Corticotropin-Releasing Hormone Inhibits Nuclear Factor-κB Pathway in Human HaCaT Keratinocytes

Abstract

Treatment of human HaCaT keratinocytes with corticotropin-releasing hormone modulates cell proliferation and expression of in£ammation markers. In this study we report that corticotropin-releasing hormone also inhibits nuclear factor-κB binding and transcriptional activity. Incubating cells in the absence of growth factors increased nuclear factor-κB activity; this effect was significantly attenuated by corticotropin-releasing hormone. Specifically, corticotropin-releasing hormone downregulated p50/p50 and p50/p65 dimers of nuclear factor-κB, diminished κB-driven CAT reporter gene activity and inhibited IκB-β degradation. Moreover, corticotropin-releasing hormone inhibited the transcription of the nuclear factor-κB responsive genes, interleukin-2 and heat shock protein 90.

It is generally accepted that corticotropin-releasing hormone (CRH), a 41 amino acid long peptide, acts as the main co-ordinator of the central response to stress (Chrousos and Gold, 1992). CRH is also produced locally in peripheral organs including the skin to regulate local homeostasis (Slominski et al, 2000b, 2001).

The outermost layer of the skin, the epidermis, plays a crucial part in maintaining internal homeostasis, and serves as a barrier between the environment and the internal milieu. Keratinocytes regulate the activity of the skin’s immune system and protect the skin from noxious stressors (Slominski et al, 2000a, 2001). HaCaT keratinocytes both produce CRH and express functional CRH-R1-α receptor (Pisarchik and Slominski, 2001; Quevedo et al, 2001; Zbytek et al, 2002). CRH signal transduction is coupled to adenylate cyclase, phospholipase C, or to voltage-gated calcium channels (Owens and Nemeroff, 1991; Slominski et al, 2001). CRH inhibits proliferation of keratinocytes and modifies expression of adhesion molecules and cytokine secretion (Slominski et al, 2000a, 2001; Quevedo et al, 2001; Zbytek et al, 2002). In this context, the epidermis may possess a high sensory capability for noxious stimuli through a local stress response pathway involving the CRH/CRH-R signaling system (Slominski et al, 2000b, 2001).

Nuclear factor (NF)-κB is an inducible and ubiquitously expressed transcription factor (Yang et al, 2000; Li and Verma, 2002). Active NF-κB complexes are dimers of various combinations of the Rel/NF-κB family of polypeptides, which includes p50, p52, v-Rel, c-Rel, RelA, and RelB. NF-κB is sequestered in cytoplasm by the binding of NF-κB inhibitor proteins, which block the nuclear localization sequences present in NF-κB. NF-κB-inducing stimuli promote dissociation of the inactive NF-κB complexes via the serine phosphorylation and degradation of IκB. These events lead to unmasking of nuclear localization sequences, thereby allowing NF-κB to enter the nucleus and bind to κB-regulatory elements.

HaCaT cells have higher constitutive levels of NF-κB activity than normal keratinocytes (Quin et al, 1999), which may be modified by ultraviolet radiation and L-ascorbic acid (Saliou et al, 1999; Tebbe et al, 2001). As NF-κB is a recognized co-ordinator of the cellular response to stress and CRH is a critical element in the epidermal response to stress, we tested whether CRH signaling was coupled to the NF-κB pathway in HaCaT cells.

MATERIALS AND METHODS

Cell culture

HaCaT keratinocytes were cultured in a Dulbecco minimal Eagle’s medium, supplemented with 10% fetal calf serum and 1% antibiotic/anti-mycotic mixture (Gibco, Invitrogen Life Technologies Carlsbad, California) as previously described (Slominski et al, 2000a).

Cells were seeded at density 10,000 cells per cm², grown for 48 h until 70% con£uency and then treated with CRH (Sigma, St. Louis, MO) as indicated in Results.

Electrophoretic mobility shift assay

Nuclei extracts were prepared as previously described (Yang et al, 2000) and were used for electrophoresis mobility shift assay. NF-κB oligonucleotide probe (Promega, Madison, Wisconsin) was end-labeled with [γ-³²P]deoxyadenosine triphosphate using T4 polynucleotide kinase and incubated with 5 μg of nuclear extract. The protein-DNA complexes were separated on 5% polyacrylamide gel. For supershift assays nuclear extracts were incubated with p50, p65, c-rel, or p52 antibodies (Santa Cruz Biotechnology Inc., Santa Cruz, California). To determine binding specifity, a 50 × excess of unlabeled oligonucleotide was used. Radioactivity was quantitated with Packard Cyclone phosphorimager, and analyzed with Optiquant™ (Perkin Elmer Life Sciences Inc., Boston, Massachusetts) and Adobe Photoshop (San Jose, California) software.

Western blot analysis of IκB-α and IκB-β levels

HaCaT cells were lyzed in RIPA buffer and clarified by centrifugation (10,000 × g, 10 min). Cell lysates (10 μg) were separated on 12% sodium dodecyl sulfate– polyacrylamide gel electrophoresis gel and transferred to the PVDF membrane. After blocking with Tris-buffered saline, Triton-X 0.05% and 5% milk, the membranes were incubated with rabbit anti-human IκB-α or IκB-β a/nity-purified polyclonal IgG (1:200, Santa Cruz Biotechnology Inc.), followed by incubation with horseradish peroxidase conjugated goat anti-rabbit IgG (1:10,000 dilution). IκB-α and IκB-β were visualized by with Supersignal West Pico Chemiluminescent Substrate (Pierce, Rockford, Illinois). The membranes were stripped and reprobed with antibody against actin (1:500, Santa Cruz Biotechnology Inc.). The chemiluminescent signal was recorded using Fluor-S MultiImager and analyzed with Quantity One software (Bio-Rad Laboratories, Hercules, California).

Reporter gene assay

Keratinocytes were transfected using Lipofectamine Plus reagent (Gibco, Invitrogen Life Technologies) with either the pUX-CAT promoterless chloramphenicol acetyltransferase construct or pUX-CAT 3XHLAκB construct, which contains three tandemly repeated copies of NF-κB site from human leukocyte antigen B7 gene (provided by Dr J. Vilcek, New York University Medical Center). CAT activity was assayed by thin-layer chromatography and quantitated by phosphorimaging using Optiquant software (Packard Cyclone, Perkin Elmer Life Sciences Inc.).

Oligonucleotide array

Total RNA was extracted using Trizol isolation kit (Gibco-BRL, Gaithersburg, Maryland). Human pathway finder-1 and human NF-κB-1 pathway GEArrays (Superarray, Inc., Bethesda, Maryland) were performed according to the manufacturer’s protocol. Abundance of transcripts was quantitated by phosphorimaging as described above and normalized to glyceraldehyde 3-phosphate dehydrogenase expression.

Heat shock protein (HSP)-90 and interleukin (IL)-2 northern blot

Twenty micrograms of total RNA was separated on 1% agarose gel, transferred to the nylon membrane and then ultraviolet cross-linked. HSP-90 (accession number X15183; 367 bp) was prepared using DNA isolated from the plasmid of the Escherichia coli clone IMAGE:5539895. The sequences of primers were as follows: sense 5′-GGTAGCTAACTCAGCCTTTGTGG, anti-sense: 5′-TGAGTTGTCTCTTAGGGCTTGAGC. Polymerase chain reaction conditions: 94°C 3′, 30 cycles (94°C 45′′, 54.7°C 45′, 72°C 1′), 72°C 4′. Amount of polymerase chain reaction reagents as described in manufacturer’s protocol ( JumpStart™ AccuTaq™, Sigma, St Louis, Missouri). IL-2 (accession number U25676; 468 bp) was prepared by polymerase chain reaction using complementary DNA isolated from stimulated lymphocytes. Primers sequences were as follows: sense: 5′-ACATTTAAGTTTTACATGCCCAAG, anti-sense: 5′-GTAAACCATTTTAGAGCCCCTAG. Polymerase chain reaction conditions and solution as above except elongation step: 52.1°C 45 s. Denatured complementary DNA probes (0.2 μg) were labeled by random priming using 50 μCi [α-³²P]deoxyadenosine triphosphate (NEN Life Science Products, Boston, Massachusetts), 5 μg random hexamers, 3 μL of 0.5 M solution of deoxynucleotide triphosphates, 5 U Klenow fragment (exonuclease^–, MBI Fermentas, Vilnius, Lithuania), 5 × Klenow buffer and water to a total volume of 50 μL. The membrane was hybridized with radioactive probe in standard conditions and visualized with Kodak X-OMAT film in −70°C for 24 h.

Statistical analysis

Data are presented as mean ± SEM, and was analyzed using one-way analysis of variance and appropriate post-hoc test using Prism 3.00 software (GraphPad Software, San Diego, California). Significant differences are denoted by p values less than 0.005.

RESULTS

NF-κb binding activity in HaCaT

HaCaT cells were incubated in serum-free Dulbecco minimal Eagle’s medium containing CRH (0–100 nM) for 0, 15, 30, or 60 min. Nuclear extracts were prepared and analyzed for NF-κB activation by electrophoresis mobility shift assay and supershift assays (Fig 1). Serum withdrawal increased formation of specific NF-κB complexes defined by supershift assays as p50/p50 and p50/p65. In contrast, CRH treatment (100 nM) significantly decreased formation of p50/p50 and p50/p65 complexes after 15 and 30 min as compared with cells not treated with CRH (Fig 1). The effects of serum withdrawal and CRH on NF-κB disappeared after 60 min.

Figure 1. CRH treatment attenuates NF-κB DNA binding activity induced upon serum withdrawal from HaCaT keratinocytes.

Nuclear extracts were prepared from HaCaT cells treated with serum-free medium with 100 nM CRH for 0, 15, 30, and 60 min and subjected to electrophoresis mobility shift assay; 0 min represents the NF-κB signal from cells not subjected to serum withdrawal. The presence of specific Rel proteins in DNA complexes in CRH-treated cells (100 nM, 15 min) was detected by supershift assay using p50 and p65 anti-sera. Cold represents extract incubated with a 50-fold excess of unlabeled κB probe. Results are representative of three separate experiments.

Western blot analysis of IκB-α and IκB-β levels

Inhibitory IκB proteins sequester NF-κB in the cytoplasm and tightly control the activity of NF-κB. To determine whether NF-κB activation by CRH re£ects IκB degradation, IκB-α and IκB-β levels were determined by immunoblotting of cell lysates prepared at various times after serum withdrawal with or without CRH addition. As shown in Fig 2, the levels of cellular IκB-α did not change upon serum withdrawal. In contrast, IκB-β levels were reduced by 63% upon serum withdrawal, and treatment with CRH attenuated this effect. These results indicated that CRH attenuated IκB-β degradation induced by serum withdrawal.

Figure 2. CRH diminishes IκB-β degradation.

HaCaT cells grown in medium supplemented with fetal calf serum (0) were shifted to serum-free medium without (−) or with 100 nM CRH (+). Total cell lysates were analyzed by immunoblotting. Results are representative of six separate experiments.

CAT assay

CRH-mediated attenuation of serum withdrawal-induced NF-κB signal was further confirmed by CAT gene reporter assay. Extracts were prepared from cells containing κB-driven vector that were treated for 30 and 60 min in serum-free medium without or with 100 nM CRH. As shown in Fig 3 serum withdrawal resulted in a 6-fold induction of κB-driven CAT activity as compared with the cells transfected with a promoterless CAT construct. In contrast, CRH attenuated the increased activity by 74% (down to 1.6-fold as compared with the cells transfected with a promoterless CAT construct).

Figure 3. CRH effects on NF-κB-dependent reporter construct.

Ha-CaT cells were transiently transfected with pUX-CAT (EV) or pUX-CAT 3XHLAκB (NF-κB) vector, incubated for 30 min in serum-free media without (control) or with 100 nM CRH, and assayed for CAT activity. The data represent the mean ± SEM (n =4) after subtraction of background, related to empty vector activity (=1).

Analysis of NF-κB dependent genes

Using commercially available oligonucleotide arrays we also tested the effect of CRH on the transcription of genes associated with NF-κB signaling pathways in cells subjected to serum withdrawal. Genes coding NF-κB, IκB kinase (IKK)-α, IKK-β, IKK-γ, IκB-α, c-rel, cellular myelocytomatosis oncogene, and NF-κB responsive genes (human endothelial leukocyte adhesion molecule, granulocyte colony stimulating factor, granulocyte-macrophage colony stimulating factor, intercellular adhesion molecule 1, interferon regulatory factor 1, small inducible cytokine A2, IL-6, IL-8, inducible nitric oxide synthase; P-selectin, tumor necrosis factor-α and -β, tumor necrosis factor-α-induced protein 1; vascular cell adhesion molecule) were not affected by CRH. Exposure of cells to CRH (100 nM), however, significantly decreased transcription of IL-2 and HSP-90 genes in comparison with cells at 30 min after serum withdrawal (data not shown). To determine by an independent means our preliminary results with oligonucleotide arrays, we examined whether CRH affected IL-2 and HSP-90 mRNA levels by northern blot. As shown in Fig 4, at 30 min after serum withdrawal there was high expression of IL-2 and HSP-90 mRNA. Treatment with CRH markedly reduced IL-2 and HSP-90 mRNA expression (70% and 99% reduction, respectively). The observed effect of CRH was transient and by 2 h after CRH treatment there was no detectable effect on IL-2 or HSP-90 mRNA levels.

Figure 4. CRH inhibits transcription of IL-2 and HSP-90 mRNA.

HaCaT keratinocytes were treated for 30 min with serum-free medium without (control) or with 100 nM CRH, and total RNA was isolated. Gene expression was estimated by northern blot technique using HSP-90 and IL-2 radioactive probes (upper panel). Ethidium bromide stain of total RNA indicates equal loading of the samples into the gel.

DISCUSSION

In previous studies we have documented that human keratino-cytes express functional CRH receptors. Activation of these receptors results in increased cyclic adenosine monophosphate production, cytosolic calcium £ux, inhibition of proliferation, modification of cell surface expression of adhesion molecules and production of cytokines (Slominski et al, 2000a, b, 2001; Quevedo et al, 2001; Zbytek et al, 2002). In addition, CRH treatment of HaCaT cells inhibited lipopolysaccharide (LPS)-enhanced IL-6 production (Zbytek et al, 2002) and decreased interferon-γ induced expression of intercellular adhesion molecule-1 (unpublished observations).

In this study using immortalized HaCaT keratinocytes we have shown that CRH can inhibit NF-κB binding and transcriptional activity. Choice of 100 nM concentration was based on preliminary electrophoresis mobility shift assay experiments, which have shown that CRH at this concentration exerts the maximal effect on serum deprivation-induced NF-κB activation. Effects of CRH on cyclic adenosine monophosphate activation in HaCaT cells also was maximal at 100 nM (Slominski et al, 2000a). CRH levels may reach relatively high levels locally (Karalis et al,1991). A 100 nM concentration is used in comparable models, such as in the model of regulation of IL-1α production in the monocyte (Pereda et al, 1995) or the model of suppression of NF-κB activity in mouse pituitary corticotropic AtT20 cells (Lezoualc’h et al, 2000). NF-κB is important in several stress-related pathways. Stimuli acting through different receptor systems converge on the NF-κB pathway. On the other hand NF-κB affects different genes depending on the cell type or presence of external stimuli (Yang et al, 2000; Li and Verma, 2002). In our model, withdrawal of serum for 15 and 30 min enhances a constitutively activated NF-κB pathway that is in turn attenuated by CRH.

The activation of NF-κB involves selectively IκB-β degradation. Kinetics of NF-κB activation and IκB degradation vary among stimulating agents and cell types (Li and Verma, 2002). IκB-β degradation is triggered by IL-1, nerve growth factor, LPS, as well as upon serum withdrawal (Thompson et al, 1995; Cosgaya and Shooter, 2001). IκB degradation is characteristic for biphasic NF-κB activation. Degradation of IκB-β is observed shortly after application of the stimulus as well as after prolonged incubation (Thompson et al, 1995). Phosphorylation of IκB proteins that precedes their degradation is mediated by IκB kinases. Multiple signaling pathways converge on the IKK complex and result in its activation. For example, the atypical protein kinase Cζ may be involved in this process (Li and Verma, 2002). Protein kinase Cζ is expression in the keratinocytes and activated by phospholipids (Efimova et al, 2002). CRH receptor signaling results in phospholipase C activation (Owens and Nemeroff, 1991). Elucidation of the pathway by which CRH is linked to IκB-β will require further studies.

CRH through its action on NF-κB, an integrator of cellular response to stress, counteracts acute stress to stabilize cellular homeostasis. In support of this hypothesis we have found that CRH altered the pattern of NF-κB-dependent gene expression. We found that CRH decreased κB-driven reporter gene transcription and selectively downregulated the NF-κB-dependent genes IL-2 and HSP-90. IL-2 and HSP-90 are produced constitutively in several immortalized cell lines, suggesting their positive role in cell proliferation (Stepanova et al, 1996; Reichert et al, 2000). Specifically, IL-2 mediates cell cycle progression through downregulation of cyclin-dependent kinase inhibitors 1A and 1B (Reichert et al, 2000). HSP-90 together with cell division cyclin 37 stabilizes cyclin-dependent kinase 4 and the cyclin D complex that mediates passage through a cell cycle restriction point (Stepanova et al, 1996). Therefore, inhibition of IL-2 and HSP-90 genes expression by CRH may be a part of a signaling cascade induced by CRH that results in inhibition of HaCaT cell proliferation (Slominski et al, 2000a, 2001). HSP-90 is also involved in LPS signal transduction (Triantafilou et al, 2001). Thus, it is also possible that modulation of HSP-90 expression may be involved in CRH-induced differential regulation of IL-6 production dependent on the presence of LPS (Triantafilou et al, 2001; Zbytek et al, 2002). Nevertheless, we have not found that short-term serum withdrawal and CRH affect mRNA levels of in£ammatory mediators in HaCaT keratinocytes. Further studies are essential to elucidate the in£uence of CRH on stress-related gene expression in the keratinocyte.

In summary, we have demonstrated that CRH can inhibit the stress-related NF-κB pathway in immortalized human epidermal keratinocytes.