Abstract
Rosacea, a chronic inflammatory skin disorder, severely impairs patients’ quality of life and imposes substantial medical burdens. Existing treatments are hampered by unsatisfactory therapeutic effects and high recurrence rates, driving urgent research into novel pathogenic mechanisms. As a common spice and cosmetic preservative, cinnamaldehyde’s safety risks in rosacea patients remain poorly characterized. This study aimed to integrate network toxicology with in vitro cellular assays to clarify its pathogenic roles and molecular pathways, providing novel theoretical support for its adverse impacts on rosacea sufferers. First, we predicted the target genes of cinnamaldehyde using the PubChem, SwissTargetPrediction, SuperPred, and ChEMBL databases. Second, we obtained relevant targets for rosacea using the GeneCards and CTD databases and identified the intersection targets with toxic substances. Then, we constructed a protein–protein interaction network of the core toxic substance combination using the STRING database and performed Gene Ontology functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis to predict possible mechanisms. Subsequently, molecular docking validation of the active ingredients with the main targets was performed using AutoDock Vina software. Finally, cinnamaldehyde was added to the rosacea-like cellular model in HaCaT cells to assess its effects and molecular mechanisms by Cell Counting Kit-8, reverse transcription quantitative polymerase chain reaction, and enzyme-linked immunosorbent assay. Database screening revealed 43 overlapping targets between the 195 predicted targets of cinnamaldehyde and the 1889 disease-related targets for rosacea, which included notable proteins such as prostaglandin‑endoperoxide synthase 2 (cyclooxygenase‑2; PTGS2), matrix metallopeptidase 9 (MMP9), and epidermal growth factor receptor (EGFR). Protein–protein interaction analysis identified PTGS2 as a pivotal hub protein, while Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis indicated that interleukin-17 signaling and lipid metabolism are vital mechanisms involved in rosacea. Molecular docking studies demonstrated strong binding affinities between cinnamaldehyde and the identified core targets. Furthermore, cell culture experiments confirmed that cinnamaldehyde exacerbates rosacea by upregulating the expression of PTGS2, MMP9, and EGFR, along with increasing the production of inflammatory cytokines such as interleukin-8 and interleukin-1 beta. In summary, our findings elucidate the potential pathogenic mechanisms by which cinnamaldehyde may induce rosacea, highlighting the roles of the core targets PTGS2, MMP9, and EGFR in mediating inflammation. This research provides new scientific insights that could inform preventive strategies and therapeutic interventions aimed at managing the toxicity associated with rosacea.
Keywords: cinnamaldehyde, molecular docking, network toxicology, rosacea
1. Introduction
Rosacea is a chronic inflammatory skin disorder characterized by facial erythema, papules, and telangiectasia, imposing significant psychological distress and socioeconomic burdens on affected individuals.[1–3] Current therapeutic strategies, including topical agents, oral antibiotics, and laser therapies, remain suboptimal due to limited efficacy, adverse effects, and high recurrence rates.[4,5] The complex interplay of genetic predisposition, dysregulated immune responses (e.g., toll‑like receptor 4/nuclear factor kappa‑B activation),[6] and environmental triggers underscores the urgent need to elucidate novel pathogenic mechanisms and therapeutic targets. Emerging evidence implicates exogenous factors, particularly cosmetic ingredients, in exacerbating cutaneous inflammation and barrier dysfunction,[7] yet their molecular contributions to rosacea progression are poorly understood.
Cinnamaldehyde, a widely used fragrance and preservative in cosmetics, has drawn attention for its pro-inflammatory properties and potential role in skin irritation.[8] In vitro experiments suggest that this compound can induce immunotoxicity in NCTC 2544 cells (a skin epithelial-like cell line originating from normal human skin) by increasing the total amount of interleukin-18.[9] Clinical observations also indicate that cinnamaldehyde, as a flavoring component, carries risks of contact allergy and photosensitivity, and excessive use in facial products can trigger erythema and dermatitis.[8] However, a systematic investigation into its target-specific effects and downstream cascades is lacking, creating a critical knowledge gap in cosmetic safety assessment and rosacea management.
Conventional reductionist approaches often fail to capture the multicomponent, multi-target nature of cosmetic ingredient interactions. This study addresses this limitation by integrating computational biology methodologies – network pharmacology, molecular docking, and multi-omics data mining – to construct a comprehensive cinnamaldehyde–rosacea interaction network. Such an approach enables the simultaneous evaluation of compound-target-pathway relationships, offering higher predictive power than traditional single-target assays. The methodology leverages curated databases (GeneCards, SwissTargetPrediction) and topological algorithms to prioritize high-confidence targets, followed by molecular dynamics simulations to validate binding affinities.
This study adopts an innovative multi-layered integration strategy to systematically analyze the pathogenic mechanisms of cinnamaldehyde through 4 key steps: first, integrating specialized databases to construct a component-disease-target network; then, using Cytoscape for topological analysis to identify hub genes; next, revealing key pathways through KEGG/GO annotation; and finally, validating the findings through molecular docking using AutoDock Vina (The Scripps Research Institute). This methodological design not only overcomes the fragmented shortcomings of traditional reductionist research but also attempts for the first time to establish a systems biology association model between cosmetic ingredients and rosacea. At the same time, cytological experiments were further conducted to validate the potential mechanism of cinnamaldehyde-induced rosacea.
Ultimately, this study aims to establish a paradigm for the safety assessment of cosmetic ingredients through systematic toxicology, bridging the gap between empirical observations and mechanistic insights. The research results will provide evidence-based guidance for developing skincare products suitable for rosacea patients. This dual focus on mechanistic discovery and translational applicability highlights the novelty of this research in dermatology and cosmetic science.
2. Methods
2.1. Prediction of toxicity targets of cinnamaldehyde
Obtain the SMILES number of cinnamaldehyde from the PubChem database[10] and import it into the SwissTargetPrediction, SuperPred, and ChEMBL databases[11] to predict its target genes.
2.2. Acquisition of rosacea-related targets
Use “Rosacea” as the keyword to screen relevant targets using the GeneCards and the CTD databases. After integrating the target data from the 2 databases, remove duplicate genes to obtain the genes related to the pathogenesis of rosacea.
2.3. Acquisition of intersection genes between rosacea and cinnamaldehyde
Import the drug and disease targets into the Microbioinformatics Platform (https://www.bioinformatics.com.cn/) to obtain the intersection targets and draw a Venn diagram.
2.4. Prediction of compound and disease component targets
After drawing a Venn diagram for visualization, use Cytoscape 3.9.1 software (Cytoscape Consortium) to construct and visualize the interactions among drugs, components, intersection targets, and diseases.
2.5. Construct the protein–protein interaction (PPI) network
Use the STRING database (http://www.string-db.org/), select “Homo sapiens” as the species, and keep the default values for the remaining parameters to construct the PPI network of the core drug combination. Import the data obtained from the above steps into Cytoscape 3.9.1. Analyze the network topology parameters and set the screening criteria so that the betweenness centrality (BC), closeness centrality (CC), and degree centrality (DC) are all greater than or equal to the median values to obtain the information on core targets.
2.6. Functional and pathway enrichment analysis
Import the screened intersection targets into the DAVID database, set the species to “Homo sapiens,” conduct Gene Ontology (GO) functional enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis, and visualize the results using the Microbiome Bioinformatics–Online Bioinformatics Analysis and Visualization Cloud Platform.
2.7. Molecular docking and visualization analysis
Use the AutoDock Vina software to perform molecular docking of cinnamaldehyde with important targets and verify its binding activity. The specific methods are as follows: download the SDF file of the 2D structure of the compound from the PubChem database, input it into Chembio3D, minimize its energy, input it into AutoDockTools 1.5.7 for hydrogen addition, calculate charges, assign charges, set rotation keys, and store it in the “pdbqt” format; download key target proteins from the protein data bank database (give priority to human-derived proteins, give priority to the original ligand groups that are highly similar in structure to the active ingredient to be docked, and select the original ligand groups with high resolution); import the protein into PyMOL, remove the original ligand and water, then import it into AutoDockTools (v1.5.7; Scripps Research Institute), perform hydrogen addition on the protein, calculate charges, assign charges, specify the atomic type, and store it in the “pdbqt” format; and apply PyMOL to analyze the interaction mode. The selected proteins for docking are the top 5 proteins in the PPI network analysis.
2.8. Material
Immortalized normal human epidermal cells (HaCaT) were purchased from Guangzhou Suyan Biotechnology Co., Ltd. High-glucose Dulbecco’s modified Eagle’s medium, trypsin digestion solution for cell culture, and penicillin-streptomycin mixed antibacterial solution were purchased from Gibco (Grand Island). Fetal bovine serum was obtained from NEWZERUM Ltd. Cinnamaldehyde was purchased from Macklin (Cat. No.: C11965500BT). Cathelicidin antimicrobial peptide LL-37 (LL37) was purchased from Sangon Biotech (Shanghai) Co., Ltd. (No. A410536). The Total RNA Extraction Mini Kit was acquired from Mabio (Cat. No.: RNT412). The HiScript II Q RT SuperMix for qPCR kit (Cat. No.: R222) and ChamQ Universal SYBR qPCR Master Mix (Cat. No.: R711) were purchased from Vazyme. The SuperKine Maximum Sensitivity Cell Counting Kit-8 (CCK-8) was obtained from Abbkine (Cat. No.: BMU106-CN). The Human IL-8/CXCL8 ELISA Kit (Cat. No.: F1551-A) and Human IL-1β ELISA Kit (Cat. No.: F0179-A) were purchased from FANKEW (Shanghai Kexing Trading Co., Ltd.). The main instruments used in this study included a carbon dioxide (CO2) cell incubator (ESCO CLM-170B-8-NF), a constant temperature water bath incubator (HWS-24, Shanghai Yiheng Instrument Co., Ltd.), a real-time quantitative PCR (qPCR) system (Bio-Rad CFX96™ Optics Module, Bio-Rad Laboratories, Inc.), and a multifunctional microplate reader (PerkinElmer Victor Nivo 3F, PerkinElmer).
2.9. Cell viability assay
The cytotoxicity of cinnamaldehyde on HaCaT cells was detected by the CCK-8 assay. Cinnamaldehyde at different concentrations (0, 0.016, 0.032, 0.048, 0.064, and 0.08 μg/mL) was prepared by dissolving it in Dulbecco’s modified Eagle’s medium. The cinnamaldehyde solution was sterilized by filtration through a 0.22 μm filter. HaCaT cells were seeded in 96-well plates and treated with different concentrations of cinnamaldehyde solution for 24 hours; then, an appropriate concentration was selected for the subsequent experiments. Next, HaCaT cells were pretreated with LL37 for 2 hours, followed by treatment with cinnamaldehyde, and cell viability was detected at 6, 12, 18, and 24 hours. In these experiments, 10 μL of CCK-8 solution was added to each well at the corresponding time points, and the cells were incubated at 37°C for 1 hour. Absorbance was measured using a microplate reader at a wavelength of 450 nm.
2.10. Reverse transcription quantitative polymerase chain reaction
Total RNA was extracted using the Total RNA Extraction Mini Kit and converted to complementary DNA using the HiScript II Q RT SuperMix for qPCR kit according to the manufacturer’s protocol. The complementary DNA was amplified using TB Green dye and quantified by a Bio-Rad CFX96 Optics Module. The expression levels of the analyzed genes were normalized to the glyceraldehyde 3-phosphate dehydrogenase mRNA level. All primer sequences used in this study are listed in Table 1.
Table 1.
Primer sequences used for RT-qPCR.
| Gene symbol (GenBank Accession No.) | Primers |
|---|---|
| PTGS2 (NM_000963) | F: TAAGTGCGATTGTACCCGGAC |
| R: TTTGTAGCCATAGTCAGCATTGT | |
| MMP9 (NM_004994) | F: GGGACGCAGACATCGTCATC |
| R: TCGTCATCGTCGAAATGGGC | |
| EGFR (NM_001346897) | F: TGCGTCTCTTGCCGGAAT |
| R: TGGTACTGGGAGCCAATGG | |
| IL-1β (NM_008361.4) | F: GCAACTGTTCCTGAACTCAACT |
| R: ATCTTTTGGGGTCCGTCAACT | |
| IL-8 (NM_001003200) | F: TGACACTGGCAAAACAATGCA |
| R: GGTCCTTTTCACCAGCAAGCT |
EGFR = epidermal growth factor receptor, IL-1β = interleukin-1 beta, IL-8 = interleukin-8, MMP9 = matrix metallopeptidase 9, PTGS2 = prostaglandin‑endoperoxide synthase 2 (cyclooxygenase‑2), RT-qPCR = reverse transcription quantitative polymerase chain reaction.
2.11. Enzyme-linked immunosorbent assay (ELISA)
The cell supernatant was collected and centrifuged at 3000 rpm for 10 minutes to remove particles and polymers. Following the manufacturer’s instructions, ELISA tests were conducted for various biomarkers. The ELISA kits for interleukin-8 (IL-8) and interleukin-1 beta (IL-1β) were obtained from FANKEW (Shanghai, China).
2.12. Data analysis
All data were legally and publicly available and, therefore, did not require additional ethical approval. The data were expressed as the mean ± standard deviation and analyzed using one-way analysis of variance, where a P value of <.05 was considered statistically significant. Statistical results were graphically presented using GraphPad Software (version 9.5; GraphPad Software, Inc.).
3. Results
3.1. Identification of cinnamaldehyde-related targets associated with rosacea
Daily cosmetics and skincare products, including various types of makeup, lotions, masks, and perfumes, are an important source of human exposure to cinnamaldehyde (Fig. 1A). Using databases such as SwissTargetPrediction, SuperPred, and ChEMBL to predict drug targets, we found that cinnamaldehyde had 195 targets. By using the GeneCards and CTD databases, searching for rosacea-related targets with “Rosacea” as the keyword, and setting the conditional cutoff value of Inference Score > 7 in the CTD database, we obtained 1889 disease targets after merging and removing duplicates. After taking the intersection of the drug’s action targets and the rosacea target genes, 43 intersection target genes of rosacea and the drug were obtained, which were the interactive target genes for the drug treatment of rosacea. The results are shown in Figure 1B.
Figure 1.

Cinnamaldehyde exposure from daily cosmetics and skincare products and its potential rosacea-related targets. Daily cosmetics and skincare products, such as makeup, lotions, masks, and perfumes, represent significant sources of human exposure to cinnamaldehyde (A, by Figdraw). Venn diagram illustrating the overlap between cinnamaldehyde-related targets and rosacea-associated genes (B).
3.2. Construction of the target network of the active ingredients of the drug
Using the drug-target data, Cytoscape 3.9.1 was used to draw a diagram to predict the target network of the drug ingredients. The results are shown in Figure 2.
Figure 2.

Cinnamaldehyde–rosacea–target map.
3.3. Construction and analysis of PPI network
The intersection gene file was imported into the STRING database; the species was selected as “Homo sapiens,” and the protein interaction network was obtained, as shown in Figure 3A. Cytoscape was used to conduct centrality analysis and evaluation on all target gene information. The BC, CC, and DC were set to be greater than or equal to the median values for screening. The median value of BC was 31.85365854, the median value of CC was 0.01431331, and the median value of DC was 12.68292683. It was found that the network had 10 nodes and 39 edges, as shown in Figure 3B and Table 2.
Figure 3.

Intersection target network diagram (A), core target PPI network diagram (B). PPI = protein–protein interaction.
Table 2.
Intersection target-related data.
| No. | Betweenness | Closeness | Degree | Gene name |
|---|---|---|---|---|
| 1 | 171.7800 | 0.0200 | 30 | PTGS2 |
| 2 | 156.7073 | 0.0189 | 27 | MMP9 |
| 3 | 133.1577 | 0.0189 | 27 | EGFR |
| 4 | 127.6838 | 0.0182 | 26 | ESR1 |
| 5 | 65.2090 | 0.0179 | 24 | NFKB1 |
| 6 | 53.6750 | 0.0169 | 22 | TLR4 |
| 7 | 39.0428 | 0.0164 | 21 | MAPK3 |
| 8 | 27.5674 | 0.0164 | 20 | CXCL8 |
| 9 | 49.5077 | 0.0159 | 19 | MAPK14 |
| 10 | 78.4029 | 0.0159 | 19 | NFE2L2 |
| 11 | 6.2907 | 0.0154 | 17 | RELA |
| 12 | 10.9935 | 0.0152 | 16 | MAPK1 |
| 13 | 6.3969 | 0.0152 | 16 | STAT1 |
| 14 | 58.5364 | 0.0147 | 15 | CYP3A4 |
| 15 | 6.4303 | 0.0147 | 14 | PRKCA |
| 16 | 16.1729 | 0.0149 | 14 | PTGS1 |
| 17 | 27.8494 | 0.0149 | 14 | XDH |
| 18 | 5.3980 | 0.0147 | 13 | NR3C1 |
| 19 | 52.3145 | 0.0143 | 12 | ALDH1A1 |
| 20 | 4.9342 | 0.0139 | 11 | ESR2 |
| 21 | 13.6929 | 0.0139 | 11 | CYP2E1 |
| 22 | 16.8570 | 0.0139 | 11 | CYP2C9 |
| 23 | 85.8384 | 0.0139 | 10 | F3 |
| 24 | 5.5802 | 0.0135 | 10 | SLC2A1 |
| 25 | 2.7515 | 0.0137 | 9 | FLT1 |
| 26 | 13.8390 | 0.0130 | 9 | CYP1A2 |
| 27 | 0.0000 | 0.0135 | 8 | TGFBR2 |
| 28 | 8.9327 | 0.0135 | 8 | HMGCR |
| 29 | 20.7513 | 0.0130 | 8 | NR1I2 |
| 30 | 14.5555 | 0.0132 | 8 | TTR |
| 31 | 3.4667 | 0.0128 | 7 | MIF |
| 32 | 2.6624 | 0.0128 | 7 | CTSS |
| 33 | 4.9190 | 0.0125 | 7 | CYP2D6 |
| 34 | 5.5143 | 0.0125 | 6 | CTSD |
| 35 | 4.2789 | 0.0128 | 6 | IDH1 |
| 36 | 4.3100 | 0.0132 | 6 | ACHE |
| 37 | 0.0000 | 0.0119 | 5 | PPP3CA |
| 38 | 0.0000 | 0.0099 | 3 | CYP2S1 |
| 39 | 0.0000 | 0.0104 | 2 | CSNK2B |
| 40 | 0.0000 | 0.0090 | 1 | PRCP |
| 41 | 0.0000 | 0.0109 | 1 | CA2 |
EGFR = epidermal growth factor receptor, ESR1 = estrogen receptor 1, MMP9 = matrix metallopeptidase 9, PTGS2 = prostaglandin-endoperoxide synthase 2 (cyclooxygenase-2), TLR4 = toll-like receptor 4.
3.4. Biological function enrichment analysis
The intersection targets were imported into the microbial database for enrichment analysis. A total of 189 GO entries were found, among which 118 were related to biological processes, 25 were related to cellular components, and 46 were related to molecular functions. The top 10 GO entries for biological processes, cellular components, and molecular functions were selected for plotting, as shown in Figure 4. A total of 120 pathways were obtained through KEGG Pathway analysis. Twenty pathways were selected according to the P value (Table 3). The main pathways involved included cancer pathways, lipid and atherosclerosis, hepatitis B, programmed death‑ligand 1 expression and programmed cell death protein 1 checkpoint pathway in cancer, advanced glycation end products – receptor for advanced glycation end products signaling pathway in diabetic complications, tuberculosis, chemical carcinogenesis–receptor activation, human cytomegalovirus infection, coronavirus disease–COVID-19, hypoxia‑inducible factor 1 signaling pathway, relaxin signaling pathway, prolactin signaling pathway, leishmaniasis, interleukin-17 (IL-17) signaling pathway, endocrine resistance, Chagas disease, C-type lectin receptor signaling pathway, Toll-like receptor signaling pathway, T helper 17 (Th17) cell differentiation, and pancreatic cancer, as shown in Table 3. A bubble chart of the top 20 KEGG metabolic pathways was drawn. The Fold Enrichment value was on the horizontal axis. The size of the bubble represented the number of genes enriched in the corresponding pathway, and the color depth represented the significance. The significantly enriched information could be visually observed (Fig. 5).
Figure 4.

GO enrichment analysis bar chart. BP = biological processes, CC = cellular components, GO = Gene Ontology, MF = molecular functions.
Table 3.
KEGG pathway information.
| ID | Term | Gene ratio | P value | Count |
|---|---|---|---|---|
| hsa05200 | Pathways in cancer | 34.8837 | 8.51E−08 | 15 |
| hsa05417 | Lipid and atherosclerosis | 27.9070 | 3.44E−09 | 12 |
| hsa05161 | Hepatitis B | 25.5814 | 3.30E−09 | 11 |
| hsa05235 | PD-L1 expression and PD-1 checkpoint pathway in cancer | 23.2558 | 2.85E−10 | 10 |
| hsa04933 | AGE-RAGE signaling pathway in diabetic complications | 23.2558 | 8.13E−10 | 10 |
| hsa05152 | Tuberculosis | 23.2558 | 1.46E−07 | 10 |
| hsa05207 | Chemical carcinogenesis–receptor activation | 23.2558 | 6.02E−07 | 10 |
| hsa05163 | Human cytomegalovirus infection | 23.2558 | 9.15E−07 | 10 |
| hsa05171 | Coronavirus disease–COVID-19 | 23.2558 | 1.41E−06 | 10 |
| hsa04066 | HIF-1 signaling pathway | 20.9302 | 3.96E−08 | 9 |
| hsa04926 | Relaxin signaling pathway | 20.9302 | 1.46E−07 | 9 |
| hsa04917 | Prolactin signaling pathway | 18.6047 | 3.66E−08 | 8 |
| hsa05140 | Leishmaniasis | 18.6047 | 7.76E−08 | 8 |
| hsa04657 | IL-17 signaling pathway | 18.6047 | 2.80E−07 | 8 |
| hsa01522 | Endocrine resistance | 18.6047 | 3.72E−07 | 8 |
| hsa05142 | Chagas disease | 18.6047 | 4.88E−07 | 8 |
| hsa04625 | C-type lectin receptor signaling pathway | 18.6047 | 5.56E−07 | 8 |
| hsa04620 | Toll-like receptor signaling pathway | 18.6047 | 7.18E−07 | 8 |
| hsa04659 | Th17 cell differentiation | 18.6047 | 7.18E−07 | 8 |
| hsa05212 | Pancreatic cancer | 16.2791 | 1.55E−06 | 7 |
AGE-RAGE = advanced glycation end products – receptor for advanced glycation end products, HIF-1 = hypoxia‑inducible factor 1, IL-17 = interleukin-17, KEGG = Kyoto Encyclopedia of Genes and Genomes, PD-1 = programmed cell death protein 1, PD-L1 = programmed death‑ligand 1, Th17 = T helper 17 cells.
Figure 5.

KEGG pathway enrichment bubble chart. KEGG = Kyoto Encyclopedia of Genes and Genomes.
3.5. Molecular docking and visualization analysis
AutoDock and AutoDock Vina were used for molecular docking. The lower the docking score, the better the binding activity between the compound and the target. Using prostaglandin‑endoperoxide synthase 2 (cyclooxygenase‑2; PTGS2), matrix metallopeptidase 9 (MMP9), and epidermal growth factor receptor (EGFR) as target genes and cinnamaldehyde as the small molecule, molecular docking was conducted, and the molecular docking results were organized, as shown in Figure 6. The binding energy is an index to measure the binding degree between the small molecule and the target protein. When the binding energy is lower than −5, the small molecule can freely bind to the target protein, and the probability of binding between the small molecule and the target protein is greater. The simulation results show that the small molecule binds well to the target protein. The docking with the lowest binding energy was selected and drawn into a diagram, as shown in Figure 7A–C.
Figure 6.

Binding energy heat map.
Figure 7.

Molecular docking simulations of cinnamaldehyde with core proteins. PTGS2 with cinnamaldehyde (A), MMP9 with cinnamaldehyde (B), EGFR with cinnamaldehyde (C). EGFR = epidermal growth factor receptor, MMP9 = matrix metallopeptidase 9, PTGS2 = prostaglandin‑endoperoxide synthase 2 (cyclooxygenase‑2).
3.6. Cinnamaldehyde regulates core target expression in the process of inducing inflammation
To validate the real effect of cinnamaldehyde on HaCaT in the context of rosacea-like dermatitis, various assays were conducted accordingly. Via the CCK-8 assay, the results showed that the cell viability of HaCaT was not significantly affected when cinnamaldehyde was used within a dosage of 0, 0.016, 0.032, 0.048, 0.064, or 0.08 µg/mL for 24 hours (Fig. 8A). Therefore, 0.064 µg/mL was selected as the safe concentration for subsequent experiments. Furthermore, after pre-treating HaCaT cells with LL37 (a key peptide in rosacea pathogenesis) for 2 hours to establish a rosacea-like inflammatory environment, cinnamaldehyde was added at a concentration of 0.064 µg/mL, and CCK-8 was used to detect cell viability at 6, 12, 18, and 24 hours (Fig. 8B). With the increase in time, the cell viability of the LL37 + cinnamaldehyde group decreased in a time-dependent manner. To identify whether the rosacea-like dermatitis induced by cinnamaldehyde was related to the core genes, reverse transcription quantitative polymerase chain reaction was performed. The results showed that cinnamaldehyde extract solution upregulated the PTGS2, MMP9, and EGFR gene expression while increasing the mRNA levels of cytokines associated with rosacea-like dermatitis, including IL-1β and IL-8 (Fig. 8C–G). To further confirm that the upregulation of inflammatory genes translated into enhanced protein secretion, we measured the levels of IL-1β and IL-8 in cell supernatants by ELISA. Consistent with the mRNA results, the LL37 + cinnamaldehyde group exhibited significantly higher secretion of both IL-1β and IL-8 compared to the LL37-alone group (Fig. 8H and I). This indicates that cinnamaldehyde not only elevates the transcription of pro-inflammatory cytokines but also sustains their translation and release at the protein level, thereby exacerbating the rosacea-like inflammatory response. Taken together, these findings indicated that the core targets PTGS2, MMP9, and EGFR played an essential role in cinnamaldehyde-induced rosacea-like dermatitis, and that cinnamaldehyde promoted a sustained inflammatory response through both transcriptional upregulation and increased protein secretion of key cytokines.
Figure 8.

The effect of different concentrations of cinnamaldehyde on the viability of HaCaT cells after 24 hours. *P < .05 versus 0 (A); the cell viability of HaCaT after 24 hours of intervention with LL37 alone or LL37 + cinnamaldehyde (B); relative mRNA expression level of IL-8, IL-1β, PTGS2, MMP9, and EGFR (C–G); the levels of cytokines secreted by different groups, including IL-1β and IL-8 (H, I). *P < .05, **P < .01, ***P < .001 versus control or 0 µg/mL; #P < .05, ##P < .01, ###P < .001 versus LL37. Data represent the mean ± SD of at least 3 independent experiments. EGFR = epidermal growth factor receptor, IL-8 = interleukin-8, LL37 = cathelicidin antimicrobial peptide LL-37, MMP9 = matrix metallopeptidase 9, PTGS2 = prostaglandin‑endoperoxide synthase 2 (cyclooxygenase‑2), SD = standard deviation.
4. Discussion
Rosacea is a chronic inflammatory skin disorder characterized by facial erythema, papules, and telangiectasia, which significantly impairs patients’ quality of life and imposes substantial psychosocial and economic burdens.[12,13] Current therapeutic strategies, including topical agents, oral antibiotics, and laser therapy, remain suboptimal due to limited efficacy, adverse effects, and high recurrence rates.[14] Elucidating novel pathogenic mechanisms and therapeutic targets is thus clinically imperative, particularly given the potential role of environmental triggers in disease exacerbation.
This study innovatively integrates network toxicology, molecular docking, and bioinformatics to systematically investigate the molecular mechanisms by which cinnamaldehyde – a widely used cosmetic ingredient with documented pro-inflammatory properties – may aggravate rosacea. It is noteworthy that the biological effects of cinnamaldehyde appear to be highly context-dependent. While several studies have reported its antioxidant and anti-inflammatory properties in models of gingival inflammation, general immune modulation, or atopic dermatitis,[15–17] some articles also propose the viewpoint that cinnamaldehyde is a known transient receptor potential vanilloid 1 agonist, and oral intake of foods containing cinnamaldehyde can lead to facial flushing and exacerbation of inflammation through this pathway.[18–20] Therefore, cinnamaldehyde may primarily play the role of a “neurovascular trigger” rather than a systemic anti-inflammatory agent in rosacea. Our multi-omics approach identified 43 overlapping targets (e.g., PTGS2 and MMP9) and implicated critical pathways such as IL-17 signaling, providing mechanistic insights into cinnamaldehyde’s potential role in disrupting skin barrier function and amplifying inflammatory responses.
The identification of 43 overlapping targets between cinnamaldehyde and rosacea, including PTGS2, MMP9, and EGFR, provides compelling molecular evidence for its potential role in exacerbating rosacea. These targets are closely associated with key pathological processes such as inflammation, immune dysregulation, and skin barrier dysfunction, which are hallmarks of rosacea.[21–23] The upregulation of these targets by cinnamaldehyde may amplify the production of pro-inflammatory cytokines like IL-1β and IL-8, further aggravating skin inflammation.
The PPI network analysis revealed PTGS2 as the most interconnected hub, with a DC of 30, underscoring its pivotal role in cinnamaldehyde’s action. This finding aligns with clinical observations of elevated prostaglandin levels in rosacea patients,[24] implicating PTGS2-mediated pathways in disease progression. The synergistic interactions among core targets like MMP9 and EGFR suggest a multi-target mechanism, where cinnamaldehyde disrupts skin homeostasis by concurrently activating inflammatory and tissue-remodeling pathways. While this network provides a plausible framework, experimental validation in animal models or patient-derived samples is critical to confirm the functional relevance of these targets and explore potential compensatory mechanisms, such as feedback loops involving anti-inflammatory mediators.
KEGG pathway analysis highlighted the enrichment of cinnamaldehyde’s targets in the IL-17 signaling pathway, which is particularly significant given the established role of Th17 immune responses in rosacea pathogenesis.[24–26] This suggests that cinnamaldehyde may exacerbate disease by promoting Th17 differentiation and subsequent cytokine release. Additionally, the involvement of lipid metabolism pathways (e.g., advanced glycation end products – receptor for advanced glycation end products[27]) points to a potential dual mechanism where cinnamaldehyde not only drives inflammation but also compromises skin barrier integrity. These findings position IL-17 and lipid metabolic markers as promising candidates for cosmetic safety assessment, though their specificity for rosacea versus other inflammatory skin conditions like acne or eczema requires further clarification.
Molecular docking results demonstrated strong binding affinities between cinnamaldehyde and key targets like PTGS2 (−7.2 kcal/mol) and MMP9 (−6.8 kcal/mol), validating the network pharmacology predictions. The variation in binding energies may reflect differential target sensitivity, with PTGS2 emerging as the most susceptible. This direct interaction likely disrupts enzymatic activity, such as PTGS2’s role in prostaglandin synthesis, contributing to inflammatory cascades. Future studies employing site-directed mutagenesis could elucidate the precise binding motifs and assess whether alternative binding modes exist, particularly for targets with lower but still significant affinity (e.g., estrogen receptor 1).
Subsequently, we validated these results through in vitro experiments. The results indicated that using cinnamaldehyde alone did not affect the viability of HaCaT cells, but in a rosacea cell model induced by LL37, the addition of cinnamaldehyde enhanced the inhibitory effect of LL37 on HaCaT cells, which suggested that cinnamaldehyde might make the disease worse for people with rosacea. Compared with the group with only LL37 added, the addition of a specific concentration of cinnamaldehyde upregulated the mRNA expression of PTGS2, MMP9, EGFR, IL-1β, and IL-8. ELISA analysis confirmed that the LL37 + cinnamaldehyde group exhibited a significantly higher release of both IL-1β and IL-8 compared to the LL37-alone group, indicating that low-dose cinnamaldehyde might worsen the inflammatory progression of rosacea. These results suggest that the use of skincare and daily chemical products containing cinnamaldehyde might make rosacea worse by upregulating EGFR, PTGS2, and MMP9 genes, leading to neurogenic inflammation, blood vessel issues, and immune problems, creating a vicious cycle. Despite the novel insights provided by this computational study, several limitations should be acknowledged. We acknowledge that validating these results is an important step in supporting the findings from our network analysis. Future research should carry out large-scale epidemiological studies to clarify the core genes and pathways so we can develop better prevention and treatment plans to reduce the harmful effects of cinnamaldehyde.
5. Conclusion
In conclusion, this integrative bioinformatics approach systematically demonstrates that cinnamaldehyde may exacerbate rosacea through multi-target modulation of IL-17 signaling, prostaglandin synthesis (via PTGS2), and matrix metalloproteinase activity. The identified core targets and pathways establish a mechanistic framework for reassessing the safety of this widely used cosmetic ingredient. These findings warrant immediate translational research, including biomarker development in clinical cohorts and formulation studies to establish safe exposure thresholds for sensitive populations.
Author contributions
Conceptualization: Ju Wen, Hongfeng Tang.
Formal analysis: Hao Huang, Ju Wen.
Funding acquisition: Hao Huang, Ju Wen.
Software: Hao Huang.
Validation: Hao Huang, Junrong Feng, Zhihong Zhang.
Writing – original draft: Hao Huang, Ju Wen.
Writing – review & editing: Hao Huang, Ju Wen.
Correction
This article was originally published without a department name for the affiliation "b": “b The Affiliated Guangdong Second Provincial General Hospital of Jinan University, Guangzhou, Guangdong Province, China.” This affiliation has now been updated online to include a department name as follows: “b Department of Dermatology, The Affiliated Guangdong Second Provincial General Hospital of Jinan University, Guangzhou, Guangdong Province, China.”
Abbreviations:
- BC
- betweenness centrality
- CC
- closeness centrality
- CCK-8
- Cell Counting Kit-8
- DC
- degree centrality
- EGFR
- epidermal growth factor receptor
- ELISA
- enzyme-linked immunosorbent assay
- GO
- Gene Ontology
- IL-1β
- interleukin-1 beta
- IL-8
- interleukin-8
- KEGG
- Kyoto Encyclopedia of Genes and Genomes
- LL37
- cathelicidin antimicrobial peptide LL-37
- MMP9
- matrix metallopeptidase 9
- PPI
- protein–protein interaction
- PTGS2
- prostaglandin-endoperoxide synthase 2 (cyclooxygenase-2)
- RT-qPCR
- reverse transcription quantitative polymerase chain reaction
- Th17
- T helper 17 cells
This project was supported by the Science and Technology Projects in Guangzhou (Grant No. 2023A03J0269); the Guangdong Basic and Applied Basic Research Foundation (Guangdong-Foshan Joint Fund Regional Incubation Programme) (Grant No. 2023A1515140031); the Guangdong Provincial Department of Education Characteristic Innovation Project (Grant No. 2021KTSCX017); the Guangdong Provincial Medical Science and Technology Research Fund (Grant No. B2025565); the Scientific Research Start Plan of The Eighth Affiliated Hospital of Southern Medical University (The First People’s Hospital of Shunde Foshan) (Grant Nos. SRSP2021036, SRSP2021005, SRSP2023030, SRSP2023004); and the Clinical Research Startup Program of Shunde Hospital, Southern Medical University (Grant No. CRSP2022003).
The authors have no conflicts of interest to disclose.
The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.
How to cite this article: Huang H, Wen J, Feng J, Zhang Z, Tang H. Assessing the rosacea potential pathogenic risks of cinnamaldehyde in daily skin care products and cosmetics using network toxicology and molecular docking with experimental validation. Medicine 2026;105:30(e49854).
Contributor Information
Hao Huang, Email: 103411594@qq.com.
Junrong Feng, Email: 978745044@qq.com.
Zhihong Zhang, Email: cherry_kiss@163.com.
Hongfeng Tang, Email: tmwthf@126.com.
References
- [1].Halioua B, Cribier B, Frey M, Tan J. Feelings of stigmatization in patients with rosacea. J Eur Acad Dermatol Venereol. 2017;31:163–8. [DOI] [PubMed] [Google Scholar]
- [2].van Zuuren EJ. Rosacea. N Engl J Med. 2017;377:1754–64. [DOI] [PubMed] [Google Scholar]
- [3].Tan J, Almeida LMC, Bewley A, et al. Updating the diagnosis, classification and assessment of rosacea: recommendations from the global ROSacea COnsensus (ROSCO) panel. Br J Dermatol. 2017;176:431–8. [DOI] [PubMed] [Google Scholar]
- [4].Thiboutot D, Anderson R, Cook-Bolden F, et al. Standard management options for rosacea: the 2019 update by the National Rosacea Society Expert Committee. J Am Acad Dermatol. 2020;82:1501–10. [DOI] [PubMed] [Google Scholar]
- [5].Chen C, Wang P, Zhang L, et al. Exploring the pathogenesis and mechanism-targeted treatments of rosacea: previous understanding and updates. Biomedicines. 2023;11:2153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Yuan X, Li J, Li Y, et al. Artemisinin, a potential option to inhibit inflammation and angiogenesis in rosacea. Biomed Pharmacother. 2019;117:109181. [DOI] [PubMed] [Google Scholar]
- [7].Kianmehr S, Jahani M, Moazzen N, Ahanchian H, Khameneh B. The potential of probiotics for treating skin disorders: a concise review. Curr Pharm Biotechnol. 2022;23:1851–63. [DOI] [PubMed] [Google Scholar]
- [8].Gunia-Krzyzak A, Sloczynska K, Popiol J, Koczurkiewicz P, Marona H, Pękala E. Cinnamic acid derivatives in cosmetics: current use and future prospects. Int J Cosmet Sci. 2018;40:356–66. [DOI] [PubMed] [Google Scholar]
- [9].Corsini E, Galbiati V, Mitjans M, Galli CL, Marinovich M. NCTC 2544 and IL-18 production: a tool for the identification of contact allergens. Toxicol Vitro. 2013;27:1127–34. [DOI] [PubMed] [Google Scholar]
- [10].Kim S, Chen J, Cheng T, et al. PubChem in 2021: new data content and improved web interfaces. Nucleic Acids Res. 2021;49:D1388–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Mendez D, Gaulton A, Bento AP, et al. ChEMBL: towards direct deposition of bioassay data. Nucleic Acids Res. 2019;47:D930–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Boaretto Netto JC, Cardoso Borges AJ, Gomes Farias T, de Souza da Fonseca A. Could photobiomodulation be used for treatment of rosacea? Lasers Med Sci. 2025;40:250. [DOI] [PubMed] [Google Scholar]
- [13].Johnson AW, Johnson SM. The role of topical brimonidine tartrate gel as a novel therapeutic option for persistent facial erythema associated with rosacea. Dermatol Ther (Heidelb). 2015;5:171–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Feldman SR, Huang WW, Huynh TT. Current drug therapies for rosacea: a chronic vascular and inflammatory skin disease. J Manag Care Spec Pharm. 2014;20:623–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Ghardashpour M, Saeedi M, Negarandeh R, et al. Anti-inflammatory and tissue repair effect of cinnamaldehyde and nano cinnamaldehyde on gingival fibroblasts and macrophages. BMC Oral Health. 2023;23:1014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Ustaoglu E, Turkoglu Z, Ulgen OA, Caytemel C, Agirgol S. Anti-inflammatory effect of cinnamaldehyde in a mouse model of 2,4-dinitrofluorobenzene-induced atopic dermatitis. Indian J Dermatol. 2023;68:170–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Benaissa A, Bouali W, Tamfu AN, et al. Inhibition of clinical multidrug-resistant Pseudomonas aeruginosa biofilms by cinnamaldehyde and eugenol from essential oils: in vitro and in silico analysis. Chem Biodivers. 2025;22:e202402693. [DOI] [PubMed] [Google Scholar]
- [18].Searle T, Ali FR, Carolides S, Al-Niaimi F. Rosacea and diet: what is new in 2021? J Clin Aesthet Dermatol. 2021;14:49–54. [PMC free article] [PubMed] [Google Scholar]
- [19].Weiss E, Katta R. Diet and rosacea: the role of dietary change in the management of rosacea. Dermatol Pract Concept. 2017;7:31–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Alia E, Feng H. Rosacea pathogenesis, common triggers, and dietary role: the cause, the trigger, and the positive effects of different foods. Clin Dermatol. 2022;40:122–7. [DOI] [PubMed] [Google Scholar]
- [21].Zhang H, Zhang Y, Li Y, et al. Bioinformatics and network pharmacology identify the therapeutic role and potential mechanism of melatonin in AD and rosacea. Front Immunol. 2021;12:756550. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Liang J, Chen Y, Wang Z, et al. Exploring the association between rosacea and acne by integrated bioinformatics analysis. Sci Rep. 2024;14:3065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Agrahari G, Sah SK, Nguyen CT, Choi SS, Kim H-Y, Kim T-Y. Superoxide dismutase 3 inhibits LL-37/KLK-5-mediated skin inflammation through modulation of EGFR and associated inflammatory cascades. J Invest Dermatol. 2020;140:656–65.e8. [DOI] [PubMed] [Google Scholar]
- [24].Steinhoff M, Vocanson M, Voegel JJ, Hacini-Rachinel F, Schäfer G. Topical ivermectin 10 mg/g and oral doxycycline 40 mg modified-release: current evidence on the complementary use of anti-inflammatory rosacea treatments. Adv Ther. 2016;33:1481–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Yang L, Shou Y, Yang Y, Xu J-H. Elucidating the immune infiltration in acne and its comparison with rosacea by integrated bioinformatics analysis. PLoS One. 2021;16:e0248650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Ali AA, Vender R, Vender R. The role of IL-17 in papulopustular rosacea and future directions. J Cutan Med Surg. 2019;23:635–41. [DOI] [PubMed] [Google Scholar]
- [27].Wang S, Wang Y, Han B, et al. Huanglian ointment alleviates eczema by maintaining the balance of c-Jun and JunB and inhibiting AGE-RAGE-mediated pro-inflammation signaling pathway. Phytomedicine. 2022;105:154372. [DOI] [PubMed] [Google Scholar]
