DNAJB2 Attenuates Rosacea Skin Inflammation and Angiogenesis by Inhibiting the Endoplasmic Reticulum Stress-mediated TLR2/Myd88/NF-κB pathway

Abstract

Endoplasmic reticulum stress (ERS) has recently been proposed as a core factor in the pathogenesis and aggravation of rosacea. The roles of ERS-related genes in rosacea are largely unknown and were investigated in this study. Rosacea microarray datasets were downloaded from the Gene Expression Omnibus (GEO) database. Differentially expressed ERS-related genes in rosacea patients vs. controls were screened using the Limma package, and LASSO regression was used to screen for characteristic genes. The infiltrating fraction was evaluated using ssGSEA. Clinical rosacea samples, age-matched healthy volunteers, and LL37-induced mice models were used to investigate the expression of DNAJB2 and its function. In the GSE65914 dataset, 17 differentially expressed ERS-related genes were screened. Of these, 13 were identified as characteristic genes predicting rosacea risk. The adaptive immune response, TLR signaling pathway, and chemokine signaling pathway were activated with a high risk of rosacea. After expression validation using the GSE155141 dataset, DNAJB2 was identified as a key gene. DNAJB2 expression was significantly decreased in both datasets, clinical samples, and the LL37-induced mice model. DNAJB2 overexpression could alleviate rosacea skin injury and inhibit expression of inflammatory cytokines and chemokines as well as angiogenesis. The infiltration levels of the majority of immune cell types were elevated in rosacea samples, and DNAJB2 overexpression inhibited CD4 + T cell infiltration, as well as Th1 and Th17 polarizing genes. Moreover, DNAJB2 could inhibit ERS marker proteins and the activated TLR2/Myd88/NF-κB pathway. DNAJB2 may be a novel target for rosacea treatment.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10753-025-02278-5.

Introduction

Rosacea is a chronic inflammatory skin disease characterized by recurrent erythema, flushing, telangiectasia, papules, pustules, and granulomatous changes, primarily affecting the central facial region [1, 2]. It predominantly affects individuals aged 30 to 50 years [3]. Patients with rosacea usually experience social and psychological stress, which significantly affects their quality of life [4, 5]. Therefore, the diagnosis and treatment of rosacea are of critical importance.

Immune system dysregulation and abnormal neurovascular function are involved in the clinical characteristics of rosacea [6–8]. Overactivated Toll-like receptor 2 (TLR2) plays a core role in the innate immunity of rosacea, which is activated by exposure to environmental stimuli that trigger the production of cytokines and antimicrobial peptides (including cathelicidin LL37) [9, 10]. Cathelicidin LL37 modulates the host immune response and initiates a cascade of inflammatory processes [11]. In addition to the infiltration of innate immune cells, the adaptive immune system contributes to rosacea pathogenesis [12, 13]. The recruitment of T cells around hair follicles and blood vessels in the dermis primarily involves Th1 and Th17 cells [14]. Th17 cells also stimulate the expression of vascular endothelial growth factor (VEGF), promoting vascular proliferation and expansion [14, 15]. Understanding the pathogenesis of rosacea is important for identifying new treatment strategies.

The endoplasmic reticulum (ER) is a specialized organelle responsible for various functions, including the coordination of protein biogenesis, folding, assembly, transport, and the degradation of defective proteins [16–19]. However, when certain proteins fail to fold correctly, they aggregate and accumulate in the ER cavity, leading to impaired biological function and disrupted homeostasis, a state known as endoplasmic reticulum stress (ERS) [20]. Prolonged or severe ERS can lead to abnormal inflammatory signaling and promote cell death [21]. ERS has been shown to occur in several skin diseases, including Darier’s disease, vitiligo, melanoma, and rosacea [22]. Moreover, rosacea-inducing factors can trigger ERS, and ERS is found as a crucial promoter of rosacea pathogenesis [23]. However, the detailed molecular mechanisms of ERS involved in rosacea remains unclear.

In this study, key genes involved in the onset and exacerbation of rosacea mediated by ERS were identified through bioinformatics analysis and validated using clinical samples and animal models. These findings aim to provide new insights into the pathogenesis and potential treatment strategies for rosacea.

Materials and Methods

Data Acquisition

Microarray datasets for rosacea, including GSE65914 (GPL570 platform) and GSE155141 (GPL16791 platform), were downloaded from the Gene Expression Omnibus (GEO) database. The GSE65914 dataset included 58 samples, with 32 (12 papulopustular rosacea and 20 healthy volunteers) used as the training set. Data from five rosacea and five controls in the GSE155141 dataset were utilized as a validation set.

Screening of Differentially Expressed ERS-Related Genes

Differential analysis between rosacea patients and healthy controls was conducted using the empirical Bayesian and linear regression provided in the Limma package (version 3.10.3), followed by Benjamini–Hochberg method-based multiple test correction. Differentially expressed genes (DEGs) were selected with a cut-off value of adjusted P < 0.05 and log2|FC|> 1. ERS-related genes were obtained from a previous study [24]. The differentially expressed ERS-related genes were screened from the DEGs.

Protein–Protein Interaction (PPI) Network

The PPIs for the differentially expressed ERS-related genes were predicted using the STRING database (version 11.5), with an interaction score threshold set at 0.15 (low confidence). The PPI network was visualized using Cytoscape software (version 3.9.2).

Identification and Evaluation of Characteristic Genes

Genes that existed in the PPI network were selected to further screen characteristic genes using LASSO Cox regression. Briefly, the GLMNet package (version 4.0–2) was employed to conduct LASSO regression with tenfold cross-validation, and genes with non-zero regression coefficients were determined as characteristic genes. A diagnostic risk model was then established based on these characteristic genes as per the formula: Risk score = β gene1*exprgene1 + β gene2*exprgene2 + . + β genen*exprgenen. In the formula, β and expr represent the regression coefficient and expression value of the gene, respectively. The risk score was calculated for all samples to further assign them into high- and low-risk groups according to the median value. To evaluate the diagnostic performance of the model and genes, we plotted ROC curves using the pROC package (version 1.18.0) for both the training and validation sets. The expression of characteristic genes was validated in an independent dataset, GSE155141, and compared between rosacea and controls using a t-test.

Gene Set Enrichment Analysis (GSEA)

To explore changes in biological functions in the two rosacea risk groups, we conducted GSEA using the clusterProfiler package. Predefined gene sets (c5.go.bp.v2022.1. Hs.symbols.gmt and c2.cp.kegg.v2022.1. Hs.symbols.gmt) in the MSigDB database were used as enrichment references, and GSEA was conducted to explore the significantly altered GO terms and KEGG pathways between the groups, with a threshold value of P < 0.05.

Immune Cell Infiltration

Use the Rgsva package (http:/Into/master.bioconductor.org/packages/release/bioc/html/GSVA.html, verison1.42.0) for each sample ssGSEA estimation was performed, and differential immune cell analysis was performed on samples from different groups.

Clinical Sample Collection

With the approval of the Medical Ethics Committee of the Second Affiliated Hospital of Xi 'an Jiaotong University and the informed consent of the patients themselves, facial skin tissues of 32 patients with rosacea and 14 age-matched healthy volunteers were collected from the dermatology department of the Second Affiliated Hospital of Xi 'an Jiaotong University.

Preparation of Animal Models

Male BALB/c mice (7–8 weeks old, 22–25 g) were purchased from Jiangsu Yaokang Biotechnology Co., Ltd. After adaptive feeding for one week, a rosacea-like dermatitis mouse model was induced using LL37 as previously described [10]. Briefly, the dorsal hair of the mice was shaved 24 h before the experiment, and the mice were randomly divided into three groups: control, model (LL37), and treatment groups (DNAJB2 overexpression). Each group had 3 animals. The rosacea-like models were induced by intradermally injecting 40 μL of LL37 peptide (320 μM) on the backs of mice twice a day for 2 days [10]. For mice in the treatment group, DNAJB2-overexpressing lentivirus vector was injected intradermally daily for three consecutive days, followed by injection with 40 μL of LL37 peptide (320 μM) twice a day for 2 days [25]. For mice in the control group, 40 μL phosphate buffer solution (PBS) was intradermally injected twice a day for 2 days. Rosacea-like skin lesions were photographed using a digital camera, and the severity of rosacea-like skin lesions was evaluated based on the redness score and region. The rosacea-like skin lesions were collected for subsequent analyses.All mice were kept under designated pathogen free (SPF) conditions, and procedures were performed in accordance with the instructions and permission of the Ethics Committee for Animal Experiments of Xi 'an Jiaotong University Health Center (No. XJTUAE2024-1908).

Histological Analysis

The skin tissues of mice were fixed in 4% paraformaldehyde solution, embedded in paraffin wax, and sliced at a thickness of 4 μm. After dewaxing, the slices were stained with hematoxylin and eosin (H&E), and the number of inflammatory cells was observed under an inverted microscope (DXS-3).

For immunofluorescence, the skin tissue was fixed in 4% paraformaldehyde overnight, embedded in paraffin wax, and then prepared into 4 μm slices. After gradient dewaxing, samples were permeabilized with 3% Tritonx-100 (Biotopped, #90002–93-1) for 10 min. After washing with PBS three times, the sample was enclosed in 5% BSA at 25 ℃ for 1 h and then was incubated with anti-CD4 and anti-CD31 antibody at 4 ℃ overnight. Next, the secondary antibody (Affinity, #S0007) was incubated at room temperature for 1 h and washed with PBS three times. The nuclei were stained with DAPI for 5 min and washed with PBS three times. The fluorescent-quenched tablets were sealed and photographed under a fluorescence microscope.

For immunohistochemistry, skin samples were fixed in 4% paraformaldehyde solution, embedded in paraffin, and cut into 4 mm slices. The slices were soaked in citric acid repair solution for 10 min and washed three times with PBS. The sheep serum was sealed for 1 h, and the primary antibody was incubated at 4 °C overnight. Sections were washed with PBS, incubated with the secondary antibody at room temperature, and stained with hematoxylin for 5 min. The images were acquired using a light microscope (Leica). The degree of DNAJB2 expression was based on the semi-quantitative method [26] and measured using the ImageJ software.

ELISA

The skin tissues were prepared into homogenate, and levels of IFN-γ and IL17 were determined using corresponding ELISA kits (Proteintech) according to the instructions.

Cell Culture

Human HaCaT keratinocytes (ATCC) were cultured in DMEM containing 10% FBS and 1% penicillin/streptomycin in a humidified atmosphere of 5% CO2 at 37 °C. HaCaT cells were treated with LL37 (8 μM) for 12 h to construct the cell model of rosacea.

Plasmid and Transfection

DNAJB2-overexpressing plasmid pcDNA3.1-DNAJB2 (DNAJB2) were obtained from RiboBio (Guangzhou, China). HaCaT cells were planted into 6-well plates and transfected with DNAJB2 using Lipofectamine 3000 (Thermo Fisher Scientific, Waltham, MA, USA).

Reverse Transcription qPCR

Total RNA was extracted using TRIzol reagent (Beijing Kangwei Century), and cDNA was synthesized using the Maxima H Minus First Strand cDNA synthesis kit (Summer Fischer Technology, Waltham, Massachusetts, USA). QPCR was performed using the ChamQ Universal SYBR qPCR premix (Vazyme Biotech, Nanjing, China) on an Applied Biosystems 7500 machine (Life Technologies). GAPDH was used as an internal reference to calculate the relative mRNA expression levels of the genes. The primer sequences are listed in Supplementary Table S1.

Western Blot

The expressions of DANJB2, GRP78, p-eIF2 α, eIF2 α, CHOP, ATF4, TLR4, myeloid differentiation factor 88 (MyD88), p-IκBα, IκBα, p65, phospho-p65 (p-p65), and β-actin were detected by western blot. Briefly, skin tissues were lysed in RIPA lysis buffer containing a protease inhibitor, and the protein concentration was quantified using a BCA protein Assay Kit (Biotopped, #Top1003). Proteins (3 μg for each sample) were separated by SDS–polyacrylamide gel electrophoresis (Bio-Rad, USA) and transferred onto methanol-activated polyvinylidene difluoride (PVDF) membranes. The PVDF membranes were then sealed in 5% skimmed milk and incubated with 5% BSA-diluted primary antibodies (1:1000 and 1:10000 for actin) at 4 °C overnight. All the antibodies used are listed in Supplementary Table S2. Next, the membranes were incubated with TBST-diluted horseradish peroxidase-conjugated secondary antibodies at room temperature for 1 h. After washing thrice with TBST, the PVDF membrane was soaked in ECL luminescent reagent, exposed, and photographed using a chemiluminescence apparatus.

Statistical Analysis

Statistical analyses were performed using GraphPad version 7.0 (GraphPad, La Jolla, CA, USA) and SPSS version 18.0 (SPSS, Chicago, IL, USA). All experiments were repeated at least three times, and data are expressed as mean ± standard deviation (SD). A t-test and ANOVA were used for comparisons between two groups and comparisons among three or more groups, respectively. Statistical significance was set at P < 0.05.

Results

DNAJB2 was Selected as a Key Gene Associated with ERS in Rosacea

We retrospectively analyzed the GSE65914 dataset, identifying 1075 differentially expressed genes in rosacea samples compared to controls, including 668 upregulated and 407 downregulated genes (Fig. 1A-B). By comparing these 1075 DEGs with 419 ERS-related genes, we identified 17 overlapping genes (Fig. 1C). Using the STRING database, we found that 13 of the 17 genes interacted with each other (Fig. 1D). IL1B was the gene with the highest degree of interaction in the PPI network. We attempted to select the more valuable characteristic genes from these 13 genes; therefore, a LASSO regression was conducted. Thirteen genes were selected based on the lambda.min values from the LASSO analysis (Fig. 1E-F). These genes included IL1B, CCND1, CXCL8, CCL2, IFIT1, NR1H3, SLC7A5, WARS1, DNAJB2, ERO1A, BID, FBXO6, and NABP1. The diagnostic model constructed using these 13 genes demonstrated excellent performance in distinguishing rosacea from controls, with an area under the curve (AUC) of 1.00 in the training set and 0.75 in the validation set (Fig. 1G-H). The expression of CCND1 and DNAJB2 was significantly decreased in rosacea samples, whereas the expression of the other genes was increased in rosacea samples in the GSE65914 training set (all P < 0.01; Fig. 1I). The expression of DNAJB2 was significantly downregulated in rosacea samples, while the other genes showed no statistical difference between rosacea and control samples in the GSE155141 validation set (Fig. 1J). These discrepancies may be attributed to the limited sample size and individual clinical variations. Nevertheless, in both the training and validation sets, DNAJB2 expression was consistently lower in rosacea samples compared to healthy controls (Fig. 1I-J). Moreover, DNAJB2 exhibited robust performance in distinguishing rosacea, with an AUC of 1.00 in both datasets (Fig. 1K-L).

Fig. 1.

DNAJB2 was identified as a key ERS-related gene and a diagnostic biomarker in rosacea. A-B Volcano plot (A) and heatmap (B) of differentially expressed genes (DEGs) between rosacea patients and healthy controls; C Venn diagram for selecting overlapping genes between DEGs and ERS-related genes; D PPI network for differentially expressed ERS-related genes, in which the red and blue nodes represent upregulated and downregulated genes; E–F LASSO regression for screening characteristic genes in rosacea. Distribution of the LASSO coefficient (E) and likelihood deviation of the LASSO coefficient distribution (F). G-H ROC curves showing the rosacea diagnostic performance of the model established using characteristic genes. I-J Boxplots showing the expression of 13 characteristic genes in rosacea and normal samples in the GSE65914 training set (I) and GSE155141 validation set (J); K-L ROC curves showing the diagnostic performance of DNAJB2 in the GSE65914 training set (K) and GSE155141 validation set (L). *P < 0.05; **P < 0.01; ****P < 0.0001

Changed Biological Functions in Individuals with High Risk of Rosacea

Biological processes such as positive regulation of cytokine production, adaptive immune response, positive regulation of cell adhesion, and cellular ion homeostasis were activated in the high-risk group (Supplementary Fig. 1A). Additionally,, KEGG pathways, such as natural killer cell-mediated cytotoxicity, TLR signaling pathway, and chemokine signaling pathway, were also activated in the high-risk group (Supplementary Fig. 1B). On the other hand, metabolic processes—such as the alditol metabolic process, as well as fatty acid, butanoate, and propanoate metabolism—were suppressed in the high-risk group (Supplementary Fig. 1A-B).

Expression of DNAJB2 in Clinical Rosacea Samples and LL37-Induced Mice Model

DNAJB2 expression was detected in our clinical samples, and results similar to those of bioinformatics analysis were obtained. Both mRNA and protein levels of DNAJB2 were reduced in rosacea skin compared to normal skin (Fig. 2A-C). Compared to the control group, significant erythema and swelling were observed in the LL37-induced mice model, indicating erythematous papules and pustular lesions (Fig. 2D). H&E staining indicated an obvious infiltration of inflammatory cells and vascular dysregulation in the skin lesions of LL37-induced mice model (Fig. 2E). Moreover, DNAJB2 mRNA and protein expression were decreased in the skin tissue of LL37 group compared to that of the control group (Fig. 2F-G).

Fig. 2.

Expression of DNAJB2 in clinical rosacea samples and mice model. A-C expression of DNAJB2 in skin tissue of clinical rosacea samples and healthy controls, including the mRNA expression of DNAJB2 determined by qRT-PCR (A), and representative images (Scale bar: 100 μm) (B) and quantitative statistics (C) of DNAJB2 protein expression determined by immunohistochemical staining; D Phenotypic presentation of mice in control (injected with PBS) and LL37 (injected with LL37) groups, n = 3; E H&E staining of skin tissue from mice in control and LL37 groups, Scale bar: 100 μm, n = 3; F-G DNAJB2 mRNA and protein expression in skin tissue from mice in control and model groups determined by qRT-PCR and western blot, n = 3. *P < 0.05, **P < 0.01

Overexpression of DNAJB2 Relieved Skin Injury, Inflammation, Chemokine and Cytokine Expression in LL37-Induced Mice Model

We successfully induced overexpression of DNAJB2 in the LL37-induced mice model (Fig. 3A). Compared to the LL37 group, overexpression of DNAJB2 significantly relieved the skin lesions, as evidenced by reduced erythema area, erythema score, and decreased inflammatory cell infiltration in H&E staining (Fig. 3B-D). We further assessed the levels of the inflammatory cytokines TNF-α and IL-6 using qRT-PCR. Both TNF-α and IL-6 levels were markedly elevated in the LL37-induced mice model, while this increase was largely reversed after overexpression of DNAJB2 (Fig. 3E-F). In the above Gene Set Enrichment Analysis (GSEA), we observed that the positive regulation of cytokine production and chemokine signaling pathways was activated with a high risk of rosacea. Therefore, we analyzed the expression of chemokines and cytokines,as well as their correlation with DNAJB2 expression. As anticipated, the expression of chemokines and cytokines, including CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9, increased in rosacea samples and showed strong negative correlations with DNAJB2 expression (Fig. 3G-H). Consistently, the expression of these chemokines and cytokines was markedly elevated in the LL37-induced mice model, but it largely decreased after the overexpression of DNAJB2 (Fig. 3I). These findings suggest that the overexpression of DNAJB2 could restrain skin injury, inflammation, and chemokine and cytokine expression in rosacea.

Fig. 3.

DNAJB2 overexpression relieved skin injury, inflammation, and chemokine and cytokine expression in LL37-induced mice model. A protein expression of DNAJB2 in skin tissue from mice in each group determined by western blot; B Phenotypic presentation and H&E staining (Scale bar: 100 μm) of mice in each group; C-D skin injury evaluated according to erythema area (Redness area, C) and erythema score (Redness score, D); E–F mRNA expression of TNF-α and IL-6 in skin tissue from mice in each group determined by qRT-PCR. G Expression of DNAJB2 and chemokines/cytokines (CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9) in rosacea and normal samples in the GSE65914 training set; H scatter diagram showing the correlation between DNAJB2 and chemokines/cytokines (CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9) in the GSE65914 training set; I mRNA expression of chemokines/cytokines (CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9) in skin tissue from mice in each group determined qRT-PCR. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001

Overexpression of DNAJB2 Inhibited Th1- and Th17-Polarized Immune Response in LL37-Induced Mice Model

Immune dysregulation plays a central role in the pathogenesis of rosacea. Thus, we evaluated immune cell infiltration in the skin tissues of rosacea patients using ssGSEA. There appeared to be excessive immune activation in the rosacea samples, manifested by elevated infiltration levels of the majority of immune cell types, such as activated CD4 + T cells, CD8 + T cells, natural killer cells, and T helper cells (including Th1, Th2, and Th17 cells) (Fig. 4A). Using immunofluorescence, we detected the infiltrating CD4 + T cells in the skin tissue of LL37-induced mice model (Fig. 4B). Consistent with the ssGSEA results, the number of infiltrating CD4 + T cells was elevated in the LL37-induced mice model (Fig. 4C). Notably, overexpression of DNAJB2 partially reduced the infiltration of CD4 + T cells (Fig. 4C). We next assessed the correlation between DNAJB2 expression and infiltrating fractions of Th1 and Th17 cells. The expression of DNAJB2 was negatively correlated with the infiltrating levels of Th1 (r = −0.74, P = 2.9e-06) and Th17 cells (r = −0.56, P = 0.00099, Fig. 4D-E). We further examined the expression of genes associated with Th1/Th17 polarization.Including mRNA levels of Th1 cell-related genes (CCR5, STAT1, and IFN-γ) and Th17-related genes (CCR6, STAT3, and IL17) in mouse skin detected by qRT-PCR, as well as the protein levels of IFN-γ and IL17 in skin homogenate analyzed by ELISA. All these genes were elevated in the LL37-induced mice model, while their expression was reduced after overexpressing DNAJB2 (Fig. 4F-I).

Fig. 4.

Overexpression of DNAJB2 inhibited Th1/Th17 immune response in LL37-induced mice model. A Infiltration levels of immune cells in rosacea and control samples analyzed by ssGSEA in the GSE65914 training set; B immunofluorescence staining of CD4 in skin tissue from mice in each group, Scale bar: 50 μm; C Quantification of CD4 + T cells determined by immunofluorescence staining. D-E scatter diagram showing the correlations of DNAJB2 expression with infiltrating levels of Th1 (D) and Th17 cells (E) in the GSE65914 training set; F mRNA expression level of Th1-related genes (STAT1, CCR5, and IFN-γ) in mice skin determined by qRT-PCR; G IFN-γ protein level in skin homogenate determined by ELISA; H mRNA expression level of Th17-related genes (STAT3, CCR6, and IL17) in mice skin determined by qRT-PCR; I IL17 protein level in skin homogenate determined by ELISA. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Overexpression of DNAJB2 Inhibited Skin Angiogenesis in LL37-Induced Mice Model

Angiogenesis plays a crucial role in the development of rosacea [27]. CD31, also known as the platelet endothelial cell adhesion molecule-1 (PECAM-1), is a well-established marker for vascular endothelial cells [28]. Immunofluorescence results revealed an increase in the number of CD31 + cells in the LL37-induced mice model, whereas overexpression of DNAJB2 reduced the number of CD31 + cells to levels similar to those observed in the control group (Fig. 5A-B). Moreover, the LL37-induced mice model showed enhanced VEGF expression, which largely decreased after DNAJB2 overexpression (Fig. 5C). Hence, the overexpression of DNAJB2 could inhibit skin angiogenesis in a rosacea-like model.

Fig. 5.

DNAJB2 overexpression inhibited skin angiogenesis. A Immunofluorescence staining of CD31 in skin tissue from mice in each group, Scale bar: 50 μm; B Quantification of CD31 immunofluorescence staining; C VEGF mRNA expression in skin tissue from mice in each group determined by qRT-PCR. n = 3; *P < 0.05, **P < 0.01

Overexpression of DNAJB2 Inhibited Inflammation in LL37-Induced HaCaT Cells

We used LL37-treated HaCaT cells to investigate the effect of DNAJB2 on rosacea in vitro. qRT-PCR and western blot showed the decreased levels of DNAJB2 mRNA and protein in LL37-treated HaCaT cells, but they were significantly increased after transfection with DNAJB2-overexpressing vector (Supplementary Fig. 2A-B). Moreover, qRT-PCR revealed elevated levels of TNF-α and IL-6 in LL37-treated HaCaT cells, which were downregulated upon DNAJB2 overexpression (Supplementary Fig. 2C-D). Also, the expression levels of chemokines and cytokines, including CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9, were increased LL37-treated HaCaT cells, but these effects were attenuated due to DNAJB2 overexpression (Supplementary Fig. 2E).

Overexpression of DNAJB2 Inhibited ERS and TLR2/Myd88/NF-κB Signaling in LL37-Induced Mice Model

Then, the effect of DNAJB2 on ERS and TLR2/Myd88/NF-κB signaling was determined in LL37-induced mice by western blot. As shown in Fig. 6A-B, the expression of GRP78, p-eIF2α, CHOP, and ATF4 was observably enhanced in LL37 group compared to that in the control group. However, the expression levels of these ERS markers were observably inhibited after overexpression of DNAJB2 compared to that in the LL37 group. Additionally, the expression levels of TLR2, MyD88, p-IκBα, and p-p65 in the LL37 group were observably elevated compared with that in the control group, but these increases were reduced after overexpressing DNAJB2 (Fig. 6C-D).

Fig. 6.

DNAJB2 overexpression inhibited ERS and TLR2/Myd88/NF-κB signaling in LL37-induced mice. A-B Western blot bands (A) and quantification (B) showing the protein expression of ERS marker proteins (GRP78, p-eIF2α, eIF2α, CHOP, and ATF4) in skin tissue from mice in each group. C-D Western blot bands (C) and quantification (D) showing the protein expression of MyD88, p-IκBα, IκBα, p-p65, p65, and TLR2 in skin tissue from mice in each group. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001

Overexpression of DNAJB2 Inhibited ERS and TLR2/Myd88/NF-κB Signaling in LL37-Induced HaCaT Cells

The effect of DNAJB2 on ERS and TLR2/Myd88/NF-κB signaling was determined in LL37-induced HaCaT cells by western blot. The expression levels of ERS marker proteins were increased in HaCaT cells treated with LL37, but were reduced due to the overexpression of DNAJB2 (Fig. 7A-B). Additionally, the expression levels of TLR2, MyD88, p-IκBα, and p-p65 were increased in HaCaT cells treated with LL37, however, they were decreased after the overexpression of DNAJB2 (Fig. 7C-D).

Fig. 7.

DNAJB2 overexpression inhibited ERS and TLR2/Myd88/NF-κB signaling LL37-induced HaCaT cells. A-B Western blot bands (A) and quantification (B) showing the protein expression of ERS marker proteins (GRP78, p-eIF2α, eIF2α, CHOP, and ATF4) in HaCaT cells from each group. C-D Western blot bands (C) and quantification (D) showing the protein expression of MyD88, p-IκBα, IκBα, p-p65, p65, and TLR2 in HaCaT cells from each group. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001

Overexpression of DNAJB2 Inhibited TLR2/Myd88/NF-κB Signaling by Suppressing ERS in LL37-Induced HaCaT Cells

ERS has been demonstrated as an activator to TLR2 signaling in several inflammatory disease [29]. Thus, we wondered whether DNAJB2 inhibited TLR2/Myd88/NF-κB signaling through affecting ERS in rosacea. An ERS agonist (Tunicamycin) was employed to investigate whether the effect of DNAJB2 on TLR2/Myd88/NF-κB signaling was mediated by ERS in LL37-induced HaCaT cells. As demonstrated by western blot, the inhibitory effects of DNAJB2 on ERS marker proteins were reversed by Tunicamycin in LL37-induced HaCaT cells (Fig. 8A-B). Moreover, the inhibitory effects of DNAJB2 on TLR2, MyD88, p-IκBα, and p-p65 were reversed by Tunicamycin in LL37-induced HaCaT cells (Fig. 8C-D). These data suggest that DNAJB2 suppressed ERS to inactivate the TLR2/Myd88/NF-κB signaling, thereby alleviating skin inflammation and angiogenesis in rosacea (Fig. 8E).

Fig. 8.

Overexpression of DNAJB2 inhibited TLR2/Myd88/NF-κB signaling by suppressing ERS in LL37-induced HaCaT cells. HaCaT cells were assigned into 4 groups: Control, LL37, LL37 + DNAJB2, and LL37 + DNAJB2 + Tunicamycin. A-B Western blot bands (A) and quantification (B) showing the protein expression of ERS marker proteins (GRP78, p-eIF2α, eIF2α, CHOP, and ATF4) in cells from different groups. C-D Western blot bands (C) and quantification (D) showing the protein expression of TLR2, MyD88, p-IκBα, IκBα, p-p65, and p65 in HaCaT cells from different groups. E A schematic diagram illustrating the involvement of DNAJB2/ERS/TLR2/Myd88/NF-κB in rosacea. n = 3; *P < 0.05, **P < 0.01, ***P < 0.001

Overexpression of DNAJB2 Inhibited Inflammation in LL37-Induced HaCaT Cells by Suppressing ERS

qRT-PCR showed that the inhibitory effects of DNAJB2 on TNF-α and IL-6 were abated by Tunicamycin in LL37-induced HaCaT cells (Supplementary Fig. 3A-B). Also, the inhibitory effects of DNAJB2 on CCL5, CCL20, CXCL10, CXCL11, CXCL12, and MMP9 were attenuated by Tunicamycin in LL37-induced HaCaT cells (Supplementary Fig. 3C).

Discussion

ERS may play a central role in the pathogenesis and progression of rosacea [23]. Various clinical triggers of rosacea that activate TLR2-cathelicidin and reactive oxygen species signaling may act as ER stressors at the molecular level [23]. For instance, skin lipid changes may disrupt the epidermal barrier function, leading to ERS; ultraviolet radiation can promote ceramide release of S1P to induce ERS; and heat, skin irritants, capsaicin, and resveratrol can activate transient receptor potential vanillic acid receptors (TRPV) channels and ERS signals [23]. ERS could activate the ATF4/TLR2/NF-κB pathway to mediate the signal transduction of downstream pro-inflammatory factors and angiogenic factors, thus contributing to the pathogenesis of rosacea [30–32]. Hence, the ERS pathway may represent a new research direction for rosacea therapy.

Here, 13 key ERS-related genes were identified in rosacea using differential analysis, PPI network analysis, and the LASSO algorithm. Rosacea risk was quantized, and GSEA indicated that an adaptive immune response, positive regulation of cytokine production, natural killer cell-mediated cytotoxicity, TLR signaling pathway, and chemokine signaling pathway were activated with a high risk of rosacea. This suggests that immune response and inflammation are related to the disease. DNAJB2 was identified as a key gene underlying the mechanisms of ERS in rosacea because it showed significantly decreased expression in rosacea, regardless of the sample size and individual sample differences.

DNAJ, also known as Heat Shock Protein 40 (HSP40), is a molecular chaperone of the Hsp70 protein that guides the conformation of the protein throughout its life cycle [33, 34]. It enhances the versatility of the Hsp70 mechanism, such as assisting in protein folding during ribosome synthesis, driving protein transport across membranes, and regulating protein–protein interactions by controlling conformational changes [34]. DNAJB2, also known as heat shock protein J1 (HSJ1), encodes a molecular chaperone member of the Hsp40 family and plays crucial roles in multiple diseases [35, 36]. The C-terminal region of DNAJB2 contains two ubiquitin-interacting motifs that mediate binding to polyubiquitin proteins and proteasomes, and promote the degradation of target proteins such as misfolded proteins through the ubiquitin–proteasome system [37]. Considering the function of DNAJB2 in protein processing and previous studies suggesting that rosacea and ERS are linked to the accumulation of misfolded proteins, we can reasonably hypothesize that the DNAJB2 may be associated with rosacea. Here, the ROC curves showed that DNAJB2 had a good ability for disease diagnosis, with an AUC of 1 in both sets. However, the precise role of DNAJB2 in rosacea has yet to be fully explored.

The pathogenesis of rosacea involves abnormalities in both innate and adaptive immunity. Abnormal congenital immune activation enhances the expression of keratinocyte-derived TLR2 and protease-activated receptor 2 (PAR2), which increases the levels of cathelicidin antimicrobial peptides (CAMP). Whereafter, CAMP is hydrolyzed by Recombinant Kallikrein 5 (KLK5) into a bioactive LL37 fragment, which can promote the increase of inflammatory factors after being released from the epidermis [38]. These factors facilitate the activity of mast cells, neutrophils, and macrophages, promote chemokine production [39], and recruit CD4 + T and B cells to accumulate in the skin [13]. This can cause skin erythema, telangiectasia, and inflammation. TLR2 can also promote the activation of NLRP1 inflammatory bodies [40], leading to pustule formation, pain,vascular response, and the release of prostaglandin E2 [41]. PAR2 activation can lead to inflammation, itching, and pain, accompanied by T lymphocyte and neutrophil recruitment, mast cell degranulation, and the further release of inflammatory cytokines and chemokines [42]. In addition, myeloid differentiation factor 88 (MyD88), downstream of TLR2, is a cytoplasmic soluble protein that plays a crucial signal transduction role in IL1 and Toll-like receptor signal transduction pathways [43]. As an important pathway affecting inflammatory factors of keratinocytes, the TLR2/Myd88/NF-κB pathway mediates the occurrence and development of inflammatory response in rosacea. When the adaptive immune system is activated, there is a cascade of further expansion of Th1 and Th17 cells and their related immune mediators [15]. The combined dysfunction of both innate and adaptive immunity underpins the pathogenesis of rosacea.

The expression of DNAJB2 was observably reduced in rosacea skin compared to normal skin, and DNAJB2 overexpression relieved skin injury in LL37-induced mice model. Our experiments suggest that DNAJB2 overexpression could inhibit expression of inflammatory cytokines and chemokines. Moreover, DNAJB2 overexpression inhibited CD4 + T cell infiltration. Th1/Th17-polarized immune cells are predominant among the T-cell responses in rosacea [14]. DNAJB2 expression was negatively correlated with the infiltration levels of Th1 and Th17 cells, and overexpression of DNAJB2 inhibited the expression of Th1- and Th17-related genes. Furthermore, DNAJB2 overexpression inhibited skin angiogenesis, as evidenced by reduced levels of CD31 (a vascular endothelial marker) [28] and VEGF. The expression of ERS marker proteins (GRP78, p-eIF2α/eIF2α, CHOP, ATF4) and TLR2, Myd88, p-IκBα, and p-p65 in LL37-induced rosacea skin was significantly higher in this study, confirming the abnormal activation of ERS classical proteins and the TLR2/Myd88/NF-κB pathway. Overexpression of DNAJB2 significantly reduced the activation of this pathway in an LL37-induced rosacea mice model.

We speculated that the overexpression of DNAJB2 can reduce the inflammation and angiogenesis of rosacea by inhibiting the TLR2/Myd88/NF-κB pathway mediated by ERS, thus alleviating the occurrence and development of rosacea. Therefore, DNAJB2 may be a key gene leading to the activation of adaptive/innate immunity in ERS, and DNAJB2 may become a new target for the treatment of rosacea. However, this should be further confirmed based on clinical data.

Conclusion

Our bioinformatics analysis of ERS-related genes in rosacea identified DNAJB2 as a key candidate. DNAJB2 is downregulated in rosacea, and its overexpression alleviates rosacea skin injury by inhibiting inflammation, angiogenesis, and excessive immune response in skin tissue. The ERS-mediated TLR2/Myd88/NF-κB pathway may be an underlying mechanism. This study demonstrated that DNAJB2 may be a new target for the treatment of rosacea.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Jiawen Wu and Hong Sun conceived the experiments; Yuxin Qing conducted the bioinformatics analysis, Zining Xu, Yuxin Qing, and Bingyang Xu performed the experiments; Yuxin Qing, Jiawen Wu, Bingyang Xu, Zining Xu, Shuhong Ye ,Yuanqin Wang, Bin Zhao and Na Wu analysed the results and wrote the article. All authors reviewed the article.

Funding

Natural Science Basic Research Program of Shaanxi Province.(Project Number: 2024JC-YBMS-632).

Data Availability

The data that support the findings of this study are available from the Xi'an Jiaotong University but restrictions apply to the availability of these data, which were used under license for this study and therefore are not publicly available. Data are however available from the corresponding author upon reasonable request and with permission of the Xi'an Jiaotong University.

Declarations

Ethical Approval

Reviewed and approved by Medical Ethics Committee of the Second Affiliated Hospital of Xi 'an Jiaotong University and met the requirements of the Helsinki Declaration.(Approval ID:2024 Ethics Review 019); Ethics Committee for Animal Experiments of Xi 'an Jiaotong University Health Center (Approval ID: XJTUAE2024-1908).

Consent to Participate and Publish

Informed consent was obtained from all individual participants included in the study.

Consent to Publish

The written informed consent of the individual and the legal guardian/next of kin of the minor has been obtained for the publication of any potentially identifiable images or data in this article.

Competing Interests

The authors declare no competing interests.