The cytokine CSBF inhibits the IL-17A and TNF-α inflammatory pathways via SUSD2-ACT1 in keratinocytes and alleviates IMQ-induced psoriasis

Abstract

Overactivation of inflammatory signaling in keratinocytes is critical for psoriatic skin inflammation, but its regulatory mechanisms remain incompletely understood. Here, we demonstrate that the cytokine CSBF inhibits both individual and synergistic proinflammatory signaling induced by IL-17A and TNF-α (IL-17A/TNF-α) in keratinocytes, playing a protective role in psoriatic inflammation. The expression of CSBF was increased in the skin lesions and serum of psoriatic patients, and IL-17A/TNF-α enhanced its production. Csbf deletion exacerbated IMQ-induced psoriasis-like skin inflammation and led to hyperactivation of IL-17A/TNF-α signaling in keratinocytes. The CSBF protein significantly ameliorated psoriatic manifestations and suppressed IL-17A/TNF-α signaling through the receptor SUSD2. Mechanistically, CSBF-SUSD2 competed with TRAF6 and TNFR1 for interaction with ACT1, inhibiting the IL-17A/TNF-α signaling pathway. Overall, the anti-inflammatory cytokine CSBF has the potential to be a therapeutic option for psoriasis by targeting keratinocytes.

Introduction

Psoriasis is a chronic, recurrent, and disfiguring inflammatory skin disease with a global prevalence rate of ~2–3%. Clinically, it is characterized by itchy, red patches with clear edges covered in scales [1, 2]. The main histopathological changes include epidermal thickening, acanthosis, and significant infiltration of leukocytes, particularly neutrophils [2]. Abnormal activation of keratinocytes caused by IL-17 and TNF-α is the core pathological mechanism of psoriasis [3–5]. Current biologic therapies targeting these inflammatory signaling pathways have shown significant efficacy in moderate-to-severe psoriasis [6]. However, these treatments are limited by a high recurrence rate, increased risk of infection, potential harm from exogenous immune complexes, high costs, and relapse after withdrawal [7, 8]. Many studies suggest that keratinocytes initiate psoriasis [9–13]. After stimulation by inflammatory cytokines such as IL-17A and TNF-α, keratinocytes express chemokines that recruit leukocytes to skin lesions. They produce more inflammatory cytokines, resulting in positive feedback that exacerbates and sustains skin inflammation and triggers other complications [14, 15]. Consequently, the timely targeting of keratinocytes and skin inflammation may offer a more effective therapeutic strategy with fewer side effects.

Colon-derived SUSD2 binding factor (CSBF, also known as GPR15LG or C10orf99) is a cytokine identified by our team [16]. CSBF inhibits colon cancer cell proliferation via its receptor SUSD2 [16]. CSBF also exhibits chemotactic activity and mediates lymphocyte trafficking to the epidermis [17] and mucosa [18–20] through the receptor GPR15. Although blocking GPR15 counteracts integrin-dependent T-cell gut homing, making it a promising target for inflammatory bowel disease [21], GPR15 is unrelated to murine psoriasis [22]. The expression of CSBF is significantly upregulated in psoriatic lesions [23], but its effects on skin tissue remain controversial. Some studies have indicated that CSBF alleviates skin diseases, such as by playing a critical role in the establishment and maintenance of the epidermal barrier and promoting wound healing [19, 24, 25]. Compared with that in wild-type mice, skin inflammation is attenuated in CSBF transgenic (genetically modified to overexpress GPR15L (CSBF)) mice following imiquimod (IMQ) challenge [26]. Other reports have suggested that CSBF is pathogenic in skin inflammation, including the promotion of keratinocyte proliferation [27], activation of sensory neurons and mast cell degranulation [28], disruption of the epidermal barrier, and induction of inflammatory gene expression in keratinocytes following transient overexpression [24]. Consequently, the precise functions and mechanisms of CSBF in skin inflammation warrant further investigation and validation.

In this study, we systematically investigated the expression, function, and mechanism of CSBF in psoriatic skin inflammation. Our findings revealed that CSBF was significantly upregulated in keratinocytes from psoriatic lesions and in the serum of patients with psoriasis. IL-17A and TNF-α increased the expression of CSBF. Using gene knockout mice and recombinant protein, we confirmed that CSBF relieved IMQ-induced psoriasis. It significantly inhibited the infiltration of inflammatory cells, especially neutrophils, in skin lesions. The protective CSBF derived from nonhematopoietic cells. Mechanistically, SUSD2 served as the functional receptor of CSBF in keratinocytes. The CSBF-SUSD2 interaction recruited the essential adaptor protein ACT1, thereby blocking the interaction of ACT1 with TRAF6 and TNFR1. This disruption effectively inhibited the inflammatory signaling of IL-17A and TNF-α, ultimately leading to the reduced expression of CXC/CC chemokines and mitigated symptoms of psoriasis. In conclusion, our study reveals that CSBF suppresses IL-17A and TNF-α signaling in keratinocytes and has substantial therapeutic efficacy in an IMQ-induced murine psoriasis model. These findings broaden the understanding of the IL-17A and TNF-α signaling pathways and provide a promising therapeutic strategy for psoriasis.

Results

CSBF is expressed mainly by keratinocytes in psoriatic lesions

Consistent with previous reports [23], analysis of the collected microarray data revealed significant upregulation of CSBF expression in psoriatic skin lesions (Supplementary Fig. 1a), which was confirmed by immunohistochemical staining (Supplementary Fig. 1b). Additionally, CSBF levels in the serum of psoriasis patients were significantly elevated compared with those in healthy controls (Fig. 1a). The expression of Csbf in skin lesions from the IMQ-induced murine psoriasis model was also markedly increased (Supplementary Fig. 1c). Furthermore, single-cell RNA sequencing (scRNA-seq) datasets of human and mouse skin tissues (GSE150672; GSE193350) [29, 30] demonstrated that CSBF was predominantly expressed by keratinocytes (Fig. 1b and Supplementary Fig. 1d) and was significantly more highly expressed in psoriatic lesions than in normal skin tissue (Fig. 1c).

Fig. 1.

CSBF is upregulated in psoriatic keratinocytes and plays a protective role in psoriasis. a The concentration of CSBF in serum from normal controls (n = 10) or psoriatic patients (n = 15). b Expression of CSBF in various cell types of psoriasis skin lesions on the basis of scRNA-seq data (GSE150672). c The expression of CSBF in keratinocytes from normal skin tissue and psoriatic skin lesions (GSE150672). The relative mRNA expression of CSBF in HaCaT cells stimulated with IL-17A (d), TNF-α (e) and IL-17A&TNF-α (f) and the concentration of CSBF in the cell supernatants (g). Representative images (h) and lesion scores (i) of dorsal backs from WT (intraperitoneally injected with normal saline; n = 8), Csbf ^−/− (intraperitoneally injected with normal saline; n = 8), and Csbf ^−/− mice that received mouse CSBF protein treatment (mCSBF, 50 μg/kg; n = 8) and were treated with IMQ for 5 days. H&E staining (j) and epidermal thickness (k) of mouse psoriatic lesions. The scale bar represents 200 μm. Flow cytometry of inflammatory cells in lesions. The numbers in the outlined areas indicate the percentage of CD45⁺ cells among live cells and the percentages of CD11c⁺ DCs, CD11b⁺ Ly6G⁺ neutrophils and CD11b⁺ F4/80⁺ macrophages among CD45⁺ cells (l), with quantification of the percentages of CD45⁺ cells (m), DCs (n), neutrophils (o) and macrophages (p) in live cells. q The relative mRNA expression of Cxcl1, Cxcl2 and Ccl2 in mouse psoriatic lesions. All samples were run in triplicate, and expression was calculated via the delta CT method relative to the housekeeping gene Actb. The data are presented as the means ± SEMs; the symbols indicate individual independent biological experiments; p values were determined via one-way ANOVA. ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. The data are representative of three (d–g) and two (h–q) independent experiments

We also analyzed publicly available microarray data (GSE53552) from psoriasis patients treated with brodalumab, a blocking antibody targeting IL-17RA [31], and found decreased expression of CSBF in the skin lesions after treatment (Supplementary Fig. 1e). These findings indicate that IL-17A signaling may modulate the expression of CSBF. On the other hand, TNF-α can act on keratinocytes to synergistically regulate the pathogenesis of psoriasis with IL-17A. These findings prompted us to investigate whether IL-17A and TNF-α regulate CSBF expression. As shown in Fig. 1d–g, upon stimulation with IL-17A and TNF-α, either individually or in combination, the expression of CSBF was significantly upregulated in HaCaT cells.

Therefore, the results show that CSBF is dramatically expressed by keratinocytes in psoriatic disease and under inflammatory stimulation, indicating that it plays an important role in psoriasis.

CSBF plays a protective role in psoriasis

To clarify the role of CSBF in psoriasis, we generated Csbf knockout (Csbf^−/−) mice on a C57BL/6J genetic background via the CRISPR/Cas9 technique (Supplementary Fig. 1f, g). Csbf^−/− mice were born normally, and the absence of Csbf expression was confirmed (Supplementary Fig. 1h, i). Both Csbf^−/− mice and wild-type (WT) mice were treated with IMQ to induce psoriasis-like lesions. Compared with WT mice, Csbf^−/− mice developed more severe skin lesions, exhibiting increased swelling, redness, and chapping (Fig. 1h, i; Supplementary Fig. 1j, 1k). Histological analysis revealed that the stratum corneum and spinous layers were thicker in the Csbf^−/− mice than in the WT mice, with significantly increased epidermal thickness (Fig. 1j, k; Supplementary Fig. 1l, m). Moreover, infiltrating CD45⁺ leukocytes in the skin lesions of Csbf^−/− mice, including dendritic cells (CD11c⁺), neutrophils (CD11b⁺ Ly6G⁺), and macrophages (CD11b⁺ F4/80⁺), was significantly increased (Fig. 1l–p; Supplementary Fig. 1n–q). These results indicate that Csbf deficiency exacerbates skin inflammation in IMQ-induced psoriasis.

To further confirm the effects of CSBF in psoriasis, eukaryotic recombinant human and mouse CSBF proteins (hereafter referred to as hCSBF and mCSBF, respectively) were expressed and purified (Supplementary Fig. 2a). In WT mice, mCSBF significantly alleviated skin inflammation after intraperitoneal injection at a dose of 25 μg/kg or 50 μg/kg. More significant therapeutic effects were observed at 50 μg/kg, but no effects were observed at 10 μg/kg (Supplementary Fig. 2b–h). Furthermore, the exacerbation of psoriasis in Csbf^−/− mice was significantly reversed by 50 μg/kg mCSBF (Fig. 1h–p). The expression of chemokines, including the neutrophil chemokines Cxcl1 and Cxcl2 and the monocyte/macrophage chemokine Ccl2, was significantly greater in the skin lesions of Csbf^−/− mice than in those of WT mice and was reduced by mCSBF (Fig. 1q). Collectively, these results demonstrate the protective role of CSBF in psoriasis.

We next delineated the cellular sources of CSBF during psoriasis via bone marrow chimeras generated from Csbf^−/− and WT mice (Supplementary Fig. 3a). When the hosts were Csbf^−/− mice, skin inflammatory manifestations were aggravated irrespective of the donor genotype (Supplementary Fig. 3b–j), indicating that CSBF from nonhematopoietic cells is required for psoriasis suppression.

CSBF inhibits individual and synergistic proinflammatory signaling of IL-17A and TNF-α in keratinocytes

As previously mentioned, Cxc/Cc chemokines in mouse skin lesions were increased by Csbf deficiency and decreased by mCSBF protein (Fig. 1q). Furthermore, analysis of publicly available scRNA-seq data from human psoriasis lesions (GSE150672) and IMQ-induced mouse psoriasis-like lesions (GSE193350) revealed that CSBF (Csbf)-positive keratinocytes expressed lower levels of CXC/CC (Cxc/cc) chemokines than CSBF (Csbf)-negative keratinocytes did (Supplementary Fig. 4a). These findings suggest that CSBF may negatively regulate the expression of CXC/CC chemokines, which are target genes of IL-17A signaling [32].

To investigate whether CSBF modulates IL-17A signaling, we performed bulk RNA-seq analysis. IL-17A induced higher levels of Cxc/Cc chemokines in primary keratinocytes derived from Csbf^−/− mice than in those derived from WT mice (Supplementary Fig. 4b). Consistently, IL-17A led to increased Cxcl1 production and increased p65 and p38 phosphorylation in Csbf^−/− mouse keratinocytes, and these effects were reversed by mCSBF (Supplementary Fig. 4c, d).

Furthermore, hCSBF significantly inhibited the production of CXCL1 and CXCL8 as well as the activation of the p65 and p38 pathways in IL-17A-stimulated HaCaT cells (Supplementary Fig. 4e–g). The effects of CSBF on IL-17A signaling in other epithelial cell lines were further examined. Knockout of CSBF in SCC25 cells (a tongue squamous cell carcinoma cell line with high endogenous CSBF) increased IL-17A signaling, which was reversed by hCSBF (Supplementary Fig. 4h). Conversely, stable expression of CSBF in TR146 cells (an esophageal squamous cell carcinoma cell line with low endogenous CSBF) attenuated IL-17A signaling (Supplementary Fig. 4i). Collectively, the above results demonstrate that CSBF inhibits IL-17A signaling.

To confirm that the relieving effects of CSBF on IMQ-induced psoriasis are mediated through the inhibition of the IL-17A signaling pathway, Csbf and Il17a double knockout (Csbf^−/−/Il17a^−/−) mice were generated by crossing Csbf^−/− mice with Il17a^−/− mice [33]. IMQ-induced psoriasis was more severe in Csbf^−/− mice than in WT and Csbf^−/−/Il17a^−/− mice, indicating that CSBF alleviated skin inflammation partially by inhibiting IL-17A signaling. Notably, compared with Il17a^−/−mice, Csbf^−/−/Il17a^−/− mice presented more severe skin inflammation (Supplementary Fig. 5), suggesting that CSBF has additional anti-inflammatory mechanisms beyond the suppression of IL-17A signaling.

To elucidate the mechanisms of CSBF, we performed single-cell RNA sequencing (scRNA-seq) of psoriatic lesions from Csbf^−/− mice treated with or without mCSBF (Fig. 2a). A functional enrichment analysis of the differentially expressed genes (DEGs) identified in keratinocytes was performed. Among the downregulated DEGs, the molecular function “cytokine activity and chemokine receptor binding” was significantly enriched according to Gene Ontology (GO) analysis (see Supplementary Table 1 for the full results) (Fig. 2b). For the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, the “IL-17 signaling pathway” was top-ranked on the basis of p values (Fig. 2c), which reaffirmed that CSBF regulated IL-17A signaling. Importantly, the “TNF signaling pathway” was also among the significantly enriched KEGG pathways (Fig. 2c). IL-17A and TNF-α have crucial synergistic proinflammatory effects, and both signals induce the expression of CXC/CC chemokines by activating the NF-κB and MAPK pathways. Therefore, we next determined the effects of CSBF on TNF-α signaling as well as the synergistic effects of IL-17A and TNF-α signaling in keratinocytes.

Fig. 2.

CSBF inhibits individual and synergistic proinflammatory signaling of IL-17A and TNF-α in keratinocytes. a Clustering results from scRNA-seq analysis of psoriatic lesions in Csbf^−/− mice injected with mCSBF (50 μg/kg) or normal saline (NS) via the Seurat package. GO (b) and KEGG (c) pathway enrichment analysis of the DEGs in keratinocytes from Csbf^−/− mice that were intraperitoneally injected with mCSBF (50 μg/kg) or normal saline and treated with IMQ for 5 days. d Expression of Cxc/Cc chemokines in keratinocytes on the basis of scRNA-seq analysis of psoriatic lesions from Csbf^−/− mice injected with mCSBF (50 μg/kg) or normal saline treated with IMQ for 5 days. Phosphorylation of p65 and p38 (e) as well as the relative mRNA expression of Cxcl1 and Ccl2 (f) in primary keratinocytes isolated from WT or Csbf^−/− mice stimulated with IL-17A, TNF-α, and IL-17A&TNF-α, with or without mCSBF pretreatment. Phosphorylation of p65 and p38 (g), relative mRNA expression of CXCL1 (h) and CXCL8 (i), and concentrations of CXCL1 (j) and CXCL8 (k) in the cell supernatants. HaCaT cells were stimulated with IL-17A, TNF-α, and IL-17A plus TNF-α for the indicated durations, with or without hCSBF pretreatment. l Phosphorylation of p65 and p38 in HaCaT cells stably expressing CSBF or control HaCaT cells (Ctrl) stimulated with IL-17A, TNF-α, and IL-17A&TNF-α for the indicated time points. The data are presented as the means ± SEMs; the symbols indicate individual independent biological experiments; p values were determined via one-way ANOVA (e, f, h–k) or two-sided unpaired Student’s t tests (g, l). ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. The data are representative of three (e, f, h–l) and five (g) independent experiments. ST time, stimulation time

In psoriatic lesions, mCSBF reduced the expression of Cxc/Cc chemokines (Fig. 2d). In primary mouse keratinocytes, knocking out Csbf increased the phosphorylation of p65 and p38 (Fig. 2e) and promoted the expression of Cxcl1 and Ccl2 (Fig. 2f) after stimulation with IL-17A and TNF-α individually or synergistically, which was significantly reversed by mCSBF. In HaCaT cells, hCSBF inhibited the activation of the IL-17A and TNF-α pathways (Fig. 2g) and decreased the expression of CXCL1 and CXCL8 (Fig. 2h–k). Stable expression of CSBF in HaCaT cells also suppressed the phosphorylation of p65 and p38 induced by IL-17A and TNF-α individually or synergistically (Fig. 2l).

Taken together, these findings suggest that CSBF inhibits both individual and synergistic proinflammatory signaling involving IL-17A and TNF-α (hereinafter referred to as IL-17A/TNF-α) in keratinocytes. Given that CSBF suppresses the activation of both IL-17A and TNF-α, it likely targets the upstream components of their signaling pathways.

SUSD2 is the functional receptor of CSBF in keratinocytes and psoriatic skin inflammation

The known receptors of CSBF include SUSD2, GPR15, and MRGX2 [16, 28, 34]. To identify the functional receptor and signaling components of CSBF in keratinocytes, an affinity purification coupled with mass spectrometry (MS)-based approach was employed [35]. Flow cytometry confirmed that hCSBF bound to HaCaT cells (Supplementary Fig. 6a). HaCaT cells were treated with hCSBF, and the whole-cell lysates were purified via affinity chromatography. Three groups were established: (1) hCSBF treatment alone, (2) hCSBF treatment followed by synergistic stimulation with IL-17A and TNF-α (designated hCSBF+stim), and (3) a control group in which hCSBF was added after lysis. The results of MS analysis revealed that SUSD2 was detected in the precipitated complex, but GPR15 and MRGX2 were not detected (Supplementary Table 2). The normalized iBAQ intensity, LFQ intensity, and protein abundance of SUSD2 were low in the control group but increased in both the hCSBF group and the hCSBF + stim group (Fig. 3a, Supplementary Fig. 6b, c), suggesting a specific CSBF-SUSD2 interaction in HaCaT cells. This interaction was further validated by co-immunoprecipitation (Fig. 3b).

Fig. 3.

SUSD2 is the functional receptor of CSBF in keratinocytes and psoriasis skin inflammation. a The scaled abundances of SUSD2 in the MS results. HaCaT cells were treated with human CSBF protein (hCSBF, with a 6 × His-2 × Strep-1 × Flag tag) with or without IL-17A and TNF-α costimulation; the cell lysate was isolated via Flag and Strep immunoprecipitation and analyzed via MS; hCSBF was added only after lysis as a control. b Co-immunoprecipitation of lysates from HaCaT cells. The cells were treated with hCSBF with or without IL-17A and TNF-α costimulation. Whole-cell lysates were immunoprecipitated with anti-Flag M2 magnetic beads and immunoblotted with the indicated antibodies. Phosphorylation of p65 and p38 (c, d) and the relative fold change in the CXCL8 concentration in the supernatants (e) of HaCaT cells transfected with common negative siRNA (si-negative), GPR15 siRNA (siGPR15-1 and siGPR15-2) or SUSD2 siRNA (siSUSD2-1 and siSUSD2-2) and stimulated with IL-17A&TNF-α for 30 min with or without hCSBF pretreatment. Representative images (f) and lesion scores (g) of dorsal backs from WT and Susd2^−/− mice that were intraperitoneally injected with mCSBF (50 μg/kg) or normal saline and treated with IMQ for 5 days (n = 10 animals per group). H&E staining (h) and epidermal thickness (i) of skin lesions from WT and Susd2^−/− mice. The scale bar represents 200 μm. Flow cytometry of inflammatory cells in lesions from WT and Susd2^−/− mice. The numbers in the outlined areas indicate the percentage of CD45⁺ cells among live cells and the percentage of CD11b⁺Ly6G⁺ neutrophils among CD45⁺ cells (j), with quantification of the percentages of CD45⁺ cells (k) and neutrophils (l) in live cells. Phosphorylation of p65 and p38 in primary keratinocytes isolated from WT and Susd2^−/− mice stimulated with IL-17A (m), TNF-α (n), or IL-17A&TNF-α (o) for 30 min, with or without mCSBF pretreatment. Relative mRNA expression of Cxcl1 in primary keratinocytes isolated from WT and Susd2^−/− mice. The differences among the responses to IL-17A, TNF-α, and IL-17A + TNF-α stimulation (p) and the differences among the responses to mCSBF (q) were analyzed. The data are presented as the means ± SEMs; the symbols indicate individual independent biological experiments; p values were determined via one-way ANOVA (a, e, g, i, k, l, q) or two-sided unpaired Student’s t tests (c, d, m–p). ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. The data are representative of three (a–e, m–q) and two (f–l) independent experiments

We next investigated whether SUSD2 mediated the inhibitory effect of CSBF on IL-17A/TNF-α signaling. In the control groups transfected with either common negative siRNA (si-negative) or GPR15-targeting siRNA (siGPR15-1 and siGPR15-2), hCSBF significantly suppressed IL-17A/TNF-α-induced p65 and p38 phosphorylation (Fig. 4c, d; Supplementary Fig. 6d, e) and CXCL8 secretion (Fig. 3e). In contrast, SUSD2 knockdown (siSUSD2-1 and siSUSD2-2) markedly attenuated the inhibitory effects of hCSBF (Fig. 3c–e), demonstrating that SUSD2 is the functional receptor of CSBF in HaCaT cells. To further validate the role of SUSD2 in psoriasis, Susd2 knockout (Susd2^−/−) mice were generated (Supplementary Fig. 6f–i). Consistent with the findings in Csbf^−/− mice, Susd2^−/− mice presented more severe skin inflammation than WT mice did. While mCSBF alleviated inflammation in WT mice, it did not affect Susd2^−/− mice (Fig. 3f–l). Compared with WT keratinocytes, Susd2^−/− mouse keratinocytes presented increased activation of IL-17A/TNF-α signaling and elevated Cxcl1 expression (Fig. 3m–p). Notably, mCSBF inhibited IL-17A/TNF-α signaling in WT but not Susd2^−/− keratinocytes (Fig. 3m–o, q). Together, these findings indicate that CSBF suppresses IL-17A/TNF-α signaling in keratinocytes via SUSD2, thereby exerting a protective role in psoriasis.

Fig. 4.

The suppressive effect of CSBF-SUSD2 on IL-17A/TNF-α signaling depends on ACT1. a The scaled abundances of ACT1 in the MS results. b Co-immunoprecipitation of lysates from HaCaT cells. The cells were treated with hCSBF with or without IL-17A and TNF-α co-stimulation. Whole-cell lysates were immunoprecipitated with anti-Flag M2 magnetic beads and immunoblotted with the indicated antibodies. Phosphorylation of p65 and p38 (c) and the concentration of CXCL8 in the supernatant (d) of HaCaT cells transfected with common negative siRNA (si-negative) or ACT1 siRNA (siACT1-1 and siACT1-2) and stimulated with IL-17A, TNF-α, and IL-17A&TNF-α for 30 min. Phosphorylation of p65 and p38 in HaCaT cells transfected with common negative siRNA (si-Negative) or ACT1 siRNA (siACT1-1 and siACT1-2) and stimulated with IL-17A (e), TNF-α (f), and IL-17A&TNF-α (g) for 30 min with or without hCSBF pretreatment; the quantitative statistical results for p-p65/t-p65 and p-p38/t-p38 (h) are shown. i The concentration of CXCL8 in the supernatant of HaCaT cells transfected with common negative siRNA (si-negative) or ACT1 siRNA (siACT1-1 and siACT1-2) and stimulated with IL-17A, TNF-α, and IL-17A&TNF-α with or without hCSBF pretreatment. The data are presented as the means ± SEMs; the symbols indicate individual independent biological experiments; p values were determined via one-way ANOVA. ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. The data are representative of three independent experiments

To identify the cellular sources by which SUSD2 plays a protective role in psoriasis, we generated bone marrow chimeras (Supplementary Fig. 7a). The severity of skin inflammation in recipient mice was similar between those receiving WT and Susd2^−/− mouse bone marrow transplants (Supplementary Fig. 7b–g), demonstrating that nonhematopoietic cell-derived SUSD2 is essential for psoriasis alleviation.

The CSBF-SUSD2-mediated suppression of IL-17A/TNF-α signaling depends on the key adaptor ACT1

ACT1, also known as TRAF3IP2 or CIKS [36], is a critical cytoplasmic adaptor molecule in IL-17A signaling [37–39] that activates the NF-κB and MAPK pathways by recruiting TRAF6 [39, 40]. MS analysis revealed that the iBAQ intensity, LFQ intensity, and abundance of ACT1 were high in the hCSBF+stim group but undetectable in the control and hCSBF groups, suggesting that ACT1 might be involved in the downstream signaling of CSBF-SUSD2 under IL-17A and TNF-α stimulation (Fig. 4a; Supplementary Fig. 6b, c). The interaction between CSBF-SUSD2-ACT1 was confirmed by co-immunoprecipitation (Fig. 4b). ACT1 participates in the signaling pathways of TNF receptor (TNFR) superfamily members, including CD40 and BAFFR [41–43], but no studies have explicitly addressed its role in TNF-α-TNFR1 signaling. To elucidate the role of ACT1 in IL-17A/TNF-α signaling and to determine whether it affects the inhibitory effect of CSBF on the IL-17A/TNF-α pathways, HaCaT cells were transfected with siRNAs targeting ACT1 (Supplementary Fig. 8a, b). Knockdown of ACT1 significantly attenuated the activation of p65 and p38, as well as the expression of CXCL8, in response to IL-17A and TNF-α individually or synergistically (Fig. 4c, d). More importantly, hCSBF markedly reduced the phosphorylation of p65 and p38 and the expression of CXCL8 induced by IL-17A/TNF-α in the si-negative group, but these inhibitory effects were abolished after ACT1 knockdown (Fig. 4e–i). Collectively, these results suggest that ACT1 is involved in TNF-α signaling and is indispensable for the anti-inflammatory effects of CSBF-SUSD2 in keratinocytes.

CSBF-SUSD2 competes with TNFR1 and TRAF6 for interaction with ACT1, which relies on the N-terminal TRAF-binding domain of ACT1 and the intracellular segment of SUSD2

To validate the interaction between ACT1 and SUSD2, we transfected HeLa cells with ACT1 and SUSD2 plasmids. These proteins interacted significantly when the cells were stimulated with IL-17A/TNF-α following hCSBF treatment (Supplementary Fig. 8c, d). In HaCaT cells, endogenous SUSD2 interacted with ACT1 after hCSBF treatment, and this interaction was further enhanced by IL-17A/TNF-α stimulation. SUSD2 did not interact with either the IL-17A receptor IL17RA or the TNF-α receptor TNFR1 (Fig. 5a, b). ACT1 interacted with TNFR1 after stimulation with TNF-α alone and in synergy with IL-17A, which was attenuated by hCSBF (Fig. 5b). While the interaction between IL-17RA and ACT1 remained unaffected, the interaction between ACT1 and the downstream molecule TRAF6 was weakened by hCSBF stimulated with IL-17A alone or in combination with TNF-α (Fig. 5b, c).

Fig. 5.

CSBF-SUSD2 competes with TNFR1 and TRAF6 to interact with ACT1. Co-immunoprecipitation of lysates from HaCaT cells after stimulation with IL-17A, TNF-α, or IL-17A + TNF-α with or without hCSBF pretreatment. Whole-cell lysates were immunoprecipitated with anti-SUSD2 (a), anti-ACT1 (b), or anti-TRAF6 (c) antibodies and immunoblotted with the indicated antibodies. Co-immunoprecipitation of cell lysates from primary keratinocytes isolated from WT and Susd2^−/− mice stimulated with IL-17A (d), TNF-α (e), or IL-17A&TNF-α (f) with or without mCSBF pretreatment. Whole-cell lysates were immunoprecipitated with anti-ACT1 and immunoblotted with the indicated antibodies. g Co-immunoprecipitation of cell lysates from primary keratinocytes isolated from WT and Susd2^−/− mice stimulated with IL-17A with or without mCSBF pretreatment. Whole-cell lysates were immunoprecipitated with an anti-TRAF6 antibody and immunoblotted with the indicated antibodies. h Domain structures of the indicated ACT1 truncations. TB1, the TRAF-binding domain at the N-terminus; HLH helix-loop-helix domain; Ubox U-box–like region; TB2 the TRAF-binding domain at the middle. Co-immunoprecipitation of lysates from HeLa cells overexpressing HA-SUSD2 together with FLAG-ACT1 or various ACT1 mutants stimulated with IL-17A (i, k) or TNF-α (j, l) with or without hCSBF pretreatment. Whole-cell lysates were immunoprecipitated with anti-Flag (i, j) or anti-HA (k, l) and immunoblotted with the indicated antibodies. Co-immunoprecipitation of lysates from HeLa cells overexpressing FLAG-ACT1 together with HA-SUSD2 or HA-SUSD2 lacking an intracellular segment (HA-SUSD2-dIC) stimulated with IL-17A, TNF-α, and IL-17A&TNF-α with or without hCSBF pretreatment. Whole-cell lysates were immunoprecipitated with anti-Flag (m) or anti-HA (n) and immunoblotted with the indicated antibodies. The data are representative of three independent experiments

Consistently, in WT mouse keratinocytes, mCSBF inhibited the interaction between ACT1 and TRAF6 following IL-17A stimulation or IL-17A and TNF-α costimulation. It also repressed the interaction between ACT1 and TNFR1 after TNF-α stimulation or IL-17A and TNF-α costimulation. However, mCSBF had no such effects on Susd2^−/− mouse keratinocytes, further indicating that SUSD2 is the functional receptor for CSBF (Fig. 5d–g). Moreover, a more pronounced interaction was detected between ACT1 and TRAF6 (Fig. 5d, f, g), as well as between ACT1 and TNFR1 (Fig. 5e, f), in Susd2^−/− mouse keratinocytes than in WT mouse cells. These results suggest that the absence of SUSD2 enhances both the interaction between ACT1 and TRAF6 in IL-17A signaling and the interaction between ACT1 and TNFR1 in TNF-α signaling, which is consistent with the increased activation of the IL-17A/TNF-α pathways in Susd2^−/− mouse keratinocytes (Fig. 3m–o).

ACT1 contains several key domains, including a helix-loop-helix (HLH) domain [36], a U-box-like (Ubox) region [44], a SEFIR domain [39], and two TRAF-binding domains located at the N-terminus (TB1) and middle (TB2) [36]. To delineate the specific domain mediating its interaction with SUSD2, a series of ACT1 truncations were constructed (Fig. 5h). Full-length ACT1 (WT) or each deletion mutant was co-expressed with SUSD2. Consistent with previous reports, the SEFIR domain is essential for the interaction between ACT1 and IL-17RA [39] (Fig. 5i). Notably, the interaction between ACT1 and SUSD2 was undetectable in cells expressing the ACT1 dTB1 mutant, underscoring the critical role of the N-terminal TRAF-binding domain (TB1) in their interaction (Fig. 5i–l). This domain was also essential for the interaction between ACT1 and TRAF6, as well as TNFR1 (Fig. 5i, j). Additionally, the interaction between SUSD2 and ACT1 relied on the intracellular segment (IC) of SUSD2 (Fig. 5m, n). These results suggest that CSBF-SUSD2 inhibits IL-17A/TNF-a signaling by competing with TRAF6 and TNFR1 for binding the TRAF-binding domain at the N-terminus of ACT1 and that the cytoplasmic tail of SUSD2 is indispensable for CSBF-SUSD2 signaling.

The CSBF protein has significant therapeutic effects and good safety in BALB/c mice with IMQ-induced psoriasis

We evaluated the therapeutic efficacy of mCSBF in an IMQ-induced psoriasis model in BALB/c mice. Consistent with findings in C57BL/6 J mice, the 10 μg/kg dose showed limited efficacy, whereas skin inflammation was visibly alleviated when the dose reached 25 μg/kg, with greater stability and pronounced improvement at 50 μg/kg (Fig. 6a–h). No adverse effects of mCSBF were observed on the histological structure of major organs, including the liver, kidney, and spleen (Fig. 6i), or on serum biochemical parameters, such as blood urea nitrogen, serum creatinine, alanine aminotransferase, and aspartate transaminase (Fig. 6j), indicating the good safety of mCSBF. Collectively, these results indicate that CSBF has therapeutic effects on IMQ-induced psoriasis in a mouse model with different genetic backgrounds, reaffirming its protective role and therapeutic potential in psoriasis.

Fig. 6.

The CSBF protein shows obvious therapeutic effects and good safety in IMQ-induced psoriasis in BALB/c mice. Representative images (a) and lesion scores (b) of dorsal backs from BALB/c mice intraperitoneally injected with normal saline or the indicated doses of mCSBF and treated with IMQ for 5 days (n = 5 animals per group). H&E staining (c) and epidermal thickness (d) of skin lesions from BALB/c mice. The scale bar represents 200 μm. Flow cytometry of inflammatory cells in lesions. The numbers in the outlined areas indicate the percentage of CD45⁺ cells among live cells and the percentage of CD11b⁺ Ly6G⁺ neutrophils and CD11b⁺ F4/80⁺ macrophages among CD45⁺ cells (e), with quantification of the percentages of CD45⁺ cells (f), neutrophils (g), and macrophages (h) in live cells. i H&E staining of livers, kidneys, and spleens from BALB/c mice that received the indicated doses of mCSBF. The scale bar represents 100 μm. j Blood urea nitrogen (BUN), serum creatinine (Scr), alanine aminotransferase (ALT), and aspartate transaminase (AST) levels in the serum of BALB/c mice that received mCSBF treatment. Representative images (k) and lesion scores (l) of dorsal backs from BALB/c mice that received IL-17A neutralizing antibody (anti-IL-17A) or mCSBF and were treated with IMQ for 5 days (n = 5 animals per group). H&E staining (m) and epidermal thickness (n) of skin lesions from BALB/c mice treated with an anti-IL-17A neutralizing antibody (anti-IL-17A) or mCSBF. The scale bar represents 200 μm. Flow cytometry of inflammatory cells in lesions. The numbers in the outlined areas indicate the percentage of CD45⁺ cells among live cells and the percentages of CD11b⁺Ly6G⁺ neutrophils and CD11b⁺F4/80⁺ macrophages among CD45⁺ cells (o), with quantification of the percentages of CD45⁺ cells (p), neutrophils (q), and macrophages (r) in live cells. The data are presented as the means ± SEMs; the symbols indicate individual independent biological experiments; p values were determined via one-way ANOVA. ns (not significant), *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. The data are representative of two independent experiments

Here, we systematically compared the curative effects of mCSBF with those of a commercially available IL-17A neutralizing antibody (anti-IL-17A). Compared with anti-IL-17A, mCSBF had comparable efficacy in alleviating psoriatic skin inflammation, as evidenced by improvements in the appearance of skin lesions, reductions in epidermal thickness, and decreases in inflammatory cell infiltration (Fig. 6k–r). There was no statistically significant difference between mCSBF and anti-IL-17A, suggesting their comparable psoriasis-relieving effects. Nonetheless, compared with the saline group, the mCSBF group exhibited slightly greater improvement and greater statistical significance. The above results demonstrate that CSBF displays better therapeutic efficacy and favorable safety in a BALB/c mouse model of psoriasis.

Discussion

CSBF is a cytokine first identified in 2014 [16], but studies on this cytokine are limited. Its functions and mechanisms in skin inflammation are not fully understood. Here, we demonstrated that CSBF was significantly increased in the keratinocytes of psoriasis lesions and in the serum of psoriatic patients. IL-17A/TNF-α (IL-17A and/or TNF-α) upregulated CSBF expression in keratinocytes. Using Csbf^−/− mice and recombinant CSBF proteins, we determined the protective role of CSBF in an IMQ-induced murine psoriasis model with different genetic backgrounds. Mechanistically, CSBF significantly inhibited IL-17A/TNF-α signaling in keratinocytes through SUSD2, which interacted with the key adaptor ACT1, disrupting its binding to both TRAF6 and TNFR1 and thereby suppressing the activation of IL-17A/TNF-α. The interaction between SUSD2 and ACT1 depends on the N-terminal TRAF-binding domain of ACT1 and the intracellular segment of SUSD2. In brief, CSBF inhibits IL-17A/TNF-α signaling in keratinocytes and is an anti-inflammatory cytokine in psoriasis.

The upregulation of anti-inflammatory molecules is not uncommon at inflammatory sites. For example, both IL-10 and IL-1 receptor antagonist (IL-1Ra) are natural anti-inflammatory molecules, and their levels are increased during infection and inflammation, thereby maintaining the balance of the immune system [45–47]. CSBF is significantly upregulated when the skin undergoes pathological changes such as psoriasis and effectively suppresses IL-17A/TNF-α signaling. The excessive secretion of CSBF might be a self-saving strategy employed by keratinocytes in response to inflammatory signals and pathological stimuli. Both human CSBF and mouse Csbf were significantly upregulated in psoriatic skin lesions, but Csbf, as well as Cxcl1, Cxcl2, and Ccl2, were expressed at higher levels in mouse keratinocytes (GSE193350 and GSE165021) than in human keratinocytes (GSE150672). This might be due to the different biological characteristics of skin inflammatory responses between humans and mice or sequencing batch effects.

Our results indicate that SUSD2 is the functional receptor of CSBF in keratinocytes that suppresses IL-17A/TNF-α activation. However, SUSD2 is also expressed in other cells, including CD8⁺ T cells [48], fibroblasts, and dendritic epidermal T cells (DETCs). SUSD2 modulates IL-2 signaling and suppresses CD8⁺ T-cell antitumor immunity by interacting with IL-2Rα through its extracellular Sushi domain [48]. scRNA-seq data analysis (GEO datasets, GSE150672 and GSE193350) revealed that SUSD2 was expressed in fibroblasts, which might be an additional nonhematopoietic target for CSBF. The GEO dataset GSE160476 revealed that SUSD2 was highly expressed in DETCs, which are a specialized γδ T-cell population residing in the murine epidermis. DETCs maintain the epidermal barrier and control inflammatory responses [49–51], and the role of CSBF-SUSD2 in DETCs requires further investigation.

ACT1 plays important roles in multiple signaling pathways, and its function is context-dependent. In B cells, ACT1 negatively regulates the signaling of CD40 and BAFFR [41, 43]. Conversely, in epithelial cells, ACT1 promotes CD40L-induced NF-κB activation [42]. ACT1 is essential for IL-17A signaling [39, 44, 52], but its effects on TNF-α signaling remain unclear. Two studies conducted in mouse embryonic fibroblasts demonstrated a trend toward reduced expression of TNF-α target genes following Act1 knockout, although the inhibition was not significant [39, 53]. Here, our results indicate that ACT1 is crucial for both individual and synergistic signaling involving IL-17A and TNF-α. ACT1 contains two TRAF-binding domains [36]. In CD40 signaling, ACT1 interacts with TRAF3 [42], and in IL-17A signaling, ACT1 initiates downstream signaling by interacting with TRAF6 [35, 39, 40]. Consistent with existing reports, our results showed that ACT1 interacted with TRAF6 following IL-17A stimulation or in combination with TNF-α. In addition, upon stimulation with TNF-α or in combination with IL-17A, ACT1 significantly interacted with TNFR1. TNF-α-TNFR1 initiates signal transduction through TRAF2 and TRAF5 [54–56]. However, whether the interaction between ACT1 and TNFR1 is related to TRAF molecules requires further investigation. ACT1 participates in IL-1/Toll-mediated signaling through interaction with TRAF6, and the amino-terminal half of ACT1 is required for its interaction with the TRAF domain [57]. Our study suggests that the N-terminal TRAF-binding domain of ACT1 mediates its interaction with TRAF6, TNFR1, and SUSD2. Consequently, CSBF-SUSD2 inhibits IL-17A/TNF-α signaling by competing with TRAF6 and TNFR1 for binding to ACT1 through this domain.

One of the major limitations of the use of biologic agents that target IL-17A signaling is the high relapse rate after withdrawal [7]. Anti-IL-17A therapies, such as brodalumab, reduce the expression of CSBF in skin lesions [31], which limits IL-17A/TNF-α signaling in keratinocytes and might be one of the underlying mechanisms for disease relapse. As an endogenous bioactive molecule that targets keratinocytes, CSBF offers unparalleled advantages in terms of specificity and safety.

In summary, our study demonstrated that CSBF is an anti-inflammatory cytokine in psoriasis. SUSD2 is its functional receptor. CSBF-SUSD2 competes with TRAF6 and TNFR1 for interaction with ACT1, suppressing IL-17A/TNF-α signaling in keratinocytes. Our findings expand the understanding of the regulatory mechanism of IL-17A/TNF-α signaling in keratinocytes and highlight the therapeutic application of CSBF recombinant protein in psoriasis.

Materials and methods

Human samples

A total of 15 paraffin-embedded sections were obtained, including 10 from patients with psoriasis (5 males and 5 females, aged 16–74 years) and 5 from normal skin controls (3 males and 2 females, aged 20–43 years), for immunohistochemical analysis. Additionally, a total of 25 serum samples were collected, comprising 15 from psoriatic patients (6 males and 9 females, aged 27–60 years) and 10 from healthy controls (5 males and 5 females, aged 33–60 years). All the samples were obtained from the Department of Dermatology at Peking University First Hospital. Written informed consent was obtained from all participants. This study was performed in accordance with the Declaration of Helsinki principles and was approved by the Research Ethics Board of Peking University First Hospital (Approval No. 2022--307).

Mice

Csbf^−/− mice and Susd2^−/− mice on a C57BL/6J genetic background were generated at the Shanghai Research Center as southern model organisms via CRISPR/Cas9 technology. The knockout strategies are shown in Fig. S1 and Fig. S2. Il17a^−/− mice were purchased from the Laboratory Animal Resources Center of Tsinghua University and generated via gene targeting in embryonic stem cells with replacement of part of exon 2 with a luciferase-IRES-eGFP cassette [33]. C57BL/6J mice and BALB/c mice were purchased from Vital River. All the mice were bred and maintained under specific-pathogen-free conditions at Peking University Health Science Center, with an ambient temperature of 21 ± 2 °C and a 12-h light‒dark cycle. Mice aged 9–12 weeks were used for all the experiments. The animal experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committees at Peking University Health Science Center (Approval No. 2022--135).

Cell culture

The cells were cultured at 37 °C and 5% CO₂. For the cell lines, HaCaT cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Cat# SH30022.01) supplemented with 10% fetal bovine serum (FBS). HeLa cells were cultured in RPMI-1640 medium (HyClone, Cat# SH30027.01) supplemented with 10% FBS. HEK293F cells were grown in SMM 293-T II expression medium (Sino Biological, Cat# M293T II-1 L). Primary mouse keratinocytes were isolated from neonatal pups and cultured as described for primary mouse keratinocyte isolation and culture below.

Transfection

For siRNA transfection, HaCaT cells were transfected via Lipofectamine RNAiMAX (Thermo Fisher, Cat# 13778150) according to the manufacturer’s instructions. For plasmid transfection, HeLa cells were transfected via jetPRIME (Polyplus, Cat# 101000046) following the manufacturer’s instructions, whereas HEK293F cells were transfected via PEI reagent (MCE, Cat# HY-K2014) according to the manufacturer’s instructions.

Purification of recombinant human and mouse CSBF proteins

Following the transfection of the plasmids encoding HSF-human CSBF/mouse CSBF (6 × His-2 × Strep-1 × Flag-tagged) into HEK293F cells, the culture supernatant was harvested after 3 days. The recombinant protein, which was tagged with a His tag, was purified via a HisTrap Excel column (Cytiva, Cat# 29048586) and an ÄKTA go chromatography system (Cytiva). Prior to purification, the HisTrap Excel column was equilibrated with a balanced buffer consisting of 20 mM sodium phosphate (pH 7.4) and 500 mM NaCl. The column was then washed with a solution containing 50 mM imidazole in the balanced buffer and eluted with 500 mM imidazole in the balanced buffer. To further increase the purity of the recombinant protein and remove residual imidazole, the eluted samples were processed via a centrifugal filter (30 kDa molecular weight cutoff, Millipore), and the lower filtrates were retained (the molecular weight of the CSBF protein was ~10 kDa). The lower filtrates were subsequently loaded onto a centrifugal filter (3 kDa molecular weight cutoff, Millipore), washed several times with PBS, and concentrated. The protein concentration was measured via NanoDrop (Thermo Fisher) and BCA protein assay kits (Thermo Fisher, Cat# A55864). The production and purity of the CSBF recombinant protein were confirmed via Coomassie blue staining (Solarbio, Cat# P1305). The recombinant proteins were stored at −80 °C.

Imiquimod (IMQ)-induced psoriasis-like skin inflammation model

For the IMQ-induced psoriasis model, 9–12-week-old mice with a weight variation of ±1 g were treated with 5% IMQ (Aldara cream, 3 M Health Care, Cat# H20160079) to induce psoriasis-like skin inflammation. The dorsal fur of each mouse was removed the day before modeling to ensure a consistent application area and extent. A fixed quantity of 45 mg of IMQ cream was applied to the back skin of the mice at the same time each day for five consecutive days. The mice in the control group were treated with an equal amount of Vaseline in the same manner. For the mCSBF treatment model, the mice received a daily intraperitoneal injection of the indicated recombinant mCSBF protein prior to IMQ application. For the anti-IL-17A treatment model, an anti-IL-17A neutralizing antibody (eBioscience, Cat# 16-7173-85) was administered intraperitoneally at a dose of 10 mg/kg on the first day of modeling.

Appearance scores of psoriasis-like skin lesions in mice

In this study, the severity of psoriasis-like skin lesions in mice was assessed via a scoring system ranging from 0 to 4: 0 indicates no skin lesions, 1 indicates mild skin lesions, 2 indicates moderate skin lesions, 3 indicates severe skin lesions, and 4 indicates very severe skin lesions. The scoring criteria were based on the following five parameters: (1) the extent of the affected area; (2) the quantity and thickness of scales; (3) the color and degree of redness and swelling of the lesions; (4) abnormal changes in skin texture, such as wrinkling, dimpling, or protrusion; and (5) the degree of compromised skin integrity, such as ulceration or breakage.

HE staining of paraffin sections

Skin lesions (1 cm × 1 cm) were fixed in 4% paraformaldehyde for 24 h at room temperature. The samples were subsequently dehydrated through a gradient of 60–100% ethanol, followed by immersion in xylene and hot wax (65 °C) to obtain paraffin-embedded tissue. The paraffin-embedded sections were 4 μm thick and incubated at 60 °C for 3 h before staining. The sections were covered with hematoxylin stain for 5 min, rinsed with tap water for 10 min, and then stained with 1% eosin for 4 min. Changes in the structure of the skin lesions were observed and photographed under a microscope. Twenty random sites of epidermal thickness were measured, and the average value was used as the epidermal thickness of the lesion for statistical analysis.

Primary keratinocyte isolation and culture

Primary mouse keratinocytes were isolated from the skin of neonatal mice aged 1–3 days. Skin tissues were incubated in a digestion mixture containing 1 mg/mL Dispase II (Sigma‒Aldrich, Cat# D4693-1G) prepared in DMEM (HyClone, Cat# SH30022.01) overnight at 4 °C. The epidermal and dermal layers were then separated via sterile forceps. The epidermis was cut into small pieces and further digested in a 0.05% trypsin solution at 37 °C for 10 min with two gentle shaking steps. The tissue suspension was filtered through a 70-μm cell strainer (Falcon). The isolated cells were washed twice with PBS and resuspended in 154CF complete medium (Gibco, Cat# M154CF500) containing 1% HKGS (Gibco, Cat# S0015) and 0.05 mM calcium chloride. The cells were seeded in 6-well plates at a density of 1 × 10⁶ cells per well. The medium was replaced after 24 h and then every other day. The cells reached ~80% confluence within 5 days and were then stimulated with recombinant mCSBF protein (100 ng/mL) and recombinant mIL-17A (PeproTech, Cat# 210-17, 100 ng/mL)/mTNF-α (PeproTech, Cat# 315-01 A, 50 ng/mL) according to experimental needs.

Bone marrow chimeras

For the generation of bone marrow chimeras, 6-week-old recipient mice were lethally irradiated with X-rays (9.5 Gy, divided into two fractions of 5 Gy or 4.5 Gy). The mice were subsequently reconstituted with 1 × 10⁷ bone marrow leukocytes from the indicated donors via intravenous (i.v.) injection. After 8 weeks of initial reconstitution, the chimeric mice were challenged with IMQ.

Flow cytometry

The skin lesion tissues were minced and transferred to a digestion mixture containing 1 mg/mL dispase II (Sigma‒Aldrich, Cat# D4693-1G), 0.5 mg/mL DNase I (Solarbio, Cat# D8071), and 2.5 mg/mL collagenase D (Roche, Cat# 11088866001). The digestion was performed at 37 °C with constant shaking at 120 rpm for 1 h. The tissue suspension was filtered through a 100-mesh sieve and centrifuged at 1600 rpm for 5 min at 4 °C. The cells were resuspended in 4 mL of RPMI-1640 medium (HyClone, Cat# SH30027.01). Next, 4 mL of 80% Percoll (GE Healthcare, Cat# 17-0891-09) was added, the mixture was gently mixed, and an additional 4 mL of 80% Percoll was carefully layered on the bottom of the cell suspension. The mixture was subsequently centrifuged at 2500 rpm for 30 min with the brake off. The leukocyte-enriched middle layer was collected, washed with 10 mL of PBS, and centrifuged at 1600 rpm for 5 min at 4 °C. The cells were then resuspended in 5 mL of PBS, filtered through a 70 μm cell strainer, and stored temporarily at 4 °C. For flow cytometry antibody staining, a single-cell suspension (~1 × 10⁶ cells) was added to a 96-well plate and centrifuged at 1800 rpm for 5 min at 4 °C. Afterward, the cells were stained with the viability dye Zombie Aqua Fixable Viability Kit (BioLegend, Cat# 423101) at room temperature for 15 min. Following a single wash with PBS, 200 μL of an antibody mixture containing anti-CD45 (BioLegend, Cat# 103128, Cat# 103106), anti-CD11c (BioLegend, Cat# 117329, Cat# 117308, Cat# 117317), anti-CD11b (BioLegend, Cat# 101206), anti-Ly6G (BioLegend, Cat# 164506, Cat# 127624, Cat# 127617), and anti-F4/80 antibodies (BioLegend, Cat# 123118, Cat# 123116) was added to each well. The cells were incubated in the dark at 4 °C for 30 min. After that, the cells were centrifuged at 1800 rpm for 5 min at 4 °C, and the supernatant was discarded. The cells were then washed with PBS, resuspended in 200 μL of PBS per well, and transferred to flow cytometry tubes for further analysis.

RNA isolation and quantitative RT‒PCR

For cellular experiments, HaCaT cells or primary mouse keratinocytes were pretreated with recombinant hCSBF protein (100 ng/mL) for 5 h and then stimulated with recombinant hIL-17A (PeproTech, Cat# 200-17, 100 ng/mL)/hTNF-α (PeproTech, Cat# 300-01 A, 50 ng/mL) for 24 h. Total RNA was extracted from skin biopsies or cells via TRIzol reagent (Gibco, Cat# 15596018). One microgram of total RNA was used for reverse transcription of cDNA via a reverse transcription kit (TransGen, Cat# AU341-02). Real-time PCR was performed with SYBR Green Endpoint. All samples were run in triplicate, and expression was calculated via the delta CT method relative to the housekeeping gene GAPDH or Actb. The primer sequences are provided in Supplementary Table 3.

Western blot and immunoprecipitation (IP)

For western blot analysis, the cells were pretreated with hCSBF protein (100 ng/mL) for 5 h, followed by stimulation with IL-17A (100 ng/mL)/TNF-α (50 ng/mL) for 30 or 60 min. For IP, the cells were pretreated with hCSBF protein (100 ng/mL) for 5 h and then stimulated with IL-17A (100 ng/mL)/TNF-α (50 ng/mL) for 30 min. Then, the cells were harvested and lysed in buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, protease inhibitors, and phosphatase inhibitors. Proteins were separated by SDS‒PAGE and electrophoretically transferred onto nitrocellulose membranes (Cytiva), after which the membranes were blotted with the indicated antibodies. For coimmunoprecipitation, the cells were lysed in buffer composed of 30 mM Tris (pH 7.4), 120 mM NaCl, 2 mM KCl, 2 mM EDTA, 10% glycerol, 10 mM chloroacetamide, 1% n-dodecyl-β-D-maltoside (DDM), protease inhibitors (Roche, Cat# 11697498001), and phosphatase inhibitors (Roche, Cat# 4906845001). The cell lysates were then incubated with 2 μg of IP antibody and precipitated with Protein G Sepharose beads (Cytiva, Cat# 17061801). The immunoprecipitated proteins were separated by SDS‒PAGE, and immunoblotting was performed with the indicated antibodies. For affinity immunoprecipitation, the cell lysates were incubated with anti-FLAG M2 magnetic beads (Sigma‒Aldrich, Cat# M8823). Following precipitation, the samples were subjected to gel electrophoresis, and immunoblotting was performed with the indicated antibodies.

Binding assay of hCSBF to HaCaT cells

The culture medium of HaCaT cells was replaced with serum-free base medium 1 h prior to the experiment. A total of 1 × 10⁶ cells were then resuspended in 1 mL of PBS containing either 5 μg of human CSBF protein (fused with a 6 × His-2 × Strep-1 × Flag tag) or 5 μg of 3 × Flag peptide (serving as a control; Sigma‒Aldrich, Cat# F4799). The samples were incubated on a rotating platform at 4 °C for 30 min. Following this step, the cells were washed twice with PBS, and APC-conjugated anti-Flag antibody was added. The cells were incubated in the dark at 4 °C for 30 min. Afterward, the cells were washed three times with PBS, resuspended in 200 μL of PBS, and analyzed via flow cytometry.

Tandem affinity purification with hCSBF

The culture medium of HaCaT cells was replaced with serum-free base medium 1 h prior to the experiment. The experiment contained three groups: Group 1 was incubated with recombinant hCSBF protein (500 ng/mL, 5 μg protein per 10 cm dish) for 1 h; Group 2 was incubated with hCSBF protein (500 ng/mL) for 1 h, followed by stimulation with hIL-17A (100 ng/mL) and hTNF-α (50 ng/mL) for 30 min; and Group 3 served as a control and received no stimulation but 3 μg hCSBF protein postcell lysis. The cells were collected with lysis buffer composed of 30 mM Tris (pH 7.4), 120 mM NaCl, 2 mM KCl, 2 mM EDTA, 10% glycerol, 10 mM chloroacetamide, 1% DDM, protease inhibitors, and phosphatase inhibitors. The lysates were then incubated on ice for 30 min [35]. The cell lysates were cleared via centrifugation at 13,000 rpm for 15 min at 2 °C. For the first affinity chromatography step, the cell lysates were incubated with 50 μL of anti-FLAG M2 magnetic beads (Sigma‒Aldrich) overnight at 4 °C. The beads were then washed three times with TBS buffer. Next, 250 μL of a 3 × Flag peptide mixture (150 μg/mL, Sigma‒Aldrich, Cat# F4799) was added, and the mixture was incubated at 4 °C for 1 h to elute the target protein from the magnetic beads. The supernatant was collected, and the elution steps were repeated. Following the first affinity chromatography step, the supernatant was incubated with 50 μL of streptavidin magnetic beads (Vazyme, Cat# N512-01) at 4 °C for 2 h. The supernatant was discarded, and the beads were washed five times with TBST buffer. The magnetic beads were resuspended in 50 μL of 2 × nonreducing protein loading buffer and boiled for 3 min. The supernatant was collected and subjected to SDS‒PAGE, and the electrophoresis was stopped when the bromophenol blue indicator reached 1 cm from the top of the separating gel. The gel was stained with Coomassie Brilliant Blue and sent for MS analysis.

Mass spectrometry

The MS analysis for this study was conducted via the MS testing platform of the Peking University Health Science Center. For filter-aided sample preparation (FASP) digestion, protein samples (200 μg) from each group were processed according to the FASP method with a minor revision. Initially, 10 μL of 50 mM Tris-(2-carboxyethyl) phosphine (TCEP) in water was added to the protein samples (50 μL) and incubated at 67 °C for 10 min. After cooling to room temperature, 10 μL of 50 mM iodoacetamide was added and incubated in darkness for 30 min. Subsequently, the samples were transferred to Vivacon 500 filtration tubes (Cat No. VNO1HO2, Sartorius Stedim Biotech, UK), mixed with 150 μL of 8 M urea in 0.1 M Tris/HCL (pH 8.5), and centrifuged at 14,000 × g for 15 min at 20 °C. This step was repeated twice. The samples were then washed three times with 200 μL of 50 mM NH₄HCO_3. Finally, 4 μg of trypsin (Promega, Madison, WI) dissolved in 100 μL of 50 mM NH₄HCO₃ was added for digestion. The protein-to-enzyme ratio was 50:1. The samples were incubated overnight at 37 °C. The released peptides were collected via centrifugation and vacuum-dried. The digested peptides were loaded on a C18 precolumn (Thermo Scientific) and separated via nano-LC‒MS/MS on an easy-LC nano-HPLC system (Thermo Scientific). Mass spectrometric analysis was performed via an Orbitrap Exploris 480 with FAIMS Pro (Thermo Scientific, Bremen, Germany). The raw files were analyzed via Proteome Discoverer v2.5 (Thermo Scientific) or MaxQuant 1.6. MS/MS spectra were searched against the UniProt Human database (UP000005640). The search parameters were set as follows: fixed modification of cysteine residues (+57 Da), variable modification of methionine oxidation (+16 Da), and full trypsin cleavage, at most one missed tryptic cleavage site, 10 ppm error tolerance in MS and 0.02 Da error tolerance in MS/MS. False discovery rates were obtained via Percolator 2, selecting identifications with a q value equal to or less than 0.01.

Single-cell RNA sequencing (scRNA-seq) of mouse skin lesions

Psoriasis lesions from Csbf^−/− mice, with or without CSBF protein treatment, were collected for scRNA-seq. The skin lesion tissues were digested with 1 mg/mL dispase II (Sigma-Aldrich), 0.5 mg/mL DNase I (Solarbio), and 2.5 mg/mL collagenase D (Roche) at 37 °C with shaking at 120 rpm for 1 h. Single-cell suspensions were subsequently obtained after being filtered through a 70 μm cell strainer, after which the adhesive and dead cells were removed. The scRNA-seq in this study was completed by Beijing Capital Biotechnology Co., Ltd. and was performed on the 10 × Genomics Chromium platform. scRNA-seq data were aligned to the hg38 genome and processed via CellRanger (version 6.0). The matrix was subsequently imported into the Seurat package (version 4.0) for further analysis. The cells were filtered according to the following parameters: nFeature_RNA > 500 and nFeature_RNA < 8000 and percent.mito < 10. Finally, a total of 28,417 cells remained (Csbf^−/−_NS: 16,364, and Csbf^−/−_mCSBF: 12,053). A total of 2000 highly variable genes were generated for performing principal component analysis dimension reduction. The datasets were integrated via FindIntegrationAnchors. Single-cell clustering was visualized via uniform manifold approximation and projection with a resolution parameter of 0.3. The marker genes for identifying keratinocytes include Krt5, Krt14, Krtdap, Krt1, Krt10, Cdk1, and Pcna. DEGs between keratinocytes from Csbf ^−/−_NS and those from Csbf^−/−_mCSBF were identified via FindMarkers (min.pct = 0.05). GO and KEGG enrichment analyses were performed on the DEGs.

Bulk RNA-seq analysis

Primary keratinocytes isolated from WT and Csbf^−/− mice were stimulated with IL-17A (100 ng/mL) for 8 h. Total RNA was then extracted via TRIzol reagent (Gibco) and subjected to RNA-seq analysis. Library construction and sequencing were performed by Shanghai Majorbio Biopharm Technology Co., Ltd., according to the manufacturer’s instructions (Illumina, San Diego, CA).

Statistical analysis

Statistical analysis was performed with GraphPad Prism 9.4. The data are presented as the means ± standard errors of the means (SEMs). The data are presented as the means ± SEMs. Significance was determined by two-tailed Student’s t tests or one-way ANOVA followed by Tukey’s test (for all pairwise comparisons) or Dunnett’s test (for comparisons against a single control group). P values < 0.05 were considered statistically significant (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001). The methods used to analyze the statistical significance (p-value), number of biological replicates, and number of repeat experiments are reported in the respective figure legends.

Author contributions

WH, PW, RL, XL, KZ, and TL designed the study; XL, KZ, XY, and YC carried out the experiments and performed the data analysis; SH, YH, HD, TL, WD, XM, and JZ helped with the experiments and manuscript writing; PW and YH helped with the bioinformatics analysis; RL and KZ provided samples of patients; XL wrote the manuscript; WH, PW, RL, KZ and SH revised the manuscript; and WH, PW and RL supervised the study. All the authors have read and approved the article.

Data availability

The raw scRNA-seq data reported in this paper have been deposited in the Genome Sequence Archive (GSA) at the National Genomics Data Center, China National Center for Bioinformation, Beijing Institute of Genomics, Chinese Academy of Sciences, and are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human under the accession number HRA006747. The raw MS data have been deposited in OMIX (https://ngdc.cncb.ac.cn/omix: accession no. OMIX005869) [58]. Published datasets reanalyzed in this study are available under accession codes GSE150672 [29] and GSE193350 [30]. This paper does not report the original code. All other data associated with this study are presented in the main text or the supplementary materials.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41423-025-01325-3.