Fungal–Bacterial Crosstalk Modulates Glucocorticoid-Primed TLR2 Signaling in the Human Skin

TL;DR

This study shows how a skin yeast interacts with bacteria to influence inflammation in skin cells under steroid treatment, offering new insights into acne and immune responses.

Contribution

The study reveals a novel mechanism of fungal-bacterial crosstalk modulating glucocorticoid-primed TLR2 signaling in human skin.

Findings

Malassezia restricta suppresses dexamethasone-enhanced TLR2 expression by Cutibacterium acnes in keratinocytes.

M. restricta modulates TLR2 signaling via p38 MAPK activation and inhibition of NF-κB p65 nuclear translocation.

Fungal-bacterial interactions shape host immune responses in steroid-treated skin, impacting acne pathogenesis.

Abstract

Cutibacterium acnes, a major skin commensal bacterium, induces inflammatory cytokine production in keratinocytes through Toll-like receptor 2 (TLR2) signaling and contributes to acne vulgaris pathogenesis. Although glucocorticoids, e.g., dexamethasone (Dex), exert anti-inflammatory effects in related treatments, prolonged glucocorticoid exposure paradoxically induces acneiform eruptions, a phenomenon referred to as steroid-induced acne. Moreover, how commensal fungi influence bacterial-driven inflammatory signaling under glucocorticoid treatment remains unclear. In this study, we investigated how the lipophilic skin yeast Malassezia restricta affects C. acnes-induced TLR2 expression under Dex treatment using normal human epidermal keratinocytes. We discovered that M. restricta selectively suppressed Dex-enhanced C. acnes-induced TLR2 expression both at the transcriptional level and cell surface. Mechanistically, M. restricta enhanced p38 MAPK phosphorylation and inhibited NF-κB p65 nuclear translocation, indicating context-dependent glucocorticoid-primed TLR2 signaling modulation rather than simple inhibition. These results demonstrate that M. restricta modulates bacterial-induced inflammatory responsiveness in keratinocytes under glucocorticoid exposure and highlight the importance of fungal–bacterial interactions in shaping host immune signaling in steroid-treated skin. Our study provides new insight into the mechanistic basis of steroid-induced acne and the polymicrobial regulation of cutaneous innate immunity.

Figures

Full text

1. Introduction

The human skin harbors a diverse microbiome comprising bacteria, fungi, and viruses that are essential for maintaining barrier function and preventing invasion by external pathogens through the sebum and keratinized layers [1,2]. Within the bacterial microbiome, Staphylococcus, Cutibacterium, and Corynebacterium are the predominant genera, although their relative abundance varies depending on the body site. In contrast, the fungal microbiome is dominated by the lipophilic yeasts Malassezia restricta and Malassezia globosa, regardless of body region [3,4,5]. Therefore, a notable feature of the skin microbiome comprises the distinct predominance of bacterial and fungal genera that differ between these microbial communities, existing in a microbial balance, the disruption of which potentially triggers inflammatory skin disorders (e.g., acne vulgaris and seborrheic dermatitis) [6,7]. Among these microbes, the Gram-positive anaerobe Cutibacterium acnes (formerly Propionibacterium acnes) resides in the hair follicles. Its cell wall components activate nuclear factor kappa B (NF-κB) and mitogen-activated protein kinase (MAPK) pathways via Toll-like receptor (TLR) on keratinocytes, inducing inflammation-driving cytokine and chemokine production [8,9].

Glucocorticoids exert potent anti-inflammatory effects by suppressing NF-κB activity and reducing inflammatory cytokine production [10,11]. However, prolonged or high-dose glucocorticoid use reportedly induces acneiform eruptions, commonly referred to as steroid acne [12,13]. Although this phenomenon appears paradoxical, it might be explained by upregulated glucocorticoid-induced TLR2 expression in keratinocytes, enhancing cellular responsiveness to inflammatory stimuli [14,15,16]. Therefore, glucocorticoids possess a dual role, i.e., anti-inflammatory and keratinocyte sensitizing effects to external stimuli through the TLR2 pathway. Steroid acne could thus be considered a paradoxical inflammatory reaction arising from a TLR2-sensitized state under immunosuppression. Furthermore, the hair follicle microbiome includes lipophilic yeasts of the genus Malassezia, coexisting with C. acnes and potentially influencing host immune responses [17]. Nonetheless, how fungal–bacterial interactions affect keratinocyte signaling under glucocorticoid exposure remains unclear. Notably, altered TLR2 expression does not necessarily display a linear correlation with disease severity; TLR2 expression rather reflects the balance between inflammatory priming and downstream signal execution under specific hormonal and microbial conditions. Importantly, our study does not assume that reduced TLR2 expression necessarily leads to disease attenuation. Instead, we propose that commensal fungi (e.g., M. restricta) actively modulate glucocorticoid-dependent host signaling, thereby reshaping inflammatory responsiveness to bacterial stimuli within a polymicrobial follicular environment.

Based on previous studies and the hereby-presented conceptual framework, Figure 1 presents a schematic overview of the proposed signaling pathways underlying dexamethasone (Dex)-mediated TLR2 expression regulation.

In this study, we established a C. acnes-M. restricta co-culture model in the presence of Dex to elucidate how microbial crosstalk affects TLR2 expression and downstream signaling, including p38 MAPK phosphorylation and NF-κB nuclear translocation, in normal human epidermal keratinocytes (NHEKs). We aimed to clarify the role of skin commensal interactions in steroid-induced acne pathogenesis and provide novel insight into host–microbe relationships under hormonal regulation.

2. Materials and Methods

2.1. Microorganisms

C. acnes NBRC107605 was cultured anaerobically on Gifu anaerobic agar medium (Nissui Pharmaceutical, Tokyo, Japan) at 37 °C for 2 days and used for experiments. M. restricta NBRC103918 was grown aerobically on modified Leeming and Notman agar medium (mLNA, peptone 10 g, glucose 10 g, yeast extract 2 g, bile salt 8 g, glyceryl monostearate 0.5 g, glycerine 10 g, olive oil 20 mL, Tween 60 5 mL, and agar 15 g per 1 L) at 32 °C for 4–5 days. Both strains were obtained from the NITE Biological Resource Center (NBRC, Chiba, Japan, https://www.nite.go.jp/nbrc/catalogue/, accessed on 5 November 2025).

2.2. Reagents

Dex and anisomycin were purchased from FUJIFILM Wako Pure Chemical Corporation (Tokyo, Japan). U0126, SP600125, and celastrol were obtained from InvivoGen (San Diego, CA, USA), and MG132 from Merck KGaA (Darmstadt, Germany).

2.3. Cell Culture

Primary NHEKs (Takara Bio, Shiga, Japan) were cultured in EpiLife™ medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with Human Keratinocyte Growth Supplement (HKGS; Gibco™, Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics (penicillin 100 IU/mL and streptomycin 100 µg/mL; Nacalai Tesque, Kyoto, Japan). Cells were maintained at 37 °C in a humidified atmosphere containing 5% CO_2_.

2.4. Cell Treatment and Infection Experiments

When NHEKs reached confluency, they were detached using ReagentPack™ Subculture Reagents (Lonza, Walkersville, MD, USA) and seeded at 3 × 10^5^ cells/mL in 24-well plates. Two hours before infection, the medium was replaced with supplement- and antibiotic-free EpiLife medium. Dex (1 μM) was added 20 min before infection. This pretreatment period was selected to establish glucocorticoid receptor-dependent signaling at the time of microbial stimulation, rather than to induce Dex-driven TLR2 protein expression. C. acnes and M. restricta (multiplicity of infection [MOI] = 50 and 10, respectively) were freshly harvested, washed, resuspended in the same medium, and used for infection. The concentration of Dex and the MOIs of C. acnes and M. restricta were determined based on preliminary optimization experiments and previous reports to induce measurable signaling responses without cytotoxicity [18,19,20,21,22,23,24]. Cells were collected at 30 min, 6 h, and 24 h post-infection for Western blotting, RNA analysis, and immunofluorescence, respectively.

Inhibitors or activators for signaling pathway analysis were added 1 h prior to infection: MG132 (a proteasome inhibitor that blocks IκBα degradation; Calbiochem, San Diego, CA, USA), celastrol (NF-κB pathway inhibitor; Sigma–Aldrich, St. Louis, MO, USA), anisomycin (p38 activator; Sigma–Aldrich, St. Louis, MO, USA), U0126 (ERK inhibitor; Cell Signaling Technology, Danvers, MA, USA), and SP600125 (JNK inhibitor; Calbiochem, San Diego, CA, USA). The concentrations of these inhibitors and activators (0.5–1 μM) were selected based on previous studies and manufacturer recommendations, demonstrating effective NF-κB and MAPK signaling pathway modulation in keratinocytes without inducing cytotoxicity.

2.5. Quantitative PCR Analysis of TLR2 Gene Expression

Total RNA was extracted from NHEKs 6 h after infection using the High Pure RNA Isolation Kit (Roche Diagnostics, Mannheim, Germany), and cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA). Quantitative PCR was performed using 2× SYBR Green PCR Master Mix (Applied Biosystems, Foster, CA, USA) with the following primers [25]: TLR2-F (5′-TCTCCCATTTCCGTCTTTTT-3′), TLR2-R (5′-GGTCTTGGTGTTCATTATCTTC-3′), GAPDH-F (5′-CCCCACACACATGCACTTACC-3′), and GAPDH-R (5′-TTGCCAAGTTGCCTGTCTT-3′). Relative gene expression levels were normalized to GAPDH that was used as an internal control. Melting curve analysis confirmed the presence of an amplification peak for each primer pair, indicating specific amplification.

2.6. TLR2 Immunofluorescence Staining

NHEKs were cultured on sterile glass slides (Matsunami Glass Industry, Osaka, Japan), placed in 24-well plates, until reaching confluency, before infection with C. acnes and M. restricta. After 24 h, the medium was removed, and the cells were washed once with phosphate-buffered saline (PBS). Cells were fixed using 4% paraformaldehyde in PBS (600 µL/well) for 5 min, washed thrice with PBS, and permeabilized with 0.2% Triton^®^ X-100 (FUJIFILM Wako Pure Chemical, Osaka, Japan) for 30 min at room temperature.

After washing, the cells were incubated overnight at 4 °C with a mouse anti-human TLR2 monoclonal antibody (Abcam, Cambridge, UK; 1:1000 dilution), followed by additional PBS washes and incubation with DyLight™ 488 goat anti-mouse IgG (Abcam, 1:2500 dilution) for 1 h at room temperature. Nuclei were counterstained with DAPI (Dojindo, Kumamoto, Japan). Samples were mounted with a commercial Mounting Medium (R&D Systems, Minneapolis, MN, USA) and visualized under a fluorescent microscope (BX61, Olympus Corporation, Tokyo, Japan).

2.7. Cytoplasmic and Nuclear Protein Extraction and Western Blotting

After Dex pretreatment for 20 min, C. acnes and M. restricta were added to the cell cultures, and the cells were harvested 30 min post-infection. Cytoplasmic and nuclear fractions were prepared using NE-PER™ Nuclear and Cytoplasmic Extraction Reagents (Thermo Fisher Scientific, Waltham, MA, USA). Proteins were separated using SDS–PAGE and probed with the following primary antibodies: phospho-p38 MAPK (Thr180/Tyr182; #9211, Cell Signaling Technology, Danvers, MA, USA), p38 MAPK (#9212, Cell Signaling Technology, Danvers, MA, USA), NF-κB p65 (ab28856, Abcam, UK), Lamin B1 (Proteintech, Rosemont, IL, USA), and GAPDH (Proteintech, Rosemont, IL, USA). HRP-conjugated goat anti-mouse or anti-rabbit IgG (Abcam, Cambridge, UK) was used as a secondary antibody. Signal detection relied on chemiluminescence.

2.8. Statistical Analysis

All experiments were performed independently at least three times. The data are represented as the mean ± standard deviation (SD). Statistical significance was analyzed using one-way ANOVA on ΔCt values followed by Dunnett’s multiple comparisons test. Normality was assessed using the Shapiro–Wilk test. For all statistical analyses, p-values of p < 0.05 were considered statistically significant.

3. Results

3.1. M. restricta Suppressed C. acnes-Induced TLR2 Expression upon Dex Treatment

First, we assessed how M. restricta could affect TLR2 gene expression in NHEKs using real-time PCR. Dex concentration (1 μM) was selected based on preliminary dose–response experiments demonstrating maximal TLR2 expression induction without detectable cytotoxicity. Stimulation with C. acnes for 6 h upregulated TLR2 mRNA expression, further enhanced by Dex (p < 0.01); however, co-culture with M. restricta significantly reduced TLR2 expression (p < 0.01) (Figure 2).

Immunofluorescence staining further confirmed the above-described discoveries: C. acnes stimulation induced moderate surface TLR2 fluorescence intensity, which was markedly enhanced in the presence of Dex. In contrast, co-culture with M. restricta reduced fluorescence intensity, confirming that this yeast strain suppressed TLR2 expression. DAPI nuclear staining revealed no apparent differences in cell density or overall cellular condition among the experimental groups (Figure 3).

To confirm that the anti-human TLR2 antibody used in this study specifically recognized its target antigen, immunofluorescence staining was performed using an anti-TLR2 antibody and a mouse IgG1 isotype control. Under the C. acnes + Dex treatment condition, we observed distinct green fluorescence, corresponding to the TLR2 localization, whereas we detected no specific staining with the IgG1 isotype control (Figure S1). These results confirm the specificity of the anti-TLR2 antibody used in this study.

Taken together, these results demonstrated that M. restricta suppressed C. acnes-induced TLR2 expression at the mRNA level in NHEKs, even in the presence of Dex, highlighting a context-dependent TLR2 signaling modulation rather than a simple receptor expression attenuation. Importantly, M. restricta stimulation alone did not induce TLR2 expression, and C. acnes + M. restricta co-stimulation in the absence of Dex neither suppressed nor enhanced TLR2 expression (Figures S2 and S3), indicating that the observed suppressive effect is not attributable to simple TLR2 co-ligation by microbial ligands. In parallel, inflammatory cytokine expression analysis revealed that Dex markedly suppressed C. acnes-induced IL-6, IL-8, and TNF-α expression, and co-culture with M. restricta did not further enhance this suppressive effect (Figure S4).

3.2. M. restricta Enhanced p38 MAPK Phosphorylation and Inhibited NF-κB Nuclear Translocation

To investigate how M. restricta affects downstream TLR2 signaling, we analyzed p38 MAPK phosphorylation and NF-κB localization using Western blotting. C. acnes stimulation increased p38 phosphorylation regardless of Dex treatment, while co-culture with M. restricta further enhanced phosphorylation intensity (Figure 4a). Our cytoplasmic and nuclear fraction analysis revealed that C. acnes or Dex treatment increased nuclear NF-κB p65 accumulation. We used GAPDH and Lamin B1 as loading controls for the cytoplasmic and nuclear fractions, respectively. However, in the presence of M. restricta, nuclear NF-κB levels markedly decreased to approximately 30% of those under the Dex-only treatment conditions, while cytoplasmic NF-κB levels remained mainly unchanged (Figure 4b). Notably, we observed enhanced p38 phosphorylation and the M. restricta-induced suppression of NF-κB nuclear translocation not only under Dex + C. acnes stimulation but also upon M. restricta stimulation alone, indicating that these signaling effects occur glucocorticoid exposure-independently. We quantified individual band intensities in the Western blot images using an image analysis software (Figure 4c,d).

Taken together, our results suggest that M. restricta modulates TLR2 expression by promoting p38 activation while inhibiting NF-κB nuclear translocation.

3.3. Effects of Inhibitors and Activators on TLR2 Expression

To elucidate the mechanisms underlying Dex-enhanced TLR2 expression, we examined the effects of various signaling inhibitors and activators. The NF-κB inhibitors MG132 and celastrol, as well as the p38 activator anisomycin, markedly reduced TLR2 gene expression. In contrast, the ERK1/2 inhibitor U0126 exerted no effect, while the JNK inhibitor SP600125 increased TLR2 expression (Figure 5). The concentrations of all the inhibitors and activators we used are indicated in the corresponding figure legends. In addition, we observed the dose-dependent effects of MG-132 and celastrol on TLR2 gene expression (Figure S5) and registered that the application of each compound alone had little to no effect on TLR2 expression (Figure S6).

These results indicate that the Dex-induced TLR2 expression is primarily regulated through the NF-κB and p38 MAPK pathways.

4. Discussion

In this study, we analyzed the interaction between skin commensal microbes under glucocorticoid treatment and demonstrated that M. restricta suppresses C. acnes-induced TLR2 expression in NHEKs. Although C. acnes stimulation markedly upregulated TLR2 expression in the presence of Dex, this increase was significantly attenuated by co-culture with M. restricta at the mRNA levels. These results suggest that commensal fungi could modulate hormone-dependent immune responses of the host, providing new insights into steroid-induced acne and rosacea-like dermatitis pathogenesis.

Dex and cortisol enhance TLR2 expression in human keratinocytes by suppressing p38 and JNK phosphorylation via MAPK phosphatase-1 (MKP-1) [15,16]. Furthermore, while glucocorticoids suppress C. acnes-induced inflammatory responses, they act synergistically with C. acnes or proinflammatory cytokines (e.g., IL-1α and TNF-α) to further enhance TLR2 expression, thereby contributing to steroid therapy-associated acneiform eruptions. Therefore, glucocorticoids exhibit dual effects: while they suppress NF-κB activity and are anti-inflammatory agents, they simultaneously enhance TLR2 gene expression, thereby creating a transiently anti-inflammatory yet “primed” state in keratinocytes, which might respond excessively to subsequent bacterial stimuli via TLR2 signaling. Steroid acne could thus be interpreted as a paradoxical inflammatory reaction occurring in this sensitized state under immunosuppression.

Interestingly, our results revealed that co-culture with M. restricta reversed this Dex-induced TLR2 upregulation, accompanied by reduced NF-κB nuclear translocation and enhanced p38 phosphorylation. Although Malassezia species reportedly induce TLR2 expression, such induction remains generally modest and context-dependent [26]. Importantly, we do not propose that TLR2 alone accounts for the entire proinflammatory response in keratinocytes, but rather that it should be considered a key regulatory node within a broader PRR network, potentially involving other receptors (e.g., TLR1 and TLR6) with detailed contributions beyond the scope of this study. Consistently, M. restricta alone elicited only minimal TLR2 expression in our system. In contrast, under glucocorticoid-primed conditions, where Dex markedly amplified C. acnes-induced TLR2 expression, M. restricta selectively attenuated this excessive response, indicating a context-dependent modulatory role rather than a simple inhibitory effect. Notably, we did not observe this modulatory effect when using heat-inactivated M. restricta or under Transwell-based non-contact co-culture conditions, indicating that viable fungal cells or direct cell–cell interactions are required for the suppression of Dex-enhanced TLR2 expression (Figure S7).

As M. restricta alone failed to induce TLR2 expression and did not modify C. acnes-induced TLR2 expression in the absence of Dex, the suppressive effect observed upon Dex treatment could not be explained by simple receptor co-ligation through TLR2. Rather than representing a protective or deleterious effect per se, such TLR2 expression modulation should be interpreted as a shift in signaling that governs inflammatory responsiveness under glucocorticoid exposure. Therefore, although M. restricta modulates p38 and NF-κB signaling irrespective of glucocorticoid exposure, TLR2 expression suppression requires a Dex-primed signaling context, highlighting a selective interaction between fungal modulation and glucocorticoid-dependent transcriptional regulation.

These results suggest that M. restricta might modulate the activation threshold of inflammatory signaling through NF-κB modulation, thereby dampening C. acnes-induced responses. As Malassezia species are lipid-dependent yeasts, their cell wall lipids or secreted metabolites could plausibly influence keratinocyte signaling pathways [26,27]. Specifically, M. restricta-derived factors such as long-chain saturated fatty acids, glycerol esters, or secretory lipases might interfere with TLR2 signaling, although the precise mechanisms remain to be clarified.

Moreover, other commensal bacteria (e.g., Staphylococcus epidermidis and Corynebacterium species) reportedly modulate host immune responses through TLR2-dependent pathways [28,29,30,31]. Along with our results, these studies support the existence of a multilayered TLR2-mediated regulatory network among skin commensals, wherein M. restricta might play an opposing, fine-tuning role.

The skin microbiome composition is easily influenced by host hormonal status, sebum content, and topical treatments [1]. Changes in Malassezia and C. acnes relative abundances during corticosteroid therapy could thus alter the inflammatory threshold, thereby affecting acneiform eruption susceptibility. Our data suggest that M. restricta suppresses C. acnes-induced TLR2 activation through NF-κB inhibition and p38 activation, thereby modulating the host response involved in steroid-induced acne formation.

This study retains several limitations. First, our experiments were conducted using an in vitro keratinocyte model, which does not fully reproduce the complex multicellular and immune interactions of the skin tissue in vivo. Second, the specific fungal components or secreted products responsible for TLR2 modulation remain unidentified. Potential candidates comprise cell wall lipids (e.g., long-chain saturated fatty acids and glycerol esters) and secretory lipases. Further biochemical and genetic analyses of these factors would be required to elucidate the relationship between fungal lipid metabolism and host signal transduction. Third, we tested only a single fungal and bacterial strain and a specific Dex concentration, which might not fully reflect clinical conditions. Future studies using three-dimensional skin models or in vivo systems are warranted to clarify the physiological relevance of Malassezia-mediated immune modulation.

5. Conclusions

In conclusion, in this study, we demonstrated that the skin commensal yeast M. restricta selectively attenuates glucocorticoid-enhanced, C. acnes-induced TLR2 expression in human keratinocytes. Rather than exerting a general inhibitory effect, M. restricta modulated TLR2 signaling in a Dex-dependent context, accompanied by enhanced p38 MAPK phosphorylation and reduced NF-κB nuclear translocation.

These results indicate that M. restricta functions as a context-dependent regulator of bacterial-driven inflammatory responsiveness under glucocorticoid exposure, highlighting the importance of fungal–bacterial interactions in shaping host innate immune signaling within the skin microbiome. Our results suggest that alterations in microbial balance under hormonal conditions might reshape host innate immune signaling and contribute to the pathophysiology of steroid-induced acne. Further studies using in vivo models or clinical samples would be required to elucidate the physiological relevance and molecular mediators of this modulatory interaction.