Intradermal delivery of lipophilic siRNAs enables prolonged skin retention and sustained gene silencing in a porcine model

Abstract

Small interfering RNAs (siRNAs) offer significant therapeutic potential; however, extrahepatic applications, particularly to the skin, remain a challenge. Limited work has explored siRNA therapies for the skin, the largest organ in the human body, where dermatological conditions affect over one-third of the population worldwide. The skin’s external location makes it easily accessible for direct, local administration. Here, we present the in vivo intradermal delivery of therapeutic siRNAs into a porcine model whose skin structure most closely resembles that of human skin, demonstrating functional, and sustained gene silencing. We characterize two siRNA conjugates in human ex vivo and porcine in vivo skin models, showing that increased hydrophobicity significantly enhances skin retention and efficacy of siRNAs. Using a validated JAK1-targeting compound, we demonstrate that local delivery of siRNA enables accumulation across multiple cell types and suppression of JAK1-dependent inflammatory pathway in human skin ex vivo. In porcine models, intradermal injections result in prolonged skin siRNA retention for more than eight weeks, limited systemic tissue exposure, and sustained gene silencing for at least one month. These results underscore the importance of tailored siRNA conjugate design for achieving optimal skin biodistribution and therapeutic efficacy, providing a foundation for siRNA-based treatments for a broad range of dermatological conditions.

Skin-targeted siRNA therapies require optimized delivery to achieve therapeutic efficacy. Here, authors show that increasing conjugate hydrophobicity enhances siRNA skin retention and gene silencing in porcine and human models while limiting systemic tissue exposure.

Introduction

RNA therapeutics are revolutionizing the way we treat diseases¹. Small interfering RNAs (siRNA) represent a class of highly programmable molecules capable of silencing their complementary mRNA targets to prevent translation of disease-causing proteins^2,3. Recent chemical advances in improving stability and in vivo delivery have enabled the translation of siRNAs into the clinic, with multiple FDA-approved drugs emerging in the past few years⁴. While this represents a significant leap forward for this modality, the clinical utility of siRNAs is currently limited to the liver². As such, recent efforts in the field are focused on expanding the utility of siRNA drugs to extrahepatic organs (e.g., muscle⁵, lung⁶, CNS⁷) to treat previously “undruggable” diseases^7,8.

One understudied organ of interest is the skin, where many diseases, such as cutaneous lupus erythematosus, psoriasis, vitiligo, alopecia areata, and atopic dermatitis, may benefit from treatment with siRNA^9,10. According to the World Health Organization, skin diseases are among the most common human health conditions, impacting as many as one-third of the population worldwide, underscoring the demand for developing novel therapeutic approaches to address unmet medical needs^11,12. As many of these conditions are primarily caused by the dysregulation of the immune system in the skin, delivering therapeutic modalities to the skin while avoiding systemic immunosuppression is critical^13–15. While targeted delivery of siRNA to the skin following systemic administration remains a challenge, the inherent accessibility of skin makes local administration a clinically viable and straightforward approach^9,10. Biologics, such as antibodies, are highly specific but are only efficient for targeting secreted or cell surface targets. Conventional small molecules require frequent dosing, which can be burdensome and limit patient compliance^{9,10,12,16,17}. In contrast, siRNA therapeutics can overcome these challenges and offer sustained gene silencing when engineered for local skin administration, offering an ideal treatment option.

We have previously shown that the delivery properties of siRNAs can be altered by modifying conjugate chemistry, with lipophilic conjugates enhancing skin delivery and efficacy^18–21. However, direct comparison of diverse siRNA conjugates for intradermal administration is lacking, and existing studies are largely confined to ex vivo tissues or rodent models^21,22. These models present limitations in accurately assessing skin retention, systemic exposure, and therapeutic durability due to the inherent differences in rodent skin physiology and the limited viability of ex vivo human tissue^15,23,24.

In this work, we investigate the pharmacokinetics and pharmacodynamics of lipophilic siRNAs in human and porcine skin models. We compare unconjugated siRNA (Unc-siRNA), amphiphilic dendritic siRNA (D-siRNA), and lipophilic docosanoic acid siRNA (DCA-siRNA). In human ex vivo skin, DCA-siRNA demonstrates enhanced cellular uptake and superior gene silencing in both the epidermis and dermis, suggesting that strong lipophilicity is ideal for siRNA skin efficacy. Targeting JAK1 expression with DCA-siRNA, a key immune regulator in skin diseases, leads to reduced downstream inflammatory chemokine release (CXCL9, 10, and 11)^21,22.

Furthermore, we evaluate the pharmacokinetics and pharmacodynamic profile of DCA-siRNA in a porcine in vivo model. Intradermal injection resulted in prolonged skin retention (at least up to eight weeks), sustained JAK1 silencing (at least one month), and minimal systemic exposure. DCA-siRNA also exhibited an excellent safety profile, with no alterations to blood markers. These findings suggest that lipophilic DCA-siRNA is a promising therapeutic strategy for skin conditions with a defined target.

Results

Intradermal injection of lipophilic siRNAs into human skin leads to multiple cell type uptake and efficacy

Building on our previous findings that lipophilic conjugation enhances siRNA delivery to the skin^18,19,21, we investigated how lipid conjugate hydrophobicity impacts cellular uptake and activity in human ex vivo skin after intradermal injection. From the library of conjugates previously assessed, docosanoic acid (DCA) demonstrated the highest skin retention and siRNA activity in mouse skin tissue, making it a compelling candidate for further evaluation¹⁸. Additionally, a recently developed amphiphilic dendrimer conjugate achieved effective skin delivery while being significantly less hydrophobic than DCA, warranting a direct comparison between the two¹⁹. For that, we assessed the cellular uptake of cyanine-3 (Cy3) labeled unconjugated (Unc-siRNA), the moderately lipophilic (dendritic, D-siRNA), and strongly lipophilic (docosanoic, DCA-siRNA) siRNAs, 24 h post-injection into skin biopsies (Fig. 1a, b). Flow cytometry allowed for analysis of cell-type-specific uptake analysis (Fig. 1c; gating strategy in Supplementary Fig. 1). While all tested siRNAs were internalized by all major skin cell populations, DCA-siRNA showed significantly increased uptake, particularly in melanocytes and immune cells (**p < 0.01), compared to D-siRNA and Unc-siRNA. Overall, DCA-siRNA uptake was significantly higher than both Unc-siRNA (**p < 0.01) and D-siRNA (***p < 0.001).

Fig. 1. Chemically stabilized siRNAs accumulate in multiple cell types and significantly silence JAK1 mRNA and downstream chemokines, following intradermal injections in fresh human skin ex vivo.

a Chemical structure of various lipid-conjugated, fully chemically modified siRNAs assessed for optimal functional delivery in skin. b Scheme of administering cyanine 3-labeled (Cy3) siRNAs into human skin biopsies (8 mm) with representative images at 24 h following intradermal injection of Cy3-labeled siRNAs (50 µl of 200 µM siRNA/biopsy, 10 nmol final dose/biopsy). c siRNA accumulation in various cell types, MFI – mean fluorescence intensity, n = 3 8 mm biopsies for all groups except D-siRNA with n = 2 (two-way ANOVA, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; non-significant comparisons not shown). d JAK1 mRNA silencing in epidermis and dermis, with different conjugates 4 days post intradermal injection. e mRNA silencing of JAK1 downstream chemokines (CXCL9,10,11) with different conjugates, 4 days post intradermal injection, followed by stimulation with IFN-γ for 24 h. (n = 4 8 mm biopsies/sample; 50 µl of 200 µM siRNA/biopsy; 10nmol final dose/biopsy; one-way ANOVA, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant; all data are presented as mean value ± SD) (siRNA cell-type accumulation was quantified via flow cytometry; mRNA quantification was via Quantigene 2.0 assay). Source data are provided as a Source Data file. Elements of this figure are created in BioRender. Fakih, H. (2026) https://BioRender.com/iu0wv7a.

To examine the influence of conjugate chemistry on siRNA efficacy, we employed our previously validated JAK1-targeting siRNA sequence (JAK1³⁰³³)²¹, and intradermally administered Unc-JAK1³⁰³³, DCA-JAK1³⁰³³ and D-JAK1³⁰³³ siRNAs into human skin biopsies. Four days post-injection, JAK1 mRNA levels were quantified using a QuantiGene 2.0 assay in both epidermal and dermal layers of the skin (Fig. 1d)²¹. All conjugates induced significant JAK1 mRNA silencing in the dermal layer. However, D-JAK1³⁰³³, while showing a silencing trend, had considerable variability in the epidermal layer, failing to achieve statistical significance (p = 0.07). This increased variability in the epidermis following an intradermal injection could be a result of the chemistry-dependent distribution away from the site of injection, leakage into the media, or issues with the separation of the various skin layers for processing (Fig. 1d).

Consistent with cellular uptake results, DCA-JAK1³⁰³³ exhibited the most potent silencing, achieving approximately 50% and 75% reduction in JAK1 mRNA in the epidermis and dermis, respectively. In contrast, Unc-JAK1³⁰³³ and D-JAK1³⁰³³ yielded comparable silencing, with approximately 25% knockdown in the epidermis and 50% in the dermis (Fig. 1d). This indicates that increased hydrophobicity enhances siRNA-mediated JAK1 mRNA silencing in human skin following intradermal delivery. Furthermore, we observed dose-dependent silencing of DCA-JAK1³⁰³³ in an escalating dose study (Supplementary Fig. 2).

JAK1 is a component of the IFN-γ signaling cascade; therefore, we sought to evaluate the impact of JAK1 silencing on downstream signaling molecules following IFN-γ stimulation^21,25. Human skin biopsies were pre-treated with siRNAs for four days and subsequently stimulated with IFN-γ for 24 h to induce chemokine expression (Fig. 1e). We then quantified mRNA levels of the IFN-γ-responsive chemokines CXCL9, CXCL10, and CXCL11. Consistent with JAK1 mRNA silencing, all siRNAs effectively suppressed chemokine expression, with DCA-JAK1³⁰³³ exhibiting the most potent inhibition, achieving approximately 70% reduction across all three chemokines (Fig. 1e). D-JAK1³⁰³³ and Unc-JAK1³⁰³³ showed comparable levels of inhibition, approximately 50%, where D-JAK1³⁰³³ demonstrated a slightly more robust inhibition than Unc-JAK³⁰³³. These findings demonstrate that siRNA-mediated JAK1 silencing not only reduces JAK1 mRNA expression but also effectively attenuates downstream chemokine signaling, which plays a pivotal role in the pathogenesis of numerous skin conditions^26–28.

Intradermal administration of conjugated JAK1³⁰³³ siRNAs in pigs demonstrates durable skin retention for at least 2 months, minimal systemic exposure, and a favorable safety profile

While JAK1 inhibitors have revolutionized inflammatory skin disease treatment, the potential for systemic toxicity can be minimized with localized delivery strategies^14,15,29. To assess systemic distribution of intradermally administered JAK1-targeting siRNAs, we utilized a porcine model, chosen for its structural similarity to human skin^17,30. Pigs received intradermal injections of D- or DCA-JAK1³⁰³³ (50 µL of 200 µM siRNA/1 cm² at an intradermal injection site; 10 nmol final dose/biopsy, a total of 80 injections per pig [40/side]; details in Supplementary Fig. 3). Skin and major organ biopsies (heart, spleen, liver, kidney) were collected at various time points, and siRNA accumulation was quantified using a peptide-nucleic acid hybridization assay (Fig. 2a)³¹. Notably, siRNAs exhibited prolonged skin retention for at least up to eight weeks, with 15 to 20-fold lower siRNA levels detected in systemic organs, indicating minimal systemic exposure (Fig. 2b). To evaluate potential immunogenicity, we measured 17 cytokines and interleukins in blood samples collected from days one to seven post-injection. No significant changes were observed for any analyte (Supplementary Fig. 4), confirming a favorable safety profile. Additionally, blood levels of platelets (thrombocytopenia), white blood cell count (inflammation), alanine aminotransferase (ALT, liver damage), and carbamide (kidney damage) had no significant changes (Fig. 2c). These findings, combined with sustained local skin retention and minimal systemic distribution, suggest that intradermal administration of hydrophobic siRNAs will have prolonged silencing activity in skin, with minimal toxicity and off-target effects.

Fig. 2. D- and DCA-conjugated siRNA show prolonged skin retention, with minimum systemic exposure and no alterations to measured safety parameters, following intradermal injection in pigs.

a Scheme of intradermal injection of human JAK1-targeting siRNAs (JAK1³⁰³³) in pig with biopsies (plasma, skin, organs) collected over time to quantify accumulation and assess safety. b D-siRNA and DCA-siRNA accumulation in skin, liver, kidney, spleen, and heart over time, n = 1–3 8 mm biopsies/group; all data points are provided in Source Data file. c Platelets, white blood cell count (WBC), ALT, and carbamide levels in pigs over time following DCA and D-JAK1³⁰³³ intradermal injections (n = 1–5 pigs/timepoint). (all organ and skin biopsies are 8 mm; 50 µl of 200 µM siRNA/1 cm² intradermal injection site; 10nmol final dose/biopsy; one-way ANOVA analysis for E., ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant; all data are presented as mean value ± SD). (siRNA accumulation in plasma, tissue and skin was quantified via LC-MS/MS; mRNA quantification was quantified via Quantigene 2.0 assay). Source data are provided as a Source Data file. Elements of this figure are created in BioRender. Fakih, H. (2026) https://BioRender.com/iu0wv7a.

Pharmacokinetic and pharmacodynamic assessment of DCA-siRNA in pigs

Building on DCA-siRNA’s superior potency in human skin ex vivo and its prolonged retention and safety in pig in vivo, we conducted a comprehensive pharmacokinetic/pharmacodynamic assessment in pigs to evaluate intradermal administration as a safe and effective therapeutic route for JAK1 skin silencing. We compared siRNA plasma circulation, tissue distribution, and skin retention following both intradermal and intravenous injections (Fig. 3a), utilizing a pig-reactive JAK1 siRNA sequence, JAK1⁸⁸³ (as our human-targeting JAK1³⁰³³ is not cross-reactive). Intradermally injected DCA-JAK1⁸⁸³ demonstrated minimal systemic exposure compared to intravenous injection, showing an absolute bioavailability of 10.5% (AUC_last) for DCA-JAK1⁸⁸³ (Fig. 3b), and significantly lower tissue accumulation across multiple organs seven days post-injection relative to the dose injected (Fig. 3c). Importantly, siRNA was only detected in the skin after intradermal injection, as intravenous administration failed to deliver sufficient siRNA to the skin (below the limit of detection by LC-MS/MS). We also observed sustained retention of DCA-JAK1⁸⁸³ in skin over time (Fig. 3d), with levels at approximately 300 µg/g at three hours post-injection, reducing to ~60 µg/g at 24 h and maintaining this level to seven days post-intradermal injection. Finally, assessing siRNA activity by measuring JAK1 mRNA levels seven days post-intradermal injection (Fig. 3e), we found a ~ 50% reduction in both epidermal and dermal JAK1 mRNA levels (**p < 0.01), confirming efficacious delivery of DCA-siRNA.

Fig. 3. Pharmacokinetics and pharmacodynamics of DCA-conjugated siRNA show minimal systemic exposure, high skin retention in skin, and strong silencing, following intradermal injection in pigs.

a Scheme of intradermal (I.D.) or intravenous (I.V.) injection of pig-targeting JAK1⁸⁸³ DCA-siRNAs into pigs with biopsies (plasma, skin, organs) collected over time to quantify siRNA in circulation, skin retention, tissue distribution, and gene silencing. b siRNA plasma levels over time following intradermal or intravenous injection up to 7 days post injection (n = 4 pigs/group). c siRNA distribution to tissues following I.D. or I.V. injection (n = 4 pigs/group, with n = 12 skin biopsies/group – 3 biopsies/pig, no siRNA was detected in skin in the I.V. group). d siRNA retention in skin following intradermal injection (n = 4 pigs/group, 1 biopsy/pig). e mRNA silencing of JAK1 in pig skin 7 days post I.D. injection (n = 3 biopsies/pig, n = 4 pigs, total of n = 12 skin biopsies/group. (all skin biopsies are 8 mm; 50 µl of 400 µM siRNA/1 cm² site; 20nmol final dose/biopsy; one-way ANOVA analysis for E., ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant; all data are presented as mean value ± SD). (siRNA accumulation in plasma, tissue and skin was quantified via LC-MS/MS; mRNA quantification was quantified via Quantigene 2.0 assay). Source data are provided as a Source Data file. Elements of this figure are created in BioRender. Fakih, H. (2026) https://BioRender.com/iu0wv7a.

Furthermore, parallel studies with the human-targeting DCA-JAK1³⁰³³ siRNA showed identical plasma, tissue, and skin accumulation profiles with an absolute bioavailability (AUC_last) of 8.35% when injected intradermally (Supplementary Fig. 5), supporting the known principle that siRNA sequences minimally influence the biodistribution of this modality³. Hematological readouts from these studies were also unaltered, supporting the safety profile of DCA-siRNA, both intradermally and intravenously (Supplementary Fig. 5). We also confirmed that DCA-NTC does not impact JAK1 mRNA in ex vivo pig skin (Supplementary Fig. 6).

DCA-siRNA supports a minimum of one-month-long silencing in porcine skin following intradermal administration

Building on promising skin retention, a favorable safety profile, and one-week efficacy observed in pigs, we further investigated the durability and downstream signaling modulation of DCA-siRNA targeting JAK1 expression. We intradermally injected the pig-reactive DCA-JAK1⁸⁸³ (50 µL of 400 µM/1 cm² site; 20 nmol final dose) and subsequently assessed JAK1 mRNA expression and downstream chemokines (CXCL9, 10, and 11 mRNA) in skin biopsies collected one-month post-injection following exposure to IFN-γ (Fig. 4a). DCA-JAK1⁸⁸³ effectively silenced JAK1 mRNA in both epidermal and dermal layers, achieving approximately 50% knockdown in the epidermis (***p < 0.001) and 75% in the dermis (***p < 0.001), compared to untreated skin, while intradermally administered PBS showed no silencing, as expected (Fig. 4b). Similarly, we observed significant reductions in CXCL9 (~50% in both epidermis and dermis, **p < 0.01 in epidermis, ***p 0.001 in dermis) and CXCL10 ( ~ 50% in epidermis *p < 0.05, ~40% in dermis p = 0.28) mRNAs. Significant CXCL11 mRNA silencing was observed only in the dermis (~50% ****p < 0.0001) (Fig. 4b). These results collectively demonstrate the potent and sustained JAK1 silencing and pathway modulation achievable with DCA-siRNA in a large animal model.

Fig. 4. Intradermally administered DCA-siRNA maintains maximum silencing up to at least one month post injection in pig.

a Schematic of intradermal injection of DCA-JAK1⁸⁸³ in a pig with biopsies collected one month post-injection for mRNA silencing quantification of JAK1 and downstream chemokines. b mRNA silencing of JAK1 (target gene) and downstream chemokines (CXCL9, 10, 11) in both epidermal and dermal layers of skin. (all skin biopsies were 8 mm (n = 4); 50 µL of 400 µM siRNA/1 cm² site; 20 nmol final dose/biopsy; one-way ANOVA, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001; ns, not significant; all data are presented as mean value ± SD). (mRNA quantification was quantified via Quantigene 2.0 assay). Source data are provided as a Source Data file. Elements of this figure are created in BioRender. Fakih, H. (2026) https://BioRender.com/iu0wv7a.

Discussion

This study demonstrates the efficacy of lipophilic siRNA conjugates for dermatological applications, highlighting the critical role of conjugate chemistry in optimizing delivery and retention. We observed that increased lipophilicity, particularly with DCA-siRNA, significantly enhanced cellular uptake and gene silencing in skin models, likely due to increased interactions with the dermal environment^9,18,21,32. Targeting JAK1, a key regulator in inflammatory skin diseases^14,29, resulted in robust silencing and downstream chemokine modulation, supporting the translational potential of this approach.

Importantly, we established the in vivo pharmacokinetics and safety profile of DCA-siRNA in pigs as a large animal model, chosen for its structural similarity to human skin³⁰. Intradermal administration of the human targeting DCA-JAK1³⁰³³ yielded prolonged skin retention (at least up to eight weeks) with minimal systemic exposure, mitigating potential systemic side effects associated with JAK inhibitors²⁹. Similar safety and skin retention levels were observed with the amphiphilic D-JAK1³⁰³³, but due to the enhanced activity observed with DCA-siRNA—likely associated with increased cellular accumulation resulting from the increased hydrophobicity—it is the preferred chemistry for intradermal applications of RNA therapeutics. It is important to highlight that while we previously observed an impact on platelet levels in mice with DCA-siRNA²¹, we do not observe any alterations to platelet levels or any other blood markers, suggesting that this phenotype may be species and/or dose-dependent and supporting the favorable safety profile of DCA-siRNA.

To assess siRNA activity in pig skin, we utilized a previously developed JAK1 siRNA sequence that targets the pig transcript²². Indeed, the pharmacokinetics of pig-targeting JAK1⁸⁸³ and human-targeting JAK1³⁰³³ demonstrated similar safety and pharmacokinetic profiles, which is an advantage of siRNA technology that allows for decoupling of drug pharmacodynamics (only dictated by siRNA sequence) and pharmacokinetics (dictated by chemical architecture). Utilizing JAK1⁸⁸³, DCA-siRNA showed potent and sustained silencing of JAK1 in pigs, as silencing activity was consistent from one-week post-injection to one-month post-injection, with even a slight improvement (~50% in dermis at one week, ~75% at one month). This data is in accordance with the skin accumulation data, where we quantified a sufficient amount of siRNA retained in skin from one week to one month. We expect the activity to last more than eight weeks based on the siRNA accumulation in skin data. Sustained JAK1 silencing and downstream pathway inhibition in pig at one-month post-injection further underscores the durability and translatability of DCA-siRNAs as a therapeutic modality for skin conditions.

The work presented does have certain limitations. Our siRNA activity validation in human skin was limited to only four days post-injection due to tissue viability concerns of ex vivo skin explants during prolonged cultures. Additionally, while intradermally injected siRNA is potent and translatable, it requires a multi-injection procedure to treat skin conditions that impact a large area of the skin’s surface. This warrants careful consideration as it might result in patient discomfort hindering applications to these conditions; however, the prolonged therapeutic effect of siRNAs (months from a single injection) may justify this administration route^3,6,33. Beyond intradermal injections, there are currently multiple advanced delivery systems being explored to administer therapeutics to the skin³⁴. These non-invasive strategies can potentially serve as alternatives for treating diseases that require coverage of much larger skin surface areas. Key examples include microneedle patches (MNs)^35,36, which create transient microchannels for enhanced drug uptake, ointments and other topical formulations, and novel delivery enhancers utilizing materials such as ionic liquids³⁷ and STAR particles³⁸. These strategies can potentially be utilized as alternatives to deliver therapeutic siRNAs to much larger skin surface areas for diseases that require bigger coverage.

Another limitation is the absence of a non-targeting control (NTC) siRNA group in the pig study. Due to the significant cost constraints of large animal studies, and the inclusion of these data into the Investigational New Drug (IND)-enabling toxicity package, PBS (the placebo used in clinical trial), was selected as the control. However, the impact of the siRNA chemistry in the form of an NTC control is mitigated by several lines of evidence: human biopsy data showed that NTC-treated and PBS-treated tissues were identical in their inability to affect JAK1 mRNA levels; additionally, the human-targeting JAK1³⁰³³ compound—which acts as an NTC in pigs since it is not cross-reactive—demonstrated no alterations in CBC, blood chemistry, or inflammatory markers at various timepoints and in two different conjugates. These findings support the on-target specificity of the JAK1 siRNA and confirm that the siRNA chemistry itself does not cause non-specific inflammation.

In conclusion, our findings demonstrate that DCA-siRNA achieves potent and durable gene silencing in both human ex vivo and porcine in vivo skin. The favorable pharmacokinetic profile observed in pigs, with sustained skin retention and minimal systemic exposure, supports the potential use of these compounds as safe and effective therapeutics for dermatological diseases. The DCA-JAK1³⁰³³ compound is currently being tested in a Phase 2a human clinical trial³⁹.

Methods

Ethics statement

All research performed complies with the relevant ethical regulations. De-identified surgically discarded fresh abdominal skin was obtained from the UMass Chan Biospecimen, Tissue, and Tumor Bank. The collection of these samples from patient donors was approved by the Institutional Review Board (IRB) at the University of Massachusetts Chan Medical School, and all participants gave written informed consent before surgical procedures. Minipig studies were conducted by SCANTOX (No. 79362) and Altasciences (No. 2181.04), under their approved IACUC protocols.

Oligonucleotide synthesis, deprotection and purification

Oligonucleotides were synthesized on a MerMade 6/12 synthesizer (Bioautomation) and AKTA Oligopilot 100 (GE Healthcare Life Sciences) following standard protocols. In brief, conjugated sense strands were synthesized at 5 to 20 μmol scales on custom-synthesized lipid-functionalized controlled pore glass (CPG) supports for DCA conjugate^5,18,40. For the dendritic sense strand, synthesis was on a CPG functionalized with UnyLinker (ChemGenes) and commercially available amidites (Cy3, C6, C12, and symmetrical branching from ChemGenes) were used to build the dendritic moiety on the 5′-end as previously described^41,42. All sense strands had a 2dT spacer between the strand and the conjugate. Antisense strands were synthesized on CPG functionalized with UnyLinker. All strands were cleaved and deprotected using 28% aqueous ammonium hydroxide solution for 16 h at 50 °C, followed by drying under vacuum at 45 °C, and resuspended in Millipore H₂O. The deprotected oligonucleotide solutions were then filtered to remove CPG residues. Oligonucleotides were purified using HPLC purification carried out on an Agilent 1290 Infinity II system as previously described^18,19,21. Briefly, Sense strands were purified over a C18 column for lipid-conjugated sense strands and over an ion-exchange column for antisense strands. Purified oligonucleotides were desalted by size-exclusion chromatography and characterized by liquid chromatography-mass spectrometry (LC/MS) analysis on an Agilent 6530 accurate-mass quadrupole time-of-flight (Q-TOF) LC/MS (Agilent Technologies).

In vivo oligonucleotide sequences

Oligonucleotide compounds and sequences used are listed in Table 1.

Table 1.

Sequence of oligonucleotides used in this work

Unc-Cy3 _sense	(Cy3)#(mC)#(mU)#(mU)(fU)(mG)(fU)(mA)(fA)(mA)(fA)(mG)(mU)(mU)(fU)#(mU)#(mA)(dT)(dT)
DCA-Cy3 _sense	(Cy3)#(mC)#(mU)#(mU)(fU)(mG)(fU)(mA)(fA)(mA)(fA)(mG)(mU)(mU)(fU)#(mU)#(mA)(dT)(dT)-DCA
D-Cy3 _sense	(dZ)(dX)(dY)(dX)(dT)(dT)(Cy3)#(mC)#(mU)#(mU)(fU)(mG)(fU)(mA)(fA)(mA)(fA)(mG)(mU)(mU)(fU)#(mU)#(mA)
Cy3_antisense	V(mU)#(fA)#(mA)(fA)(fA)(fC)(mU)(fU)(mU)(fU)(mA)(fC)(mA)(fA)#(mA)#(fG)#(mA)#(mA)#(mG)#(mA)#(fA)
Unc-JAK13033_sense	(mG)#(mU)#(mA)(fC)(mA)(fA)(mA)(fG)(mG)(fA)(mA)(mU)(mC)(fU)#(mG)#(mA)(dT)(dT)
DCA-JAK13033_sense	(mG)#(mU)#(mA)(fC)(mA)(fA)(mA)(fG)(mG)(fA)(mA)(mU)(mC)(fU)#(mG)#(mA)(dT)(dT)-DCA
D-JAK13033_sense	(C12)(SB)(C6)(SB)(dT)(dT)(mG)#(mU)#(mA)(fC)(mA)(fA)(mA)(fG)(mG)(fA)(mA)(mU)(mC)(fU)#(mG)#(mA)
JAK13033_antisense	V(mU)#(fC)#(mA)(fG)(fA)(fU)(mU)(fC)(mC)(fU)(mU)(fU)(mG)(fU)#(mA)#(fC)#(mU)#(mU)#(mC)#(fA)#(mU)
DCA-JAK1883_sense	(mG)#(mA)#(mA)(fA)(mC)(fA)(mU)(fU)(mG)(fA)(mA)(mU)(mA)(fA)#(mG)#(mA)(dT)(dT)-DCA
JAK1883_antisense	V(mU)#(fC)#(mU)(fU)(fA)(fU)(mU)(fC)(mA)(fA)(mU)(fG)(mU)(fU)#(mU)#(fC)#(mU)#(mG)#(mG)#(mU)#(fU)
NTC_sense	(mU)#(mU)#(mG)(fA)(mC)(fA)(mA)(fA)(mU)(fA)(mC)(mG)(mA)(fU)#(mU)#(mA)(dT)(dT)-DCAv1
NTC_antisense	V(mU)#(fA)#(mA)(fU)(fC)(fG)(mU)(fA)(mU)(fU)(mU)(fG)(mU)(fC)#(mA)#(fA)#(mU)#(mC)#(mA)#(fU)#(mU)

(#: PS backbone, m: 2′-o-methyl, f: 2′-fluoro, C12: hexaethylene spacer, C6: triethylene spacer, SB: symmetrical branching, V:(E)-vinylphosphonate, DCAv1: Docosanoic acid functionalized CPG.).

Testing siRNA efficacy in human ex-vivo skin

Freshly discarded post-panniculectomy abdomen skin was obtained from UMass Chan Biospecimen, Tissue, and Tumor Bank or was purchased from BIOIVT and ZENBIO. The human specimens were hydrated with PBS, and subcutaneous fat was excised prior to treatment. Then, the skin is submerged in an antifungal solution diluted in PBS to prevent any fungal infection or growth, followed by another PBS wash. Following that, siRNA is administered into the ex vivo skin and incubated in Iscove′s Modified Dulbecco′s Media (Sigma-Aldrich; #I3390) supplemented with 10% fetal bovine serum, 100 U/mL penicillin-streptomycin, and 0.05 μM 2-mercaptoethanol (termed complete media).

For intradermal testing, 8 mm skin biopsies were punched, and 10 nmols of siRNA was intradermally injected with a volume of 50 µL. The punches were cultured for 96 h at 37 °C in a 24-well plate with 2 mL of complete media per punch. To induce CXCL9, 10, and 11 expression, after 96 h of siRNA treatment, skin biopsies were treated with new media containing 10 ng/mL recombinant human IFN-γ protein (R&D System, #285-IF-100) and 10 ng/mL recombinant human TNF protein (R&D System, #210-TA-020). mRNA expression was measured by the Quantigene 2.0 assay.

Dissociation of epidermis and dermis skin biopsies

The skin biopsies were incubated in 0.5 mL of 30 mg/mL dispase II (Roche, #04942078001) in Dulbecco’s Modified Eagle’s Media (Corning Cellgro; #10-013CV) for 1 h at 37 °C. Epidermis and dermis were manually separated prior to being placed in RNAlater® for 24 h as previously reported²¹. Epidermis and dermis were placed in 0.75 to 1 mL of homogenizing solution (Invitrogen; QS0517) containing 10 µL of 10 mg/mL proteinase K (Invitrogen, #AM3546) and were mechanically dissociated and homogenized. Specifically for the mechanical dissociation, the epidermis was dissociated using QIAGEN Tissue Lyser II as previously described following Quantigene protocols²¹, and the tougher dermis layer mechanical dissociation was carried out using manual dissociation probe or using the Bullet blender 5E Gold+ (Next Advance; #BB5EAUP) to ensure disruption of the dense extracellular matrix in skin tissue. Homogenates were then incubated at 55 °C for 30 min, followed by centrifugation at 14,000 × g for 5 min. The clear supernatant was then collected for subsequent assays. Silencing of the target mRNA expression was measured through the Quantigene 2.0 assay.

mRNA quantification

The target mRNA levels were measured by using Quantigene 2.0 assays (Affymetrix). All QuantiGene detection probe sets were ordered from ThermoFisher. For human skin, the following probes were used: human JAK1 (SA-50455), human CXCL9 (SA-12372), human CXCL10 (SA-50393), human CXCL11 (SA-50464), and human ACTB (SA-10008) as a housekeeping gene. For pig skin, the following probes were used: porcine JAK1 (SF-4295068), pig CXCL9 (SF-4296654), pig CXCL10 (SF-4295763), pig CXCL11 (SF-4299676), and pig ACTB (SF-4295627) as a housekeeping gene.

Human skin flow cytometry

Skin punch biopsies (4 mm) were injected with 50 µL of 100 µM Cy3-labeled siRNAs formulated in 1x PBS. Samples were then cultured for 24 h in 2 mL/punch of Iscove’s Modified Dulbecco’s Medium (IMDM, Sigma; #I3390–500ML) supplemented with 10% fetal bovine serum (Gibco; #26140), 1x penicillin–streptomycin (Sigma; #P4333–100ML), and 50 µM of 2-mercaptoethanol (Sigma; #M3148). Skin biopsies were digested in IMDM media containing 1.6 mg/mL collagenase, type 4 (Worthington-biochem; #LS004186). Briefly, skin biopsies were diced and minced into small pieces and incubated at 37 °C for 3 h at 200 rpm in gentleMACS C tubes (Miltenyi Biotec, #130-093-237) on a MACS dissociator. The suspended cells were filtered through a 70 µm cell strainer (Fisher Scientific) and stained with the following dye and fluorescent antibodies: Human TruStain FcX (Biolegend, #422302), Sytox Blue (Thermo Fisher), BV605 anti-human CD45 (Biolegend), APC anti-human CD117 (c-kit) (Biolegend), PE-VIO770 anti-human CD140 (Biolegend) and BRIGHT-FTIC anti-Human CD49f (Biolegend). The samples were analyzed using BioRad ZE5 (BioRad), and data were processed using FlowJo 10 software. Gating of cell types were as follows: Immune cells, CD45+; Fibroblasts, CD45- CD140+; Endothelial cells, CD45- CD49f+; Keratinocytes, CD45- Ckit-; Melanocytes, Ckit+.

Peptide nucleic acid (PNA) hybridization assay

Tissue biodistribution levels of the siRNA were measured using a PNA hybridization assay as previously described. Briefly, the accumulation of the JAK1³⁰³³ was quantified using a custom Cy3-labeled fluorescent PNA oligonucleotide probe (Alexa488-OO-GTACAAAGGAATCTGA-KK, PNABio) that is fully complementary to the antisense strand. Pre-weighted tissue punches were placed in 500 μL homogenizing solution (Invitrogen; QS0517) containing 0.2 mg/mL proteinase K (Invitrogen, #AM3546) in a QIAGEN collection microtube holding a 3-mm tungsten bead. The tissues were then homogenized for 10 min under 30 Hz frequency using a QIAGEN TissueLyser II. Homogenized tissues were incubated at 55 °C for 30 min and centrifuged at 1000 × g for 10 min. Sodium dodecyl sulfate was precipitated from homogenates by using 3 M potassium chloride, followed by a centrifugation for 15 min at 5000 × g. siRNAs in the supernatant were then collected and hybridized to the PNA probe at 95 °C, followed by slow cool down. Anion-exchanged chromatography was used to analyze the sample mixtures on an Agilent 1260 Infinity quad-pump HPLC with a 1260 FLD fluorescent detector, using a DNAPac PA100 column (ThermoFisher) for peak separation. Gradient and buffers were used as previously described^31,43. Cy3 fluorescence was monitored, and peaks were integrated. Final concentrations of oligos were determined by using a calibration curve generated by spiking known quantities of siRNA into tissue lysates from untreated animals.

Pig models for siRNA pharmacokinetics and pharmacodynamics

Pig experiments for in vivo intradermal and intravenous studies were performed by SCANTOX for the pharmacokinetics and accumulation of JAK1³⁰³³, and by Altasciences for pharmacokinetics and pharmacodynamics (JAK1³⁰³³ and JAK1⁸⁸³). Briefly, siRNAs were formulated in 1x PBS at 200 or 400 µM and administered intradermally or intravenously for the indicated doses. Intravenous dose was 10.76 mg/kg

For JAK1³⁰³³ pharmacokinetics, a study was conducted in 10 male Göttingen SPF minipigs (n = 5 per group, 4 months of age), where areas of treatment were tattooed prior to injection (50 µl of 200 µM siRNA/1 cm² site; 10 nmol final dose/biopsy; 8 mm biopsy). Detailed biopsy location and collection are provided in Supplementary Fig. S3. Skin biopsies and blood/serum samples were collected over a 56-day period. Animals were euthanized at varying time periods over a 56-day period to obtain terminal heart, spleen, liver and kidney biopsies for systemic distribution. All biopsies were flash frozen and reconstituted in RNAlater® 24 h prior to PNA hybridization analysis as described above. CBC and blood chemistry measurements were performed by SCANTOX.

For JAK1⁸⁸³ pharmacokinetics and pharmacodynamics, a study was conducted in 8 Göttingen SPF minipigs, n = 2 pigs/sex for I.D. JAK1³⁰³³ and JAK1⁸⁸³, a total of n = 4 pigs per siRNA; similar setup for I.V. injection. Areas of treatment were tattooed prior to injection (50 µL of 400 µM siRNA/1 cm² site; 20 nmol final dose/biopsy; 8 mm biopsy). No sex-dependent differences were observed; therefore, the data were based on combined female and male results. Skin biopsies and blood/plasma samples were collected over a one-week period. Animals were euthanized one week post administration, and tissues were collected to assess systemic biodistribution. All biopsies were flash frozen, and skin biopsies were reconstituted in RNAlater® 24 h prior to mRNA quantification using Quantigene Assay 2.0. siRNA amounts in plasma, tissue, and skin were quantified by Altasciences via LC-MS/MS. JAK1 mRNA silencing in pigs for a one-month duration was conducted similarly, with the addition of the 24-h stimulation protocol to assess downstream chemokine measurements. For the one-month silencing study, a Göttingen SPF male minipig was injected intradermally (50 µL of 400 µM siRNA/1 cm2 site; 20 nmol final dose/biopsy; 8 mm biopsy; total of 4 biopsies per group), and samples were collected and processed as mentioned previously.

Pharmacokinetic parameters were determined using a non-compartmental approach within Phoenix® WinNonlin® software (version 8.0; Certara USA, Inc., Princeton, New Jersey), which is appropriate for both intravenous bolus and extravascular routes of administration. The linear trapezoidal method with linear interpolation was applied to estimate parameters from individual plasma and skin (biopsy) concentration-time data. Descriptive statistics for these parameters were calculated for males, females, followed by sexes combined, as no sex-dependent differences were observed. AUC_last is defined as the area under the concentration versus time curve from the start of dose administration to the time after dosing at which the last quantifiable concentration was observed. Absolute bioavailability is defined as the percentage of the administered siRNA drug that reaches the systemic circulation intact.

Cytokine level measurement

Pigs were injected subcutaneously as described in Supplementary Fig. S3. Blood was collected pre-dosing, and on days one, two, and seven post-intradermal injection. Serum was separated and analyzed for cytokine concentration measurement using Porcine 17-plex Discovery Panel (CD31, GM-CSF, IFN-gamma, IL-1 alpha, IL-1 beta, IL-1ra, IL-2, IL-4, IL-6, IL-8, IL-10, IL-18, MMP-1, PDGF-BB, Serpin E1, TNF, IL-12). Serum samples were analyzed as described in the manufacturer’s protocol via the UMass Metabolic Disease Research Center.

Statistics & reproducibility

The data was analyzed and graphed using the GraphPad Prism 9 software (GraphPad Software, Inc.). Details of methods for statistical analysis and sample sizes are included in each figure legend. P-values of less than 0.05 were considered significant for comparisons between groups. Most studies had at least 3 biologically independent replicates. No data points were excluded unless there was a failed assay measurement. No technical replicates were used, and all studies had at least one biological replicate. Randomization was followed. No blinding was necessary, hence not utilized in the experiments reported.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Author contributions

H.H.F., M.Z.U., Q.T., J.F.A., J.E.H., A.K., C.B., C.B.P., and T.P. conceived the project. H.H.F., M.Z.U., Q.T., J.F.A., A.K., S.F., S.J., C.B., C.B.P., and J.O.G. contributed to the experimental design. S.J. was the sole point of contact responsible for overseeing the conduct of studies at Altasciences. H.H.F., M.Z.U., and Q.T. contributed experimentally to human ex vivo studies, including flow cytometry and gene silencing. H.H.F., Q.T., M.Z.U., M.O.R., and K.Y.G. contributed experimentally to ex vivo human skin stimulation and downstream signaling. M.Z.U., V.R., K.Y.G., and R.G. contributed experimentally to siRNA accumulation in porcine organs. M.O.R. and T.J.R.O. contributed experimentally to porcine silencing, stimulation and downstream signaling. M.Z.U., Q.T., H.H.F., M.O.R., T.J.R.O., C.D., R.C.F., C.L., and A.S. helped with overall skin processing and experimental setup. H.H.F. synthesized all siRNA compounds used in the manuscript, with H.H.F. and H.F.S. contributing to the design of the dendritic siRNA conjugation. H.H.F., Q.T., M.Z.U., C.B., J.F.A., and A.K. wrote the manuscript. All authors provided feedback and approved the manuscript.

Peer review

Peer review information

Nature Communications thanks Amogh Vaidya and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

All data generated in this study are provided in the Supplementary Information and Source Data file. A Source Data file is provided with this paper. Source data are provided with this paper.

Competing interests

The authors declare the following competing financial interest(s): The University of Massachusetts Chan Medical School holds patent or filed patent applications for the modulation of JAK1 with RNAi-based technologies (patent application number 18393044; Oligonucleotides for IFN-γ signaling pathway modulation; status pending), docosanoic acid conjugate (patent application number 17377632; Conjugated oligonucleotides for tissue-specific delivery) and dendritic conjugate for skin delivery (Patent application number 18592943; status pending). A.K. serves on the scientific advisory board of Alys Pharmaceuticals. J.E.H. owns equities in Rheos Medicines; is a founder of Alys Pharmaceuticals and Villaris Therapeutics. Select authors of this publication are listed as inventors on RNAi technology patents (HHF, M.Z.U., Q.T., J.F.A., J.E.H., A.K., M.Z.U., R.F., and K.G.) that have been licensed to biotech and pharmaceutical companies. The following patents are licensed by Aldena Therapeutics Ltd.: patent application number 18393044 (Oligonucleotides for IFN-γ signaling pathway modulation) and patent application number 17377632 (Conjugated oligonucleotides for tissue-specific delivery). C.B.P., C.B., J.O.G., and T.P. are executive employees of Alys Pharmaceuticals. SF is employed by Certara. SJ is employed by Altasciences. The remaining authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-026-68993-1.