Transdermal siRNA delivery via biomineralized nanoparticle-incorporated microneedles modulates cuproptosis-ferroptosis interaction for psoriasis therapy

Abstract

Psoriasis is a chronic immune-mediated skin disorder driven by abnormal keratinocyte proliferation and inflammation, and the dysregulation of copper transport is increasingly recognized as a key metabolic driver and potential therapeutic target in psoriasis. Herein, we identified solute carrier family 31 member 1 (SLC31A1) as a pivotal molecular switch connecting cuproptosis and ferroptosis, two interconnected forms of regulated cell death that synergistically promote psoriatic pathology. Upregulated SLC31A1 induces copper accumulation and elevates α-ketoglutarate (α-KG), activating KDM5B-dependent histone demethylation and repressing FTH1 transcription, thereby amplifying ferroptotic damage and inflammation. To therapeutically target this axis, we developed a nanoparticle-incorporated microneedle system (CaP-siSlc31a1@MN) enabling localized, efficient and minimally invasive siRNA delivery through the psoriatic barrier. The dissolvable microneedles with favorable mechanical performance ensured precise epidermal deposition, while biomineralized calcium phosphate (CaP) nanoparticles facilitated intracellular uptake and siRNA release. In vitro and in vivo studies confirmed that CaP-siSlc31a1@MN effectively silenced Slc31a1, inhibited cuproptosis and ferroptosis, suppressed IL17A-driven inflammation and restored epidermal homeostasis. Overall, this study introduces a first-in-class transdermal gene-silencing nanoplatform that integrates metabolic regulation with anti-inflammatory therapy for precision psoriasis treatment.

Graphical abstract

Highlights

• •SLC31A1 was identified as a key regulator linking cuproptosis and ferroptosis.
• •Silencing SLC31A1 restored FTH1, blocking Cu-overload-driven metabolic inflammation.
• •Biomineralized CaP-siSlc31a1 were integrated into microneedles for siRNA delivery.
• •The microneedle inhibited epidermal deposition, Th17 infiltration and IMQ dermatitis.
• •This work introduced a translatable microneedle-based gene-silencing strategy.

1. Introduction

Psoriasis is a chronic, immune-mediated inflammatory skin disease resulting from complex interactions between genetic susceptibility and environmental triggers, affecting approximately 2-3% of the global population [1,2]. Beyond its high prevalence, psoriasis imposes a substantial physical, psychological and socioeconomic burden, severely impairing overall life quality [3]. Histopathologically, it is characterized by hyperproliferation and aberrant differentiation of keratinocytes, accompanied by extensive infiltration of immune cells, particularly those driven by the IL23/Th17 axis, into the dermis and epidermis [4,5]. This immunometabolic imbalance sustains a vicious “inflammation-proliferation” feedback loop, fueling chronic lesion recurrence [6,7]. Although current treatment strategies, including topical corticosteroids, phototherapy and systemic biologics, have revolutionized disease management [8,9], they remain limited by several fundamental challenges: 1) the incomplete understanding of the upstream drivers linking inflammatory signaling to cellular metabolic reprogramming; 2) poor transdermal bioavailability due to the abnormally thickened stratum corneum of psoriatic plaques, which severely restricts local drug penetration [10]. Consequently, existing therapies often require long-term systemic administration, leading to reduced efficacy and undesirable side effects [11] (see Scheme 1)

Scheme 1.

Schematic illustration of CaP-siSlc31a1@MN microneedle for psoriasis therapy via the SLC31A1/α-KG/KDM5B/H3K4me3/FTH1 axis. In psoriatic lesions, SLC31A1 upregulation induces copper accumulation, elevating α-KG levels and activating the demethylase KDM5B, which reduces H3K4me3 enrichment at the FTH1 promoter. This epigenetic repression downregulates FTH1, triggers ferroptosis, and exacerbates skin inflammation. The CaP-siSlc31a1@MN delivered CaP nanoparticles into localized subcutaneous regions, achieving effective cellular uptake and SLC31A1 silencing, restoring copper homeostasis and simultaneously inhibiting cuproptosis and ferroptosis, thereby alleviating IMQ-induced psoriasiform dermatitis.

Recent advances in the biology of regulated cell death have reshaped our understanding of the molecular events underlying psoriasis. In addition to impaired apoptosis, emerging inflammatory cell death modalities, such as necroptosis, pyroptosis and ferroptosis, have been implicated in disease progression [12,13]. Ferroptosis, an iron-dependent, lipid peroxidation-driven form of regulated cell death, has attracted particular attention due to its dual role in oxidative stress amplification and immune activation [14]. Excessive ferroptosis in keratinocytes promotes the release of damage-associated molecular patterns (DAMPs) and proinflammatory cytokines, further activating Th17-mediated responses [15,16]. Pharmacological inhibition of ferroptosis has been shown to alleviate epidermal hyperplasia and reduce inflammatory cytokine expression in preclinical models, suggesting that targeting ferroptotic pathways may hold therapeutic potential [17]. However, the upstream regulatory cues and their interplay with other death modalities, particularly in the context of chronic cutaneous inflammation, remain largely undefined.

Parallel to ferroptosis, copper-induced cell death, or cuproptosis, has recently emerged as a mitochondria-associated death pathway triggered by intracellular copper overload. Accumulating evidence highlights a critical role for dysregulated metal ion homeostasis in chronic inflammatory diseases [18]. Clinical studies have reported significantly elevated serum copper levels in psoriatic patients, correlating positively with disease severity and systemic inflammatory burden [19,20]. Mechanistically, excessive copper disrupts Fe-S cluster assembly and induces mitochondrial protein lipoylation, leading to oxidative stress, metabolic dysfunction and inflammatory activation, pathological features that closely mirror those observed in psoriasis [21]. Intriguingly, cuproptosis and ferroptosis exhibit substantial mechanistic convergence. Copper accumulation can potentiate ferroptosis by depleting glutathione (GSH) and destabilizing glutathione peroxidase 4 (GPX4) [22], whereas ferroptotic stress reciprocally alters the expression of copper transport and storage proteins [23]. This interdependence suggests the existence of a coordinated cell death network integrating metal ion metabolism and inflammatory signaling. Whether such crosstalk contributes to psoriatic pathogenesis, and the molecular mediators bridging these two pathways, remains an open question.

Among the copper-handling proteins, solute carrier family 31 member 1 (SLC31A1) serves as the primary high-affinity importer of Cu⁺ ions and a key regulator of intracellular copper homeostasis [24,25]. Our group has focused on elucidating the pathological role of SLC31A1-related signaling pathways and demonstrated that SLC31A1 expression is markedly upregulated in psoriatic lesions. Notably, SLC31A1 acts as a pivotal molecular link between cuproptosis and ferroptosis in psoriasis, offering new insights into how disturbances in metal homeostasis exacerbate inflammatory and metabolic dysfunction [26]. In light of these findings, silencing SLC31A1 using small interfering RNA (siRNA) offers a precise therapeutic strategy to simultaneously modulate cuproptotic and ferroptotic activity. Nevertheless, the clinical translation of siRNA therapeutics is hindered by inherent barriers such as enzymatic degradation, poor cellular uptake and inefficient penetration across psoriatic hyperkeratosis [27]. Although lipid nanoparticles and viral vectors have been explored as siRNA delivery platforms, issues such as cytotoxicity, immunogenicity and poor tissue selectivity persist [28,29]. In contrast, calcium phosphate (CaP) nanoparticles, inspired by natural biomineralization, have re-emerged as an attractive non-viral carrier owing to their excellent biocompatibility, pH-responsive degradation and ability to facilitate endosomal escape. CaP nanoparticles thus represent a safe and efficient vector for siRNA delivery and gene silencing [30,31]. To address the cutaneous barrier, microneedle (MN) technology offers a minimally invasive transdermal delivery approach that can perforate the stratum corneum, enabling direct deposition of therapeutic agents into the viable epidermis and dermis [32,33]. Integrating CaP nanocarriers with dissolvable MNs synergistically overcomes the dual obstacles of transdermal and intracellular delivery, enabling localized, efficient and sustained gene silencing with minimal systemic exposure.

Herein, we elucidate the previously unrecognized crosstalk between cuproptosis and ferroptosis in psoriasis and identify SLC31A1 as a central regulator of this pathogenic interface. We further design a CaP-siSlc31a1-loaded dissolving microneedle system (CaP-siSlc31a1@MN) capable of precise transdermal siRNA delivery and localized gene silencing. This intelligent delivery platform effectively penetrates the psoriatic stratum corneum, achieves effective cellular uptake, and restores metal ion homeostasis by suppressing SLC31A1-mediated dual cell death activation. Our study not only establishes the SLC31A1/α-KG/KDM5B/H3K4me3/FTH1 axis as a key mechanistic link between cuproptosis and ferroptosis in psoriasis but also introduces a minimally invasive, nanocarrier-assisted microneedle therapy that integrates mechanistic precision with translational feasibility, offering a new paradigm for the treatment of inflammatory skin diseases.

2. Results and discussion

2.1. Cuproptosis and ferroptosis are co-activated in psoriasis

To validate the co-activation of cuproptosis and ferroptosis in psoriasis and their clinical relevance, we first analyzed RNA sequencing data from psoriasis patients. In human psoriasis lesions, we observed a marked dysregulation of cuproptosis-related genes, characterized by significant downregulation of genes such as ATP7A, ATP7B, LIPT1, FDX1 and LIAS, alongside a pronounced upregulation of the copper influx transporter SLC31A1 (Fig. 1A), suggesting aberrant cuproptosis activation in psoriatic skin. Consistent with the human data, IMQ-induced psoriatic mice also displayed a marked disturbance in cutaneous copper metabolism. Quantitative analysis revealed significantly elevated epidermal copper levels (Fig. 1B) accompanied by upregulated SLC31A1 expression (Fig. 1D), suggesting enhanced copper influx into keratinocytes (KCs). Meanwhile, lipoylated dihydrolipoamide S-acetyltransferase (DLAT), a key component of the pyruvate dehydrogenase complex, was notably increased (Fig. 1C). Conversely, the expression of iron-sulfur cluster assembly proteins such as FDX1 and LIAS were markedly reduced (Fig. 1D), reflecting disrupted mitochondrial redox homeostasis and impaired Fe-S cluster biogenesis. Given that FDX1 catalyzes the reduction of Cu²⁺ to Cu⁺ and participates in lipoylation reactions, its suppression likely represents a feedback adaptation to copper-induced stress. Further corresponding analysis confirmed decreased expression of iron-sulfur cluster-related genes (Fdx1, Lias) and copper exporters (Atp7a, Atp7b), with elevated Slc31a1 (Fig. S1), collectively affirming disrupted copper homeostasis and cuproptosis activation.

Fig. 1.

(A) Comparison of qRT-PCR heatmaps of cuproptosis related genes between normal people and psoriasis (GSE13355). (B) Copper levels in ear skin tissues from BALB/c mice with IMQ-induced psoriasis-like dermatitis on the ear (n = 3). (C) The lipoylation level of DLAT in ear skin tissues from different mouse groups, and (D) Protein levels of SLC31A1, iron-sulfur cluster proteins (FDX1 and LIAS). (E) ROS levels in ear skin tissues from different mouse groups. Flow cytometric quantification of (F) Fe²⁺ accumulation and (G) lipid peroxidation products levels in epidermal cells from mouse ears across groups (n = 3). (H) GSH levels in ear skin tissues from different mouse groups (n = 3). Spatial transcriptomics analysis (GSE225475) of (I) SLC31A1 and (J) FTH1 in psoriasis. (K) Spatial transcriptomic analysis of the correlation between SLC31A1 and FTH1 in psoriasis (GSE225475). IHC detection and statistical analysis of (L) SLC31A1 and (M) FTH1 expression in skin tissues from healthy individuals and psoriasis (n = 5). Data are presented as the mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

Given the close interaction between cuproptosis and ferroptosis, we next assessed ferroptosis-related indicators in the mouse model. The psoriatic lesions exhibited a prototypical ferroptotic phenotype, characterized by a remarkable accumulation of reactive oxygen species (ROS) (Fig. 1E), significantly increased levels of ferrous ions and lipid peroxidation products (Fig. 1F and G), as well as a profound depletion of intracellular GSH (Fig. 1H). These biochemical alterations collectively reflect a state of oxidative imbalance and lipid peroxidative stress, hallmarks of ferroptosis activation. Building on our prior finding that SLC31A1 negatively regulates FTH1, a key ferroptosis-suppressing protein responsible for iron sequestration and detoxification [34,35], we next examined their relationship in the psoriatic context. Strikingly, FTH1 expression was markedly diminished at both the mRNA and protein levels in IMQ-induced psoriatic mice (Fig. S2). The downregulation of FTH1 may weaken intracellular iron buffering capacity, facilitating Fe²⁺ overload and promoting lipid peroxidation, which further reinforces ferroptotic cell death. To verify whether this regulatory axis is conserved in humans, we conducted spatial transcriptomic analysis of psoriatic patient skin. The data revealed a pronounced inverse correlation between SLC31A1 and FTH1 expression, with SLC31A1 upregulated and FTH1 markedly suppressed within the epidermal compartment (Fig. 1I–K). This spatially restricted pattern implies that KCs are the major cellular site of copper-iron dyshomeostasis, linking aberrant SLC31A1-driven copper influx to ferroptotic sensitivity. Furthermore, immunohistochemical (IHC) staining validated these transcriptional trends, showing strong SLC31A1 positivity and reduced FTH1 immunoreactivity in both human psoriatic epidermis (Fig. 1L and M) and IMQ-treated mouse skin (Figs. S3 and S4). Collectively, these findings reinforced the co-activation of cuproptosis and ferroptosis as a pathogenic feature of psoriasis and identify SLC31A1-mediated copper influx as a pivotal upstream trigger driving this dual-cell-death mechanism.

2.2. Fabrication of CaP-siSlc31a1 for efficient cellular uptake and gene silencing

Given that the SLC31A1-FTH1 axis bridges cuproptosis and ferroptosis, we next aimed to therapeutically modulate this pathway through nanocarrier-mediated delivery of Slc31a1 siRNA (siSlc31a1). The siSlc31a1-loaded CaP nanoparticles (CaP-siSlc31a1) were synthesized via a biomineralization approach [36,37]. Briefly, siSlc31a1 was added to Dulbecco's Modified Eagle Medium (DMEM) containing bovine serum albumin (BSA), followed by the addition of CaCl₂. The abundant phosphate groups in DMEM coordinated with Ca²⁺, leading to the gradual formation of fine calcium phosphate nuclei that served as nucleation sites for biomineralization. With extended incubation, these nuclei matured into CaP-siSlc31a1 nanoparticles. To assess the siRNA loading capacity of the CaP nanocarrier, FAM-labeled siSlc31a1 was employed to measure fluorescence intensity within the nanoparticles using a fluorospectrophotometer, enabling determination of drug-loading content and encapsulation efficiency. At low feeding concentrations of FAM-siSlc31a1 (1 and 2 μM), the nucleation rate exceeded the growth rate due to sufficient carrier availability, resulting in nearly 100% encapsulation efficiency as most siRNA molecules were entrapped within the nanostructure (Fig. 2A). However, when the siRNA concentration was increased to 4 μM, the active binding sites on the carrier surface became saturated, leading to a decrease in the amount of siRNA bound per nanocarrier. Taking into account both loading capacity and encapsulation efficiency, 2 μM was identified as the optimal siSlc31a1 loading concentration for subsequent experiments (Fig. 2B). Transmission electron microscopy (TEM) revealed that the obtained CaP-siSlc31a1 were uniform and spherical, while dynamic light scattering (DLS) indicated a hydrodynamic diameter of approximately 97.4 nm (Fig. 2C). Energy-dispersive X-ray spectroscopy (EDS) confirmed the co-localization of C, O, N, P, and Ca elements, verifying the successful synthesis of CaP-siSlc31a1 (Fig. S5).

Fig. 2.

(A) Encapsulation efficiency and (B) loading capacity of CaP-FAM-siSlc31a1 (n = 3). (C) TEM image and hydrodynamic diameter of CaP-siSlc31a1. (D) Fluorescence intensity of free FAM-siSlc31a1 and CaP-FAM-siSlc31a1 under different pH conditions. (E) Fluorescence images and (F) TEM micrographs of CaP-siSlc31a1 after incubation at various pH values. (G) Confocal images showing intracellular localization and lysosomal escape of FAM-labeled CaP-siSlc31a1 in KCs, and (H) time-dependent Pearson's correlation coefficients between FAM and LysoTracker signals (n = 3). (I) Representative fluorescence images and (J) quantitative analysis of cellular uptake of FAM-siSlc31a1, CaP-FAM-siSlc31a1 and Lipo-FAM-siSlc31a1 in KCs (n = 3). Relative mRNA expression of Slc31a1 in KCs treated with different formulations under (K) normal and (L) Il17a-induced inflammatory conditions (n = 3). (M) Cell viability of KCs after treatment with various formulations (n = 3). Data are presented as mean ± SD. ns, not significant; *p < 0.05, ***p < 0.001.

We next investigated the acid-responsive degradation behavior using FAM-siSlc31a1-loaded nanoparticles. The CaP-FAM-siSlc31a1 nanoparticles were incubated under different pH conditions, and fluorescence intensity was measured to evaluate CaP degradation and the corresponding siRNA release. At pH 7.5, CaP-FAM-siSlc31a1 exhibited markedly lower fluorescence intensity at 520 nm compared with free FAM-siSlc31a1 (Fig. 2D), suggesting that CaP encapsulation induced a fluorescence quenching effect. In contrast, at pH 5.5, fluorescence intensity increased substantially, approaching that of free siRNA, which indicated acid-triggered degradation of CaP and subsequent siRNA release. Moreover, no free siRNA was detected in the supernatant of centrifuged CaP-FAM-siSlc31a1 (Fig. 2E), confirming efficient encapsulation and high stability under physiological conditions. Consistently, TEM analysis of CaP-siSlc31a1 incubated in PBS (pH 5.5) for 0.5 and 2 h revealed progressive structural disintegration, further validating its pH-responsive behavior (Fig. 2F).

Considering the demonstrated acid-responsive degradability of CaP-siSlc31a1, we further investigated its intracellular fate to determine whether this pH-sensitive property facilitates efficient cellular internalization and lysosomal escape. Primary mouse KCs were incubated with FAM-labeled nanoparticles for 0.5-6 h and visualized using confocal laser scanning microscopy (CLSM). The intracellular green fluorescence gradually intensified, peaking at 2 h, indicating efficient cellular internalization (Fig. 2G). The cellular uptake of nanoparticles was also quantitatively analyzed by flow cytometry. Liposomes, which are widely used as gene delivery vectors in clinical applications, were employed as a positive control (Lipo-siSlc31a1) to evaluate transfection efficiency. As expected, CaP-siSlc31a1 exhibited a comparable level of cellular uptake to that of Lipo-siSlc31a1 at 2 h (Fig. 2I and J), as determined by measuring the mean fluorescence intensity of treated cells. In addition, lysosomes were stained with LysoTracker Red, and Pearson's correlation coefficients were calculated to assess lysosomal escape (Fig. 2G and H). During the first hour, CaP-siSlc31a1 nanoparticles were primarily localized within lysosomes, however, after 6 h, the overlap between green (nanoparticles) and red (lysosomes) fluorescence decreased, and the correlation coefficient declined from 0.84 to 0.38, demonstrating effective lysosomal escape. This lysosomal escape capability of CaP-siSlc31a1 can be attributed to proton uptake and subsequent nanoparticle decomposition [38].

Owing to effective cellular uptake and timely lysosome escape, CaP-siSlc31a1 achieved substantial Slc31a1 mRNA knockdown compared with Lipo-siNT, CaP-siNT, and Lipo-siSlc31a1 (Fig. 2K). Under Il17a-induced psoriatic inflammatory conditions, gene silencing efficiency was further enhanced (Fig. 2L), which can be ascribed to cytokine-induced increases in cellular permeability that promote nanoparticle uptake and siRNA release [39]. Interestingly, CaP-siSlc31a1 exhibited markedly lower cytotoxicity than Lipo-siSlc31a1 (Fig. 2M). Even at equivalent siRNA concentrations, CaP-siSlc31a1 maintained over 80% cell viability, indicating excellent biocompatibility. This reduced toxicity is likely due to the biodegradable and non-cationic nature of the CaP carrier, which minimizes membrane disruption compared with liposomal systems.

2.3. CaP-siSlc31a1 suppresses psoriatic inflammation and cuproptosis/ferroptosis via epigenetic festoration of FTH1 expression

To further investigate the therapeutic potential and mechanism of CaP-siSlc31a1 in psoriasis, we established a cellular model mimicking the psoriatic inflammatory microenvironment by treating mouse primary KCs with Il17a, a cytokine known to initiate and exacerbate psoriasis [40]. Il17a stimulation markedly upregulated the mRNA expression of multiple psoriasis-related mediators, including Cxcl1, Il6, Cxcl3, Tnf, Il23 and S100a8 (Fig. 3A), confirming the successful establishment of the model. Treatment with CaP-siSlc31a1 significantly suppressed the expression of these inflammatory factors, and nearly restored the Il17a-induced downregulation of FTH1 to normal levels (Fig. 3B and C). At the functional level, CaP-siSlc31a1 treatment also led to a notable reduction in intracellular Fe²⁺ accumulation and lipid peroxidation (Fig. 3D and E), both of which are key hallmarks of ferroptosis. These findings demonstrated that CaP-siSlc31a1 effectively inhibits SLC31A1 and alleviates psoriasis-related inflammation by restoring FTH1 expression and suppressing ferroptosis.

Fig. 3.

(A) Heatmap showing the transcriptional levels in mouse primary KCs under different treatment conditions. (B) FTH1 protein and (C) mRNA expression levels in mouse primary KCs after 24 h of treatment with Il17a, CaP-siNT, or CaP-siSlc31a1 (n = 3). (D) Intracellular Fe²⁺ levels, (E) lipid peroxidation levels and (F) α-KG levels in mouse primary KCs from various groups (n = 3). (G) Heatmap illustrating the transcriptional profiles of inflammatory factors and chemokines across different treatment groups. (H) H3K4me3 protein levels in mouse primary KCs under different treatment conditions. (I) Predicted H3K4me3 binding sites in the FTH1 promoter region identified using the Cistrome Data Browser, which were used to design primers for ChIP assays. (J) ChIP-qPCR analysis showing H3K4me3 enrichment at the FTH1 promoter in human primary KCs treated with or without CaP-siSlc31a1. Data are presented as mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001.

We next explored the mechanism by which CaP-siSlc31a1 regulates FTH1 expression. α-ketoglutarate (α-KG), a key metabolite linking cuproptosis and ferroptosis, has been reported to suppress FTH1 expression [21]. In Il17a-stimulated KCs, α-KG levels were elevated (Fig. 3F), and exogenous α-KG treatment further enhanced the expression of inflammatory cytokines, including Tnf, Cxcl1, Cxcl3, Il23, Il6 and Il1b (Fig. 3G). Importantly, CaP-siSlc31a1 intervention not only reduced α-KG levels but also reversed the α-KG-induced upregulation of psoriasis-related inflammatory factors, indicating its ability to influence psoriasis development and progression through modulation of α-KG. Furthermore, qRT-PCR and Western blot analyses revealed that increasing concentrations of α-KG treatment led to a progressive decrease in Fth1 expression in KCs (Fig. S6). Notably, CaP-siSlc31a1 effectively counteracted the α-KG-mediated downregulation of Fth1 expression (Fig. 3G), confirming that CaP-siSlc31a1 alleviates psoriatic-like inflammation and ferroptosis in KCs by suppressing α-KG and upregulating FTH1.

It has been reported that α-KG, as an essential cofactor of lysine demethylase 5B (KDM5B), suppresses FTH1 transcription by reducing histone H3K4 trimethylation (H3K4me3) at the FTH1 promoter region [41]. We hypothesized that CaP-siSlc31a1 might exert its therapeutic effects in psoriasis through negative regulation of this pathway. To validate this mechanism, we knocked down KDM5B in KCs (Fig. S7) and observed a marked increase in FTH1 expression at both mRNA and protein levels (Fig. S8), accompanied by elevated global H3K4me3 levels (Fig. S9), indicating that KDM5B negatively regulates FTH1 expression by modulating H3K4me3 modification. Furthermore, we examined the relationship between H3K4me3 and α-KG under inflammatory conditions. Treatment with Il17a and α-KG reduced H3K4me3 protein levels in primary mouse KCs, while CaP-siSlc31a1 treatment attenuated this decrease (Fig. 3H), supporting the mechanism by which CaP-siSlc31a1 modulates H3K4me3 via α-KG. To determine whether H3K4me3 directly binds to the FTH1 promoter, we performed bioinformatic analysis to predict its enrichment sites and designed corresponding chromatin immunoprecipitation (ChIP) primers (Fig. 3I). ChIP assays confirmed that H3K4me3 directly binds the FTH1 promoter, and this binding was significantly enhanced following CaP-siSlc31a1 intervention (Fig. 3J). These findings identified the SLC31A1/α-KG/KDM5B/H3K4me3/FTH1 axis as a key pathogenic pathway in psoriatic cuproptosis and ferroptosis, and demonstrated that CaP-siSlc31a1 mitigates psoriatic inflammation through epigenetic restoration of FTH1 expression and suppression of cuproptotic/ferroptotic signaling.

2.4. Preparation, characterization and skin penetration performance of CaP-siSlc31a1@MN

After confirming the therapeutic potential of CaP-siSlc31a1 in psoriasis, we employed micromolding technology to encapsulate CaP-siSlc31a1 into MNs, thereby fabricating a separable microneedle, designated CaP-siSlc31a1@MN, for efficient transdermal delivery to psoriatic lesions (Fig. 4A). Briefly, siSlc31a1 was homogeneously mixed with a 20% MN precursor solution and dispensed into a polydimethylsiloxane (PDMS) mold. Vacuum and centrifugation were applied sequentially to ensure uniform nanoparticle deposition within the microneedle tips. Subsequently, a 50% hyaluronic acid (HA) solution was added as the backing layer, followed by drying and demolding to obtain CaP-siSlc31a1@MN. Optical microscopy and scanning electron microscopy (SEM) revealed an intact 10 × 10 array of conical microneedles on the base layer, each measuring 550 μm in height, 250 μm in base diameter, and featuring a tip width of approximately 10 μm (Fig. 4B). To visualize nanoparticles distribution within the MNs, FAM-labeled CaP-siSlc31a1 was incorporated into the tips, and Rhodamine B (RhB) was used to label the backing layer. Fluorescence imaging confirmed the localization of FAM-labeled nanoparticles at the microneedle tips (Fig. 4C). EDS elemental mapping further verified the presence of C, N, O, Ca, and P elements, verifying successful nanoparticle loading (Fig. 4D).

Fig. 4.

(A) Schematic illustration of the preparation process of CaP-siSlc31a1@MN. (B) Photographic, optical microscopy, and SEM images of CaP-siSlc31a1@MN. (C) Fluorescence images of CaP-FAM-siSlc31a1@MN. (D) Elemental mapping showing the surface distribution of elements in CaP-siSlc31a1@MN. (E) Force-displacement curves of MN and CaP-siSlc31a1@MN under different applied forces. (F) H&E-stained pinhole in skin tissue after administration of microneedle on the mouse ear. (G) Microscopy images showing the time-dependent degradation of CaP-siSlc31a1@MN within mouse skin. (H) Fluorescence images of vertical mouse skin sections at various depths following application of RhB-loaded CaP-siSlc31a1@MN. (I) Fluorescence images and (J) quantitative analysis of ears showing the local retention and biodistribution of nanoparticles in mouse ears after application of free ICG MN and ICG-labeled CaP-siSlc31a1@MN.

Adequate mechanical strength is essential for MNs to penetrate the thickened psoriatic stratum corneum. The mechanical performance was evaluated using a universal testing machine, revealing that both blank MNs and CaP-siSlc31a1@MN exhibited a force-displacement relationship with critical fracture forces of 0.52 N and 0.15 N per needle, respectively, which were well above the minimum threshold required for skin penetration (0.045 N/needle) (Fig. 4E) [42]. Moreover, CaP-siSlc31a1@MN achieved a penetration depth of approximately 188 μm (Fig. 4F), demonstrating robust mechanical strength suitable for in vivo application.

Given their mechanical robustness, we next examined the drug release behavior of CaP-siSlc31a1@MN in psoriatic mice. Patches were applied to mouse ears and removed after 0, 5, 10, 30 and 60 min. Progressive tip blunting and partial dissolution were observed, with approximately 50% dissolution at 30 min (Fig. 4G). This rapid dissolution is attributed to the high hydrophilicity of HA, facilitating efficient nanoparticles release in the subcutaneous tissue. To assess penetration depth, RhB was incorporated into the needle tips. CLSM imaging revealed red fluorescence extending through the vertical skin layers to a depth of ∼180 μm, consistent with the puncture depth observed in hematoxylin-eosin (H&E) staining (Fig. 4H). These findings indicated that CaP-siSlc31a1@MN effectively penetrates the psoriatic stratum corneum and enables rapid nanoparticle diffusion into surrounding tissues, supporting efficient and minimally invasive drug delivery. As MNs primarily function via localized drug administration, we further evaluated the biodistribution of CaP-siSlc31a1@MN in a psoriatic mouse model. Indocyanine green (ICG)-loaded CaP-siSlc31a1@MNs were applied to the ears of psoriatic mice and removed after 2 h. In vivo fluorescence imaging demonstrated sustained local retention of ICG at the application site for over 96 h (Fig. 4I and J), with negligible fluorescence signals in major organs (Fig. S10). These results confirmed that CaP-siSlc31a1@MN enables prolonged local drug retention, minimizing systemic exposure and potential off-target toxicity.

2.5. The CaP-siSlc31a1@MN alleviated IMQ-induced psoriasis-like dermatitis in vivo

To assess the therapeutic efficacy of CaP-siSlc31a1@MN, a psoriasis-like mouse model was established through topical application of IMQ. The left ears of Balb/c mice were treated with IMQ once daily for seven consecutive days, and microneedles were applied on day 1, 3 and 5. Psoriatic mice were randomly divided into three treatment groups: CaP-siNT@MN, siSlc31a1@MN, and CaP-siSlc31a1@MN (Fig. 5A). Two additional groups, healthy controls and IMQ-induced but untreated mice, served as references. IMQ treatment induced pronounced epidermal thickening, erythema, scaling and inflammatory cell infiltration (Fig. 5B–D), confirming successful establishment of the psoriasis model. The CaP-siNT@MN group displayed skin lesions comparable to those in the IMQ group, suggesting negligible therapeutic benefit. In contrast, both siSlc31a1@MN and CaP-siSlc31a1@MN treatments substantially ameliorated psoriatic manifestations to varying extents. Quantitative analysis of lesion area and Psoriasis Area and Severity Index (PASI) scores (Fig. 5E) demonstrated a significant reduction in disease severity in the siSlc31a1@MN group, while CaP-siSlc31a1@MN treatment achieved superior outcomes and favorable body weight recovery (Figs. S11 and S12). This improvement was likely due to the CaP nanocarrier enhancing intracellular siRNA delivery. Moreover, qRT-PCR analysis revealed that Slc31a1 expression was markedly elevated in the IMQ and CaP-siNT@MN groups relative to the normal control, whereas both siSlc31a1@MN and CaP-siSlc31a1@MN effectively downregulated Slc31a1 transcription, with the latter showing the strongest suppression (Fig. 5F).

Fig. 5.

(A) Schematic illustration of the IMQ-induced psoriasis-like mouse model and the treatment regimen using microneedles. (B) Representative H&E-stained images of mouse ear skin sections. (C) Quantitative analysis of epidermal thickness among different treatment groups (n = 5). (D) Disease severity and (E) PASI scores recorded throughout the treatment period (n = 5). (F) The qRT-PCR analysis of Slc31a1 mRNA expression in mouse ear epidermis across groups (n = 5). (G) Flow cytometric analysis of Th17 cell proportions among CD4⁺ T cells and (H) representative flow cytometry plots from each group (n = 5). (I) Immunofluorescence staining and quantitative evaluation of epidermal proliferation markers (J) PCNA and (K) P63 in psoriatic skin lesions (n = 5). Data are presented as the mean ± SD. ns, not significant, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

We next evaluated inflammatory cell infiltration and epidermal proliferation in psoriatic lesions to further elucidate the therapeutic effect of CaP-siSlc31a1@MN. Both the IMQ and CaP-siNT@MN groups exhibited markedly increased Th17 cell infiltration compared with the control group (Fig. 5G and H), confirming robust inflammatory activation in the psoriatic microenvironment. In contrast, siSlc31a1@MN and CaP-siSlc31a1@MN treatments significantly reduced the number of Th17-positive cells, with the latter showing a more pronounced inhibitory effect, suggesting that the CaP nanocarrier markedly enhances the in vivo gene-silencing efficiency of siRNA. Furthermore, immunofluorescence staining revealed that the expression of proliferation markers PCNA and P63 was substantially elevated in the IMQ and CaP-siNT@MN groups, consistent with the hyperproliferative phenotype of psoriatic epidermis. Treatment with siSlc31a1@MN reduced the expression of both markers, while CaP-siSlc31a1@MN treatment resulted in an even stronger suppression (Fig. 5I–K), indicating effective restoration of epidermal homeostasis. In summary, CaP-siSlc31a1@MN effectively alleviated IMQ-induced psoriatic inflammation and epidermal hyperplasia, suppressing Th17 cell infiltration, and restoring epidermal homeostasis, demonstrating enhanced siRNA delivery and superior therapeutic efficacy.

2.6. In vivo therapeutic mechanisms of CaP-siSlc31a1@MN on psoriasis

Our previous in vitro investigations revealed that CaP-siSlc31a1 efficiently downregulated SLC31A1 expression, thereby reducing intracellular copper accumulation and suppressing cuproptosis. Concurrently, SLC31A1 silencing decreased intracellular α-KG levels, leading to the inhibition of histone demethylase KDM5B and a subsequent increase in H3K4me3 modification at the Fth1 promoter, which collectively attenuated ferroptosis. The coordinated regulation between these two programmed cell death pathways synergistically alleviated oxidative stress and inflammatory responses within the psoriatic microenvironment, markedly improving IMQ-induced psoriasiform dermatitis. Building upon these in vitro findings, we further assessed the in vivo therapeutic mechanism of CaP-siSlc31a1@MN using an IMQ-induced psoriasis mouse model. Compared with healthy controls, IMQ-treated mice exhibited dramatically elevated transcriptional levels of inflammatory mediators (Il17c, Tnf, Il1b, S100a9, Cxcl1, and Cxcl2), validating successful model establishment. Similarly, the CaP-siNT@MN treatment exerted minimal influence on inflammatory cytokine expression. In contrast, both siSlc31a1@MN and CaP-siSlc31a1@MN significantly downregulated psoriasis-associated inflammatory genes, with the CaP-siSlc31a1@MN group showing the most pronounced anti-inflammatory effect (Fig. 6A).

Fig. 6.

(A) Heatmap illustrating the transcriptional levels of inflammatory cytokines and chemokines in the epidermal tissues of mouse ears from different experimental groups. (B) Quantification and statistical analysis of copper levels in ear skin tissues across groups (n = 3). (C) Quantification and statistical analysis of Fe²⁺ levels in ear skin tissues across groups (n = 5). (D) Assessment of lipid peroxidation levels and corresponding statistical analysis in ear skin tissues from different groups (n = 5). (E) Measurement and statistical analysis of α-KG levels in ear skin tissues across groups (n = 5). (F) qRT-PCR analysis of Fth1 mRNA expression in the epidermal tissues of mouse ears from different groups (n = 5). (G) qRT-PCR analysis of Kdm5b mRNA expression in the epidermal tissues of mouse ears from different groups (n = 5). (H) Western blot analysis of SLC31A1 and FTH1 protein levels in ear skin tissues; β-actin served as the loading control. (I) Western blot analysis of H3K4me3 modification levels in ear skin tissues; histone H3 served as the control. Data are presented as the mean ± SD. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001.

Given the potent therapeutic efficacy of CaP-siSlc31a1@MN in mitigating psoriatic progression, we next examined its influence on cuproptosis- and ferroptosis-related biomarkers. Compared with normal controls, both the IMQ and CaP-siNT@MN groups exhibited elevated copper and α-KG accumulation and increased levels of ferroptosis-related factors, including Fe²⁺ and lipid peroxides, accompanied by a notable reduction in Fth1 and Kdm5b expression, consistent with our in vitro observations. Conversely, CaP-siSlc31a1@MN markedly suppressed copper accumulation and α-KG buildup, consequently lowering intracellular Fe²⁺ content, reducing lipid peroxidation, and preventing ferroptotic progression (Fig. 6B–G). At the protein level, Western blot analysis further confirmed that CaP-siSlc31a1@MN effectively inhibited SLC31A1 expression while enhancing H3K4me3 enrichment at the Fth1 promoter and upregulating FTH1 protein levels (Fig. 6H and I). These in vivo findings validated the central role of the SLC31A1/α-KG/KDM5B/H3K4me3/FTH1 signaling axis in the regulation of psoriasis pathogenesis and underscored the remarkable therapeutic potential of CaP-siSlc31a1-loaded microneedle as a targeted and localized strategy for psoriasis management.

3. Conclusion

In this study, we elucidated a previously unrecognized pathogenic interaction between cuproptosis and ferroptosis in psoriasis mediated by the SLC31A1/α-KG/KDM5B/H3K4me3/FTH1 signaling cascade. Upregulated SLC31A1 disrupted copper homeostasis, promotes α-KG accumulation and drives KDM5B-dependent epigenetic repression of FTH1, thereby amplifying ferroptotic and inflammatory responses. To counteract this process, we developed a CaP-siSlc31a1-loaded dissolvable microneedle (CaP-siSlc31a1@MN) that enables efficient transdermal delivery and precise gene silencing within psoriatic lesions. This localized therapy not only rebalanced copper and iron metabolism but also restored epidermal integrity and significantly reduced disease severity in vivo. Our findings establish SLC31A1 as a crucial metabolic-epigenetic regulator of psoriatic inflammation and highlight the potential of CaP-based microneedle systems as a next-generation platform for nucleic acid therapeutics in chronic inflammatory skin diseases.

4. Materials and methods

4.1. Materials

Keratinocyte Growth Medium was purchased from ScienCell (USA). MagZol reagent was supplied from Magen Co., Ltd (Guangzhou, China). The siSlc31a1 (GCATGATGATGATGCCTAT) was obtained from RiboBio Co., Ltd (Guangzhou, China). Collagenase Type IV, DNase I and dispase II were obtained from Sigma-Aldrich Co., Ltd (USA). Copper Assay Kit and GSH Assay Kit were purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Histone Extraction Kit was purchased from Active Motif Co., Ltd (USA). HiScript Q RT Kit was purchased from Yeasen Co., Ltd (Shanghai, China). Bovine serum albumin (BSA), glucose-free Dulbecco's modified Eagle's medium (DMEM), Calcium chloride dihydrate (CaCl₂⋅2H₂O) were supplied from Procell Co., Ltd (Beijing, China). The 5-Carboxyfluorescein (FAM), Rhodamine B (RhB) and indocyanine Green (ICG) were purchased from Aladin Co., Ltd (Shanghai, China). Hyaluronic acid (HA, Mw∼10 kDa) was provided by MeilunBio Co., Ltd (Dalian, China). Imiquimod (IMQ) was purchased from Sichuan Med-Shine Pharmaceutical Co., Ltd (Chengdu, China). Human recombinant IL17A and mouse recombinant IL17A were bought from BioLegend Co., Ltd (USA). Antibodies against DLAT, FDX1, LIAS, β-actin, PCNA and P63 were obtained from Proteintech Co., Ltd (Wuhan, China). SimpleChIP® Enzymatic Chromatin IP Kit (Magnetic Beads) and antibody against FTH1, H3K4me3 and H3 were obtained from Cell Signaling Technology, lnc. (Danvers, MA, USA). Antibody against SLC31A1 was obtained from Abcam plc (Cambridge, UK).

4.2. Spatial transcriptomics data processing and analysis

The public spatial transcriptomics dataset GSE225475 (containing both psoriatic lesional [PP] and non-lesional [NS] skin samples) was retrieved from the Gene Expression Omnibus (GEO) database. Data processing and analysis were conducted using R (v4.3.0) and the Seurat package (v4.3.0.1). First, raw 10 × Visium data for each sample were loaded using the Load10X_Spatial function, and sample metadata (e.g., GSM ID, group, and donor information) was extracted. The data was structured into Seurat objects, and a representative psoriatic lesional sample (e.g., spt_pp4 in the script) was selected for subsequent analysis.

Data normalization and feature selection were performed using the SCTransform method, where the total Unique Molecular Identifier (UMI) count per spot (nCount_Spatial) was regressed out as a technical covariate to minimize technical variation. The normalized data then underwent Principal Component Analysis (PCA), followed by Uniform Manifold Approximation and Projection (UMAP) for dimensionality reduction (based on the top 30 principal components). Unsupervised clustering of the spatial spots was performed using the FindNeighbors and FindClusters functions with a resolution set to 0.3.

The spatial expression patterns of the target genes, SLC31A1 and FTH1, were visualized by overlaying their expression levels onto the hematoxylin-eosin (H&E)-stained tissue image using the SpatialFeaturePlot function. To quantify the relationship between these two genes within the psoriatic lesion, the normalized expression data from all spatial spots were extracted. The Spearman correlation coefficient was then calculated between the expression levels of SLC31A1 and FTH1 to assess their significant negative correlation.

4.3. Western blot

Proteins were extracted from KCs and epidermal tissues for Western blot analysis. Total proteins were isolated using RIPA lysis buffer containing a protease inhibitor cocktail, while histones were specifically extracted with a commercial Histone Extraction Kit. Protein concentrations were determined by bicinchoninic acid assay, and equal amounts of protein were separated by SDS-PAGE and transferred to PVDF membranes. The membranes were probed with the following primary antibodies: anti-FTH1, anti-SLC31A1, anti-DLAT, anti-FDX1, anti-LIAS, anti-H3K4me3, anti-histone H3, and anti-β-actin. Protein bands were visualized using a Bio-Rad gel imaging system.

4.4. Immunohistochemistry staining analysis

The 6-weeks-old Male Balb/C mice (20 ± 2 g) were procured from Hunan Slack Jingda Experimental Animal Co., Ltd. The animal research protocol received ethical approval from the Institutional Animal Care and Use Committee of Central South University (approval ID: 202310040), with experimental procedures performed following the regulatory standards established by the Chinese authorities on laboratory animal welfare. Immunohistochemical (IHC) staining was carried out according to the following scheme. Firstly, Paraffin-embedded skin specimens underwent systematic processing through formalin fixation and dehydration. Tissue sections then were incubated with primary antibodies, followed by HRP-conjugated secondary antibodies and DAB. This methodology was dual-implemented across murine ear tissue and human psoriatic lesional skin samples to ensure experimental reproducibility.

4.5. Isolation and culture of mouse or normal human primary keratinocytes (KCs)

Mouse primary KCs were isolated from neonatal wild-type mice. Normal human skin samples were obtained from foreskins of several individuals (age ≤20 years) collected during circumcision procedures. The acquisition of skin samples was approved by the Ethics Committee of Xiangya Hospital, Central South University (approval ID: 202308636). The separated skin was incubated in a solution containing 2 mg/mL dispase II at 4 °C for 16 h, after which the epidermis was separated using forceps. The epidermal tissue was then digested in 1 mL of 0.25% trypsin-EDTA solution at 37 °C for 10 min. Subsequently, 2 mL of DMEM supplemented with 10% fetal bovine serum was added to neutralize the trypsin/EDTA activity. The KCs were collected by centrifugation at 1000 rpm for 3 min at room temperature and subsequently cultured in Keratinocyte Growth Medium.

4.6. Quantitative real-time PCR (qRT-PCR)

Total RNA was extracted from cells or tissues using MagZol reagent (Magen, #R4801-01). cDNA was synthesized by reverse transcription using the HiScript Q RT Kit (Yeasen, #11141ES60). qRT-PCR was then performed using the RT-PCR-Q3 system (Thermo Fisher Scientific) with 2 × SYBR Green qPCR premix (low ROX). The reaction mixture contained 0.5 μL each of forward and reverse mouse primers as described (Table S1). Values were normalized to β-actin.

4.7. Preparation and characterization of CaP-siSlc31a1

Firstly, siSlc31a1 (10 nmol) was dissolved in 100 μL Diethyl pyrocarbonate (DEPC)-treated water. Subsequently, 100 mg BSA was added to 10 mL glucose-free DMEM via sonication. The siSlc31a1 solution was then introduced to the BSA-DMEM mixture and vortexed thoroughly, and 100 μL CaCl₂ (1 M) was incrementally added. After incubation at 37 °C overnight, the nano-precipitate was collected by centrifugation at 15000 rpm for 20 min (Centrifuge 5430R, Eppendorf, Germany). Finally, the nanoparticles were concentrated and resuspended in PBS and stored at 4 °C until use.

The morphological characteristics and elemental composition of CaP-siSlc31a1 NPs were evaluated using a transmission electron microscopy (TEM, Tecnai G2 F20, FEI, USA). Hydrodynamic size distributions were determined via dynamic light scattering with a Malvern Zetasizer (Nano ZS90, Malvern Panalytical, UK). The spatial distribution of elements was investigated through an energy-dispersive X-ray spectroscopy (EDS, GeminiSEM 30F0, ZEISS, Germany).

4.8. Drug loading and encapsulation efficiency

Encapsulation efficiency (EE) was investigated via a fluorescence spectrophotometry method. During nanoparticle preparation, FAM-siSlc31a1 was introduced to label the nanoparticles for the preparation of CaP-FAM-siSlc31a1. Specifically, different concentrations of siSlc31a1 (1, 2, 4 and 8 μM) were incrementally added, and CaP-FAM-siSlc31a1 NPs were separated with the Centrifuge (15060 rpm, 20 min). Then, the fluorescence intensity of total siRNA (F_total) and fluorescence signal intensity in supernatant (F_free) were quantified using a fluorescence spectrophotometer (FL 6500, PerkinElmer, USA). A dose-response curve was generated by correlating siRNA input concentrations with corresponding EE values: EE = (F_total - F_free)/F_total × 100%.

4.9. Cellular uptake and lysosomal escape

The KCs were seeded into a 6-well plate and cultured overnight at 37 °C with 5% CO₂ to achieve 80-90% confluency. For siRNA delivery, cells were treated with FAM-siSlc31a1, CaP-FAM-siSlc31a1 NPs and Lipo-FAM-siSlc31a1 respectively for 2 h. Then, cells were collected after removing the supernatant with PBS and detected intracellular fluorescence expression with a flow cytometry (S3-FACSARIA Ⅲ, BD, USA). For lysosomal escape, KCs were seeded in confocal dishes and incubated with CaP-FAM-siSlc31a1 NPs for different time points (0.5 h, 1 h, 2 h, 4 h and 6 h). Then, the cells were stained by Hoechst 33342 (nuclear) and Lyso-Tracker red (lysosomal). Ultimately, treated cells were captured with a Confocal Laser Scanning Microscope (CLSM, ZEISS LSM900, Germany).

4.10. Cell viability assay

KCs were seeded in 96-well plates and cultured for 24 h (2000 cells per well). Cells were then treated with Lipo-siNT, CaP-siNT NPs, Lipo-siSlc31a1 and CaP-siSlc31a1 NPs separately and incubated for 48 h. Following two PBS washes, 100 μL MTT reagent (5 mg/mL) was introduced to cells and incubation for 4 h at 37 °C. Formazan crystals were solubilized by adding 100 μL DMSO per well, with absorbance quantified at 570 nm with a microplate reader (EPOCH-SN, Biotek, USA).

4.11. Tissue processing and flow cytometry

Ear skin tissues were collected from post-modeling mice (approximately 3 × 3 mm per sample) and cut into small pieces. The skin samples were digested in DMEM containing 2 mg/mL Collagenase Type IV and 100 μg/mL DNase I, then minced with scissors while oscillating at 37 °C for 60 to 90 min. Digestion was terminated by adding DMEM supplemented with 10% FBS. The incompletely digested tissue was further homogenized using a syringe and filtered through a 70 μm cell strainer. Cells were washed with 10 mL PBS and centrifuged (2000 rpm, 5 min) using a 4 °C centrifuge. Subsequently, single-cell suspensions were stimulated for 5 h at 37 °C in 1640 medium containing Cell Stimulation Cocktail and 10% FBS, followed by staining with fluorescent antibodies: APC/Cyanine7 anti-mouse CD45 Antibody, PerCP/Cyanine5.5 anti-mouse CD4 Antibody and APC anti-mouse IL17A Antibody for flow cytometric analysis. Data were analyzed using FlowJo software (BD Biosciences, Le Pont de Claix, France).

4.12. Chromatin immunoprecipitation assay

Chromatin immunoprecipitation (ChIP) assays were performed using the SimpleChIP® Enzymatic Chromatin IP Kit according to the manufacturer's protocol. Cells were cross-linked, lysed, and subjected to sonication. Antibodies used for ChIP were H3K4me3 (9751T, CST, 2 μg/test) and IgG (provided in the kit, 2 μg/test). Chromatin was immunoprecipitated using protein A/G beads to isolate protein-DNA complexes. The binding sites of H3K4me3 on the FTH1 promoter were predicted using the Cistrome Data Browser database (http://cistrome.org/db/). Purified DNA was analyzed by qRT-PCR using primers specific for the FTH1 promoter. The qRT-PCR primers for the FTH1 promoter used in the ChIP assay are listed (Table S2).

4.13. Preparation and characterization of CaP-siSlc31a1@MN microneedle

To prepare CaP-siSlc31a1@MN, the solution of CaP-siSlc31a1 was added to MN to form MN (20 wt%) precursor solution containing NPs. Then, 100 μL precursor solution was poured into a PDMS microtip mold, and centrifugated (4000 rpm, 10 min) three cycles to form needle tips. After that, 100 μL MN (50 wt%) matrix solution was poured into the bottom of the mold to form the backing layer. Finally, the CaP-siSlc31a1@MN was demolded from the mold after air-drying at 30 °C for 24 h. Besides, the fluorescent-labeled MNs were synthesized via mix FAM, RhB or ICG-labeled CaP-siSlc31a1 into the precursor solution. the size and morphology of MNs were captured by a stereomicroscope (MZ62, Mshot, China) and a scanning electron microscope (SEM, Phenom ProX G6, Netherlands), and the elemental distribution of the needle tips was recorded by EDS. The mechanical properties of MN and CaP-siSlc31a1@MN were characterized using a universal texture tensile testing machine (CMT6103, China). The fluorescence images were captured by the CLSM.

4.14. In vivo penetration capability and the drug delivery performance

In short, MN patches were applied in the right ear of mice for different time points (0, 5, 10, 30 and 60 min), and then the MNs were immediately removed from the ear. The MNs were immediately observed under an optical microscope (Scope.A1, ZEISS, Germany). Besides, RhB-labeled CaP-siSlc31a1@MN inserted into the ear skin for 2 min and removed quickly, and then the skin was collected for frozen section and observed via the inverted fluorescence microscope. What's more, the skin also fixed with 4% paraformaldehyde for H&E staining to observe the micropores. For the accumulation and biodistribution behavior of CaP-siSlc31a1 NPs under skin, ICG-labeled CaP-siSlc31a1@MN were applied to the PSO mice ear and removed after 2 h. The accumulation of CaP-FAM-siSlc31a1 was observed via an Xenogen IVIS imaging system (S12-FMT400010, PerkinElmer, US).

4.15. Mice and treatments

The IMQ-induced psoriasis-like skin inflammation model is a widely accepted mouse model for psoriasis. In this study, this model was used to investigate the effects of siSlc31a1/CaP@MN on the morphological features of IMQ-induced psoriatic mice. To this end, mice were divided into 5 groups, with 5 animals per group (n = 5). The groups were designated as follows: normal control group (Con), IMQ control group (IMQ), standard microneedle group (IMQ + CaP-siNT@MN), SLC31A1 siRNA-loaded microneedle group (IMQ + siSlc31a1@MN), and SLC31A1 siRNA-CaP-loaded microneedle group (IMQ + CaP-siSlc31a1@MN). IMQ (20 mg per ear per day) was applied evenly to both the inner and outer surfaces of the ears (control mice received Vaseline). The modeling period lasted for 7 days in total. On the day before IMQ treatment, as well as on Days 2, 4, and 6, each mouse received CaP-siSlc31a1@MN application on the right ear for several hours. Skin lesions and the body weight of each mouse were observed and recorded daily. The severity of ear skin condition was assessed daily using the clinical Psoriasis Area and Severity Index (PASI), evaluating thickness, scaling, and erythema. Ear thickness was measured using a vernier caliper (Mitutoyo, Japan). Scaling and erythema were scored on a scale of 0-4: 0, none; 1, slight; 2, moderate; 3, marked; 4, very marked. On Day 7, mice were euthanized to collect skin tissues. For histological analysis, skin samples were fixed in 4% paraformaldehyde and stained with H&E for microscopic examination.

4.16. Immunofluorescence staining analysis

Immunofluorescence staining of ear tissues was performed as follows: Ear skin specimens from treated mice were surgically excised and paraffin-embedded. Tissue sections underwent standard processing including xylene dewaxing, graded ethanol rehydration, and heat-induced antigen retrieval using citrate buffer (pH = 9.0). Non-specific binding sites were blocked by incubating sections with 10% normal goat serum in PBS for 1 h at room temperature. Immunolabeling was conducted by incubation with primary antibodies PCNA and P63 overnight at 4 °C in a humidity-controlled chamber. After thorough washing, sections were incubated with species-matched fluorescent secondary antibodies for 60 min at 37 °C in dark conditions. Nuclear counterstaining was achieved using DAPI followed by visualizing via a CLSM.

4.17. Statistical analysis

All statistical analyses were performed using GraphPad Prism 9 (GraphPad Software, San Diego, CA, USA). For datasets with non-normal distribution, statistical significance between values was determined using unpaired t-tests or one-way ANOVA with Dunnett's post hoc test. Correlations between measured variables were examined using Spearman's rank correlation analysis. All data are presented as the mean ± standard deviation (SD). Asterisks (*) indicate the degree of statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

CRediT authorship contribution statement

Lu Hao: Data curation, Formal analysis, Methodology, Writing – original draft. Pian Yu: Data curation, Funding acquisition, Software, Writing – review & editing. Rongxuan Yan: Data curation, Methodology, Writing – original draft. Kaixuan Li: Conceptualization, Data curation, Software. Zhisheng Luo: Investigation, Software. Shijun Xiang: Methodology. Yilan Wang: Supervision. Chi Fang: Formal analysis. Guanming Wang: Formal analysis. Sihui Ma: Software. Cong Peng: Funding acquisition, Methodology, Validation. Shuo Hu: Funding acquisition, Supervision, Visualization. Peng Liu: Conceptualization, Funding acquisition, Investigation, Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.