Efficacy of Glycyrrhetinic Acid in the Treatment of Acne Vulgaris Based on Network Pharmacology and Experimental Validation

Glycyrrhetinic acid (GA) is a saponin compound, isolated from licorice (Glycyrrhiza glabra), which has been wildly explored for its intriguing pharmacological and medicinal effects. GA is a triterpenoid glycoside displaying an array of pharmacological and biological activities, including anti-inflammatory, anti-bacterial, antiviral and antioxidative properties. In this study, we investigated the underlying mechanisms of GA on acne vulgaris through network pharmacology and proteomics. After the intersection of the 154 drug targets and 581 disease targets, 37 therapeutic targets for GA against acne were obtained. A protein–protein interaction (PPI) network analysis highlighted TNF, IL1B, IL6, ESR1, PPARG, NFKB1, STAT3 and TLR4 as key targets of GA against acne, which is further verified by molecular docking. The experimental results showed that GA inhibited lipid synthesis in vitro and in vivo, improved the histopathological damage of skin, prevented mast cell infiltration and decreased the level of pro-inflammatory cytokines, including TNF-α, IL-1β and IL-6. This study indicates that GA may regulate multiple pathways to improve acne symptoms, and the beneficial effects of GA against acne vulgaris might be through the regulation of sebogenesis and inflammatory responses.


Introduction
Acne vulgaris is a persistent, cutaneous and inflammatory disorder of pilosebaceous follicles, which is characterized by clinical manifestations of papules, pustules or nodules, seborrhea, inflammatory lesions and various degrees of scarring due to cyst formation [1].The epidemiology of acne vulgaris has revealed the prevalence in teenagers is more than 60%, which is the highest prevalence of all ages [2].Although it remains unknown, the main acne pathogenesis has been identified to be related to an abundance of Propionibacterium acnes (P.acnes), high androgen levels, excessive secretion of sebum and hair follicle hyperkeratosis.The interaction of these pathogenic factors results in the formation of comedones and the development of inflammation [3].Androgen excess can cause acne by triggering sebaceous glands (SGs) in the skin to overproduce sebum, which clogs pores and hair follicles, influences inflammation of the skin and feeds P. acnes, thereby resulting in acne vulgaris [4].SGs are composed of sebocytes, which produce different pro-inflammatory cytokines (TNF-α, IL-1β, IL-6, IL-8), accumulate lipids and release the

Network Pharmacology Analysis of GA against Acne Vulgaris
We first identified 37 targets through the intersection potential between pharmacological targets of GA and acne vulgaris-related targets (Figure 1B).As shown in Figure 1C, the PPI network was constructed by the STRING platform to show 37 nodes and 356 interactions, with an average node degree of 17.The results of the PPI were imported into Cytoscape to show the top 10 hub targets with the highest degree.The results suggest that tumor necrosis factor (TNF), interleukin-1β (IL-1β), interleukin-6 (IL-6), Estrogen Receptor alpha (ESR1), peroxisome proliferator-activated receptor gamma (PPAR-γ), Nuclear factor kappa B subunit 1 (NFκB1), signal transducer and activator of transcription 3 (STAT3) and Toll-like receptor 4 (TLR4) may be the most critical targets of Glycyrrhetinic acid against acne vulgaris (Figure 1D).
pally implicated in the regulation of interleukin-8 production, inflammatory response and cellular response to lipopolysaccharide.Steroid binding and estrogen response element binding were highlighted as the most important MFs, while extracellular space, chromatin, receptor complexes and extracellular region were identified as the most enriched terms for CC.Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis identified 72 pathways, and the top 30 pathways were visualized in Figure 1F.Surprisingly, the intersecting targets were significantly associated with steroid hormone biosynthesis and adipocytokine and TNF and IL-17 signaling pathways, suggesting that GA may have a regulatory effect on sebaceous lipogenesis and inflammatory response.

Molecular Docking
PPAR-γ, TNF, IL-1β, IL-6, TLR and NFκB1 were identified as potential core targets of GA in the regulation of acne vulgaris.The AutoDock docking tool was used for molecular docking between GA and six potential core targets.The binding results are shown in Figure 2 and Table 1.All the binding energies are negative, indicating that GA Next, the Gene Ontology (GO) analysis enriched 199 results in biological processes (BPs), 17 results in cellular component (CC) and 39 results in molecular function (MF) (p < 0.05), and the first 10 items were visualized (Figure 1E).Interestingly, BPs are principally implicated in the regulation of interleukin-8 production, inflammatory response and cellular response to lipopolysaccharide.Steroid binding and estrogen response element binding were highlighted as the most important MFs, while extracellular space, chromatin, receptor complexes and extracellular region were identified as the most enriched terms for CC.Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis identified 72 pathways, and the top 30 pathways were visualized in Figure 1F.Surprisingly, the intersecting targets were significantly associated with steroid hormone biosynthesis and adipocytokine and TNF and IL-17 signaling pathways, suggesting that GA may have a regulatory effect on sebaceous lipogenesis and inflammatory response.

Molecular Docking
PPAR-γ, TNF, IL-1β, IL-6, TLR and NFκB1 were identified as potential core targets of GA in the regulation of acne vulgaris.The AutoDock docking tool was used for molecular docking between GA and six potential core targets.The binding results are shown in Figure 2 and Table 1.All the binding energies are negative, indicating that GA and protein can bind spontaneously, and the lower the binding energy, the easier GA will bind to the protein.PPAR-γ may be the core target of GA against acne, with the strongest binding effect, because the binding energy of PPAR-γ to GA was −11.0 kcal/mol.and protein can bind spontaneously, and the lower the binding energy, the easier GA will bind to the protein.PPAR-γ may be the core target of GA against acne, with the strongest binding effect, because the binding energy of PPAR-γ to GA was −11.0 kcal/mol.

GA Inhibits Dex-Induced Lipogenesis and Inflammation in Human SZ95 Sebocytes
The cell viability of SZ95 treated with different concentrations of GA (5, 10, 50, 100, 500, 1000 µg/mL) was first determined by MTT compared with salicylic acid (SA), an anti-inflammatory and keratolytic agent [20], which is used to remove follicular clogs and to reduce comedones [21].As shown in Figure 3A,B, the GA and SA had no effect on cell viability in SZ95 cells in a concentration ranging from 5 to 1000 µg/mL.
Then, human SZ95 sebocytes were co-treated with Dex or DMSO and GA or SA for 48 h to determine the potential regulatory effects of GA on sebaceous lipids and inflammatory response.The evidence showed that dexamethasone (Dex) induced sebaceous lipid production, which could be regulated through the expression of the transcription factor sterol response element-binding protein 1 (SREBP-1) [22].As shown in Figure 3C-H, Dex induction not only increased the levels of triglyceride (TG), SREBP-1 and PPAR-γ involved in fatty acid storage and glucose metabolism but also up-regulated the proinflammatory cytokine expression (TNF-α, IL-1β and IL-6).However, these changes were improved with GA treatment in a concentration-dependent manner.Moreover, GA had a stronger inhibition of the sebaceous lipogenesis and inflammatory response than SA at the same dose of 500 µg/mL in Dex-induced SZ95 cells.

GA Inhibits Dex-Induced Lipogenesis and Inflammation in Human SZ95 Sebocytes
The cell viability of SZ95 treated with different concentrations of GA (5, 10, 50, 100, 500, 1000 µg/mL) was first determined by MTT compared with salicylic acid (SA), an anti-inflammatory and keratolytic agent [20], which is used to remove follicular clogs and to reduce comedones [21].As shown in Figure 3A,B, the GA and SA had no effect on cell viability in SZ95 cells in a concentration ranging from 5 to 1000 µg/mL.
Then, human SZ95 sebocytes were co-treated with Dex or DMSO and GA or SA for 48 h to determine the potential regulatory effects of GA on sebaceous lipids and inflammatory response.The evidence showed that dexamethasone (Dex) induced sebaceous lipid production, which could be regulated through the expression of the transcription factor sterol response element-binding protein 1 (SREBP-1) [22].As shown in Figure 3C-H, Dex induction not only increased the levels of triglyceride (TG), SREBP-1 and PPAR-γ involved in fatty acid storage and glucose metabolism but also up-regulated the pro-inflammatory cytokine expression (TNF-α, IL-1β and IL-6).However, these changes were improved with GA treatment in a concentration-dependent manner.Moreover, GA had a stronger inhibition of the sebaceous lipogenesis and inflammatory response than SA at the same dose of 500 µg/mL in Dex-induced SZ95 cells.

Effect of GA on a Mouse Model of Acne Vulgaris
To explore the anti-lipid and anti-inflammatory effects of GA compared to those produced by SA, BALB/c mice, treated with oleic acid (OA) and then injected with P. acnes, were used as an in vivo model for acne vulgaris.We noted that the acne injection initiated an obvious inflammatory response in the ear skin of the model group (MC group).As shown in Figure 4A, ear edema, erythema and rough skin surface symptoms were clearly observed in the MC group, while these symptoms were clearly ameliorated in the GA-treated group.Then, a histopathological examination of the acne-like mouse ear skin lesions was also performed.As shown in Figure 4B,E, we found an increase in the epithelial cells within the funnel-shaped part of the hair follicle structure and a thickened layer of the stratified squamous epithelium in the acne-like mouse ear skin lesions.Indeed, the thickness of the epidermal layer of the acne-like mouse ear skin lesions increased to 75.56 ± 2.35 µm, compared to that in the normal mice (42.23 ± 1.13 mm µm), whereas GA treatment decreased the P. acnes-induced epidermal thickness to 60.03 ± 1.89 (L), 53.53 ± 2.06 (M) and 47.51 ± 1.12 (H) µm, respectively.SA treatment reduced the epidermal thickness to 60.73 ± 0.89 µm at a dose similar to the high dose of GA.Meanwhile, the H&E staining results revealed that the P. acnes injection increased the number of infiltrated inflammatory cells, which is further confirmed by the staining with toluidine blue.As shown in Figure 4C,D, mast cells were markedly enhanced in the dermis layer of the ear skin of the P. acnes-induced BALB/c mice compared with those in the normal mice (BC group).On the contrary, GA treatment significantly suppressed the number of mast cells.In addition, the weight of the P. acnes-induced ear skin lesions on the BALB/c mice, which can indicate the signs of skin edema, was measured.As shown in Figure 4F, the P. acnes-induced ear skin lesions resulted in ear edema, compared with that in the BC group.On the contrary, the ear swelling was obviously ameliorated after treatment with GA, which showed a better effect than SA in inhibitory activity.These

Effect of GA on a Mouse Model of Acne Vulgaris
To explore the anti-lipid and anti-inflammatory effects of GA compared to those produced by SA, BALB/c mice, treated with oleic acid (OA) and then injected with P. acnes, were used as an in vivo model for acne vulgaris.We noted that the acne injection initiated an obvious inflammatory response in the ear skin of the model group (MC group).As shown in Figure 4A, ear edema, erythema and rough skin surface symptoms were clearly observed in the MC group, while these symptoms were clearly ameliorated in the GA-treated group.Then, a histopathological examination of the acne-like mouse ear skin lesions was also performed.As shown in Figure 4B,E, we found an increase in the epithelial cells within the funnel-shaped part of the hair follicle structure and a thickened layer of the stratified squamous epithelium in the acne-like mouse ear skin lesions.Indeed, the thickness of the epidermal layer of the acne-like mouse ear skin lesions increased to 75.56 ± 2.35 µm, compared to that in the normal mice (42.23 ± 1.13 mm µm), whereas GA treatment decreased the P. acnes-induced epidermal thickness to 60.03 ± 1.89 (L), 53.53 ± 2.06 (M) and 47.51 ± 1.12 (H) µm, respectively.SA treatment reduced the epidermal thickness to 60.73 ± 0.89 µm at a dose similar to the high dose of GA.Meanwhile, the H&E staining results revealed that the P. acnes injection increased the number of infiltrated inflammatory cells, which is further confirmed by the staining with toluidine blue.As shown in Figure 4C,D, mast cells were markedly enhanced in the dermis layer of the ear skin of the P. acnes-induced BALB/c mice compared with those in the normal mice (BC group).On the contrary, GA treatment significantly suppressed the number of mast cells.In addition, the weight of the P. acnes-induced ear skin lesions on the BALB/c mice, which can indicate the signs of skin edema, was measured.As shown in Figure 4F, the P. acnes-induced ear skin lesions resulted in ear edema, compared with that in the BC group.On the contrary, the ear swelling was obviously ameliorated after treatment with GA, which showed a better effect than SA in inhibitory activity.These data showed that treatment with GA could efficiently ameliorate the acne symptoms including epidermal thickening and swelling of the skin, as well as inflammatory infiltration in P. acnes-induced BALB/c mice, which is possibly attributed to the recovery of skin damage caused by P. acnes.Moreover, it is worth noting that the SGs were enlarged and the skin's hair follicles became blocked (Figure 4B,G) due to the hyperkeratinization of the hair follicles in P. acnes-induced BALB/c mice, compared with those in the normal mice.However, these alterations were ameliorated with the treatment of GA, especially for the high dose, displaying GA-treated skin tissues similar to those of the BC group in terms of the structure of the skin's epidermal layer and the sebaceous glands, as well as the epidermal thickness of the entire ear.
Molecules 2024, 29, x FOR PEER REVIEW 6 data showed that treatment with GA could efficiently ameliorate the acne symptom cluding epidermal thickening and swelling of the skin, as well as inflammatory infi tion in P. acnes-induced BALB/c mice, which is possibly attributed to the recovery of damage caused by P. acnes.Moreover, it is worth noting that the SGs were enlarged the skin's hair follicles became blocked (Figure 4B,G) due to the hyperkeratinizatio the hair follicles in P. acnes-induced BALB/c mice, compared with those in the no mice.However, these alterations were ameliorated with the treatment of GA, espec for the high dose, displaying GA-treated skin tissues similar to those of the BC grou terms of the structure of the skin's epidermal layer and the sebaceous glands, as we the epidermal thickness of the entire ear.

GA Suppresses Sebogenesis and Pro-Inflammatory Cytokines in a Mouse Model of Acne Vulgaris
To uncover the possible molecular mechanisms of the anti-acne actions of GA, we first examined the role of GA on sebogenesis in a mouse model of acne vulgaris.The sebogenesis-related factors, including sterol regulatory element-binding protein 1 (SREBP-1), 11βhydroxysteroid dehydrogenase (11β-HSD), peroxisome proliferator-activated receptor (PPAR)-γ and glucocorticoid receptor (GR), also named as NR3C1 (nuclear receptor subfamily 3, group C, member 1), in the P. acnes-induced acne-like mouse ear skin lesions were determined by immunohistochemical staining (Figure 5A,C) and ELISA assays (Figure 5E-H).We noted that the expression of PPAR-γ, SREBP-1, 11βHSD and NR3C1 in the acne-like mouse ear skin lesions were obviously higher than those observed in the normal mice, whereas GA significantly down-regulated the expression of these proteins.
expression of these proteins.
Then, the proliferation of sebaceous gland cells and lipid droplet proteins such as perilipin (PLIN) 2 were also explored through immunohistochemical staining of the P. acnes-induced mouse ear skin lesions (Figure 5B,J).As shown in Figure 5I,J, the results demonstrated that the expression of PLIN 2 and PCNA was obviously higher than that noticed in the normal mice, whereas GA treatment decreased the expression levels of these proteins in a concentration-dependent manner.
To investigate the anti-inflammatory effects of GA treatment in P. acnes-induced mouse ear skin, we analyzed pro-inflammatory cytokines such as TNFα, IL-1β and IL-6, which are well known to be up-regulated in acne skin lesions [23,24].As shown in Figure 5K-M, a significant increase in inflammatory cytokine (IL-6, TNFα and IL-1β) release was observed in the ear skin lesions of the BALB/c mice induced by P. acnes (p < 0.001), whereas topical treatment of GA showed a concentration-dependent downregulation of cytokines and stronger inhibitory effects on cytokines than SA.Then, the proliferation of sebaceous gland cells and lipid droplet proteins such as perilipin (PLIN) 2 were also explored through immunohistochemical staining of the P. acnes-induced mouse ear skin lesions (Figure 5B,J).As shown in Figure 5I,J, the results demonstrated that the expression of PLIN 2 and PCNA was obviously higher than that noticed in the normal mice, whereas GA treatment decreased the expression levels of these proteins in a concentration-dependent manner.
To investigate the anti-inflammatory effects of GA treatment in P. acnes-induced mouse ear skin, we analyzed pro-inflammatory cytokines such as TNFα, IL-1β and IL-6, which are well known to be up-regulated in acne skin lesions [23,24].As shown in Figure 5K-M, a significant increase in inflammatory cytokine (IL-6, TNFα and IL-1β) release was observed in the ear skin lesions of the BALB/c mice induced by P. acnes (p < 0.001), whereas topical treatment of GA showed a concentration-dependent downregulation of cytokines and stronger inhibitory effects on cytokines than SA.

GO and KEGG Pathway Analysis
GO enrichment analysis of DEGs in GA/BC groups (Figure 7A) and BC/BC groups (Figure S1A,B) was first conducted against the GO database by using three independent ontologies, including molecular functions, cellular component categories and biological processes.The GO analysis demonstrated that the major enriched cellular components in the GA group were centriolar satellite, microtubule organizing, center nuclear envelope, chromatin and centrosome.DEPs in the GA group were associated with the regulation of molecular functions, such as metal ion binding, ATP binding, type 1 cannabinoid receptor binding, purine nucleotide binding and pyrimidine nucleotide binding.These DEPs participated in the biological processes of the immune response, tarsal gland development, lagging strand elongation, leading strand elongation and regulation of lipid kinase activity.It is worth noting that there was an obvious difference in GO enrichment among BC, MC and GA groups, which was associated with acne pathology.Based on KEGG pathway enrichment analysis, DEPs in the GA group were mainly the PPAR, AMPK and Wnt signaling pathways (Figure 7B).

GO and KEGG Pathway Analysis
GO enrichment analysis of DEGs in GA/BC groups (Figure 7A) and BC/BC groups (Figure S1A,B) was first conducted against the GO database by using three independent ontologies, including molecular functions, cellular component categories and biological processes.The GO analysis demonstrated that the major enriched cellular components in the GA group were centriolar satellite, microtubule organizing, center nuclear envelope, chromatin and centrosome.DEPs in the GA group were associated with the regulation of molecular functions, such as metal ion binding, ATP binding, type 1 cannabinoid receptor binding, purine nucleotide binding and pyrimidine nucleotide binding.These DEPs participated in the biological processes of the immune response, tarsal gland development, lagging strand elongation, leading strand elongation and regulation of lipid kinase activity.It is worth noting that there was an obvious difference in GO enrichment among BC, MC and GA groups, which was associated with acne pathology.Based on KEGG pathway enrichment analysis, DEPs in the GA group were mainly the PPAR, AMPK and Wnt signaling pathways (Figure 7B).

Validation of Proteins Identified by Proteomics Analyses
Through the analysis of the above results, the proteins identified by proteomics analyses (Apoa2, SCD-1, CYP7A1, Cyk and TRAF-6), which play an important role in acne pathology and are altered in the intervention by GA treatment (Figure 8F), were further analyzed by ELISA.Gyk is a key enzyme in fatty acid esterification, and the activation of the PPAR pathway triggers an increase in Gyk expression, further promoting triglyceride synthesis.Apoa2, SCD-1, CYP7A1 and p-akt are downstream factors of the PPAR pathway.TRAF6 is a key mediator of NF-κB signaling pathways, thereby regulating downstream gene expression.As shown in Figure 8A-E, Apoa 2, SCD-1, CYP7A1, Gyk, and P-akt were significantly up-regulated in the MC group compared to the BC group (p < 0.05).Conversely, SCD-1, CYP7A1, Cyk and P-akt were significantly downregulated in the GA group compared to the MC group.However, the expression of Apoa 2 in the GA group was not statistically different from that of the MC group.

Validation of Proteins Identified by Proteomics Analyses
Through the analysis of the above results, the proteins identified by proteomics analyses (Apoa2, SCD-1, CYP7A1, Cyk and TRAF-6), which play an important role in acne pathology and are altered in the intervention by GA treatment (Figure 8F), were further analyzed by ELISA.Gyk is a key enzyme in fatty acid esterification, and the activation of the PPAR pathway triggers an increase in Gyk expression, further promoting triglyceride synthesis.Apoa2, SCD-1, CYP7A1 and p-akt are downstream factors of the PPAR pathway.TRAF6 is a key mediator of NF-κB signaling pathways, thereby regulating downstream gene expression.As shown in Figure 8A-E, Apoa 2, SCD-1, CYP7A1, Gyk, and P-akt were significantly up-regulated in the MC group compared to the BC group (p < 0.05).Conversely, SCD-1, CYP7A1, Cyk and P-akt were significantly down-regulated in the GA group compared to the MC group.However, the expression of Apoa 2 in the GA group was not statistically different from that of the MC group.

Discussion
Increasing evidence has emerged that inflammation manifests throughout the development of acne lesions [25].P. acnes is known to be an important pathogenic factor for the development of acne.Colonization of the pilosebaceous follicle by P. acnes stimulates the expression of IL-1β, TNF-α, IL-6 and IL-8 through activating Toll-like receptors (TLRs) and induces the expression of nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) in human sebocytes, thereby amplifying the inflammatory cascade and deteriorating acne vulgaris [26,27].In this study, the network pharmacological and proteomic analyses were used to explore the mechanism of GA in treating acne vulgaris.A total of 37 potential targets were predicted by network pharmacology, among which TNF, IL-1B, IL-6, ESR1, PPARG, NFκB1, STAT3, and TLR4 may be the key targets of GA in the treatment of acne.Molecular docking showed that GA exhibited high affinities with TNF, IL-1β, IL-6, PPAR-γ, NFκB1, and TLR4.IL-1β is a potent inducer of pro-inflammatory cytokine production such as IL-6 and IL-8 in sebocytes (the sebaceous gland cells), which accumulate lipids and lipid droplets (LDs) and release sebum.IL-1β mRNA and IL-1β are abundant in inflammatory acne lesions [28].IL-6 is one of the major acne lesionassociated cytokines and is predominant in the inflammatory lesions of acne vulgaris [29].TNF-α is closely involved in mediating inflammatory and innate immune responses.The experimental results showed that there is an obvious up-regulation of TNF-α, IL-1β and IL-6 in Dex-induced SZ95 sebocytes and in the lesional skin of BALB/c mice in-

Discussion
Increasing evidence has emerged that inflammation manifests throughout the development of acne lesions [25].P. acnes is known to be an important pathogenic factor for the development of acne.Colonization of the pilosebaceous follicle by P. acnes stimulates the expression of IL-1β, TNF-α, IL-6 and IL-8 through activating Toll-like receptors (TLRs) and induces the expression of nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2) in human sebocytes, thereby amplifying the inflammatory cascade and deteriorating acne vulgaris [26,27].In this study, the network pharmacological and proteomic analyses were used to explore the mechanism of GA in treating acne vulgaris.A total of 37 potential targets were predicted by network pharmacology, among which TNF, IL-1B, IL-6, ESR1, PPARG, NFκB1, STAT3, and TLR4 may be the key targets of GA in the treatment of acne.Molecular docking showed that GA exhibited high affinities with TNF, IL-1β, IL-6, PPAR-γ, NFκB1, and TLR4.IL-1β is a potent inducer of pro-inflammatory cytokine production such as IL-6 and IL-8 in sebocytes (the sebaceous gland cells), which accumulate lipids and lipid droplets (LDs) and release sebum.IL-1β mRNA and IL-1β are abundant in inflammatory acne lesions [28].IL-6 is one of the major acne lesion-associated cytokines and is predominant in the inflammatory lesions of acne vulgaris [29].TNF-α is closely involved in mediating inflammatory and innate immune responses.The experimental results showed that there is an obvious up-regulation of TNF-α, IL-1β and IL-6 in Dex-induced SZ95 sebocytes and in the lesional skin of BALB/c mice induced by P. acnes, while GA obviously prevented the secretion of IL-1β, IL-6 and TNF-α in vitro et in vivo.
Proliferating cell nuclear antigen (PCNA) was originally characterized as an antigen expressed in the cell nucleus during the DNA synthesis in the S-phase of the cell cycle.The PCNA was immunohistochemically stained in this study to investigate the proliferation of sebaceous gland cells.We noted that PCNA-positive cells significantly increased in P. acnesinduced mouse ear skin lesions, whereas GA treatment could obviously reduce the expression of PCNA in a dose-dependent way, suggesting that GA had an inhibitory activity on the proliferation of sebaceous gland cells.Indeed, the sebaceous enlargement was observed in P. acnes-induced acne-like ear skin lesions.However, we found that the size of SGs obviously decreased after GA treatment.LDs, also known as lipid bodies, are lipid-rich cellular organelles, which modulate the storage, transportation and metabolism of lipids and play a crucial role in the lipid homeostasis system.In adipocytes and sebocytes, the lipid droplet proteins, such as the perilipin (PLIN) family proteins segregate lipids by protecting lipid droplets from lipase action, dominate the levels of comedogenic free fatty acids and play important roles in lipid metabolism and lipid-related pathologies, including acne vulgaris.PLIN2, as a major member of the PLIN family proteins in sebocytes, is implicated in the accumulation of sebaceous lipids and the regulation of SG size.The experimental results demonstrated that PLIN2 was more intensely expressed in the acne-like ear skin lesions.However, GA treatment obviously reduced the expression level of PLIN2 in the P. acnesinduced acne-like ear skin lesions in a concentration-dependent manner.
SREBP-1 is a major transcriptional factor, primarily responsible for the metabolism of fatty acids, which plays a key role in lipid synthesis [30,31].PPAR-γ is recognized for its role in lipid metabolism and the storage of fatty acids, which stimulates adipogenesis and lipid uptake [32].NR3C1 binds to cortisol and other glucocorticoids and increases the expression levels of anti-inflammatory mediators in the nucleus or suppresses the expression of pro-inflammatory mediators in the cytosol.11βHSD catalyzes the conversion of inactive cortisone into active cortisone (glucocorticoid), which can be mediated by GR to stimulate sebocyte proliferation and promote lipid secretion from sebaceous glands [32].Indeed, in this study, the experimental results revealed that the sebogenesis-related factors such as SREBP-1, PPAR-γ, NR3C1 and 11βHSD obviously increased, resulting in the lipid accumulation in P. acnes-induced acne-like ear skin lesions.However, GA treatment clearly inhibited the expression level of these factors in a concentration-dependent manner.Similar results were also found in Dex-induced SZ95 cells.
Based on the TMT-labeled quantitative proteomic analysis, 23 DEPs before and after GA treatment were identified in the GA/BC groups.Among them, 15 DEPs were downregulated, and 8 DEPs were up-regulated.The hierarchical clustering analysis of these DEPs demonstrated that the proteins in the acne-like ear skin in the GA group were closer to those in the BC group, whereas the proteins in the P. acnes-induced acne-like ear skin in the MC group were detached from those in the BC/GA groups.
The KEGG pathway enrichment analysis showed GA treatment might involve multiple pathways including PPARs, biosynthesis of unsaturated fatty acids, AMPK, autophagy and Wnt, in which the DEPs, including acyl-CoA desaturase 1 (SCD-1), vang-like protein 1 (VANGL1), Nuclear receptor-binding factor 2 (NRBF2), Dual specificity phosphatase 14 (Dusp14), Alpha-1,6-mannosyl-glycoprotein 1-beta-N-acetylglucosaminyltransferase (Mgat1), Tenomodulin (Tnmd) and Vang-like protein 1 (Vangl1), were significantly altered (p < 0.05).These DEPs play a crucial role in the pathogenesis of acne vulgaris.We found that Scd1, Nrbf2, Mgat1 and Tnmd were significantly expressed in the PPAR signaling pathway.PPARs (PPARα, -β/δ and -γ) are a group of nuclear receptor proteins that are expressed in keratinocytes, sebocytes and hair appendages and function as transcription factors regulating the expression of lipogenic genes, the innate immune system and inflammatory responses.Apart from its anti-inflammatory properties, PPAR activation promotes sebocyte lipogenesis and increases lipid accumulation in human sebocytes [33].Nrbf2 interacts with PPAR ligands and regulates PPAR-related lipid synthesis [34].Mgat1 has a key role in hepatic TG accumulation.The inhibition of MGAT1 expression levels efficiently reduces lipid accumulation through the overexpression of PPARγ in HepG2 cells [35].Tnmd, a new family of type II transmembrane glycoproteins, can accelerate adipogenic differentiation and significantly increase the mRNA level of PPAR-γ [36].SCD-1, known as a key enzyme in the biosynthesis of unsaturated fatty acids pathway, is involved in the formation of lipid droplets and the accumulation of triglycerides [37].In our analysis, Nrbf2 in the MC group is down-regulated, while Scd1, Mgat1 and Tnmd are up-regulated, compared with the BC group.Conversely, after treatment with GA, the Nrbf2 is up-regulated, while the Scd1, Mgat1 and Tnmd are down-regulated.This is consistent with our prediction that core genes are mainly involved in signaling pathways related to lipid synthesis and inflammation.

Collecting Potential Targets of GA
Potential targets of GA were obtained from the SwissTargetPrediction (http://www.swisstargetprediction.ch,accessed on 22 June 2023), Super-PRED (https://prediction. charite.de/subpages/target_prediction.php,accessed on 22 June 2023) and Similarity ensemble approach (https://sea.bkslab.org/,accessed on 22 June 2023) [38,39].Then, the targets collected by the three platforms were summarized in Microsoft Excel, and duplicate values were removed.Finally, the Uniprot ID corresponding to the collected target names was imported into the UniProtKB (https://www.uniprot.org/,accessed on 22 June 2023) protein database to acquire the gene names of the standardized total proteins.

PPI Analysis
Venny 2.1.0online tools (https://bioinfogp.cnb.csic.es/tools/venny/index.html,accessed on 22 June 2023) were used to analyze the intersections of acne vulgaris-related targets and GA targets.Potential targets were imported into STRING (www.string-db.org/,accessed on 22 June 2023) and analyzed to provide a protein-protein interaction (PPI) network diagram [40].The degrees between targets in the PPI were analyzed using Cytoscape software (version 3.9.0) to give 10 hub targets with the highest degrees.

Molecular Docking
Six potential core protein targets were selected for docking with GA.The GA and proteins were pre-processed with AutoDockTools 1.5.7, and then molecular docking was performed with AutoDock Vina (version 1.2.0) [41].The affinity fraction represents the binding energy between the GA and the protein.The smaller the affinity fraction, the stronger the interaction between the GA and the protein.

Cell Culture
SZ95 cells were grown in Sebomed basal medium with 10% fetal bovine serum, 1% penicillin/streptomycin and human epidermal growth factor (5 ng/mL) and incubated at 37 • C in a humidified atmosphere with 5% CO 2 .

Cell Viability Essay
Cell viability was assessed using the MTT assay.Briefly, SZ95 cells were transferred to 96-well plates (5 × 10 3 cells/well), GA or SA in PBS was added at final concentrations of 0.5, 1.5, 10, 50, 100, 500 and 1000 µg/mL and allowed to incubate for 24 h.Then, MTT solution was added to each of the wells and maintained for 4 h.Finally, the MTT was removed and DMSO was added to solubilize the formazan precipitate.The optical density (OD) in each well was measured at 570 nm with a microplate reader (Bio-Tek, Shoreline, WA, USA).

Enzyme-Linked Immunosorbent Assay (ELISA)
SZ95 sebocytes were seeded in a 96-well flat-bottomed plate (5 × 10 3 cells/well) and incubated at 37 • C for 24 h in an incubator containing 5% CO 2 .After washing with phosphate-buffered saline (PBS), Different concentrations (100, 250 and 500 µg/mL) of GA or SA (500 µg/mL) were added and followed by incubation for 24 h.Following lysis with RIPA lysis buffer and PMSF, the lysates were centrifuged at 12,000× g at 4 • C for 10 min to give the supernatant.The mouse ear skin tissues were homogenized with RIPA lysate buffer for 5 min using a tissue grinder and then centrifuged at 4000× g at 4 • C for 10 min to produce the supernatant.
The levels of TG, SREBP-1, PPAR-γ, IL-1, IL-6 and TNF-α were determined using commercial ELISA kits, according to the manufacturers protocol, respectively.The OD was measured at 450-550 nm using a microplate reader (Bio-Tek Infinite M200 Pro, Shoreline, WA, USA).The quantitative determination of total protein in the samples was performed using the BCA protein assay method.

Mice Treatments
BALB/c mice (18-22 g, 7-8 weeks, male) were obtained from Guangzhou Laboratory Animal Center (Guangzhou, China) and kept under pathogen-free conditions at 23 ± 2 • C, 50 ± 10% humidity, with filtered air and 12 h/12 h light:dark cycles and fed a regular diet and sterilized water.The mice were randomly divided into five groups (n = 6): blank control group (BC), acne model group (MC), high-dose GA-treated group (5 mg/mL, HGA), medium-dose GA-treated group (2.5 mg/mL, MGA), low-dose GA-treated group (1 mg/mL, LGA) and SA-treated group (5 mg/mL, SA).Each mouse's ear skin was evenly smeared with 50 µL of 80% oleic acid (OA) and injected with 50 µL of a P. acnes solution (1 × 107 colony-forming units/µL) once per day for 14 days, except for the BC group.After 4 h, 100 µL of GA or SA solution was applied to the ear skin once daily for 14 days.All the mice were euthanized on day 15.The ear skin was removed and kept at −80 • C for further analysis.

Histopathological Examination
Paraffin-embedded tissue of the mouse ear skin was prepared as described in previous studies and then subjected to H&E staining [42].The H&E stained tissue slides were analyzed by an optical microscope (BX53, Olympus, Tokyo, Japan).

Immunohistochemistry
The paraffin sections of mouse ear skin were placed in a citrate buffer (0.01 M) at 121 • C for 20 min and 1% BSA for 20 min.The sections were treated with primary antibodies (anti-HSD11b1, anti-PLIN2, anti-NC3R1 and anti-PNCA) for 16 h at 4 • C and secondary antibody (goat anti-rabbit IgG-HRP) for 1 h at 4 • C. All the slides were photographed under an optical microscope.The positive chromogenic optical density values were quantitatively determined by IPP 6.0 software.

Establishment of a Target Database for Acne Vulgaris
The target sites of acne vulgaris were collected by typing "Acne Vulgaris" into the search box on the GeneCards website (www.genecards.org,accessed on 15 October 2023), and the results were exported.

Protein Digestion and TMT Labeling
Diphtheria-tetanus toxoid (DTT, 5 mM) was added to the supernatants from the same mouse ear skin tissue, obtained as above, and the samples were incubated at 55 • C for 30 min.Then, 10 mM iodoacetamide was added and protected from light at room temperature for 15 min.The protein extracts were precipitated with six times the volume of cold acetone, maintained at −20 • C overnight and followed by centrifugation (8000 rpm) at 4 • C for 10 min.
The collected protein was dissolved in 100 µL of 200 mM triethylammonium bicarbonate (TEAB) with 1 mg/mL Trypsin-TPCK (1/50 ratio) and kept overnight at 37 • C. Each freezedried sample was labeled with a TMT reagent, according to the manufacturer's protocol.

LC-MS/MS Analysis
The samples were loaded onto a narrow diameter Acclaim PepMap RSLC RP-C18 column (75 µm × 50 cm, 3 µm), which was coupled to a QExactive system of mass spectrometry.The peptide sample was analyzed using Phase A (CH 3 CN-H 2 O, 99.9-0.1,v/v) and Phase B (CH 3 CN-H 2 O-HCO 2 H, 80-19.9-0.1, v/v/v) at a flow rate of 300 µL/min under a gradient of 2-28% Phase B for 0~50 min; 28-42% Phase B for 50~60 min; 42-90% Phase B for 60~65 min; and 90% Phase B for 65~75 min.The MS scans were set to the full scanning charge-mass ratio m/z range of 350-1500 at the resolution of 60,000 with the automatic gain control value of 3 × 10 6 and carried out on 20 of the highest peaks.All the MS/MS maps were obtained through high-energy collision cracking in the data-dependent positive ion mode, with collision energy at 32 eV, the maximum ion injection time (80 ms), the dynamic exclusion time (30 s) and the automatic gain control value of 2 × 10 5 at a resolution of 45,000.

Proteomic Bioinformatic Analysis
The protein ID obtained by LCMS was converted into the UniProt ID.Proteins with Score Sequest HT > 0, unique peptide ≥ 1 and no null expression values were screened from the original data retrieved from the UniProt database (http://ebi.ac.uk/GOA/, accessed on 15 October 2023).The trusted proteins were acquired after median normalization and log2 logarithmic conversion.On the basis of the trusted proteins, two criteria were chosen to calculate the difference between the samples.Fold change (log2 (fold change) = mean of the experimental groupmean of the control group) was used to evaluate the multiples of expression change between the samples.The p-value from the t-test shows the significance of the differences between the samples.The proteins meeting the criteria of fold change ≥ 1.2 or fold change ≤ 1/1.2 and p-value < 0.05 were set as DEPs (Tables S1 and S2).In the same way, proteins were set as trusted proteins if fold change ≥ 1.2 or fold change ≤ 1/1.2 (Tables S3 and S4).Then, GO annotations of the DEPs were accomplished with a community-based bioinformatics database (http://www.geneontology.org,accessed on 15 October 2023).KEGG pathway enrichment was performed with the online KEGG database (http://www.genome.jp/kegg/,accessed on 15 October 2023) [43].

Statistical Method
All the data were calculated using GraphPad Prism 7.0 and shown in the form of the mean ± standard deviation (SD).The differences among the group means were compared using an ANOVA test followed by the LSD multiple comparisons test.A p-value of less than 0.05 was determined to be statistically significant.

Ethics Statement
The ethical assessments related to the use of animal studies were approved by the Institutional Animal Care and Use Committee (IACUC), Center for Animal Experiments, Guangdong University of Technology (Guangzhou, China).

Conclusions
In conclusion, we predicted 37 potential GA targets, based on the network pharmacological and the TMT-labeled quantitative proteomic analysis.Among them, TNF, IL1B, IL6,

Figure 1 .
Figure 1.Network pharmacology analysis of GA and acne vulgaris.(A) Simplified molecular input line entry system (SMILES) for GA.(B) Venn diagram of target interaction between GA and acne vulgaris.(C) Interaction network of 37 targets of GA against acne vulgaris.(D) Top 10 targets by degree ranking.(E) GO analysis of intersecting genes.(F) Top 30 pathway terms presented by enriching the intersected targets in KEGG.

Figure 1 .
Figure 1.Network pharmacology analysis of GA and acne vulgaris.(A) Simplified molecular input line entry system (SMILES) for GA.(B) Venn diagram of target interaction between GA and acne vulgaris.(C) Interaction network of 37 targets of GA against acne vulgaris.(D) Top 10 targets by degree ranking.(E) GO analysis of intersecting genes.(F) Top 30 pathway terms presented by enriching the intersected targets in KEGG.

Figure 4 .
Figure 4. Effect of GA on a mouse model of acne vulgaris.Blank control group (BC), acne m group (MC), high-dose GA-treated group (HGA, 5 mg/mL), medium-dose GA-treated g (MGA, 2.5 mg/mL), low-dose GA-treated group (LGA, 1 mg/mL) and SA-treated group ( mg/mL).(A) Photographs of mouse acne ear skin.(B) H&E staining of mouse ear skin (magn tion ×200).The black arrow represents the SGs.The brown arrow represents the epidermal t ness.The red boxes represent enlarged and blocked sebaceous glands (magnification ×2000).Number of mast cells in the toluidine blue-stained tissue sections.(E) Mouse ear epidermal t ness.(F) Mouse ear weight.(G) Half of the short and long diameters of SGs.All the values ar mean ± SD (n = 6), * p < 0.05, ** p < 0.01 and *** p < 0.001, compared with the MC group.## 0.001, compared with the control group.

Figure 4 .
Figure 4. Effect of GA on a mouse model of acne vulgaris.Blank control group (BC), acne model group (MC), high-dose GA-treated group (HGA, 5 mg/mL), medium-dose GA-treated group (MGA, 2.5 mg/mL), low-dose GA-treated group (LGA, 1 mg/mL) and SA-treated group (SA, 5 mg/mL).(A) Photographs of mouse acne ear skin.(B) H&E staining of mouse ear skin (magnification ×200).The black arrow represents the SGs.The brown arrow represents the epidermal thickness.The red boxes represent enlarged and blocked sebaceous glands (magnification ×2000).(C,D) Number of mast cells in the toluidine blue-stained tissue sections.(E) Mouse ear epidermal thickness.(F) Mouse ear weight.(G) Half of the short and long diameters of SGs.All the values are the mean ± SD (n = 6), * p < 0.05, ** p < 0.01 and *** p < 0.001, compared with the MC group.### p < 0.001, compared with the control group.

A
total of 6483 proteins were obtained through quantitative proteome analysis, of which 6348 proteins were quantitatively analyzed.The differentially expressed proteins (DEPs) are shown in Figure6A,B by volcano plots.Of these, 33 up-regulated and 55 down-regulated proteins were screened in the acne model group (MC) versus the blank control group (BC) in accordance with the DEP standards of a fold change of over 1.2-or < 1/1.2-fold, as well as p < 0.05.In addition, there were a total of 23 DEPs in the GA-treated groups versus the acne model group (MC).Among them, 8 DEPs were upregulated, and 15 DEPs were down-regulated.In addition, hierarchical clustering analysis of these DEPs shown in Figure6Crevealed the pathogenesis of acne was closely related to differential protein expression levels, including nuclear receptor-binding factor 2 (NRBF-2), cannabinoid receptor interacting protein 1a (CRIP1a) and acyl-CoA desaturase 1 (SCD-1), Phospholipase DDHD2 and Proteoglycan 4 (Prg4).nification200×, scale bar = 100 µm); the yellow, green, orange and blue arrows indicate the dance of 11β-HSDb1, PLIN2, NC3R1 and PCNA proteins.(E-H) 11β-HSDb1, PPAR-γ, SR and NC3R1 expression levels of mouse ear skin tissues.(I,J) PLIN2 and PCNA expression lev mouse ear skin measured by ELISA assay.(K-M) IL-1β, IL-6 and TNF-α expression of mou skin measured by ELISA assay.All the values are the mean ± SD (n = 6), * p < 0.05, ** p < 0.01 < 0.001, compared with the MC group.## p < 0.05, ### p < 0.001, compared with the control gr2.6.Differentially Expressed Proteins (DEPs) AnalysisA total of 6483 proteins were obtained through quantitative proteome analys which 6348 proteins were quantitatively analyzed.The differentially expressed pro (DEPs) are shown in Figure6A,B by volcano plots.Of these, 33 up-regulated an down-regulated proteins were screened in the acne model group (MC) versus the b control group (BC) in accordance with the DEP standards of a fold change of over 1 < 1/1.2-fold, as well as p < 0.05.In addition, there were a total of 23 DEPs in the treated groups versus the acne model group (MC).Among them, 8 DEPs were regulated, and 15 DEPs were down-regulated.In addition, hierarchical clustering a sis of these DEPs shown in Figure6Crevealed the pathogenesis of acne was close lated to differential protein expression levels, including nuclear receptor-binding fa (NRBF-2), cannabinoid receptor interacting protein 1a (CRIP1a) and acyl-CoA de rase 1 (SCD-1), Phospholipase DDHD2 and Proteoglycan 4 (Prg4).

Figure 6 .
Figure 6.Proteomic analysis of skin (A) volcano plot of DEPs in BC group vs. MC group.Th izontal coordinate of the volcano map is log2 and the ordinate is −log10 (p-value).Blue an dots indicate up-regulated and down-regulated proteins, respectively.Darker colors indicate significant differences, and gray dots represent proteins with p-values ≥ 0.05.(B) Hierarchica tering heat map of 23 DEPs between GA group and MC group.(C) The protein expression is tered according to the amount of protein expression.The colors from red to blue indicate hi pression levels to low expression levels in the DEPs.

Figure 6 .
Figure 6.Proteomic analysis of skin (A) volcano plot of DEPs in BC group vs. MC group.The horizontal coordinate of the volcano map is log2 and the ordinate is −log10 (p-value).Blue and red dots indicate up-regulated and down-regulated proteins, respectively.Darker colors indicate more significant differences, and gray dots represent proteins with p-values ≥ 0.05.(B) Hierarchical clustering heat map of 23 DEPs between GA group and MC group.(C) The protein expression is clustered according to the amount of protein expression.The colors from red to blue indicate high expression levels to low expression levels in the DEPs.

Figure 7 .
Figure 7. GO and KEGG enrichment analysis.(A) GO analysis of DEPs.The horizontal coordinates represent the enrichment score, and the vertical coordinates represent the top 20 terms information of BP/CC/MF, respectively.The color of the bubbles changes from blue to red.The redder the color, the smaller the p value and the more significant the enrichment.Larger bubbles contain more protein.(B) KEGG analysis of DEPs.

Figure 7 .
Figure 7. GO and KEGG enrichment analysis.(A) GO analysis of DEPs.The horizontal coordinates represent the enrichment score, and the vertical coordinates represent the top 20 terms information of BP/CC/MF, respectively.The color of the bubbles changes from blue to red.The redder the color, the smaller the p value and the more significant the enrichment.Larger bubbles contain more protein.(B) KEGG analysis of DEPs.

Table 1 .
Molecular docking of GA and core target proteins.

Table 1 .
Molecular docking of GA and core target proteins.