 days of culture (7 Day) to allow for senescence. Fixed cultures were immunostained with CLIC4 antibodies. Duplicate sets of cells were subjected to assay by β-gal SA, and the number of senescing cells (7 Day) is presented as a fold increase over the control (1 Day; lower panel). (C) Mouse keratinocytes were incubated in the serum-depleted medium for 16 hours (0 min), rinsed and treated with the serum-containing medium (30 min). (D,E) Mouse keratinocytes were treated with TGFβ (1 ng/ml; D) or TPA (25 ng/ml; E) for the times indicated and immunostained for CLIC4. Insets in Fig. 4C-E represent DAPI nuclear staining. (F) Formalin-fixed human breast skin and acetone-fixed mouse skin frozen sections were stained with CLIC4 antibodies by immunohistochemical methods as described in the Materials and Methods. The mouse skin section was from a hyperplastic area to enhance the resolution of strata. Red and green arrows indicate keratinocytes in the basal layers (blue nucleus) and the differentiating layers (brown nucleus, CLIC4 stain) respectively. In all immunofluorescent and immunohistochemical staining experiments, secondary antibody alone was used as controls.

Fig. 5.

Nuclear CLIC4 induces the expression of keratinocyte differentiation markers, and intracellular chloride influences keratinocyte differentiation. (A) Mouse keratinocytes in 0.05 mM Ca2+ medium were pre-treated with null, wild-type CLIC4 (Cyt) or nuclear-targeted CLIC4 (Nuc) adenovirus for 12 hours, and cell lysates were analyzed by immunoblots using K1, K10, CLIC4 and actin antibodies. Gray arrow indicates the endogenous CLIC4 (Endo-CLIC4) and the black arrows indicate the two exogenous CLIC4 bands. (B) Mouse keratinocytes were cultured in 0.05 mM Ca2+ medium and double-immunostained with CLIC4 and K1 antibodies. Examples of spontaneously differentiating keratinocytes are indicated by white arrows. All immunostained cells were analyzed by confocal microscopy as described in the Materials and Methods. Insets show DAPI staining. (C) Mouse keratinocytes in 0.05 mM Ca2+ medium were infected with adenoviruses encoding β-gal, wild-type CLIC4 (Cyt-CLIC4), CLIC4 (Nuc-CLIC4) or nuclear-targeted GFP (Nuc-GFP) for 16 hours, treated with MQAE fluorescent dyes and assayed by confocal microscopy. (Insets) The chloride ion content of the β-galactosidase, Cyt-CLIC4 or Nuc-CLIC4 Ad infected cells treated with NPPB was visualized by confocal microscopy. (D) The fluorescence intensity was also measured in similarly treated cells. Fluorescence on the y axis is shown in arbitrary units and decreasing values (quenching) indicates greater Cl content. (E) Mouse keratinocytes were treated with increasing amount of NPPB (gradient triangle, 200 μM for 1× and 2 mM for 10×) for 2 hours prior to calcium (0.5 mM) induction. After 24 hours, cell lysates were analyzed by immunoblots using K10, CLIC4 and actin antibodies.

Nuclear-targeted CLIC4 causes Cl flux into the nucleus in keratinocytes, and Cl flux is required for expression of Ca2+-induced K10

CLIC proteins are involved in flux of chloride ions in various intracellular compartments (Tulk et al., 2002). To investigate if nuclear-specific CLIC4 alters chloride ion flux, the chloride ion-sensitive fluorescent dye MQAE was used. With this dye, the fluorescent light emission is inversely related to the Cl concentration of the MQAE-containing intracellular compartment, and the fluorescent output can be measured semi-quantitatively and visualized by fluorometer and confocal microscopy. Fig. 5C indicates that targeting CLIC4 to the nucleus of keratinocytes via an adenovirus reduces fluorescence of nuclear MQAE detected by confocal microscopy, indicating influx of chloride ion into the nucleus. By contrast, expression of β-galactosidase, cytoplasmic CLIC4 or targeting GFP to the nucleus does not cause significant MQAE quenching, indicating that the CLIC4 result is not a consequence of overexpression of any exogenous protein in the cytoplasm or nucleus. This dye quenching effect can be blocked by treating the cells with a chloride channel inhibitor NPPB [5-nitro-2-(3-phenylpropylamino) benzoic acid] prior to Ad infection (Fig. 5C, first, second and third panel insets). In this experimental setting, overloading the culture medium with NaF (Cl depletory agent) was used as a control to maximize the fluorescence. When NaF-treated cells were washed and reloaded with excess NaCl, the fluorescence was quenched again in the nucleus of nuclear-targeted CLIC4 cells but not in the cytoplasmic CLIC4 control, suggesting that the fluorescence is quenched by the nuclear CLIC4 (data not shown). When whole cell MQAE fluorescence is measured by fluorometry, the cells expressing Nuc-CLIC exhibit significant quenching of MQAE compared with the cells expressing wild-type (Cyt) CLIC4 (Fig. 5D). To investigate whether chloride ion flux is required for Ca2+-induced keratinocyte differentiation, cells were treated with a chloride channel inhibitor prior to Ca2+ induction. Treatment of keratinocytes with NPPB significantly reduced the expression level of K10 upon Ca2+ induction (Fig. 5E). At the highest concentration of NPPB (2 mM), both K10 and CLIC4 are reduced, a change that was associated with morphological evidence of toxicity.

CLIC4 is a down-stream target of AP1

AP1 signaling is involved in Ca2+- and TPA-induced keratinocyte differentiation through the regulation of transcription of multiple maturation markers (Eckert and Welter, 1996; Angel et al., 2001). In previous work we have shown that the expression of a highly specific AP1 dominant negative recombinant protein A-Fos blocks expression of several keratinocyte differentiation markers (Rutberg et al., 1997). A-Fos binds to Jun family members with high affinity but does not transactivate, thus serving as a pan AP1 inhibitor for keratinocytes (Gerdes et al., 2006). Inhibition of AP1 activity by A-Fos expression also blocks the upregulation of CLIC4 protein in keratinocytes induced to differentiate by Ca2+ (Fig. 6A,B). Inhibition of CLIC4 transcripts coincides with inhibition of Ca2+-induced transcripts for involucrin and K1 as demonstrated by quantitative PCR (Fig. 6B). CLIC4 expression is regulated by elements 5′ from the transcription start site of the first exon (Fernandez-Salas et al., 2002). Contained within this region are functional binding sites for p53 and Myc (Fernandez-Salas et al., 2002; Shiio et al., 2006). Further analysis reveals two consensus AP1 binding sites at –1507 to –1499 (site A) and –840 to –832 (site B; Fig. 6D). A luciferase reporter driven by the active human CLIC4 promoter (Fig. 6C) incorporating these sites is induced by high Ca2+ medium when transfected into mouse keratinocytes, and this is prevented by A-Fos. Analogous sites for AP1 regulation are detected in the mouse CLIC4 promoter (Fig. 6D). When either site A or B, or both, in the human CLIC4 promoter is mutated, luciferase reporter activity in response to 0.5 mM Ca2+ is abrogated (Fig. 6E). In addition, nuclear extracts from keratinocytes bind to each of these sites, and binding is enhanced in extracts derived from keratinocytes induced to differentiate by Ca2+ (Fig. 6F). Mutation of either site A or site B abrogates binding to that site, and A-Fos expression in keratinocytes prevents binding to either site. Together, these data reveal the functional importance of AP1 in the regulation of CLIC4 and indicate that AP1 activity is sufficient to increase CLIC4 expression in keratinocytes induced to differentiate by Ca2+.


Flux of different ions that are specifically modulated by membrane transport proteins including channels, pumps and carriers sets an environment that favors the initiation of differentiation programs in keratinocytes and other cell types (Hennings et al., 1983a; Mauro et al., 1993; Debska et al., 2001; Ronnov-Jessen et al., 2002). The murine intracellular chloride channel protein CLIC4 was identified as a differentially expressed gene in p53-wild-type versus p53-null keratinocytes that were induced to differentiate by Ca2+. Subsequently, CLIC4 was proved to be a direct downstream target of p53 in keratinocytes that was also regulated independently of p53 (Fernandez-Salas et al., 1999; Fernandez-Salas et al., 2002). We now report that CLIC4 is directly involved in mouse and human keratinocyte differentiation, and nuclear translocation of CLIC4 is an integral part of the Ca2+-induced differentiation response. Thus, expression of keratinocyte differentiation markers (K1, K10 and filaggrin) and a rapid G1 cell cycle arrest are strongly associated with CLIC4 upregulation and nuclear translocation. Reduction of CLIC4 by antisense or shRNA approaches prevents expression of these differentiation markers and prevents cell growth arrest associated with differentiation. CLIC4 is also frequently found in the nucleus of differentiating cells from mouse and human epidermis in vivo. The findings reported here also suggests that nuclear translocation of CLIC4 is associated with increased Cl ion content in the nucleus, and this ionic flux could be related to the nuclear events that are associated with activation of intrinsic differentiation programs. The signals downstream from Ca2+ that regulate expression and translocation of CLIC4 in differentiating keratinocytes remain to be delineated. However, our data suggest that PKCδ is an important intermediary in the regulation of CLIC4, and CLIC4 expression is more directly regulated by AP1 factors, joining a host of other differentiation-induced AP1-regulated markers in keratinocytes (Eckert and Welter, 1996; Angel et al., 2001). Further studies are required to determine if PKCδ is working through AP1 to regulate CLIC4 expression. The human CLIC4 promoter encompasses two AP1 consensus binding sites (–1507 to –1499; gTGAaTCAg and –840 to –832; gTGAgTCAa). These sites are conserved in a similar region upstream from the first exon in the mouse CLIC4 promoter. Mutation of these sites abrogates the activation of the CLIC4 promoter in response to Ca2+, indicating they are functional regulatory sites, and ELISA-based EMSA assays confirm binding activity. AP1 factors now join with p53 and Myc as key regulators of this multifunctional protein (Fernandez-Salas et al., 2002; Shiio et al., 2006).

Fig. 6.

CLIC4 is regulated by AP1 in keratinocytes undergoing Ca2+-induced differentiation. (A) Mouse keratinocytes in 0.05 mM Ca2+ medium were infected with β-gal or A-Fos adenovirus for 12 hours prior to induction with 0.5 mM Ca2+ medium for 24 hours. Duplicate sets of cells were analyzed by real-time quantitative RT-PCR using involucrin, K1 and CLIC4 gene-specific primer sets. Bars represent standard error. (B) After 24 hours, the treated cells from A were analyzed by immunoblotting using K10, CLIC4 and actin antibodies. Asterisks (*) in A-Fos lanes of both CLIC4 and K10 blots indicate `undetectable' CLIC4 and K10 bands that are within the linear exposure range of chemiluminescence. With longer exposures CLIC4 and K1 are detected in the A-Fos group (data not shown). (C) Triplicate sets of mouse keratinocytes were transfected with a reporter plasmid containing a 2.5 kb human CLIC4 promoter reporter construct prior to the treatment as described in A. The promoter responses were analyzed by measuring luciferase activity 16 hours after Ca2+ induction. (D) Schematic diagram of mouse and human CLIC4 promoters; solid black arrow indicates transcription start site, solid gray circles indicate AP1 binding sites and locations 5′ upstream from the transcription start site. Two AP1 consensus sites are shown in the human promoter (A: –1507 and B: –840). (E) Luciferase reporter constructs containing the wild-type (WT), AP1-A site mutant (Mt A), AP1-B site mutant (Mt B), both AP1-A and AP1-B sites mutant (Mt A+B) of human CLIC4 promoter and an empty reporter construct (pGL3; Mock) were used to transfect mouse primary keratinocytes. Triplicate sets of transfected cells were induced with 0.5 mM calcium, and the promoter activities were measured by luciferase assay and presented as a fold increase in the promoter activities in high calcium versus low calcium-treated cells. (F) Biotinylated DNA oligos containing the wild-type (WT A and WT B) or the mutant (Mt A and Mt B) AP1 consensus binding sites derived from human CLIC4 promoter sequence were used to detect AP1 binding activities from the nuclear extracts generated from low or high calcium-treated cells in the absence or presence of A-Fos expression. AP1 proteins and the DNA oligo complex formed in the nuclear extracts were detected by ELISA-based assays as described in the Materials and Methods section, and the signals were normalized by the protein concentration of the nuclear extracts.

CLIC family members have been associated with cell differentiation and development in other models. CLIC4 is highly expressed when mouse 3T3-L1 cells differentiate into adipocytes (Kitamura et al., 2001). CLIC4 is also upregulated during the TGFβ1-mediated differentiation of fibroblasts into myofibroblasts in breast cancer cells (Ronnov-Jessen et al., 2002). Similarly, chloride channel activity is required in TGFβ-induced differentiation of human lung fibroblast to myofibroblast (Yin and Watsky, 2005). EXC-4, a homologue of CLIC proteins, plays a critical role in digestive/excretory tubulogenesis in C. elegans (Berry et al., 2003), and CLIC4 is a key factor in VEGF-induced tubular morphogenesis of endothelial cells (Bohman et al., 2005). Tubulogenesis involves many cellular processes, including differentiation, apoptosis, cytoskeletal organization and ion fluxes (Zegers et al., 2003). CLIC1 is also involved in differentiation induced by insulin/IGFI-mediated signaling in human hematopoietic cells (Saeki et al., 2005). These reports and our studies support that importance of CLICs in several cell differentiation programs. Whether this is related to their chloride channel functions remains to be determined, but the multiple levels of regulation of CLIC4 and its subcellular trafficking suggest this protein is important in maintaining homeostasis.

Ion flux that is specifically modulated by membrane transport proteins including channels, pumps and carriers has been associated with differentiation in other cell types. hERG K+ channel activation is mediated by integrin signals and is a determining factor for osteoclastic differentiation (Arcangeli et al., 2004). Several anion channels are also reported to be involved in differentiation in multiple cell types; these channels include swelling-activated Cl channels, volume-activated Cl channels, voltage-dependent intracellular chloride channel ClCs and P-glycoprotein mdr-1 (Nilius et al., 1996; Nilius and Droogmans, 2003; d'Anglemont de Tassigny et al., 2003; Remillard and Yuan, 2005). Chloride ion flux is also relevant to cancer since re-expression of detachment-inducible chloride channel (mCLCA5) suppresses growth of metastatic breast cancer cells (Beckley et al., 2004), and swelling-activated Cl current is associated with neuroendocrine differentiation of prostate cancer cells (Lemonnier et al., 2005).

The work of Mauro et al. and Bikle et al. have implicated several channels and ion transporters in the regulation of epidermal differentiation (Mauro et al., 1993). Ca2+ is a well established regulator of keratinocyte maturation, and a Ca2+ gradient across the living epidermis in vivo supports the significance of Ca2+ in epidermal homeostasis. The presence of a Ca2+ sensing signaling receptor on keratinocytes has provided significant insight into the mechanisms involved in Ca2+ signaling (Komuves et al., 2002). Ablation of this receptor appears to work in concert with channels for other ions that may be the proximal regulators of the differentiation signal. For example, Ca2+-induced differentiation activates K+ conductance, and K+ channel inhibitors block keratinocyte maturation (Hennings et al., 1983b; Mauro et al., 1997). Of particular interest to our studies are reports that Ca2+ activates voltage gated Cl channels in keratinocytes (Mauro et al., 1990), and the activated Cl current is associated with differentiation. Understanding the relationship of ion transport to Ca2+ signaling in keratinocytes remains an important task since reports of Ca2+ signaling abnormalities in psoriasis (McKenzie et al., 2003), Hailey-Hailey (Hu et al., 2000) and Darier's disease (Sakuntabhai et al., 1999) suggest that modulation of downstream signaling could be an approach to therapy.

Nuclear translocation of CLIC4 as a coincident event in keratinocyte differentiation was particularly striking in our study. The presence of CLIC4 in the nucleus amplifies the expression of differentiation markers. We previously reported that CLIC4 translocates to the nucleus by the interaction of a nuclear localization signal with nuclear transport proteins (Ran, importin and NTF2) (Suh et al., 2004). Thus, nuclear translocation is a physiological response. The function of CLIC4 in the nucleus has not been clarified at this time. CLIC4 binds with multiple partners, including actin, tubulin, dynamin, AKAPs and 14-3-3 isoforms (Suginta et al., 2001; Berryman and Goldenring, 2003). These multiple protein-protein interactions imply that CLIC4 has a broader molecular function beyond a chloride channel/regulator, and these interacting proteins, and others to be identified, may contribute to the physiological functions of CLIC4. Immunoelectron microscopy indicates that nuclear CLIC4 resides in the nuclear membrane and the nucleoplasm so that several functions may be served by translocation (Suh et al., 2004).

In the nuclear membrane, CLIC4 could be involved in chloride channel activity. In support of this possibility, qualitative analyses of Cl flux with MQAE dye indicate nuclear CLIC4 is associated with accumulation of Cl that would subsequently acidify the microenvironment. The nucleus may be particularly sensitive to pH changes as acidification could alter chromatin structure and the binding efficiency of transcription factors (Eastman, 1995; Li and Weinman, 2002). For example, Runx transcription factors are sensitive to Cl content in the nucleus for binding to target sequences (Backstrom et al., 2002), and Runx activation is associated with multiple cellular responses, including TFGβ signals, cell cycle regulation and terminal differentiation in several cell types (Ito and Miyazono, 2003; Blyth et al., 2005). In fact, Runx3 is expressed in epidermis and is required for normal hair follicle development (Raveh et al., 2005).

Additional studies will be required to define the functions of CLIC4 in the nucleus and the signals that send it there. It is clear that CLIC4 contributes an essential regulatory role in epidermal homeostasis. Both in vivo and in vitro, by virtue of the level of expression and subcellular localization, CLIC4 is important for keratinocyte proliferation and terminal differentiation. Nuclear CLIC4 in particular is associated with keratinocyte growth arrest under conditions of differentiation, senescence and inhibitory growth factors suggesting a broad role of nuclear CLIC4 in growth control of skin keratinocytes.

Materials and Methods


Bisindolylmaleimide I, Go 6976 and rottlerin were obtained from Calbiochem (San Diego, CA). 12-O-tetradecanoyl-phorbol-13-acetate (TPA) was obtained from Sigma (St Louis, MO).

Cell culture

Keratinocytes were isolated and cultured as described previously (Dlugosz et al., 1995). Human keratinocytes from foreskin were isolated and cultured in EpiLife keratinocyte medium as described by the manufacturer (Cascade Biologics, Portland, OR). For Ca2+ and TGFβ induction, keratinocytes were treated with medium containing Ca2+ levels higher than 0.1 mM (typically 0.5 or 1 mM) or TGFβ (1 ng/ml). To induce serum-mediated differentiation in human keratinocytes, Ca2+-chelexed 5% FBS was added to EpiLife medium. To promote and assay senescence of primary mouse keratinocytes, the β-galactosidase senescence assay (β-gal SA) was used as described previously (Vijayachandra et al., 2003).

Construction of shRNA for CLIC family members

Human cDNA sequences of CLIC members were used as a query on public web-based programs (The RNAi Web) to determine preliminary sites and then pick one sequence (18-22 bp) from each group that does not overlap with other CLIC members by manual cDNA sequence alignment and use of other publicly available bioinformatic programs (e.g. siDirect). RNA interference compatibility for mouse sequences was also considered, and the choice of sequences was manually determined based on cDNA sequence alignments of mouse and human CLIC members. Sequences are as follows: nonspecific shRNA (sR-NS) 5′-gttctccga acgtgtcacg-3′, CLIC1 shRNA (sR-1) 5′-gtctgattgagcttgtgttg-3′, CLIC4 shRNA (sR-4) 5′-gctgaaggaggaggacaaaga-3′, CLIC5 shRNA (sR-5) 5′-gacactgatctctcagac-3′. All shRNA sequences have 5′-uucaagaga-3′ sequence as a loop prior to the direct inverse repeat sequences to make the short-hairpin loop. The siRNA version of CLIC4 shRNA was used and its efficiency in knockdown of CLIC4 was reported in previous works (Bohman et al., 2005; Shiio et al., 2006). These shRNA sequences were inserted into pENTR-H1/TO vectors (BLOCK-iT system; Invitrogen) and verified as described by the manufacturer. Lipofectamine and Plus reagents were used for transfection as described by the manufacturer (Invitrogen).

Production and purification of adenoviruses

Construction of antisense, sense or organelle-specific CLIC4 adenoviral vectors was described elsewhere (Suh et al., 2004). The use of dominant negative PKCδ and A-Fos adenovirus (Ad) were described previously (Li et al., 1999; Bonovich et al., 2002). The adenoviral vector without a recombinant insert (null virus) or with β-gal was used as a control in experiments where adenoviruses were used. All the adenovirus was plaque purified, amplified as described by Clontech and purified by two rounds of CsCl centrifugation by Gene Expression Laboratory/NCI Frederick, Adenovirus Core Facility. To be consistent throughout the experiments, 10 multiplicity of infection (MOI) was used in all adenoviral infections, except for MOI dose studies.

Immunoblot analysis

Protein expression was analyzed by immunoblot as described previously (Suh et al., 2004). Monospecific rabbit antibodies (Abs) against mouse K1 (1:2000), K10 (1:2000) and filaggrin (1:2000) were used to detect specific bands, as described previously (Dlugosz and Yuspa, 1993). The following antibodies were also used: Anti-PKCδ Ab (1:500) was from Sigma. Anti-β-actin mouse monoclonal antibody was from Chemicon International Inc. (Temecula, CA). For densitometry analyses, ImageMaster TotalLab-v1.11 (Amersham-Pharmacia/GE Healthcare Bioscience, NJ) was used to quantify the fold changes of CLIC4 expression by normalizing against the actin signals. All immunoblots analyzed were within the linear range of chemiluminescence exposure.

Immunofluorescent staining

Keratinocytes were stained as described in the previous report (Suh et al., 2004). Keratinocytes were exposed to 0.05 mM or higher Ca2+ medium for 24 hours after infection with null, wild-type CLIC4 (Cyt-CLIC4) and nucleus-targeted CLIC4 (Nuc-CLIC4) Ad for 17 hours. Cells were subjected to immunofluorescence analysis with anti-K1or anti-CLIC4 antibodies at 1:200 dilutions followed by appropriate secondary antibodies conjugated to FITC or Texas Red (Vector Laboratories). The specificity of the CLIC4 and K1 primary antibodies has been described in previous publications (Dlugosz and Yuspa, 1993; Fernandez-Salas et al., 1999). Cells stained with secondary antibody alone or purified control rabbit IgG were used as controls, and no specific immunofluorescent staining was detected. For detection of chloride flux, MQAE (Molecular Probes/Invitrogen) was used as described by the manufacturer.

Cell cycle analysis

Mouse primary keratinocytes (5×105) were infected with Null or AS-CLIC4 for 17 hours and further incubated with or without additional Ca2+ for 24 hours or 48 hours. 2 hours before harvesting the cells, 10 μM 5-bromo-2′-deoxyuridine (BrdU) was added to the medium and the trypsinized cells were analyzed by flow cytometry as described by the manufacturer (BD Bioscience, FACScan). Data acquisition and analysis were performed using Cell Quest software.

Immunodetection of AP1-oligo-transcription factor complex

An ELISA-based detection method that is equivalent of EMSA/SuperShift assay was used to quantify AP1 binding to the biotinylated oligos that encode the sequences derived from the human CLIC4 promoter with putative AP1 binding consensus sequences. The oligo pairs generated from the AP1-A site (capitalized, –1507) sequences were 5′-biotin-ggcttagatctgctttgaaacTGATTCAccatttggttcctttttg-3′ for the wild type (WT A) and 5′-biotin-ggcttagatctgctttgaaacGACTCAGccatttggttcctttttg-3′ for the mutant (Mt A). The oligo pairs generated from the AP1-B site (capitalized, –840) sequences were 5′-biotin-gaagttgggaggagtagctgTGAGTCAagtattatggaaagtcag-3′ for the wild type (WT B) and 5′-biotin-gaagttgggaggagtagctgGACGCAGagtattatggaaagtcag-3′ for the mutant (Mt B). PAGE-purified sense and antisense oligos were hybridized to form double strand DNA fragments and were incubated at a final concentration of 1.0 nM with the nuclear extracts at room temperature for 2 hours. As positive control for the experiments, the procedure was repeated with recombinant Fos-Jun mixture (data not shown, Active Motif, Calsbad, CA) and nuclear extracts of Jurkat (data not shown, untreated vs TPA treated) or U937 (data not shown, untreated vs TPA treated). These nuclear extracts and the nuclear extraction kit were purchased from Active Motif. The mixture was then added to streptavidin-coated plates (Pierce) that were pretreated with blocking buffer [1% BSA, 100 μg/ml sperm DNA, 1× APK (Ventana, AZ), 0.5% PMSF and 0.5 mM DTT], incubated at 4°C overnight and washed several times with wash buffer [1% BSA, 10 μg/ml Sperm DNA, 1× APK, 0.5% PMSF, Complete Protease Cocktail (Roche), and 50nM DTT]. A mixture of Fos and Jun polyclonal antibodies (both from Santa Cruz Biotechnology)were added at 50 ng/ml to the wells, after which they were incubated for 2 hours at room temperature, extensively washed with the wash buffer and further incubated with donkey anti-rabbit HRP (GE Healthcare Bioscience, Piscataway, NJ) at 1:50,000 dilution for 2 hours. The wells were extensively washed with the wash buffer, chemiluminescent reagent (Dura, Pierce, IL) was added and the signal was measured using an Infinite M200 plate reader (Tecan, CA). Triplicate samples were processed for all experiments.

Detection of chloride ion flux and semi-quantitative measurements

Nuclear chloride content was measured semi-quantitatively and qualitatively detected by confocal microscopy using the chloride-ion-sensitive fluorescent dye N-(6-methoxyquinolyle) acetoethyl ester (MQAE; Molecular Probes/Invitrogen) based on the methods described by the manufacturer and Maglova et al. (Maglova et al., 1998a; Maglova et al., 1998b). The chloride channel inhibitor NPPB was purchased from BioMol (Plymouth Meeting, PA). Briefly, cells were incubated with serum-free culture medium containing 10 mM MQAE for 2 hours and gently but quickly washed several times with Rinse-Buffer (20 mM HEPES, 100 mM NaCl, 5 mM KCl, 1 mM MgCl2, 0.05 mM CaCl2, 10 mM glucose, pH 7.4). Background fluorescence was determined by confocal microscopy after incubating the cells with the Quenching-Buffer [Rinse-Buffer without NaCl, 100 mM NaNO3, 10 μM nigericin (Biomol), 10 μM valinomycin (Biomol)] and used for normalization. To visualize chloride ion flux qualitatively, keratinocytes grown on 60 mm plates infected with adenovirus expressing β-gal, wild-type CLIC4 (Cyt-CLIC4) and nuclear-targeted CLIC4 (Nuc-CLIC4) for 16 hours were treated as stated above, and fluorescent images were captured by using the 40× Achroplan objective on the Zeiss-510 confocal microscope.

Fluorescence from the washed cells was measured using a Molecular Devices fMax fluorometric plate reader with parameters set at 350 nm for excitation and 460 nm for emission. 355 nm excitation filter and 460 nm emission filters (Molecular Devices) were used. Raw data were collected from the plate reader in triplicate and SoftMax Pro and Excel programs were used to calculate the average value that is represented in two digit numbers as arbitrary units for simplicity.

Immunostaining of skin sections

Adult skin samples from humans or mice were formalin fixed, or snap-frozen and then acetone fixed, respectively. For immunohistochemical staining of fixed sections, slides were deparafinized with HistoChoice (Amresco, Solon, OH), and the antigen was retrieved with Target Retrieval Solution-High pH (Dako, Carpinteria, CA) as described by the manufacturer. Envision-secondary antibody conjugated to horseradish peroxidase (Dako) and DAB (Pierce) were used to detect the antigen as described by the manufacturer. Slides were analyzed with bright-field microscopy using Leitz-DMRB (Leica, Bannockburn, IL) and OpenLab (Improvision, Lexington, MA) software. Tissue sections stained with secondary antibody alone or purified control rabbit IgG were used as controls, and no specific staining was detected.

Production of recombinant CLIC1 and CLIC4 proteins and in vitro kinase assay

Plasmid pGEX-6P or pGEX-4T (Amersham Bioscience, NY) containing CLIC4 or CLIC1 cDNA, respectively, was used to produce the recombinant GST-CLIC4 or GST-CLIC1 protein, and the proteins were purified as described by the manufacturer. The purified proteins were phosphorylated in vitro with several recombinant and kinase-active human PKC isoforms (Invitrogen) as described by the manufacturer. The kinase reaction mixture was resolved by SDS-PAGE, stained and visualized with Coomassie Blue dye, and then the gel was exposed to Phosphor Screen (GE Healthcare Life Bio-Science, Piscataway, NJ). The phosphorylated CLIC proteins were detected using Storm imager (Molecular Dynamics).

Real-time PCR analysis

Bio-Rad iQ iCycler and Gene Expression Macro (version 1.1) from Bio-Rad were used to measure transcript expression levels. cDNA (diluted 1:50 in the final volume reaction) was measured using iQ SYBR Green Supermix (Bio-Rad) and included experimental duplicate reactions. The primer sequences were as follows: L32 (forward) 5′-tggttttcttgttg-3′, (reverse) 5′-gggtgcggagaagg-3′, mouse CLIC4 (forward) 5′-tccttgagttatactcccagaacc-3′ (reverse) 5′-gtgagtttcactaacaaacgctc-3′. Mouse keratin 1 and mouse involucrin primers were obtained from Superarray (Frederick, MD). Relative standard curves were generated from log input (serial dilutions of pooled cDNA) versus the threshold cycle (Ct). The linear correlation coefficient (r2) was 0.81, 0.99 and 0.99 for CLIC4, K1 and involucrin respectively. The slope of the standard curve was used to determine the efficiency of target amplification (E) using the equation E=(10–1/slope–1)×100. Similar high efficiency was obtained for all primers allowing for the comparative Ct method to be used. Relative quantitation was used to calculate the 2–(ΔΔCt) formula where ΔΔCt represents the difference corrected for L32 used as internal control (Livak and Schmittgen, 2001). Electrophoresis analysis of amplified product from real-time PCR showed a single band.

CLIC4 promoter construction and reporter assay

3.5 kb of human CLIC4 promoter (accession number: NT_004610, on human chromosome 1p36.11 region, with VEGA coordinates contig 24,944,435-25,043,402, start of transcript location is AL445648.18.1.133587) was cloned by genomic PCR and used as an insert to ligate with pGEM-Teasy (Promega). The primer pairs used for the genomic PCR were 5′-ttaacatagactaatttatcaatgatccta-3′ for forward and 5′-ccgtgctgctcgctggactgtccggcgtcg-3′ for reverse, and Expand High Fidelity PCR kit (Roche) was used. Separate primer sets 5′-gagaGCTAGCgtgtaccatgagctgtcctctgagccaggaatatagccataaacaaaaacc-3′ (NheI-engineered site capitalized) for forward and 5′-cttggtgtgtttcaggctctgagctagcccttgg-3′ for reverse was used to subclone the PCR fragment into pGlow-TOPO (Invitrogen). A 2.5 kb (–2720 bp to –369 bp) fragment was further isolated by NheI digestion, subcloned into pGL3 (Promega) and used for reporter assays. Mouse primary keratinocytes (2×106/plate) were transfected with the reporter plasmid in triplicate using Lipofectamine (Invitrogen), treated with or without A-Fos Ad or 0.5 mM Ca2+ overnight, and the luciferase activities were measured using the Enhance Luciferase Assay Kit as described by the manufacturer (BD Pharmingen, San Jose, CA). The normalization of the promoter assays was done by cell number and protein concentration. Putative AP1 sites were determined by Genomatrix software within the promoter fragment and designated AP1-A (–1507) and AP1-B (–840) upstream from the putative transcription start site. To determine whether these two sites are responsible for CLIC4 promoter activity during Ca2+ shift, the two sites were mutated separately (Mt A or Mt B) or together (Mt A+B) by deleting five bases (capitalized) out of seven AP1 consensus sequence (Mt A: TGATTca and Mt A: TGAGTca).


This research was supported by the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research. We thank Joanna Anders for assistance with references, Ulrike Lichti for differential fractionations of mouse skin, Narayan Bhat of the Science Applications International Corporation, Inc. (SAIC), Frederick, NCI, Molecular Biology, Gene Expression Laboratory for adenoviral amplification, and Barbara Taylor of the CCR FACS Core Facility. M. Mutoh was the recipient of JSPS Research Fellow in Biomedical and Behavioral Research at NIH (2003-2005).

  • Accepted May 29, 2007.


View Abstract