Clinical data of patients with AD alone or AD comorbid with mental disorders—including 6 patients in the AD-alone group and 6 patients in the AD-comorbid mental disorders group—were collected from the Second Affiliated Hospital of Harbin Medical University from July 2024 to November 2024 to investigate the features of ERS and inflammation via quantitative polymerase chain reaction (qPCR) and western blotting. The baseline clinical characteristics of all enrolled patients, including gender, age, AD disease duration, and disease severity scores (SCORAD, EASI, IGA, VAS) as well as HADS anxiety and depression scores, are summarized in Supplementary Table 1. This study was conducted following the Declaration of Helsinki, and all participants signed written informed consent.

Diagnostic Criteria for AD: Diagnosis was based on the Chinese Guidelines for the Diagnosis and Treatment of Atopic Dermatitis, requiring pruritus as a core symptom plus at least two of the following: flexural eczematous lesions, personal or familial history of atopic diseases (e.g., asthma, allergic rhinitis), increased serum total IgE or peripheral blood eosinophil percentage, and history of skin dryness.

Diagnostic Criteria for Comorbid Psychiatric Disorders: Two attending psychiatrists made the diagnosis based on the International Classification of Diseases, 11th Revision (ICD-11), including major depressive disorder (Hamilton Depression Rating Scale [HAMD-17] score $\ge$17) and generalized anxiety disorder (Hamilton Anxiety Rating Scale [HAMA] score $\ge$14).

Exclusion criteria included comorbidity with autoimmune skin diseases or infectious skin diseases; use of systemic glucocorticoids, immunosuppressants, or psychotropic drugs within 1 month before enrollment; presence of severe hepatic, renal, or cardiovascular diseases or malignancy; pregnancy or lactation; and inability to cooperate with sample collection.

Healthy controls were enrolled if they reported no history of skin or psychiatric diseases, had not used immunomodulatory, psychotropic, or glucocorticoid drugs in the past 3 months, and had normal indicators, including liver and renal function. A 3 mm $\times$ 3 mm normal skin tissue sample was collected from nonexposed areas of each healthy control.

The university ethics committee approved the in vivo protocol design and procedures. In this study, 4–6-week-old specific-pathogen-free BALB/c mice (Vital River, SPF level) of both sexes were randomly allocated, with 8 mice allocated to each experimental group to ensure statistical validity. Mice were randomly assigned to each experimental group using a random number table method to eliminate grouping bias. A depression model was developed via 28 days of chronic unpredictable mild stress (CUMS) [18], comprising a series of stressors applied randomly. The stressors were horizontal shaking, tilting of the cage, food and water deprivation, tail clipping, heat exposure, and cold water swimming. Each stressor was applied once per day, and the sequence was varied to prevent habituation to the stressors. The AD model was developed on day 14 of the depression model by applying a 0.5% 2,4-dinitrofluorobenzene (DNFB) (Sigma-Aldrich, D9136) solution for sensitization. This was followed by 0.2% DNFB applications on days 5, 8, and 14 to induce dermatitis. Adenovirus-mediated MANF overexpression (negative control overexpression [oeNC] and MANF overexpression [oeMANF]) was performed via subcutaneous injection of 100 µL viral solution at the dermatitis site in the established depression model, establishing comorbidities, comorbidities + oeNC, and comorbidities + oeMANF groups. Further, the 4-phenylbutyric acid (4-PBA) as an ERS inhibitor (500 mg/kg) was administered at two separate dosages for the AD, comorbidities, and comorbidities+inhibitors [18] groups. The animal experiment was conducted from November 2024 to January 2025, with a total duration of 3 months. The CUMS-induced depression model lasted 28 days, followed by 21 days of AD induction and intervention, with behavioral and pathological assessments conducted during the final 7 days. At the end of the animal experiment, all mice were euthanized via intraperitoneal injection of sodium pentobarbital (100 mg/kg, 50 mg/mL, Sigma-Aldrich). The skin and brain tissue were dissected within 10 min for tissue collection after euthanasia to maintain tissue integrity.

The total number of scratching bouts was counted and utilized to quantify pruritus severity. Dermatitis severity was assessed using a previously established scoring system, including five key parameters [19]: (1) edema/papules; (2) exudate/scabs; (3) epidermal erosion; (4) lichenoid change; and (5) dryness of the unaffected skin. Each item was scored on a scale (typically 0–3) according to severity, with the total dermatitis score obtained by summing the scores of all five parameters. Each parameter was scored on a scale from 0 to 3, indicating no signs or symptoms, mild manifestations, moderate severity, and severe manifestations, respectively.

The sucrose preference test (SPT) assesses an animal’s emotional state by measuring its preference for sweet water, including a sucrose solution. Before testing, mice were given free access to both 1% sucrose solution and tap water. After 12 h of food and water deprivation, mice were provided with preweighed bottles containing sucrose solution and water for a 2-h testing period [20]. Fluid consumption was recorded, and the SPT was calculated utilizing the following formula:

$$\text{Sucrose Preference}(\%)=\frac{\text{Volume of sucrose solution consumed}}{\text{Total fluid intake}}\times 100$$

Total RNA was extracted from treated cells or tissues using TRIzol reagent (Invitrogen, 5596026), and cDNA was synthesized with a Transcriptor First Strand cDNA Synthesis Kit (Invitrogen, N8080234, CA, USA). qPCR was performed on an ABI Step-One Plus real-time PCR system using LightCycler 480 SYBR Green I Master Mix (azyme, SGD2203, Shanghai, China). Relative RNA expression levels of GRP78, P-IRE1, Caspase-12, Bax, Bcl-2, ATF4, CHOP, and P-EIF-2$\alpha$ were analyzed by the 2^{-Δ⁢Δ⁢Ct} method with GAPDH as the internal control. Primer sequences are listed in Supplementary Table 2.

Total proteins were extracted from cultured cells using RIPA lysis buffer (Sigma-Aldrich, 20-188, St. Louis, MO, USA). Protein concentration was determined using a BCA protein assay kit (Pierce™, Thermo Fisher Scientific, Waltham, MA, USA). Equal amounts of protein (30 µg) were separated by SDS-PAGE (Bio-Rad, Hercules, CA, USA) with protein markers (Beyotime, P0071, Shanghai, China; Yeason, 20350ES72, Shanghai, China), then transferred onto polyvinylidene difluoride (PVDF) membranes. After blocking with 5% nonfat milk in TBST, membranes were incubated overnight at 4 °C with primary antibodies: GRP78 (Proteintech, 10768-1-AP, Rosemont, IL, USA), P-IRE1 (CST, 3294S, Danvers, MA, USA), Bax (CST, 2772S, Danvers, MA, USA), Bcl-2 (CST, 2876S, Danvers, MA, USA), ATF4 (CST, 11815S, Danvers, MA, USA), CHOP (CST, 2895S, Danvers, MA, USA), p-eIF2$\alpha$ (CST, 9721S, Danvers, MA, USA), and Caspase-12 (CST, 9822S, Danvers, MA, USA). Membranes were then incubated with secondary antibody (Abcam, ab205718, 1:10,000, Cambridge, UK) for 60 min. Protein bands were detected and quantified using the Odyssey Clx infrared imaging system (LI-COR Biosciences, Lincoln, NE, USA).

Mice tissue samples were washed in PBS (Beyotime, ST476, Shanghai, China) for 30 min and then dehydrated through graded ethanol (Sigma-Aldrich, 459836-100ML). The tissues were cleared in xylene (Sigma-Aldrich, 339056, St. Louis, MO, USA), and stored at –20 °C. After sectioning (4 µm), sections were treated with 3% hydrogen peroxide (Sigma-Aldrich, 88597, St. Louis, MO, USA) for 10 min, followed by antigen retrieval in citrate buffer (Sigma-Aldrich, C9999, St. Louis, MO, USA) and blocking with 3% bovine serum albumin (Servicebio, GC305010-50g, Wuhan, China). Primary antibodies against Caspase-12 (Abcam, ab23352, Cambridge, UK), GRP78 (Cell Signaling Technology, 3183, Danvers, MA, USA), and P-IRE1 (Abcam, ab124945, Cambridge, UK) were incubated overnight at 4 °C. After washing, sections were incubated with secondary antibody (Cell Signaling Technology, 7074, Danvers, MA, USA) for 30 min at 37 °C. Further, 3,3’-diaminobenzidine (Sigma-Aldrich, D4293, St. Louis, MO, USA) was used for color development, followed by counterstaining with Mayer’s hematoxylin (Abcam, ab220365, Cambridge, UK). Sections were dehydrated, cleared in xylene, and mounted with neutral balsam (Sigma-Aldrich, 36406, St. Louis, MO, USA).

The HE staining kit (Vectorlabs, H-3502, Burlingame, CA, USA) was used for HE staining. The fixative volume was 4–10 times the tissue volume. After fixation and dehydration, sections were trimmed, placed face down in labeled embedding cassettes, and redehydrated. Tissues were then embedded, cut into 5-µm sections, and stained with Harris HE.

The In Situ Cell Death Detection Kit, Fluorescein, which is TUNEL technology, POD (Roche, Basel, Switzerland), was performed on sections from three animals per group for tissue injury analysis in the dermatitis site. Tissues were directly analyzed under a fluorescence microscope with an excitation wavelength of 530 nm and a detection wavelength of 630 nm green.

Paraffin sections were successively immersed in xylene I and II (Selleck, 2650-17-1, Houston, TX, USA) for 10 min each. Dehydration was performed through a series of ethanol (Sigma-Aldrich, 459836-100ML) concentrations, washed with PBS (Beyotime, ST476, Shanghai, China), and stained in toluidine blue (Selleck, 92-31-9, Houston, TX, USA), followed by washing in tap water for 2 min to remove excess dye. Differentiation was performed in 95% ethanol, followed by graded ethanol dehydration, xylene clearing, and mounting.

The samples were fixed overnight at 4 °C using PBS (Beyotime, ST476, Shanghai, China) and then washed four times with PBS. Subsequently, the samples were fixed for 2 h at 4 °C using 1% osmium tetroxide solution (Sigma-Aldrich, 251755, St. Louis, MO, USA) and then washed four times with PBS. The samples were dehydrated using an ethanol gradient. After centrifugation at 2000 g for 10 min, bacterial pellets were fixed in 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.4), embedded in 2% agarose, postfixed in 1% osmium tetroxide at 25 °C overnight, dehydrated with gradient ethanol, and embedded in Durcupan resin (Sigma-Aldrich, St. Louis, MO, USA). 55-nm sections were examined using a JEM-1200 transmission electron microscope (JEOL, Tokyo, Japan) equipped with a 4-K Eagle digital camera (FEI, Hillsboro, OR, USA).

Mice were individually placed in a water tank (24 $\pm$ 1 °C, depth 15–20 cm) for 5 min. The total immobility time was recorded. Immobility was defined as floating passively with minimal movement to maintain balance. Mice were dried and warmed immediately after testing.

Mice were suspended vertically by the distal 1 cm of the tail for 5 min. The immobility time was quantified, defined as the absence of active struggling movements.

Mice were placed in the central zone of the EPM (50 cm height) facing a closed arm and allowed to explore freely for 5 min. The percentage of time spent in the open arms was calculated as the primary index of anxiety-like behavior.

Data are presented as mean $\pm$ standard deviation. GraphPad Prism 8.0.2 (GraphPad Software, San Diego, CA, USA) was used for statistical analyses. Normality and homoscedasticity were verified (Shapiro–Wilk/Levene’s tests) for parametric test eligibility. Two-group comparisons used Student’s t-test, and multi-group comparisons used one-way ANOVA with least significant difference (LSD) post-hoc test. Confounding variables (behavioral variability, disease severity) were controlled by covariate adjustment and stratification. For multiple endpoints, false discovery rate (FDR) correction was applied to control false-positive results. p 0.05 was considered statistically significant.

The ERS gene expression in clinical patients between the AD and comorbidities groups was assessed, revealing significantly increased GRP78, P-IRE1, and Caspase-12 expressions in the comorbidities group (p 0.01) (Fig. 1A). The proapoptotic gene of Bax increased in the comorbidities group, whereas the anti-apoptosis gene of Bcl-2 significantly decreased compared with the AD group (p 0.05) (Fig. 1B). Meanwhile, the results of western blot and IHC revealed the significantly increased proteins of GRP78, P-IRE1, and Caspase-12 in the comorbidities group (p 0.05) (Fig. 1C,D). At protein expression levels, Bax was higher in the comorbidities group, whereas Bcl-2 was significantly decreased (p 0.01) (Fig. 1E), indicating that the activated apoptosis pathway may be aggravating underlying cell death in patients with comorbidities.

Fig. 1.

Endoplasmic reticulum stress and apoptosis are increased in atopic dermatitis (AD) patients with psychiatric comorbidities. (A) mRNA levels of GRP78, P-IRE1, and Caspase-12 in skin tissues from AD and comorbid AD patients. (B) mRNA levels of Bax and Bcl-2 in the two patient groups. (C) Protein levels of GRP78, P-IRE1, and Caspase-12 detected by Western blot. (D) Immunohistochemistry of ER stress markers (scale bar = 50 µm) and quantitative analysis. (E) Protein levels of Bax and Bcl-2. Data are mean $\pm$ SD (n = 6). *p 0.05, **p 0.01.

Subsequently, the behavioral and dermatopathological alterations were assessed in the comorbidities, comorbidities + oeNC, and comorbidities + oeMANF groups. The SPT revealed that, by day 28, the behavioral dermatitis scoring presented that the comorbidities + oeMANF group demonstrated milder dermatitis symptoms, with significantly lower dermatitis scores compared with the comorbidities group (p 0.01) (Fig. 2A,B, Table 1). The oeMANF treatment group significantly increased the SPT compared with the comorbidities group, indicating an improvement in behavioral symptoms (p 0.01) (Fig. 2C). HE staining results indicated reduced skin pathology and decreased inflammatory response in the oeMANF treatment group compared with the other two groups (Fig. 2D). Further, toluidine blue staining demonstrated a marked mast cell infiltration reduction and skin hypertrophy alleviation in the oeMANF treatment group (Fig. 2E).

Fig. 2.

Mesencephalic astrocyte-derived neurotrophic factor (MANF) overexpression improved behavior and skin lesions in comorbidities. (A) Dermatitis severity scores in comorbid, comorbid + negative control overexpression (oeNC), and comorbid + MANF overexpression (oeMANF) mice. (B) Scratching bouts in each group. (C) Sucrose preference test for depressive-like behaviors. (D) Hematoxylin and eosin (HE) staining of mouse skin tissues (scale bar = 100 µm). (E) Toluidine blue staining for mast cell infiltration (scale bar = 100 µm). Data are mean $\pm$ SD (n = 6). **p 0.01.

The underlying regulatory role of MANF in skin changes in comorbid mice under ERS and its associated pathways were explored. The oeMANF significantly reduced the expression of ERS-related factors GRP78, P-IRE1, and Caspase-12 at both mRNA and protein levels in an AD mouse model with comorbid mental disorders (comorbidities) (p 0.01) (Fig. 3A,B). Specifically, compared with the control group (comorbidities) and the empty vector control group (comorbidities + oeNC), the oeMANF group demonstrated a significant decrease in the expression of these markers (all p-values 0.001 or lower), indicating that MANF may have a protective effect against ERS, thereby decreasing cell damage or death caused by such stress. Further, immunohistochemical analysis revealed the significantly decreased GRP78, P-IRE1, and Caspase-12 protein in dermatitis tissues (p 0.01) (Fig. 3C–E). In another group of ERS-related factors, qPCR demonstrated the significantly decreased ATF4, CHOP, and P-EIF-2$\alpha$ expression (p 0.01) (Fig. 4A). Further, western blot results revealed a significant reduction of these protein levels in the comorbidities + oeMANF group (p 0.01) (Fig. 4B). The protein reduction is associated with ERS alleviation and improved AD with mental comorbidities. Finally, regarding changes in the morphology of ER, our electron microscopy results indicated significant expansion of the ER in the comorbidities group. Conversely, the swelling of the rough ER in the comorbidities + oeMANF group was significantly improved, with the ER morphology approaching normal and demonstrating a more elongated, tubular structure (Fig. 4C). Notably, quantitative morphometric analysis of ER morphology was not performed in this study; only qualitative evaluation based on transmission electron microscopy (TEM) images was conducted. The reduction in these protein levels indicates that oeMANF alleviates ERS, protects cells, and improves AD and comorbid mental disorders.

Fig. 3.

MANF overexpression alleviated endoplasmic reticulum stress (ERS) and improved pathology in comorbidities models. (A) mRNA levels of GRP78, P-IRE1, and Caspase-12 in mouse skin. (B) Protein levels of ER stress-related markers. (C–E) Immunohistochemistry and quantification of GRP78, P-IRE1, and Caspase-12 (scale bar = 50 µm). Data are mean $\pm$ SD (n = 3). *p 0.05, **p 0.01, ns, no significant difference.

Fig. 4.

MANF overexpression alleviated ERS and improved pathology in comorbidities models. (A) mRNA levels of ERS marker protein (ATF4, CHOP, and P-eIF2$\alpha$) expression levels. (B) Protein of ERS pathway protein expression (ATF4, CHOP, and P-eIF2$\alpha$). (C) Transmission electron microscopy analysis of ER morphology in skin tissue of different groups (scale bar = 500 nm). Data are mean $\pm$ SD (n = 3). **p 0.01, ns, no significant difference.

Further analysis of apoptosis in mouse skin tissue revealed significantly lower skin apoptosis levels in the comorbidities + oeMANF group than in the comorbidities group (p 0.01) (Fig. 5A). Further, qPCR results indicated a significant reduction in Bax gene expression (p 0.01) (Fig. 5B) and a significant increase in Bcl-2 gene expression in the comorbidities + oeMANF group. Conversely, no significant changes in the two genes were observed in the comorbidities + oeNC group. Further western blot assays exhibited a significant reduction in Bax protein expression (p 0.05) and a significant increase in Bcl-2 protein in the comorbidities + oeMANF group (p 0.01) (Fig. 5C). The IHC results revealed no significant differences in protein expression in the comorbidities + oeNC group relative to the comorbidity group. In addition, the IHC results revealed significant changes in the expression of Bax and Bcl-2 proteins in the comorbidities + oeMANF group compared with the comorbidity group (Fig. 5D).

Fig. 5.

MANF overexpression reduces skin cell apoptosis in comorbid mice. (A) Terminal deoxynucleotidyl transferase dUTP Nick-End labeling (TUNEL) staining and apoptotic index (scale bar = 200 µm). (B) mRNA levels of Bax and Bcl-2. (C) Protein levels of Bax and Bcl-2. (D) Immunohistochemistry and quantification of Bax and Bcl-2 (scale bar = 50 µm). Data are mean $\pm$ SD (n = 3). **p 0.01, ns, no significant difference.

To investigate the regulatory effects of ERS on the apoptotic pathway proteins in the skin tissues of comorbid mice, we first used the TUNEL staining to observe skin tissue apoptosis. Skin apoptosis levels were higher in the comorbidities group compared with the AD group (p 0.05) (Fig. 6A) and markedly reduced in the comorbidities + inhibitors group. Subsequently, qPCR results revealed that Bax was increased in the comorbidities group, whereas Bcl-2 was decreased (p 0.01) (Fig. 6B), compared with the AD group. However, Bax was decreased, whereas Bcl-2 was increased in the comorbidities + inhibitors group, consistent with the results in the western blot assay (p 0.05) (Fig. 6C). Further IHC analysis revealed that Bax protein expression was significantly increased, whereas Bcl-2 expression was significantly decreased in the comorbidities group (p 0.05) (Fig. 6D). Compared with the AD group, the comorbidities + inhibitors group demonstrated a decrease in Bax protein and an increase in Bcl-2 protein.

Fig. 6.

4-phenylbutyric acid (4-PBA)-mediated inhibition of ER stress reduces apoptosis in comorbid mice. (A) TUNEL staining and apoptotic index (scale bar = 200 µm). (B) mRNA levels of Bax and Bcl-2. (C) Protein levels of Bax and Bcl-2. (D) Immunohistochemistry and quantification of Bax and Bcl-2 (scale bar = 50 µm). Data are mean $\pm$ SD (n = 3). *p 0.05, **p 0.01.

To explore how MANF regulates ER stress and apoptosis in skin and brain tissues, rescue experiments were performed in four groups: AD, Comorbidity, Comorbidity + oeMANF, and Comorbidity + oeMANF + oeIRE1$\alpha$.

In skin tissues, the Comorbidity group showed markedly upregulated caspase‑12, GRP78, and p-IRE1$\alpha$ (Fig. 7A). MANF overexpression significantly reduced these ER stress markers, whereas IRE1$\alpha$ overexpression abolished this effect. MANF overexpression also improved behavioral performance (Fig. 7B–D) and reduced dermatitis scores and scratching frequency (Fig. 7E), which were reversed by IRE1$\alpha$ overexpression.

Fig. 7.

MANF suppresses IRE1$\alpha$-mediated ER stress and improves behavioral phenotypes. (A) Protein levels of Caspase-12, GRP78, P-IRE1$\alpha$ in mouse skin tissues. (B) Forced swim test for depressive-like behaviors. (C) Tail suspension test for depressive-like behaviors. (D) Elevated plus maze test for anxiety-like behaviors. (E) Dermatitis severity scores and scratching frequency in each group. Data are mean $\pm$ SD (n = 3). **p 0.01, ^#⁢#p 0.01.

In skin tissues, the Comorbidity group displayed severe morphological injury, increased mast cell accumulation, and elevated apoptosis (Fig. 8A–C). MANF overexpression alleviated these pathological changes. Meanwhile, the ratios of cleaved caspase‑3/pro caspase‑3 (Fig. 8D) and Bax/Bcl‑2 (Fig. 8E) were increased in the Comorbidity group; MANF overexpression normalized these ratios, and these effects were blocked by IRE1$\alpha$ overexpression.

Fig. 8.

MANF alleviates skin pathology and inhibits apoptosis via IRE1$\alpha$. (A) HE staining of mouse skin tissues (scale bar = 200 µm). (B) Toluidine blue staining of mast cells (scale bar = 200 µm). (C) TUNEL staining of apoptosis (scale bar = 50 µm). (D) Protein levels of pro-caspase 3 and cleaved-caspase 3 detected by Western blot. (E) Immunohistochemistry and quantification of Bax and Bcl-2 in skin tissues (scale bar = 200 µm). Data are mean $\pm$ SD (n = 3). **p 0.01, ^#⁢#p 0.01.

In brain tissues, the Comorbidity group exhibited elevated caspase‑12, GRP78, and p-IRE1$\alpha$ (Supplementary Fig. 1A), increased cleaved caspase‑3 (Supplementary Fig. 1B), and a higher Bax/Bcl‑2 ratio (Supplementary Fig. 1C). MANF overexpression significantly ameliorated these changes, which were fully reversed by IRE1$\alpha$ overexpression.

Collectively, MANF inhibits IRE1$\alpha$-mediated ER stress and apoptosis in both skin and brain tissues, thereby alleviating skin lesions and psychiatric disorders in the comorbid mouse model.