The reagents used in this research included: octadecanol, sodium lauryl sulfate and ethyl nicotinate (Tianjin Kemiou Chemical Reagent Co., Ltd.); petrolatum (Shandong Dexinkang Medical Technology Co., Ltd.); liquid paraffin, isopropyl myristate (IPM), glycerol and acetone (Tianjin Tianli Chemical Reagent Co., Ltd.); compound dexamethasone acetate cream (China Resources Sanjiu Medical & Pharmaceutical Co., Ltd.); olive oil (Zhejiang Meizhiyuan Biotechnology Co., Ltd.); 2,4-dinitrochlorobenzene (DNCB; Chengdu Kelong Chemical Co., Ltd.); Cnidium monnieri (Shaanxi Xingshengde Pharmaceutical Co., Ltd.); ELISA Kit IL-6 (cat. no. MM-0163M1; Jiangsu Meimian Industrial Co., Ltd.); and ELISA Kit IL-17 (cat. no. MM-0170M1; Jiangsu Meimian Industrial Co., Ltd.); goat serum (cat. no. SL038; Beijing Solarbio Science & Technology Co., Ltd.); anti JAK2/STAT3 (cat. no. ab195055/ab31370; Abcam; HRP-labeled secondary antibody for goat anti-rabbit (cat. no. ab205718; Abcam); DAB staining (cat. no. DA1010; Beijing Solarbio Science & Technology Co., Ltd.). Additionally, the equipment used includes a GC-MS System (Agilent Technologies, Inc.), a PHS-3C pH meter (Shanghai INESA Scientific Instrument Co., Ltd.) and a rotational viscometer (cat. no. LC-NDJ-55; Shanghai Li-Chen Bang Xi Instrument Technology Co., Ltd.).

Male Kunming KM mice weighing 20-22 g were supplied by Chengdu Dasuo Laboratory Animal Co., Ltd. and the license number was SCXK (Chuan) 2025-0030. A total of 80 were purchased, among which 60 were used for pharmacodynamic tests (25 days) and 20 for safety tests (21 days). For the pharmacodynamic study, mice were allocated as described below. For serum-based analyses, six animals per group (n=6) were included based on sample availability following blood collection and the corresponding data are presented accordingly. Animal experiments were approved by the Shaanxi University of Chinese Medicine Laboratory Animal Ethics Committee (approval no. SUCMDL20250512001). Mice were housed in a laboratory with standard temperature (22±2˚C) and humidity (55±10%) under a 12-h light/dark cycle and underwent a 7-day acclimatization period. During the experiment, fresh drinking water was added and replaced daily. Animals were monitored daily for general health, behavior and signs of distress. Humane endpoints included severe weight loss (>20%), persistent ulceration, or inability to access food and water, upon which animals would be sacrificed immediately. After the experiment, 20 mice used for safety tests were sacrificed by cervical dislocation and 60 mice used for pharmacodynamics tests were intraperitoneally injected with 1% pentobarbital sodium (100 mg/kg). After the successful induction of anesthesia was confirmed by the righting reflex, ~1 ml of blood was collected from the eyeball and then sacrifice was performed by cervical dislocation. After the cervical dislocation sacrifice of the 80 mice, the breathing and pupil dilation conditions were observed to determine if the mice were dead.

To minimize irritation and volatility, CMVO was incorporated into an ointment base. Oil-in-water (O/W) and water-in-oil (W/O) emulsion types were both screened to identify formulations with high gloss, fine smooth particles, light texture and excellent stability.

Formulations (Table I) were prepared by weighing the oil and aqueous phases separately. Each phase was heated separately in an 80˚C water bath until dissolved. The two phases were continuously stirred to achieve uniform mixing. Once they reached the same temperature, the aqueous phase was slowly poured into the oil phase while stirring continuously in the same direction until emulsification was complete. The emulsion was then cooled with continued stirring to room temperature until the cream formed.

A single-factor experiment was designed to investigate the optimal quantities of the aqueous and oil phases, as well as the base components with significant influence. Subsequently, a response surface experiment was conducted to derive the optimal formulation.

Based on a literature review, four ointment formulations were selected for comparison (30-33). Formulation 1 and Formulation 2 were O/W types, while Formulation 3 and Formulation 4 were W/O types. Specific ingredients are shown in Table I.

Preliminary quality assessment of the four selected prescriptions indicated that Prescription 2 demonstrated the most effective results.

To further refine the formulation ratio, a five-level single-factor experiment (n=3) was conducted to evaluate six components in the ointment formulation. Factors were labeled as follows: (A) octadecanol, (B) petrolatum, (C) liquid paraffin, (D) IPM, (E) sodium lauryl sulfate and (F) glycerin. The specific experimental design is shown in Table II.

Components markedly influencing cream stability were selected for re-evaluation. Three levels were chosen for each factor: (A) petrolatum, (B) IPM and (C) sodium lauryl sulfate. The Box-Behnken response surface method was employed using Design-Expert 13 software (Stat-Ease, Inc.) to arrange the experimental design (n=3). The design results are shown in Table III.

The cream was evaluated based on four major criteria, including the cream's appearance, droplet size, physical stability and moisturizing properties (34-36) and scored according to established standards with a maximum composite score of 100 points. The breakdown was as follows: Appearance and texture: 30% (each sub-criterion accounts for 5%); physical stability: 30% (each sub-criterion accounts for 10%); emulsion droplet size: 20% (each sub-criterion accounts for 10%); and moisturizing properties: 20%.

The cream was expected to have a uniform texture and color, with appropriate viscosity for smooth application and absorption, free from noticeable particles or separation. Appearance characteristics were subdivided into five indicators: Sensory evaluation, gloss, particle size, skin feel and spreadability, each scored according to established standards. Specific criteria are detailed in Table IV.

The physical stability of the cream was evaluated through three indicators: Centrifugal stability, heat resistance and cold resistance. Scoring criteria are shown in Table V.

For this test, 1 ml of the cream sample was placed into a 1.5 ml centrifuge tube and centrifuged at 735 x g for 30 min (25±2˚C). Oil-water separation was observed and scores were determined accordingly.

For this test, 7.5 g of the cream sample was placed in a 25 ml beaker, sealed with plastic wrap and placed in a 60˚C constant-temperature oven for 6 h. After cooling to room temperature, signs of oil-water separation, demulsification, or flocculation were observed.

For this test,7.5 g of the cream sample was placed in a sealed bag, refrigerated at 4˚C for 2 h, then frozen at -20˚C for 24 h. After returning to room temperature, oil-water separation, demulsification, or flocculation was observed.

Smaller particle size and more uniformly distributed droplets indicated higher cream quality. First, a temporary mount was prepared by taking a small amount of the cream sample, spreading it evenly onto the center of a microscope slide and then covering it with a coverslip. Next, a drop of glycerol was placed at the center of the coverslip. Finally, the prepared samples were observed using a bright-field light microscope at a magnification of x400. Five non-overlapping fields per sample were selected for particle size measurement by scanning the slide in a systematic manner (from left to right and top to bottom) to minimize selection bias. Droplet size and distribution were evaluated according to the criteria shown in Table VI.

A cream sample (1 g) was weighed onto a microscope slide. After reweighing, the slide was placed in a 37˚C constant-temperature oven for 8 h. After removal and reweighing, the moisture loss rate and final score were calculated.

At room temperature, 1 g of the cream was weighed, 10 g of purified water was added and the mixture was stirred evenly to form a white emulsion. The pH value was then measured using a pH meter. Three parallel determinations were conducted.

Using a rotational viscometer, the 4th rotor was selected and the rotational speed was adjusted to 6 rpm. The instrument was started and the data were read once stabilized.

Volatile oil from C. monnieri was extracted using steam distillation. The raw material was ground into powder and 300 g of C. monnieri powder, which was identified by Professor Yonggang Yan from the College of Pharmacy, Shaanxi University of Chinese Medicine as the dried ripe fruit of C. monnieri (L.) Cuss. of the Apiaceae family, was placed into a 2,000 ml round-bottom flask with 1,300 ml of distilled water. After thorough mixing, the essential oil collection apparatus was assembled. Extraction was conducted via steam distillation for 6 h, ceasing when the product yield no longer increased. After cooling, the essential oil was collected, with an average yield of 1.06% (v/w). Three independent batches were extracted and GC-MS analysis demonstrated consistent chemical profiles among batches. The instrument used in Gas Chromatography-Mass Spectrometry (GC-MS) was a 7890GC/5977MS, (Agilent Technologies, Inc.). The EI ion source had an ionization energy of 70 eV and an initial temperature of 40˚C. It then gradually rose to 90˚C, followed by an increase at a rate of 5˚C per minute to 260˚C, which was maintained for 1 min. Subsequently, it was raised at a rate of 10˚C per minute to 260˚C; this process lasted for 6 min. The carrier gas used was high-purity helium with a flow rate of 1 ml/min.

A total of 60 male specific pathogen-free (SPF)-grade mice were randomly divided into six groups of 10 mice each: Control, model, positive and low-, medium- and high-dose groups of Cnidium monnieri essential oil ointment. All animals were subjected to the same experimental procedures for model establishment and treatment. However, for serum-based analyses, a subset of animals (n=6 per group) was used and the corresponding data are presented accordingly. One day prior to modeling, depilation was performed on the dorsal region of mice in all groups except the normal group, covering an area of ~2 cm x 2 cm. On Day 1, all mice except the normal group were sensitized with 100 µl of a 7% DNCB solution in acetone:olive oil (4:1). A booster dose was administered on Day 2 and the challenge was repeated on Day 3. Starting on Day 4, except for the control group, mice were challenged with 30 µl of a 0.5% DNCB acetone-olive oil solution on the back every other day for a total of four challenges. After successful establishment of the eczema model, the normal group received daily application of 100 µl acetone-olive oil solution on the back. The model group received 100 µl of an unloaded cream formulation (100 mg/kg/d), the positive control group received 100 µl of compound dexamethasone acetate ointment (100 mg/kg/d) and the three groups treated with CMVOC received 100 µl of the corresponding dose of cream formulation (the dosage of CMVO in the cream): Low dose (100 mg/kg/d), medium dose (200 mg/kg/d) and high dose (400 mg/kg/d). Each group received twice-daily applications for 15 consecutive days. The precise dosage of the drug was derived by integrating multiple previous studies on local essential oil formulations, preliminary experiments and clinical dose conversions using the fingertip unit method (37-40). On the final day of the experiment, mice were fasted and deprived of water for 12 h. Blood samples were then collected from the eyeballs, skin samples were obtained from the dorsal region and changes in skin condition were observed. The severity of skin lesions was recorded on days 4, 7, 10, 13 and 16. Following the scoring principles of the EASI clinical scoring system (41), erythema, edema/thickening, scratches/erosions and crusting were selected as evaluation criteria. Scores were assigned on a scale of 0-3, where 0 indicated none (not present even upon close inspection), 1 indicated mild (related lesions visible upon close inspection), 2 indicated moderate (lesions directly visible) and 3 indicated severe (symptoms very pronounced). Based on the score, if the average score reaches 2 points and the model group remains above 2 points, it can be determined that the modeling is successful.

A total of 20 male SFP-grade KM mice were randomly divided into four groups of five mice each: Control group, experimental group 1 (low-dose group), experimental group 2 (medium-dose group) and experimental group 3 (high-dose group) (all consisting of the CMVOC). After random grouping, all mice had their backs and the sides of their spines shaved (~2 cm x 2 cm). At 24 h later, skin scratches were made on the left side of the skin, while the right side was scratched in an equal shape to cause slight bleeding on the skin surface. The medication was applied immediately after the scratches. The normal group was applied with pure water and the experimental groups were applied with low-, medium- and high-dose groups of CMVOC. After administration, the mice were covered with gauze and fixed with adhesive tape. The medication was applied once a day. At 24 h later, the skin was washed with distilled water. After 1 h of cleaning, the mice were observed for 7 consecutive days for redness, swelling, or other phenomena. After stopping the medication, the mice were observed for another 7 days (42).

According to the scoring criteria in Table VII, the condition of the mouse's back skin was scored. The intensity of irritation was determined using the average scoring formula for the animals, according to Table VIII.

Mouse serum samples were tested using ELISA kits from Jiangsu Enzyme Immunoassay Industrial Co., Ltd. IL-6 and IL-17 levels were determined according to the instruction manual. Intra-assay and Inter-assay coefficient of variation (CV) <15%. Standard curves were established and the concentration of each sample group was calculated. The standard curves for i) IL-17 and ii) IL-6 were as follows: i) y=0.0047x + 0.0659 R²=0.9957; ii) y=0.0039x + 0.1644 R²=0.991. Blood samples were collected from mice for serum preparation. During sample processing, four serum samples in the control group (K2, K3, K5 and K6) were affected by hemolysis due to technical issues during orbital blood collection and were therefore excluded from ELISA analysis.

As a result, six valid serum samples remained in the control group. Accordingly, ELISA analyses were conducted using six samples per group. For each treatment group, six serum samples were included in the analysis.

Serum levels of IL-6 and IL-17 were measured using commercial ELISA kits according to the manufacturers' instructions. All assays were performed following standard protocols and no samples were excluded based on experimental outcomes.

Fresh skin tissue was fixed in 4% paraformaldehyde solution for 24 h (25±2˚C), routinely embedded in paraffin (25±2˚C) and sectioned into 4 µm-thick slices (-20˚C). The sections underwent dewaxing (65˚C) and hydration with a gradient of ethanol (100, 95, 80 and 75%), followed by rinsing with distilled water. Sections were immediately immersed in hematoxylin stain for 5 min, rinsed with tap water for 1 min, deparaffinized with hydrochloric acid ethanol for 30 sec, soaked in tap water for 15 min and then placed in eosin solution for 2 min. After staining, sections were re-washed with graded ethanol and clarified with xylene. Finally, neutral resin was applied to mount the sections (25±2˚C). Once dried, the sections were ready for bright-field light microscopic examination (43). The tissue morphology was observed under a bright-field light microscope at a magnification of x200 and images were captured. A total of three non-overlapping fields per sample were selected for cell counting by scanning the slide in a systematic manner (from left to right and top to bottom) to minimize selection bias.

The pretreatment procedure was identical to that for H&E staining: Tissue fixation, dehydration, clearing, paraffin embedding, sectioning and dewaxing. The prepared sections were placed in a 0.5% toluidine blue solution and incubated at 55˚C for 20 min. After 2-3 rinses with water, the sections were placed in hydrochloric acid alcohol for 2-3 sec for decolorization, followed by rinsing with water. Dehydration was performed using alcohol and the sections were then stained with 0.25% eosin for 8 sec (25±2˚C). Finally, routine procedures of decolorization, dehydration, clearing and mounting were carried out sequentially. After air-drying, the sections were observed under a microscope (44).

Mouse skin tissue sections of 2-5 µm were processed dewaxing, rehydration, xylene and graded ethanol treatments. Next, antigen retrieval was performed using the citric acid antigen retrieval buffer (pH 6.0) and then using an electric pressure cooker. Following three washes with TBS, the sections were incubated with 0.1% Triton X-100 for 10 min, reacted with 0.3% H₂O₂ in methanol at room temperature (25±2˚C) for 30 min then blocked with 10% goat serum (cat. no. SL038; Beijing Solarbio Science & Technology Co., Ltd.) in PBS for 10 min at room temperature. When adding the primary antibody, gently shake off the blocking solution. Sections were incubated with primary antibody of JAK2/STAT3 (cat. no. ab195055/ab31370; Abcam) diluted 1:200 in PBS onto the slide at 4˚C for 24 h. The slide was placed in PBS (pH 7.4) and shaken three times, each time for 5 min. After slightly drying the slide, the secondary antibody of HRP-labeled goat anti-rabbit (cat. no. ab205718; Abcam) corresponding to the species of the secondary antibody was added to the slide at a ratio of 1:500 to cover the tissue and incubated at room temperature for 50 min. The slide was placed in PBS (pH 7.4) and shaken three times, each time for 5 min. Following DAB staining at room temperature (25±2˚C; cat. no. DA1010; Beijing Solarbio Science & Technology Co., Ltd.), under microscopic observation until brownish-yellow staining appeared, the sections were rinsed with tap water to terminate the color development reaction. Sections were counterstained with hematoxylin (25±2˚C, 3 min), dehydrated, cleared and mounted. Processed sections were examined under a bright-field light microscope (45,46).

Experimental data in the present study were processed using GraphPad Prism 10.4.2 software (Dotmatics). Results from single-factor experiments and response surface experiments in the cream base screening were expressed as means, with n=3. For mouse data in the eczema model experiments, serum-based analyses were performed using six animals per group (n=6).

Quantitative data from animal studies were expressed as mean ± standard deviation. Prior to statistical analysis, data were assessed for normality using the Shapiro-Wilk test and for homogeneity of variances using Levene's test.

For datasets meeting the assumptions of normality and homogeneity of variance, statistical comparisons among multiple groups were performed using one-way analysis of variance followed by Tukey's post hoc test. For datasets with unequal variances or unequal sample sizes, Welch's ANOVA followed by Games-Howell multiple comparisons test was applied. This approach ensures that statistical analyses accurately reflect the experimental design, regardless of unequal variances or deviations from normality.

Histological and immunohistochemical data, including epidermal thickness, mast cell counts, H&E staining, toluidine blue staining and immunohistochemistry, were quantified using ImageJ-win64 software (National Institutes of Health). By including all pre-defined animals and using appropriate statistical methods, the analyses were both transparent and unbiased. P<0.05 was considered to indicate a statistically significant difference.

The active components of Cnidium monnieri essential oil were analyzed using GC-MS. The total ion chromatogram is presented in Fig. 1. By comparing with the NIST database and retention indices, 21 active compounds were identified as listed in Table IX, with their chemical structures depicted in Fig. 2.

Fig. 3 illustrates the effect of varying base component levels on the final evaluation scores. Octyldodecanol, IPM and sodium lauryl sulfate markedly influenced the scores. Among these, octyldodecanol (4.40 g) provided optimal cream stability; lower levels led to oil-water separation, while higher levels resulted in hard, difficult-to-spread creams. IPM (0.80 g) exhibited optimal skin affinity; lower concentrations caused excessive greasiness, while higher levels disrupted the cream structure, leading to a granular texture. Sodium lauryl sulfate (0.40 g) achieved maximum emulsifying efficacy; lower concentrations prevented cream formation, while higher levels increased viscosity. By contrast, petrolatum, liquid paraffin and glycerin exhibited markedly lower variability, with no significant changes observed across their five respective concentration levels. As lubricants and moisturizers, their moderate dosages exerted no discernible effect on the cream formulation. The experimental variables comprised (A) octadecanol, (B) petrolatum, (C) liquid paraffin, (D) IPM, (E) lauryl sodium sulfate and (F) glycerin, with detailed experimental procedures and results outlined in Tables X and XI.

A Box-Behnken response surface design was conducted using Design-Expert 13 software. The experimental details and results are shown in Table XII.

Based on the experimental scores, regression analysis yielded the following multiple quadratic regression equation: Y=92.02 + 0.8238^*A + 0.4338^*B-0.3250^*C-0.2150^* AB-0.0575^*AC-0.1725^*BC-3.18^*A²-3.02^*B²-2.32^*C².

The results of the analysis of variance are presented in Table XIII. The model was significant (P<0.0001), with a non-significant lack-of-fit term (P=0.7267>0.05), indicating that the model fitted well with minimal interference from other factors. Additionally, R²=0.9932 and adjusted R²=0.9844 further demonstrated the model's high accuracy. The F-value revealed the following influence order: octadecanol (A) > IPM (B) > sodium lauryl sulfate (C).

Based on the multiple quadratic regression equation, Design-Expert 13 was used to generate contour plots and response surfaces, which visually depicted the effect of factor interactions on cream scores, as illustrated in Fig. 4. These interactions further elucidated the intricate relationships within the formulation matrix. In the AB interaction, cream scores initially rose and then declined with increasing octadecyl alcohol content. Similarly, when IPM content rose, octadecanol scores followed an up-and-down pattern, suggesting a non-significant interaction between octadecyl alcohol and IPM. The highest score was attained at an octadecanol content of 4.00 g and an IPM content of 0.60 g. For the AC interaction, the score initially increased and subsequently decreased with rising sodium lauryl sulfate content. When sodium lauryl sulfate content increased, the trend of octadecyl alcohol scores remained unchanged, similarly indicating a non-significant interaction between the two. The highest score was recorded at an octadecanol content of 4.40 g and a sodium lauryl sulfate content of 0.40 g. In the BC interaction, both IPM and sodium lauryl sulfate content exhibited scores that initially increased and then decreased, pointing to a non-significant interaction between them. The highest score was achieved at an IPM content of 0.80 g and a sodium lauryl sulfate content of 0.40 g. Response surface prediction pinpointed the optimal formulation for preparing CMVO ointment: octadecanol at 4.45 g, IPM at 0.81 g and sodium lauryl sulfate at 0.39 g, yielding a comprehensive score of 92.10 points.

The total assessment scores are detailed in Table XIV. The average scores were 92.46±0.26, 92.48±0.2 and 91.48±0.27, closely aligning with the model-predicted value of 92.10. This demonstrated a strong concordance between the experimental results and the predicted values.

As presented in Fig. 5, these are the appearances of four cream formulations. In the selection of a suitable cream, priority should be given to those with uniform texture, appropriate consistency and the absence of noticeable particles or separation.

After centrifugation at 735 x g for 30 min (25±2˚C), none of the formulations exhibited oil-water separation, with no significant differences observed among groups. These findings demonstrate that the formulation exhibited good centrifugal stability.

Following storage at 60˚C for 6 h and subsequent cooling to room temperature, the cream maintained a uniform texture and color. It remained smooth, refreshing and easy to apply, without any signs of thinning, clumping, or oil-water separation, demonstrating the cream's good heat resistance.

After refrigeration at 4˚C for 2 h and freezing at -20˚C for 24 h, the cream exhibited no significant changes in color or texture upon returning to room temperature. It remained light, smooth and easy to apply, with no signs of thinning, clumping, or oil-water separation, confirming the cream's strong cold resistance.

Particle size measurements of the drug-loaded ointment revealed that 3 to 5 particle sizes were marked in each group across low, medium and high dosage formulations (n=3). The average particle size was ~85.50±15.33 nm, meeting the preparation standard (ointment particle size <180 nm). The particle size distribution is detailed in Fig. 6 and Table XV.

The pH of the cream was measured using a PHS-3C pH meter at different drug loading doses, with results shown in Table XV. After analysis, it can be seen that the average pH values for the low-, medium- and high-dose groups were 6.92±0.00, 6.91±0.02 and 6.91±0.01, respectively. Overall, the cream was weakly acidic, aligning with the pH range suitable for skin application.

Viscosity measurements conducted at room temperature, with results shown in Table XV, revealed similar viscosity values across different doses, with no significant difference observed. Moreover, this viscosity falls within the range suitable for external skin application, ensuring the cream's applicability and stability while preventing phenomena such as stratification.

A scoring system for assessing the severity of skin lesions in mice was established based on the EASI criteria, as detailed in Table XVI. On Day 4, mice exhibited mild eczema symptoms. From Days 7 to 10, symptoms intensified but remained below severe levels due to the administration of the medication. By Days 13 to 16, significant improvement was observed, indicating the medication's effectiveness in alleviating eczema symptoms. As illustrated in Fig. 7, following DNCB induction, mice displayed clear signs of eczema. However, treatment with CMVO ointment mitigated skin tissue damage. Visual inspection revealed that the control group's skin appeared pale pink without abnormalities, while the model group exhibited pronounced skin lesions and swelling, confirming successful DNCB-induced eczema modeling. All treatment groups showed marked improvements in skin condition. Although both the model and treated groups displayed skin lesions, redness, swelling and scabbing, the model group showed the most pronounced lesions and swelling. In the positive control group, scabs had completely fallen off, revealing pale pink skin without swelling or hyperemia. Mice in the low-dose cream group still exhibited slight redness and swelling, with limited desquamation of crusted areas. The medium-dose cream group exhibited partial crust desquamation, exposing pale pink skin, as did the high-dose cream group.

As depicted in Fig. 8, mild erythema emerged in the high-dose group of the experimental group after seven days of use. Calculations revealed average scores per animal of 0.04 in the high-dose control group and 0.26 in the skin-lesion group, both below the 0.5 threshold. Based on these local irritation findings, the test was deemed non-irritating according to the scoring criteria. The specific scoring results are outlined in Table XVII.

As shown in Fig. 9, the model group exhibited markedly elevated levels of the pro-inflammatory cytokines IL-6 and IL-17 compared with the normal group. By contrast, the positive control group and the medium- and high-dose groups of CMVO ointment showed marked reductions in these proinflammatory factors compared with the model group.

Figs. 10 and 11 depict the results of H&E staining and toluidine blue staining of mouse skin tissue, reflecting relevant pathological conditions. H&E staining results revealed that the skin structure in the normal group was intact, with no inflammatory cell infiltration in the dermis. By contrast, the model group showed thickening of the granular and spinous layers, intercellular edema and inflammatory cell infiltration in the dermis. Treatment with different drugs led to thinner granular and spinous layers and a significant reduction in intercellular edema and inflammatory cell infiltration in the dermis.

Toluidine blue staining highlighted clear differences in mast cell distribution among the groups. The control group had very few, sparsely distributed mast cells, while the model group exhibited numerous densely clustered mast cells. After drug administration, the positive group demonstrated a significant reduction in mast cell numbers with a more scattered distribution. Among the different dosage groups, mast cell counts decreased progressively with increasing drug concentration.

Quantitative analysis of epidermal thickness and mast cell counts is presented in Fig. 12. Compared with the control group, the model group exhibited increased epidermal thickness (P<0.0001) and mast cell counts (P<0.0001). Compared with the model group, the treatment groups exhibited reduced epidermal thickness and mast cell counts (P<0.0001, P<0.001, P<0.05), reflecting the inflammatory status in each group.

Immunohistochemical staining revealed JAK2 expression as a brownish-yellow color in the cytoplasm, while STAT3 was detected in both the cytoplasm and nucleus, also displaying a brownish-yellow hue. In the normal group, JAK2 expression was weak, confined to the basal layer of the epidermis and hair follicles, while STAT3 remained inactive with minimal detectable staining. Compared with the normal group, the model group exhibited thickened stratum spinosum, with strong cytoplasmic JAK2 expression appearing as brownish-yellow. Infiltrating inflammatory cells in the superficial dermis exhibited brownish-gray staining. STAT3 showed both cytoplasmic brownish-yellow positivity and brownish-gray granular nuclear staining in inflammatory cells of the superficial dermis. Overall, JAK2 and STAT3 expression levels were markedly elevated in the model group. By contrast, all treatment groups demonstrated markedly reduced positive expression levels, narrower cytoplasmic distribution, lighter staining intensity and disappearance of STAT3 nuclear positive signals, indicating effective suppression of pathway activation following treatment.

Analysis of results is shown in Fig. 13. Compared with the normal control group, the model group exhibited markedly upregulated JAK2 and STAT3 signaling (P<0.001), indicating altered molecular expression of this pathway. Combined with H&E and toluidine blue staining results, these findings suggested that the eczema model successfully induced inflammation. Compared with the model group, the positive drug group and different doses of CMVOC administration groups showed significant differences in the molecular expression of JAK2 and STAT3 (P<0.05, P<0.01, P<0.001). These results suggested that the JAK2-STAT3 signaling pathway may be one of the mechanisms by which CMVOC alleviates eczema.

As illustrated in Fig. 14, JAK2 expression exhibited a brownish-yellow cytoplasmic staining pattern. In the normal group, JAK2 expression was minimal and distributed between the epidermis and dermis. Compared with the normal group, the model group showed such as upregulated JAK2-positive expression, with brownish-yellow cytoplasmic staining in epidermal spinous layer cells and abundant inflammatory cell infiltration in the dermis. Compared with the model group, the positive group exhibited downregulated JAK2 expression, with fewer brownish-yellow cytoplasmic cells and inflammatory cells in the dermis. Similarly, all low-, medium- and high-dose groups treated with CMVO ointment showed downregulated JAK2 signal expression, with significant differences compared with the model group, particularly in the medium- and high-dose groups.

As shown in Fig. 15, STAT3 expression appeared as a brownish-yellow cytoplasmic staining. In the normal group, STAT3 signaling was minimal, presenting as a pale yellow hue throughout the cytoplasm. Compared with the normal group, the model group exhibited positive STAT3 expression, with deepened cytoplasmic staining in epidermal spinous layer cells and dark brown nuclei in the dermis and epidermis. Compared with the model group, the positive control group and the low-, medium- and high-dose groups of CMVO ointment exhibited lighter staining in the epidermal spinous layer, disappearance of dark brown nuclei and weaker positive STAT3 signaling expression.