ORIGINAL ARTICLE

Glucocorticoids Enhance Toll-Like Receptor 2 Expression in Human Keratinocytes Stimulated with Propionibacterium acnes or Proinflammatory Cytokines Michio Shibata1, Masako Katsuyama1, Tomoko Onodera1, Ritsuko Ehama1, Junichi Hosoi1 and Hachiro Tagami2

Toll-like receptors (TLRs) on keratinocytes are important cell surface receptors involved in the innate and acquired immune response to invading microorganisms. In vulgaris, TLR2 activation by Propionibacterium acnes (P. acnes) may induce skin inflammation via induction of various proinflammatory molecules that stimulate the invasion of inflammatory cells. Although corticosteroids themselves exert immunosuppressive or anti-inflammatory effects, it is well known clinically that systemic or topical glucocorticoid treatment provokes an acneiform reaction. Nevertheless, the effect of steroids on TLR2 expression in human keratinocytes remains unknown. Here, we found that the addition of glucocorticoids, such as dexamethasone and cortisol, to cultured human keratinocytes increased their TLR2 gene expression. Moreover, these glucocorticoids markedly enhanced TLR2 gene expression, which was further stimulated by P. acnes, tumor necrosis factor-a, and IL-1a. Gene expression of mitogen-activated protein kinase (MAPK) phosphatase-1 was also increased by the addition of dexamethasone. By using several inhibitors and activators, we found that TLR2 gene induction by glucocorticoids was mediated by the suppression of p38 MAPK activity following induction of MAPK phosphatase-1. These findings strongly suggest that steroid-induced TLR2 together with P. acnes existing as normal resident flora plays an important role in the exacerbation of acne vulgaris as well as in possible induction of corticosteroid-induced acne or in that of -like dermatitis. Journal of Investigative Dermatology (2009) 129, 375–382; doi:10.1038/jid.2008.237; published online 14 August 2008

INTRODUCTION reactions mediated mainly by NF-kB and mitogen-activated The immunology of the innate immune system has been a protein kinase (MAPK) pathways. TLR2 is well known to act rapidly progressing research area in the last decade. This as a receptor for Gram-positive Propionibacterium acnes important host defense system for animals against external (P. acnes). P. acnes, which colonizes the upper pilosebaceous organisms includes various mechanisms such as the Toll-like unit, is primarily responsible for inflammatory acne vulgaris. receptor (TLR) family, the complement activation system, and It acts through the induction of cytokine secretion and release antibacterial peptides. TLR family comprises a group of cell of substances such as proteases and hyaluronidases (Kim surface receptors that are important for both innate and et al., 2002; Jugeau et al., 2005). TLR2 activation by P. acnes acquired immune responses against invading microorganisms induces skin inflammation in acne lesions by the activation of (Akira et al., 2006). Many constituents of these pathogens, various inflammatory genes that encode cytokines, chemo- collectively known as the pathogen-associated molecular kines, and cell adhesion molecules with a resulting invasion pattern, bind to TLRs to induce various inflammatory of immune cells (Kim et al., 2002). Several reports have shown that TLR2 expression is enhanced by glucocorticoids in bronchial or other epithelial 1Shiseido Research Center, Yokohama, Japan and 2Tohoku University School of Medicine, Sendai, Japan cells (Imasato et al., 2002; Shuto et al., 2002; Hermoso et al., Correspondence: Dr Michio Shibata, Shiseido Research Center, 2-12-1 2004; Homma et al., 2004). TLR2 may also be involved in Fukuura, Kanazawa-ku, Yokohama 236-8643, Japan. susceptibility to lower respiratory tract infections upon E-mail; [email protected] inhalation of corticosteroids (Homma et al., 2004). However, Abbreviations: MAPK, mitogen-activated protein kinase; MKP-1, MAPK there have been no reports on the effect of glucocorticoids on þþ phosphatase-1; P. acnes, Propionibacterium acnes; PBS(), Ca - and TLR2 expression in epidermal keratinocytes. Mg þþ-free phosphate-buffered saline; TLR, Toll-like receptor; TNF-a, tumor necrosis factor-a Cortisol, the main glucocorticoid in humans, is considered Received 24 December 2007; revised 15 April 2008; accepted 20 June 2008; a stress hormone because the blood cortisol level increases published online 14 August 2008 under conditions of psychosocial and physical stress (Evolahti

& 2009 The Society for Investigative Dermatology www.jidonline.org 375 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

et al., 2006). Animal and human studies strongly suggest an not occur at a significantly higher level (Figure 1d). We used important immunoregulatory role for the hypothalamic- phorbol 12-myristate 13-acetate-treated THP-1 cells as a pituitary-adrenocortical axis and the sympathoadrenomedul- positive control for TLR2 expression (Haehnel et al., 2002). lary system (Buske-Kirschbaum et al., 2002). Patients suffering TLR2 was found to be constitutively expressed in THP-1 cells from allergic inflammation have significantly attenuated with a high copy number even in the absence of TNF-a or cortisol responses to psychosocial stress (Buske-Kirschbaum dexamethasone (data not shown). TNF-a had no additive et al., 2002). Immobilization stress influences cutaneous effect on TLR2 expression, but dexamethasone remarkably functions by changing the levels of neuropeptides and enhanced TLR2 expression (Figure 1f). hormones, and it suppresses the contact hypersensitivity Next, we used the anti-human TLR2 monoclonal antibody response in mice (Singh et al., 1999). It is known that cortisol clone 1030A5.138 to detect TLR2 protein expression in suppresses the immune system via inhibition of the production keratinocytes (Figure 2). TLR2 staining intensity was ob- of inflammatory cytokines and chemokines (Hosoi et al., viously increased in cells stimulated with TNF-a in the 2001). Although corticosteroids themselves exert immunosup- presence of dexamethasone, indicating that enhancement by pressive and anti-inflammatory effects, it is well known glucocorticoid treatment occurred at both the gene and clinically that systemic or topical glucocorticoid treatment protein levels. In contrast, there were no fluorescence signals provokes an acneiform reaction (Plewig and Kligman, 1973; in samples stained with isotype control antibody (Figure 2e). Fung and Berger, 2000). Moreover, the prolonged use of potent topical corticosteroids on the face often results in Effect of the steroid receptor antagonist RU486 papulopustular eruptions accompanied by erythema (steroid In order to elucidate the mechanism of the effect of rosacea; Leyden et al., 1974) or persistent erythematous glucocorticoids on TLR2 expression in human keratinocytes, eruptions consisting of tiny papulopustules with a distribution we used the glucocorticoid receptor antagonist RU486, also mainly around the mouth (; Cotterill, 1979). known as mifepristone, (Flores et al., 2006), to determine if In this study, we found that two glucocorticoids, dex- the enhancement of TLR2 expression was mediated via the amethasone and cortisol, increased TLR2 gene expression in glucocorticoid receptor (Figure 1e). The addition of 1 mM cultured human keratinocytes without additional stimulation. RU486 to the culture completely blocked the enhancement Moreover, these glucocorticoids markedly enhanced TLR2 of TLR2 expression induced by dexamethasone but did not gene expression in keratinocytes that had been stimulated affect TNF-a-induced TLR2 expression, indicating that the with P. acnes or the inflammatory cytokines, tumor necrosis effect of dexamethasone was mediated by the glucocorticoid factor-a (TNF-a) and IL-1a. receptor.

RESULTS The effect of glucocorticoids on P. acnes-induced TLR2 The effects of glucocorticoids on TLR2 gene expression in expression of keratinocytes TNF-a- or IL-1a-stimulated human keratinocytes We assessed TLR2 gene expression in keratinocytes when We confirmed mRNA expression of TLR2 in cultured human viable P. acnes and dexamethasone were added to the keratinocytes by SYBR Green-based quantitative PCR and culture (Figure 3). For these studies, we used several different then determined the effect of dexamethasone on TLR2 strains of P. acnes, including JCM6425 (4.1 107 colony- expression (Figure 1a). Relatively low but detectable levels forming units per well), JCM6473 (4.4 108 colony-forming of TLR2 expression were observed in untreated keratinocytes. units per well), and ATCC11827 (4.0 107 colony-forming In contrast, expression was increased in the presence of 0.1 units per well). These P. acnes strains slightly increased TLR2 and 1 mM dexamethasone by four- to fivefold after 6 hours. expression of keratinocytes in the absence of dexamethasone. TNF-a also increased TLR2 expression by threefold. Most However, all of them enhanced keratinocyte TLR2 expression strikingly, dexamethasone increased TLR2 expression by over in the presence of dexamethasone by approximately fourfold 20-fold in keratinocytes stimulated with TNF-a. After 1 hour when compared to induction by P. acnes strains alone. We of incubation, TLR2 expression was low, but stimulation with did not find any differences in the degree of enhancement of dexamethasone and TNF-a still produced significant effects. TLR2 expression among the three different strains, although On the other hand, TNF-a-induced IL-8 expression was the numbers of added bacteria varied. Again, RU486 significantly inhibited by dexamethasone. IL-8 was strongly completely inhibited the effect of dexamethasone on the induced by TNF-a at 1 hour, and expression was decreased at TLR2 expression of keratinocytes. 6 hours (Figure 1b). Under similar experimental conditions, dexamethasone also enhanced TLR2 expression that had The effects of inhibitors and activators of the MAPK pathway been stimulated by IL-1a, another proinflammatory cytokine We used several inhibitors and activators of the MAPK (Figure 1c). Cortisol or hydrocortisone, the primary active pathway to elucidate the mechanism underlying the en- form of glucocorticoids in humans, was added to the hancement of TLR2 expression by glucocorticoids in human keratinocyte culture instead of dexamethasone after the keratinocytes (Figure 4). Ro-31-8220, an inhibitor of MAPK second medium change with cortisol-free medium. Cortisol phosphatase-1 (MKP-1) mRNA expression and an activator of treatment prominently enhanced TNF-a-stimulated TLR2 c-Jun N-terminal kinase (Beltman et al., 1996), completely expression, although the effect appeared to be less than that inhibited the induction of TLR2 expression by dexametha- of dexamethasone because induction by 0.1 mM cortisol did sone and TNF-a. Anisomycin, an activator of p38 MAPK and

376 Journal of Investigative Dermatology (2009), Volume 129 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

50 ∗∗ 120 ∗ ∗ ∗ 40 100

80 30 ∗∗ 60 ∗∗ 20 ∗∗ ∗∗ ∗∗ 40 ∗ ∗ ∗ 10

IL-8 expression (x fold) IL-8 expression 20 TLR2 expression (x fold) TLR2 expression

0 0 − − − − − − − − Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex Dex M M M M M M M M M M M M M M M M 1 μ 1 μ 1 μ 1 μ 1 μ 1 μ 1 μ 1 μ 0.1 μ 0.1 μ 0.1 μ 0.1 μ 0.1 μ 0.1 μ 0.1 μ 0.1 μ − TNF-α − TNF-α −−TNF-α TNF-α 1-hour incubation 6-hour incubation 1-hour incubation 6-hour incubation

12 12 18 NS 10 ∗∗ ∗∗ ∗∗ 16 NS 10 10 ∗∗ ∗∗ NS 14 8 8 8 12 ∗∗ ∗∗ ∗∗ ∗∗ 6 NS 10 6 6 8 ∗ 4 4 4 6 ∗ 4 TLR2 expression (x fold) TLR2 expression

TLR2 expression (x fold) TLR2 expression (x fold) TLR2 expression 2 2 2 (x fold) TLR2 expression 2

0 0 0 0 − −

− − μ − − − − − 0.1 1 0.1 1 M Dex Dex Dex Dex Dex Dex RU+D

Cortisol RU486 RU486 α α α TNF- IL-1 − TNF-α − TNF-α − TNF-α − TNF-α TNF-

Figure 1. The effect of glucocorticoids on TLR2 or IL-8 gene expression in proinflmmatory cytokine-stimulated human keratinocytes and THP-1 cells. 1 Keratinocytes were stimulated with 10 ng ml TNF-a for 1 or 6 hours in the presence or absence of 0.1 or 1 mM dexamethasone (Dex). After stimulation, total RNA was prepared from the cells and cDNA was synthesized followed by quantitative PCR. (a) TNF-a-stimulated TLR2 gene expression, (b) TNF-a-stimulated IL-8 gene expression, (c) IL-1a-stimulated TLR2 gene expression, (d) the effect of cortisol on TLR2 gene expression, (e) the effect of RU486 on TLR2 upregulation by dexamethasone, and (f) the effect of dexamethasone on TLR2 gene expression in PMA-treated THP-1 cells. *Po0.05, **Po0.01, NS; not significant, n ¼ 4. c-Jun N-terminal kinase, also inhibited the increase in TLR2 DISCUSSION expression generated by the dexamethasone and TNF-a In this study, we found that glucocorticoids remarkably treatment. On the other hand, the p38 MAPK inhibitor enhanced P. acnes- and cytokine-stimulated TLR2 expression SB202190 only modestly increased TLR2 expression. More- in human cultured keratinocytes. Low levels of TLR2 over, the c-Jun N-terminal kinase inhibitor SP600125, the expression were detectable even in nontreated keratinocytes, MEK1/2 inhibitor U0126 that causes downstream extracel- but expression was significantly increased in the presence of lular signal-regulated kinase inhibition, and the NF-kB glucocorticoids. Treatment with proinflammatory cytokines inhibitor caffeic acid phenethyl ester (CAPE) did not affect such as TNF-a and IL-1a also increased TLR2 expression TLR2 expression. by several folds. The combination of glucocorticoids with P. acnes or proinflammatory cytokines synergistically in- MKP-1 gene expression creased TLR2 gene expression. Recently, Kis et al. (2006) Finally, we studied MKP-1 gene expression in human reported that the synthetic corticosteroid budesonide en- keratinocytes by quantitative PCR. Dexamethasone increased hances TLR2 gene expression in both unstimulated and MKP-1 gene expression by 7.4-fold as compared to nontreated lipopolysaccharide-treated keratinocytes. These data strongly controls, and combination of dexamethasone with TNF-a suggest that cortisol, which is known to be a stress hormone, further increased MKP-1 gene expression by 15-fold as can exacerbate inflammatory reactions in acne vulgaris compared to the controls (Figure 5). Moreover, dexamethasone lesions. It is well known clinically that prolonged systemic significantly enhanced MKP-1 expression in keratinocytes or topical steroid treatment not only produces fresh develop- stimulated with the three different strains of P. acnes (Figure 6). ment of acne, that is, , but also exacerbates In contrast, RU486 completely inhibited the effect of already present acne lesions (Plewig and Kligman, 1973; dexamethasone on MKP-1 expression (Figure 5 and 6). Fung and Berger, 2000).

www.jidonline.org 377 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

30 ab NS NS 25 NS ∗∗ 20 ∗∗ 15 ∗∗

10

c d (x fold) TLR2 expression 5

0 − − − − − − − Dex TNF Dex+TNF Dex+TNF Dex+TNF Dex+TNF Dex+TNF Dex+TNF Dex+TNF CAPE U0126 SB202190 SP600125 Anisomycin e Ro-31-8220 Figure 4. Effect of inhibitors and activators on TLR2 gene expression in TNF-a-stimulated human keratinocytes in the presence of dexamethasone. Keratinocytes were treated with 1 mM Ro-31-8220, anisomycin, SB202190, SP600125, U0126, or CAPE, and stimulated with 10 ng ml1 TNF-a for 6 hours in the absence or presence of 1 mM dexamethasone. After stimulation, total RNA was prepared from the cells and cDNA was synthesized followed by quantitative PCR for TLR2. **Po0.01, NS; not significant, n ¼ 4.

Figure 2. Immunostaining of TLR2 protein on the surface of human keratinocytes stimulated with TNF-a. Keratinocytes were stimulated with 1 10 ng ml TNF-a for 24 hours in the presence or absence of 1 mM 20 dexamethasone. After stimulation, the cells were fixed with ethanol and TLR2 NS protein was stained with mouse anti-human TLR2 monoclonal antibody clone 1030A5.138 and Alexa Fluor 488 Goat Anti-mouse IgG. (a) No treatment, (b) ∗∗ ∗∗ 10 ng/ml TNF-a,(c)1mM dexamethasone, (d) dexamethasone and TNF-a,(e) isotype control. Scale bar ¼ 50 mm. 15

∗∗ 10 NS NS NS ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ ∗∗ 10 8

6 NS

4 (x fold) MKP1 expression 5

2 TLR2 expression (x fold) TLR2 expression 0 − − − − − 0 Dex Dex Dex Dex Dex − − RU486 Dex TNF- α TNF- α RU486+Dex RU486+Dex RU486+Dex

TNF-α JCM 6425 JCM 6473 ATCC11827 Dex+TNF Dex+TNF RU486 Figure 3. The effect of dexamethasone on TLR2 gene expression in P. acnes-stimulated human keratinocytes. Keratinocytes were stimulated with Figure 5. The effect of dexamethasone and RU486 on MKP-1 gene three P. acnes strains for 6 hours in the absence or presence of 1 mM expression in TNF-a-stimulated human keratinocytes. Keratinocytes were 1 dexamethasone and/or 1 mM RU486. After stimulation, total RNA was stimulated with 10 ng ml TNF-a for 6 hours in the absence or presence of prepared from the cells and cDNA was synthesized followed by quantitative 1 mM dexamethasone and 1 mM RU486. After stimulation, total RNA was PCR for TLR2. **Po0.01, NS; not significant, n ¼ 3. prepared from the cells and cDNA was synthesized followed by quantitative PCR for MKP-1. **Po0.01, NS; not significant, n ¼ 4.

Moreover, a causative link between emotional stress and stress in university students may significantly influence the acne has long been postulated without clear-cut evidence. progression of acne (Chiu et al., 2003). In general, women Emotional stress from external sources such as examination who suffer from acne often notice that it gets worse during

378 Journal of Investigative Dermatology (2009), Volume 129 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

30 NS ant-TLR4 neutralizing antibodies, among different strains of ∗∗ ∗∗ NS P. acnes. We have examined the differences in TLR2-inducing ∗∗ ∗∗ activity among the three different P. acnes strains in the 20 presence of dexamethasone. However, there were no NS ∗∗ ∗∗ remarkable differences in TLR2 expression. Jarrousse et al. 10 (2007) reported that the membrane fraction from P. acnes strain IP53113 increased TLR2 expression in human kerati-

MKP1 expression (x fold) MKP1 expression nocytes. Additional studies are needed to determine the 0 − − − − − differences in effects between different strains of P. acnes. Dex Dex Dex Dex Dex Most of all, the demonstration of TLR2 overexpression in RU486 biopsy specimens obtained from different acne patients,

RU486+Dex RU486+Dex RU486+Dex particularly those under corticosteroid treatment, could TNF-α JCM 6425 JCM 6473 ATCC11827 reinforce the results of these in vitro experiments. We also found that TLR2 mRNA expression normalized Figure 6. The effect of dexamethasone on MKP-1 gene expression in P. acnes-stimulated human keratinocytes. Keratinocytes were stimulated with against glyceraldehyde-3-phosphate dehydrogenase expres- three P. acnes strains for 6 hours in the absence or presence of 1 mM sion in keratinocytes were relatively low compared that of dexamethasone and/or 1 mM RU486. After stimulation, total RNA was THP-1 (data not shown). It is therefore reasonable that prepared from the cells and cDNA was synthesized followed by quantitative immune cells such as macrophages are mainly responsible PCR for MKP-1. **Po0.01, NS; not significant, n ¼ 3. for inflammatory reactions, whereas keratinocytes act as a trigger for inflammatory reactions against infecting bacteria. We also suggest another possibility that a suppressive menses or periods of psychosocial stress. A pathogenetic link mechanism keeps the expression of TLR2 at low levels in between mental stress and acne mediated by neuroendocrine keratinocytes and could suppress inflammatory reactions regulation of sebocytes has been suggested (Zouboulis and against resident microflora in the skin. Although we Bohm, 2004). Our present data provide additional evidence repeatedly tried to detect the TLR2 protein by western that indirectly supports the exacerbation of acne by blotting, we could not obtain any meaningful signals likely psychosocial stress. due to low TLR2 expression in keratinocytes. In this study, IL-8 expression in keratinocytes was strongly MKP-1, which is also known as dual specificity phospha- induced at 1 hour after the application of TNF-a but tase-1, inactivates MAPK by dephosphorylation, thereby decreased by 6 hours (Figure 1b), suggesting negative feed- contributing to the anti-inflammatory effects of glucocorti- back regulation by inhibitor proteins produced after stimula- coids (Abraham et al., 2006). There are several reports tion by proinflammatory cytokines. These data indicate that indicating that glucocorticoids induce MKP-1 expression in dexamethasone has opposite effects on TLR2 and IL-8 gene cell strains other than keratinocytes and that MKP-1 depho- expression. sphorylates and deactivates p38 MAPK, resulting in the It is known that constituents of P. acnes such as activation of TLR2 gene expression (Lasa et al., 2002; peptidoglycan and lipoteicoic acid activate TLR2, sub- Hermoso et al., 2004; Homma et al., 2004; Sakai et al., sequently activating intracellular inflammatory signals. In this 2004). In our present experiments using human keratinocyte study, we assessed TLR2 gene expression induced by three cell cultures, we found that the dexamethasone treatment different strains of P. acnes, JCM6425, JCM6473, and significantly increased MKP-1 gene expression, and that ATCC11827. P. acnes can be classified into two serotypes combination of dexamethasone with TNF-a resulted in much (I and II) that are distinguished by the agglutination test based stronger induction of MKP-1 expression than that by on the reaction to galactose in the cell walls, and five biotypes dexamethasone alone. Hence, we tried to assess the (B1–B5) on the basis of fermentation tests of ribose, erythritol, intracellular signaling mechanism of glucocorticoid- and sorbitol. Sero-biotype IB1(ATCC11827), IB3(ATCC6919, enhanced TLR2 gene expression by using several inhibitors which is the same as JCM6425), and IIB2(ATCC11828, which and an activator. All of the data suggest that TLR2 gene is the same as JCM6473), are more frequently isolated from expression was negatively regulated by p38 MAPK, which facial skin than other types (Kishishita et al., 1979). Thus, we was in turn inactivated by MKP-1. The TLR2/TLR1 dimer is chose these three different strains of P. acnes for our known to recognize triacyl lipoproteins of bacteria (Akira experiments. They slightly increased TLR2 expression in the et al., 2006). In this study, there was no significant effect of absence of dexamethasone, although TNF-a and IL-1a with- inhibitors of NF-kB, extracellular signal-regulated kinase, or out P. acnes also increased TLR2 expression even in the c-Jun N-terminal kinase, although these pathways might be absence of dexamethasone. These findings suggest that the related to TLR2 gene expression. Rather, our data strongly cellular signaling induced by P. acnes is different from that suggest that MKP-1 and p38 MAPK are closely associated induced by proinflammatory cytokines. In this study, we with glucocorticoid-enhanced TLR2 expression (Figure 7). employed living bacteria, which made it difficult to use Taken together, our results suggest that glucocorticoids exactly the same number of bacteria in each experiment. enhance TLR2 expression in human keratinocytes, thereby Nagy et al. (2005) reported that there is a large difference in inducing deleterious inflammatory changes in acne vulgaris. hBD-2 induction activity, which is inhibited by anti-TLR2 and Furthermore, the present findings suggest that steroid-induced

www.jidonline.org 379 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

IL-1α TNF-α P. acnes Cell treatment Keratinocytes were seeded in six-well culture plates at a concentration of 2 105 cells per well in 2 ml of medium and cultured for 3 days at IL-1R TNFR TLR2 TLR1 37 1C in a humidified atmosphere containing 5% CO2 in air. After 3 GC+receptor days, the medium was changed to 2 ml of cortisol-free medium. On the next day, the medium was changed again to 2 ml of fresh cortisol-free NF-κB pathway MAPK pathway medium, and glucocorticoids and/or various compounds such as RU486 SB202190, SP600125, U0126, caffeic acid phenethyl ester, Ro-31- ERK JNK p38 MKP-1 8220, and anisomycin were added to the culture. For the experiment CAPE U0126 Ro-31-8220 using viable P. acnes cultures, the medium was changed to 1 ml of fresh No significant No significant SP600125 cortisol-free medium, and dexamethasone was added to the culture. effect SB202190 effect No significant Anisomycin After a 20-minute incubation, viable P. acnes or the inflammatory effect cytokines TNF-a or IL-1a were added to the cultures and cells were TLR2 gene expression incubated for 1 or 6 hours in a humidified atmosphere of 5% CO2 in air. THP-1 cells were seeded in six-well culture plates at 1 106 cells per Figure 7. Regulation of TLR2 expression in human keratinocytes. well in 2 ml of medium, and 3.2 mlof100mM phorbol 12-myristate 13-acetate were added for differentiation to macrophage-like cells. Two TLR2 in epidermal and follicular keratinocytes may play an days after, nonattached cells were removed by aspiration, and the important role in not only acne, but also steroid rosacea and adherent cells were washed with PBS() twice. About 2 ml of medium perioral dermatitis (du Vivier and Stoughton, 1975). and dexamethasone was added to the culture. After a 20-minute incubation, TNF-a was added to the cultures and cells were

MATERIALS AND METHODS incubated for 6 hours in a humidified atmosphere of 5% CO2 in air. Materials After incubation, cells were extracted with Isogen (Nippon Gene, Normal human epidermal keratinocytes were obtained from Kurabo Tokyo, Japan), and total RNA was prepared according to the (Osaka, Japan). SB202190, SP600125, U0126, CAPE, Ro-31-8220, manufacturer’s protocol. Total RNA was dissolved in 50 mlofRNase- and anisomycin were obtained from Calbiochem (Darmstadt, free water. The RNA concentration was determined by measuring the Germany). Dexamethasone and cortisol were obtained from absorbance at 260 nm with a NanoDrop ND-1000 (NanoDrop Sigma-Aldrich (St Louis, MO). TNF-a and IL-1a were purchased Technologies, Wilmington, DE). cDNA was synthesized by incubating from R&D Systems (Minneapolis, MN). The TLR2 monoclonal 1 mgoftotalRNAina20ml reaction mixture containing cDNA synthesis antibody clone 1030A5.138 was purchased from Santa Cruz buffer (Invitrogen, Carlsbad, CA), 0.5 mg of oligo-dT primer (Invitrogen), Biotechnology (Santa Cruz, CA). 10 mM dithiothreitol, 2.5 mM dATP, 2.5 mM dTTP, 2.5 mM dGTP, 2.5 mM dCTP, 50 U of RNase inhibitor (Takara, Otsu, Japan), and 200 U of Cell culture Superscript II RNase H- reverse transcriptase (Invitrogen) at room Keratinocytes were cultured in the serum-free keratinocyte growth temperature for 10 minutes followed by 42 1C for 50 minutes. Sterile

medium EpiLife (Kurabo) supplemented with 0.06 mM.CaCl2, 0.4% (v/v) water (80 ml) was added to the reaction tube after the reaction, and this bovine pituitary extract, 0.1 ng ml1 human recombinant epidermal cDNA solution was used for quantitative PCR. growth factor, 10 mgml1 insulin, 0.5 mgml1 hydrocortisone (cortisol), 50 mgml1 gentamycin, and 50 ng ml1 amphotericin B. The monocytic Quantitative PCR cell line THP-1 (American Type Culture Collection, Manassas, VA) was The amount of cDNA was determined by real-time PCR using the cultured in RPMI1640 (GIBCO BRL, Carlsbad, CA) containing 10% SYBR Green method with an ABI PRISM 7900HT Sequence 1 fetal bovine serum, 2 mML-glutamine, 100 U ml trypsin, Detection System (Applied Biosystems, Foster City). A DNA standard 100 mgml1 streptomycin, and 250 ng ml1 amphotericin B. for each gene tested was obtained by cloning the PCR amplicon into a TA TOPO cloning vector (Invitrogen). A 2.5 ml sample of cDNA or P. acnes 102–10 8 copies of DNA standards were reacted in 25 ml of reaction P. acnes strains JCM6425 and JCM6473 were obtained from the mixture comprised of PCR buffer with Platinum SYBR Green qPCR

Japan Collection of Microorganisms, RIKEN BioResource Center SuperMix-UDG with ROX (Invitrogen) containing MgCl2, dNTPs, (Wako, Japan), and strain ATCC11827 was from American Type and Taq DNA polymerase, and 0.3 mM forward and reverse primers. Culture Collection. These bacteria strains were cultured in modified The sequences of the primers for TLR2, MKP-1, IL-8, and the internal GAM Broth medium (Nissui Pharmaceutical, Tokyo, Japan) for 3 control gene glyceraldehyde-3-phosphate dehydrogenase were days at 37 1C in an anaerobic jar without shaking. For stimulation of designed from previously reported mRNA sequences (Table 1). keratinocytes, bacterial cells were harvested by centrifugation at PCR conditions were as follows: 50 1C for 2 minutes, 95 1C for 2000 g for 10 minutes at 4 1C. The precipitate was washed with 10 minutes, 40 cycles of 95 1C for 15 seconds, and 60 1C for Ca þþ- and Mg þþ-free phosphate-buffered saline (PBS()) three 30 seconds. A dissociation curve step was added after the PCR times. Bacteria were suspended in cortisol-free EpiLife medium at reaction according to the manufacturer’s instruction to check for approximately 108 colony-forming units ml1, and 1 ml of bacteria correct PCR amplification. The amount of cDNA in the unknown culture was then added to the keratinocyte culture. The number of samples was calculated from the standard curve of each gene. bacteria in the culture was determined by counting colonies after Expression of each gene was normalized against expression of serial 10-fold dilution and culture on solid medium. glyceraldehyde-3-phosphate dehydrogenase.

380 Journal of Investigative Dermatology (2009), Volume 129 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

Table 1. PCR primers Forward Reverse Size (bp) GenBank accession no.

hTLR2 ggccagcaaattacctgtgt ttctccacccagtaggcatc 298 NM_003264 hMKP-1 tcctgccctttctgtacctg acccttcctccagcattctt 232 NM_004417 hIL-8 tcagagacagcagagcacaca aatcaggaaggctgccaag 127 NM_000584 hGAPDH gaaggtgaaggtcggagtc gaagatggtgatgggatttc 226 NM_002046

Detection of TLR2 expression by fluorescence microscopy and the sympathetic adrenomedullary system to stress in patients with Keratinocytes were seeded in four-well Lab-Tek II chamber slides atopic dermatitis. J Clin Endocrinol Metab 87:4245–51 (Life Technologies; no. 154526) at a concentration of 2.5 104 cells Chiu A, Chon SY, Kimball AB (2003) The response of skin disease to stress: per well in 0.5 ml medium and cultured for 2 days at 37 1Cina changes in the severity of acne vulgaris as affected by examination stress. Arch Dermatol 139:897–900 humidified atmosphere containing 5% CO2 in air. After 2 days, the Cotterill JA (1979) Perioral dermatitis. Br J Dermatol 101:259–62 medium was changed to 1 ml cortisol-free medium. On the next day, du Vivier A, Stoughton RB (1975) Tachyphylaxis to the action of topically the medium was changed again to cortisol-free fresh medium, and applied corticosteroids. Arch Dermatol 111:581–3 dexamethasone was added. After a 20-minute incubation, TNF-a Evolahti A, Hultcrantz M, Collins A (2006) Women’s work stress and was added to the cultures and cells were incubated for 24 hours in a cortisol levels: a longitudinal study of the association between the humidified atmosphere of 5% CO2 in air. After incubation, cells psychosocial work environment and serum cortisol. J Psychosom Res were washed with PBS() and fixed with ethanol for 15 minutes 61:645–52 followed by immunostaining with the anti-TLR2 antibody. After Flores BH, Kenna H, Keller J, Solvason HB, Schatzberg AF (2006) Clinical and removal of ethanol, cells were washed with PBS() three times and biological effects of mifepristone treatment for psychotic depression. Neuropsychopharmacology 31:628–36 incubated for 1 hour at room temperature in 200 ml of normal goat Fung MA, Berger TG (2000) A prospective study of acute-onset steroid acne serum for blocking. Then, slides were incubated with the TLR2 associated with administration of intravenous corticosteroids. Dermatol- monoclonal antibody clone 1030A5.138 at 4 1C overnight. The ogy 200:43–4 slides were washed with PBS() three times and incubated with an Haehnel V, Schwarzfischer L, Fenton MJ, Rehli M (2002) Transcriptional Alexa 488-conjugated goat anti-mouse IgG(H þ L) (Invitrogen) for regulation of the human toll-like receptor 2 gene in monocytes and 1 hour at room temperature. After washing with PBS(), the slides macrophages. J Immunol 168:5629–37 were mounted in ProLong Gold antifade reagent (Invitrogen). The Hermoso MA, Matsuguchi T, Smoak K, Cidlowski JA (2004) Glucocorticoids and tumor necrosis factor alpha cooperatively regulate toll-like receptor slides were observed with a fluorescence microscope. 2 gene expression. Mol Cell Biol 24:4743–56 Homma T, Kato A, Hashimoto N, Batchelor J, Yoshikawa M, Imai S et al. Statistical analysis (2004) Corticosteroid and cytokines synergistically enhance toll-like Experimental data are presented as the mean±SD. Statistical receptor 2 expression in respiratory epithelial cells. Am J Respir Cell Mol Biol 31:463–9 differences between the two experimental groups were calculated Hosoi J, Tanida M, Tsuchiya T (2001) Mitigation of stress-induced suppression by two-tailed Student’s t-test. Statistical comparisons between more of contact hypersensitivity by odorant inhalation. Br J Dermatol than two groups were carried out by ANOVA and Scheffe’s test. 145:716–9 Po0.05 was considered to be statistically significant. Imasato A, Desbois-Mouthon C, Han J, Kai H, Cato AC, Akira S et al. (2002) Inhibition of p38 MAPK by glucocorticoids via induction of MAPK CONFLICT OF INTEREST phosphatase-1 enhances nontypeable Haemophilus influenzae-induced The authors state no conflict of interest. expression of toll-like receptor 2. J Biol Chem 277:47444–50 Jarrousse V, Castex-Rizzi N, Khammari A, Charveron M, Dreno B (2007) Zinc salts inhibit in vitro Toll-like receptor 2 surface expression by ACKNOWLEDGMENTS keratinocytes. Eur J Dermatol 17:492–6 We thank Dr Toru Tsuchiya for helpful advice. Jugeau S, Tenaud I, Knol AC, Jarrousse V, Quereux G, Khammari A et al. (2005) Induction of toll-like receptors by Propionibacterium acnes. Br J Dermatol 153:1105–13 REFERENCES Kim J, Ochoa MT, Krutzik SR, Takeuchi O, Uematsu S, Legaspi AJ et al. (2002) Abraham SM, Lawrence T, Kleiman A, Warden P, Medghalchi M, Activation of toll-like receptor 2 in acne triggers inflammatory cytokine Tuckermann J et al. (2006) Antiinflammatory effects of dexamethasone responses. J Immunol 169:1535–41 are partly dependent on induction of dual specificity phosphatase 1. Kis K, Bodai L, Polyanka H, Eder K, Pivarcsi A, Duda E et al. (2006) J Exp Med 203:1883–9 Budesonide, but not tacrolimus, affects the immune functions of normal Akira S, Uematsu S, Takeuchi O (2006) Pathogen recognition and innate human keratinocytes. Int Immunopharmacol 6:358–68 immunity. Cell 124:783–801 Kishishita M, Ushijima T, Ozaki Y, Ito Y (1979) Biotyping of Propionibacter- Beltman J, McCormick F, Cook SJ (1996) The selective protein kinase C ium acnes isolated from normal human facial skin. Appl Environ inhibitor, Ro-31-8220, inhibits mitogen-activated protein kinase Microbiol 38:585–9 phosphatase-1 (MKP-1) expression, induces c-Jun expression, and Lasa M, Abraham SM, Boucheron C, Saklatvala J, Clark AR (2002) activates Jun N-terminal kinase. J Biol Chem 271:27018–24 Dexamethasone causes sustained expression of mitogen-activated Buske-Kirschbaum A, Geiben A, Hollig H, Morschhauser E, Hellhammer D protein kinase (MAPK) phosphatase 1 and phosphatase-mediated (2002) Altered responsiveness of the hypothalamus-pituitary-adrenal axis inhibition of MAPK p38. Mol Cell Biol 22:7802–11

www.jidonline.org 381 M Shibata et al. Glucocorticoids Enhance TLR2 in Keratinocytes

Leyden JJ, Thew M, Kligman AM (1974) Steroid rosacea. Arch Dermatol Shuto T, Imasato A, Jono H, Sakai A, Xu H, Watanabe T et al. 110:619–22 (2002) Glucocorticoids synergistically enhance nontypeable Nagy I, Pivarcsi A, Koreck A, Szell M, Urban E, Kemeny L (2005) Distinct Haemophilus influenzae-induced Toll-like receptor 2 expression strains of Propionibacterium acnes induce selective human beta- via a negative cross-talk with p38 MAP kinase. J Biol Chem 277: defensin-2 and interleukin-8 expression in human keratinocytes through 17263–70 toll-like receptors. J Invest Dermatol 124:931–8 Singh LK, Pang X, Alexacos N, Letourneau R, Theoharides TC (1999) Acute Plewig G, Kligman AM (1973) Induction of acne by topical steroids. Arch immobilization stress triggers skin mast cell degranulation via cortico- Dermatol Forsch 247:29–52 tropin releasing hormone, neurotensin, and substance P: a link to Sakai A, Han J, Cato AC, Akira S, Li JD (2004) Glucocorticoids synergize with neurogenic skin disorders. Brain Behav Immun 13:225–39 IL-1beta to induce TLR2 expression via MAP Kinase Phosphatase-1- Zouboulis CC, Bohm M (2004) Neuroendocrine regulation of sebocytes—a dependent dual Inhibition of MAPK JNK and p38 in epithelial cells. BMC pathogenetic link between stress and acne. Exp Dermatol 13(Suppl 4): Mol Biol 5:2 31–5

382 Journal of Investigative Dermatology (2009), Volume 129