Nucleic Acid Therapy for the Skin

Abstract

Advances in sequencing technologies have facilitated the identification of the genes and mechanisms for many inherited skin diseases. While targeted nucleic acid therapeutics for diseases in other organs have begun to be deployed in patients, the goal of precise therapeutics for skin diseases has not yet been realized. First, we review current and emerging nucleic acid-based gene editing and delivery modalities. Next, current and emerging viral and nanoparticle vehicles for the delivery of gene therapies are reviewed. Finally, specific skin diseases that could benefit optimally from nucleic acid therapies are highlighted. Through adopting the latest technologies and addressing specific barriers related to skin biology, nucleic acid therapeutics have the potential to revolutionize treatments for patients with skin disease.

INTRODUCTION

Skin diseases were amongst the earliest medical conditions to be described in medical texts due to the obvious visual appearance of many dermatologic conditions, allowing for their early recognition (Hartmann, 2016). Despite the noticeable appearance of many skin diseases, including numerous genodermatoses, the underlying genetic causes for any skin disease were not known until the 1990s, when researchers first mapped mutations in KRT14 to the underlying pathology of epidermolysis bullosa simplex (Coulombe et al., 1991). At the turn of the century with the first draft of the human genome (Venter et al., 2001), clinicians and researchers were able to map and identify disease-causing gene variants more rapidly. Today, next generation sequencing allows us to quickly identify these disease-causing variants (Chiu et al., 2021). With this ability to rapidly diagnose the genetic cause of a patient’s disease, the obvious next hurdle is how to use this knowledge to advance the treatment of patients with genetic skin diseases. A number of pre-clinical and clinical therapies focused on gene therapy hope to address this barrier and thus have the potential to treat some of the genetic skin disorders that affect millions of people worldwide (Karimkhani et al., 2017).

Genetic skin diseases, or genodermatoses, encompass a wide range of clinical manifestations, morbidities, and incidences. Some of the best recognized genodermatoses are ones which have devastating consequences. Dystrophic epidermolysis bullosa is just such a genodermatosis, which affects between 1:60,000 to 1:500,000 people. Patients suffer from debilitating chronic wounds which are complicated by scarring, secondary infections, and aggressive squamous cell carcinoma, frequently resulting in death by early adulthood (Bardhan et al., 2020). With no curative treatment, the disease is managed primarily by injury avoidance and supportive care, which has spurred much work towards development of gene therapies for the disease. In contrast, ichthyosis vulgaris, despite being one of the most common genodermatoses affecting about 1:100 to 1:250 people, has a favorable prognosis among affected patients who have normal lifespans and gradual improvement in symptoms with age (Thyssen et al., 2013). While symptoms are usually well managed with topical emollients and irritant avoidance, patients with ichthyosis vulgaris are at a higher risk of developing asthma, allergy, and atopic dermatitis, making its clinical impact on patients quite significant (Irvine et al., 2011). While most efforts related to gene therapy have focused on the deadliest genodermatoses, breakthroughs in the development of nucleic acid therapies for severe skin diseases could eventually offer novel treatments for patients with common, ‘milder’ genodermatoses, like ichthyosis vulgaris.

In this review, we discuss the current state of nucleic acid therapy for genodermatoses. First, we cover the different types of nucleic acid therapeutics being developed. Next, we discuss the vectors used for their therapeutic delivery. Finally, the skin diseases for which gene therapies are currently being developed are described. Ultimately, our goal is to highlight both the successes of gene therapy and challenges that must be overcome for these promising gene therapy technologies to make their way into the clinic.

OVERVIEW OF NUCLEIC ACID THERAPIES

The first nucleic acid therapy approved by the US Food and Drug Administration (FDA) was an anti-sense oligonucleotide (ASO) (Roehr, 1998), which is comprised of a short synthetic nucleic acid sequence that binds to a complementary target mRNA sequence. Binding results in post-transcriptional inhibition of protein synthesis by modulating mRNA processing, splicing, translation, or degradation through various unique pathways (Figure 1). Delivered as single-stranded oligonucleotides, injection of naked ASOs in vivo results in their rapid clearance and degradation by various nucleases (Eckstein, 2014). Consequently, much work has focused on chemically modifying the nucleic acids or adding other modifications to increase their stability and potency (Cummins et al., 1995, Roberts et al., 2020). ASOs are eventually degraded over time and do not offer a permanent genetic solution, which requires their frequent re-dosing, weekly for some drugs, to carry the same effect. Nonetheless, there are several FDA-approved ASO-based therapies for a variety of genetic diseases (Table 1), such as Duchenne muscular dystrophy (Dzierlega and Yokota, 2020) and hereditary transthyretin-mediated amyloidosis (Benson et al., 2018), with a few ongoing Phase III clinical trials using ASOs for skin diseases such as melanoma and hereditary angioedema.

Figure 1. Mechanisms and delivery modalities for nucleic acid therapeutics.

Endocytic uptake of ASOs can modulate mRNA expression by sterically inhibiting translation, by RNAse H-mediated degradation, or by modulating splicing. For siRNAs, exogenously delivered dsRNAs are processed by Dicer into siRNAs (or directly delivered as siRNAs) which are then loaded onto the RNA-induced Silencing Complex (RISC) to allow Ago2-mediated target mRNA degradation. Gene editors, delivered as ribonucleoproteins or indirectly as mRNA, can directly modify DNA. As a delivery vector, AAV and HSV undergo receptor-mediated endosomal uptake, followed by endosomal escape and nuclear import to form extra-chromosomal episomes that allow transgene expression. Lentivirus undergoes receptor-mediated fusion and endocytosis to release its viral RNA into the cytoplasm. Viral RNA is reverse transcribed (RT) into viral DNA which then integrates into the host genome for transgene expression. As a nonviral vector, LNPs can deliver various nucleic acid cargos via endosomal uptake and escape into the cytoplasm.

Table 1.

Relevant FDA-Approved In Vivo Nucleic Acid Therapies

Mechanism	Modality	Target Organ	Target Gene	Disease	Delivery Vehicle	Drug/Year	Study
Gene Knockdown	siRNA	Liver	TTR	hereditary transthyretin-mediated amyloidosis	LNPs, intravenous	Patisiran (Onpattro) 2018	(Adams et al., 2018)
	siRNA	Liver	ALAS1	acute hepatic porphyria	Tri-GalNAc conjugation of siRNAs, intravenous	Givosiran (Givlaari) 2019	(Balwani et al., 2020)
	siRNA	Liver	HAO1	primary hyperoxaluria type 1	Tri-GalNAc conjugation of siRNAs, intravenous	Lumasiran (Oxlumo) 2020	(Garrelfs et al., 2021)
	siRNA	Liver	PCSK9	heterozygo us familial hyperchole sterolemia	Tri-GalNAc conjugation of siRNAs, intravenous	Inclisiran (Leqvio) 2021	(Raal et al., 2020, Ray et al., 2017, Ray et al., 2020, Wright et al., 2021)
	ASO	Nervous System	SOD1	amyotrophic lateral sclerosis	ASO, intrathecal	Tofersen (Qalsody) 2023	(Miller et al., 2022)
	ASO	Nervous System	SMN1	spinal muscular atrophy	ASO, intrathecal	Nusinersen (Spinraza) 2016	(Finkel et al., 2017, Mercuri et al., 2018)
	ASO	Liver	TTR	hereditary transthyretin-mediated amyloidosis	ASO, subcutaneous	Inotersen (Tegsedi) 2018	(Benson et al., 2018)
	ASO	Liver	APOB	Homozygous familial hyperchole sterolemia	ASO, subcutaneous	Mipomersen (Kynamro) 2013	(Stein et al., 2012, Thomas et al., 2013)
	ASO	Skeletal and cardiac muscle	DMD	Duchenne muscular dystrophy	ASO, intravenous	Eteplirsen (Exondys 51) 2016, golodirsen (Vyondys 53) 2019, viltolarsen (Viltepso) 2020, casimersen (Amondys 45) 2021	(Brogna et al., 2019, Clemens et al., 2020, Frank et al., 2020, Wagner et al., 2021)
Gene Replace ment	Viral DNA Episome	Skin	COL7 A1	epidermolysis bullosa	HSV-1, topical	Beremagene geperpavecsvdt (Vyjuvek) 2023	(Guide et al., 2022, Gurevich et al., 2022)
	Viral DNA Episome	Cardiac and Skeletal Muscle	DMD	Duchenne muscular dystrophy	AAV, intravenous	Delandistrogene moxeparvovec (Elevidys) 2023	(Mendell et al., 2020, Mendell et al., 2023, Zaidman et al., 2023)
	Viral DNA Episome	Liver	F8	Hemophilia A	AAV, intravenous	Valoctocogene roxaparvovec-rvox (Roctavian) 2023	(Ozelo et al., 2022)
	Viral DNA Episome	Liver	F9	Hemophilia B	AAV, intravenous	Etranacogene dezaparvovec (Hemgenix) 2022	(Pipe et al., 2023)
	Viral DNA episome	Nervous System	SMN1	spinal muscular atrophy 1	AAV, intravenous	Onasemnogene abeparvovec (Zolgensma) 2019	(Mendell et al., 2017)
	Viral DNA episome	Eye	RPE65	Leber congenital amaurosis (biallelic RPE65 mutations)	AAV, subretinal injection	Voretigene neparvovec (Luxturna) 2017	(Maguire et al., 2009, Maguire et al., 2008)

Short interfering RNAs (siRNAs) are short sequences of double-stranded RNA that utilize the endogenous RNA interference (RNAi) pathway via the RNA-induced silencing complex (RISC) to silence target genes. These synthetic RNA sequences can distinguish single-nucleotide differences in target sequences, making them ideal for allele-specific gene silencing in dominant gain-of-function or dominant-negative diseases (Hickerson et al., 2008). Since siRNAs are delivered as double-stranded RNA (dsRNA) sequences and remain bound to the RISC intracellularly, they stably exert a greater long-term effect requiring less frequent dosing compared to ASOs. Similarly to ASOs, siRNAs do not offer a permanent genetic solution and again would require lifetime redosing. In addition, dsRNAs are highly hydrophilic, which limits cellular uptake and delivery, and immunogenic, which can limit dosage. For these reasons, current work is investigating modifications to improve cellular permeability, delivery, and activity. There are currently six FDA-approved siRNA therapies (Table 1), including nedosiran and lumasiran (Garrelfs et al., 2021) for primary hyperoxaluria type 1, and patisiran (Adams et al., 2018) and vutrisiran (Adams et al., 2023) for hereditary transthyretin-mediated amyloidosis, with a few early-stage clinical trials using siRNA for skin diseases.

In the early 2000s, researchers discovered a bacterial immune system against viruses that utilized clustered regularly interspaced short palindromic repeats (CRISPR) elements encoding CRISPR RNA (crRNA) that could bind to DNA sequences of viral bacteriophages (Mojica et al., 2005). CRISPR-associated nucleases (Cas) used the crRNA as a guide sequence to cleave the targeted complementary DNA sequences of these viral bacteriophages for antiviral defense. In their Nobel Prize-winning work in the early 2010s, Jennifer Doudna, Emmanuelle Charpentier, and researchers demonstrated that this CRISPR-Cas system could be engineered for site-specific DNA cleavage (Jinek et al., 2012), and others demonstrated that this system could work in mammalian cells, not as an antiviral defense system, but as a gene editor (Cong et al., 2013, Mali et al., 2013). In a mammalian cell, a programmable guide RNA (gRNA) directs the Cas nuclease to its complementary sequence site in the genome to introduce a double-stranded break (DSB)(Cong et al., 2013). Editing occurs when mammalian cells attempt to repair the DSB, generally through the non-homologous end joining (NHEJ) pathway or the homology-directed repair (HDR) pathway (Chang et al., 2017). In NHEJ, often considered an error-prone repair pathway, insertions and deletions (indels) can be introduced at the target site. These indels can be exploited for gene disruption editing purposes to cause frameshifts and disrupt gene function (Davis and Chen, 2013). In HDR, which occurs in dividing cells such as basal keratinocytes, a set of HDR-specific proteins can use sister chromatids as a template for precise DNA repair (Kocher et al., 2017). This repair process can be exploited by supplying an exogenous repair template that contains sufficient homology overlap with the surrounding sequence around the DSB, however HDR competes with the more-efficient NHEJ repair process, such that most edits are indels (Lin et al., 2014). The use of CRISPR-Cas9 nucleases for gene editing has found its way to ongoing clinical trials for treating various human diseases including sickle cell disease and beta-thalassemia (Frangoul et al., 2021, Locatelli et al., 2022), various cancers (Liu et al., 2023), and more. Despite these promising trials, there are some limitations. First, DSBs can not only result in indels, but also translocations, chromothripsis, and other large chromosomal rearrangements, that may have deleterious oncogenic consequences aside from being an undesired edit (Leibowitz et al., 2021). Second, gene disruption approaches only work for some genetic diseases in which the body can tolerate loss of the edited gene, but many disease-causing variants occur in essential genes.

Base editors exploit the ability of CRISPR-Cas to precisely target specific sites in the genome, but instead of using the Cas nuclease to create a DSB and have cell-repair processes mediate editing, rely on a deaminase enzyme fused to a catalytically dead Cas nuclease or Cas nickase (nCas) and chemical modification to mediate editing (Gaudelli et al., 2017, Komor et al., 2016). Since their invention in 2016, multiple base editors have been developed that can allow all base pair substitutions within a target genomic sequence (Chen et al., 2024, Kurt et al., 2021, Zhao et al., 2021). In addition to being highly efficient, base editors do not induce doublestranded breaks. While most suitable for correcting point mutations, single nucleotide editing of splice sites or promoter regions can introduce greater effects (Chai et al., 2023, Zeng et al., 2020). Advancements in base editing efficiency, safety, and delivery have enabled clinical trials to begin using base editors to edit a few gene targets (Kingwell, 2022). Prime editors are another gene editing tool that again rely on the DNA-targeting ability of nCas, but mediate editing with a fused engineered reverse transcriptase without the need for double-stranded breaks (Anzalone et al., 2019). An extended gRNA, called a prime editing gRNA, contains a reverse transcription template encoding the desired edit. Being completely programmable, any desired edit, including base pair transitions, transversions, and insertions and deletions of small lengths (up to about 50 bp) can be introduced (Steinbeck et al., 2024).

Gene editors enable permanent changes to the genome of a cell, which can propagate in dividing cells, allowing for the potential for treatments to be more durable. However, questions remain on limiting potential cancer-inducing off-target editing (Enache et al., 2020, Haapaniemi et al., 2018), potential immunogenicity of these bacterial nucleases (Charlesworth et al., 2019), and efficiency in creating the desired edit in the desired cell. An additional challenge towards the widespread deployment of gene editing is the requirement for more personalized therapies depending on the mutations present. This may raise the cost and slow the adoption of CRISPR-based therapies compared to “off-the-shelf” therapies that are the same for all patients (Urnov, 2021).

An alternative to permanently correcting the genome is to simply replace the missing or mutated gene. Gene replacement therapy involves directly replacing the missing gene and its encoding protein. To evade the activation of the innate immune system (Sun et al., 2013, Wu et al., 2013) and to promote transcription, the replacement DNA must be delivered to the nucleus either as a stable extra-chromosomal episome or through integration into the host genome. Episomal DNA replacement tends to be more efficient, but the episome is sometimes lost by dilution through cell division (Naso et al., 2017). In contrast, direct gene replacement into the genome allows for replication in all daughter cells, potentially allowing curative treatment, but increases the risk of mutagenesis. Gene replacement therapy would be most effective in skin diseases in which particular genes are lost due to loss-of-function mutations, such as epidermolysis bullosa and related congenital blistering diseases.

Gene replacement therapy could also be accomplished by supplying messenger RNA (mRNA) encoding a desired protein to target cells. Delivery of RNA tends to be more efficient than the delivery of DNA as it has the advantage of not requiring delivery to the nucleus. In addition, some forms of RNA have been found to be less immunogenic than DNA. Substitution of the mRNA base uridine with naturally occurring, modified pseudouridine greatly reduces the host inflammatory response – a boon that was utilized in the COVID-19 mRNA vaccines (Kariko et al., 2005). The primary disadvantage of RNA gene therapy is the short half-life of RNA, which typically requires repeated dosing to maintain its therapeutic effect. Efforts are ongoing to extend the pharmacokinetic window of mRNA through the use of circular RNA (circRNA), which resist exonuclease cleavage (Wesselhoeft et al., 2018, Yang and Wang, 2021, Yang et al., 2017, Zhao et al., 2019). However, circRNA still face limitations including decreased translation efficiency and the activation of innate immune responses, so circRNA have not yet been tested in clinical trials (Roberts et al., 2020, Roehr, 1998). Thus, additional studies are necessary before circRNA can be tested in clinical trials.

METHODS OF DELIVERY

One of the skin’s primary functions is to form a barrier against water loss and exogenous factors. The outermost layer of stratum corneum is composed of keratin-filled corneocytes in a dense lipid matrix, which presents challenges in permitting trans-epidermal delivery of nucleic acids. In many conditions (e.g., epidermolysis bullosa) with damaged skin, the natural barrier is disrupted as a part of disease pathogenesis, which provides an opportunity for delivery into the deeper epidermis and dermis. However, direct application of naked nucleic acids to the skin results in low uptake, requiring the use of other vectors to package and deliver nucleic acid therapeutics (Figure 1). As the largest organ in the body accounting for 8% of the total body mass and a surface area of approximately 2 square meters, the skin’s large surface area presents difficulties in safely delivering drugs while minimizing systemic uptake or side effects, especially in diseases that affect a high body surface area.

Viruses have naturally evolved to deliver their genetic cargo to many different cell types. This innate ability to deliver genetic material makes it an attractive vector for nucleic acid therapeutics. Among the many viral vectors that have been trialed in humans, a few have emerged as the most promising vectors for skin diseases. However, several ongoing challenges will need to be addressed for continued in vivo applications, including manufacturing cost and difficulty, and immunogenicity of the vector, which can limit dosing amount and frequency.

Herpes simplex virus (HSV) is an enveloped, non-integrating, double-stranded DNA virus. While wild-type viruses are highly immunogenic, recombinant variants that are replication-defective and less immunogenic have been developed for human gene therapy efforts (Epstein, 2009). HSV has many advantages as a viral vector, including broad tissue tropism, high transduction efficiency for both dividing and nondividing cells, and a large DNA message carrying capacity (up to 150 kilobases). However, HSV does not integrate into cellular DNA, but rather forms a stable extra-chromosomal episome, which is not replicated and thus may be diluted by cell division over time.

Retroviruses, including the most commonly deployed lentiviruses, are enveloped viruses that deliver RNA cargo that is reverse-transcribed and permanently integrated into the genome of an infected cell. While lentiviruses can both accept a larger genetic cargo and transduce a broad range of both dividing and nondividing cells, their potential for non-specific genomic integration, genotoxicity, and challenges related to manufacturing, have limited wider use in gene therapy (Milone and O’Doherty, 2018). In one notable success, autologous stem cells from a patient with junctional epidermolysis bullosa were cultured, retrovirally infected with the rescue gene, and expanded ex vivo before being transplanted back on to the patient (Hirsch et al., 2017).

Adeno-associated virus (AAV) is a nonenveloped, single-stranded DNA virus that contains a <5kb linear genome (Atchison et al., 1965, Hoggan et al., 1966). In gene therapy applications, recombinant AAV infects a target nucleus and forms a double-stranded DNA episomal concatemer (Duan et al., 1998). The most attractive features of AAV are its ability to infect quiescent cells, broad tissue tropism, low integration levels, and its relatively low immunogenicity, thus making it a popular choice for various ongoing gene therapy clinical trials and current FDA-approved gene replacement therapies (Table 1) (Kuzmin et al., 2021). Some limitations of AAV for skin gene therapy include its smaller cargo size, inability to integrate stably into the genome (Samulski et al., 1989), and inefficiencies associated with the infection of large areas of intact skin. Thus, AAV has not yet been widely demonstrated to be an effective vector for skin gene therapy.

Lipid nanoparticles (LNPs) are synthetic vesicles that have become increasingly effective and popular as a nonviral delivery method for nucleic acids. They are classically composed of four compounds: ionizable amino lipids, phospholipids, cholesterol, and poly(ethylene glycol)-lipids (Hou et al., 2021). LNPs have several advantages including their ability to encapsulate long nucleic acid sequences, ease of synthesis and production, and their ability to be administered repeatedly.

When delivered intravenously (IV), most LNPs are taken up by the liver, which has made it a functionally useful delivery tool for therapeutics in the liver as seen in several ongoing clinical trials and FDA-approved drugs (Table 1). This is likely due to formation of an ApoE-enriched protein corona in systemic circulation, which results in its preferential uptake in the liver through the LDL-receptor (Monopoli et al., 2012). To overcome this physiological barrier, some key advances have been made to engineer LNPs that can target other tissues. Addition of a supplemental fifth component to the traditional four-component LNP mixture allows selective organ targeting (SORT) of the lung or spleen with IV administration through an endogenous targeting mechanism of action (Cheng et al., 2020, Dilliard et al., 2021). Lung SORT LNPs, which bind vitronectin in the plasma, can efficiently target lung epithelium for gene correction in mouse models of cystic fibrosis (Bulcaen and Carlon, 2024, Sun et al., 2024). Alternatively, an active targeting mechanism can be employed based on attachment of ligands to the surface of nanoparticles (Dilliard and Siegwart, 2023). For example, conjugating anti-CD5 antibody fragments to the surface of four-component LNPs can allow targeting of T cells following IV administration (Rurik et al., 2022). While these approaches have not yet been applied to allow for specific delivery to the skin, they suggest a potential for skin-specific enrichment of LNPs following systemic administration. Alternatively, similar to aerosolized administration that has allowed for the effective delivery of LNP to the respiratory system (Lokugamage et al., 2021, Sun et al., 2024) direct topical administration might be a parallel approach to allow for delivery to the skin. However, transdermal transport of LNPs is limited by the numerous skin layers including the rate limiting stratum corneum layer (Barua and Mitragotri, 2014).

For delivery of nucleic acid therapeutics, LNPs have several advantages over viruses. LNPs can readily be manufactured at large-scale production, which reduces potential cost. For gene editing approaches, delivery of LNPs results in transient expression of editor nucleases following mRNA translation, which can reduce off-target editing when compared to the longer-lasting expression in viral delivery methods. While transient expression may not be ideal for gene replacement therapy strategies, because LNPs are a synthetic mixture of lipids, they are thought to have reduced immunogenicity which allows for repeated dosing. However, the long-term toxicity and potential side effects of repeated administrations of synthetic LNPs have not been fully explored. In addition, new formulations that can target specific tissues, like the skin, have yet to be developed. Thus, challenges for the widespread adoption of LNP mediated delivery of gene therapy persist.

GENETIC SKIN DISEASES

The pace of development of gene therapy for specific diseases has hastened in the past several years. Onpattro has been FDA-approved for the treatment of familial amyloid polyneuropathy (Adams et al., 2018), Luxturna for Leber’s congenital amaurosis (Maguire et al., 2009, Maguire et al., 2008), and Lyfgenia and Casgevy for sickle cell anemia (Frangoul et al., 2021). Despite the successes of nucleic acid therapy in organ systems such as the central nervous system (Mendell et al., 2017), skeletal muscle (Mendell et al., 2020, Mendell et al., 2023, Zaidman et al., 2023), liver, blood, and eye, its success in skin diseases remains relatively limited (Table 1). Herein, we discuss examples of nucleic acid therapy in skin diseases, challenges that have prevented their deployment in other diseases, and new approaches and potential directions for the future.

Epidermolysis bullosa (EB) is a clinically heterogenous group of heritable skin diseases characterized by skin fragility and the formation of blisters upon minimal mechanical trauma. All types of EB are caused by pathogenic variants in genes that encode proteins necessary for dermis-epidermis structural integrity.

Dystrophic EB (DEB) is most commonly caused by mutations in the COL7A1 gene encoding type VII collagen protein, which forms a homotrimer that acts as a main component of anchoring fibrils in attaching the dermis and the epidermis. Dominant negative mutations that result in abnormal, functionally impaired protein trimers cause dominant dystrophic epidermolysis bullosa (DDEB), whereas recessive mutations that result in reduced or a lack of homotrimers cause recessive dystrophic epidermolysis bullosa (RDEB). For RDEB, studies have shown that simply restoring functional type VII collagen expression to keratinocytes or fibroblasts or both can restore formation of anchoring fibrils at the dermal-epidermal junction. Chen and colleagues used lentivirus to restore type VII collagen in DEB cells in vitro (Chen et al., 2002). Grafting of a mixed-cell slurry of edited keratinocytes and dermal fibroblasts into mice resulted in the correct localization of type VII collagen at the basement membrane and formation of anchoring fibrils. Expression of type VII collagen was maintained for at least two months in the mice models, possibly due to intended lentiviral integration within stem cells. In an application to human patients, epidermal keratinocytes from a patient with juvenile epidermolysis bullosa were transduced with retrovirus to restore expression of LAMB3 protein. Transduced keratinocytes were expanded into skin grafts and then were autologously transplanted onto the patient over months. Full engraftment and epidermal regeneration were achieved in one month with normal localization of adhesion machinery, and about 80% of the patient’s total body surface area was eventually restored by the transgenic epidermis and maintained for at least 21 months (Hirsch et al., 2017). However, surgical skin grafting presents its own set of complications that are magnified in large area grafts (Adams and Ramsey, 2005). Similarly, in a clinical trial (NCT01263379), autologous keratinocytes from patients were transduced with retrovirus carrying the full length COL7A1 gene (Siprashvili et al., 2016, So et al., 2022). When grafted onto patients, type VII collagen expression was restored and correctly localized to anchoring fibrils, but expression and wound healing decreased over the following year.

Given the challenges of ex vivo approaches, attempts at topical gene therapy for DEB have been made (Guide et al., 2022, Gurevich et al., 2022). In November 2023, the FDA approved the first topical gene therapy, Beremagene-geperpavec (Vyjuvek), for the treatment of DEB. HSV-1 encoding full-length type VII collagen is mixed with an excipient gel and then applied to a local wound area. Transduction of local keratinocytes and fibroblasts allows type VII collagen to be expressed and form anchoring fibrils. In clinical trials, 65% (n=31) of treated wounds achieved complete closure compared to 26% of placebo-treated wounds at their primary endpoint of 22–26 weeks (Guide et al., 2022). Despite the presence of antibodies against HSV-1 at baseline in some of the patients, repeat dosing was tolerated and few adverse effects were observed. However, as DEB is characterized by multiple wound sites and fragility of the entire skin barrier, it is unclear whether the body can tolerate multiple simultaneous doses and whether a cumulative limit on Vyjuvek dosing exists. Furthermore, as EB is a lifelong disease, questions remain on how long expression of type VII collagen is maintained as the viral extra-chromosomal episome is diluted out in dividing skin cells and fibroblasts. As the expression of transgenic collagen cannot be easily regulated, special attention will need to be given to the viral dose and the number of corrected gene copies supplied to a given cell, as overexpression of type VII collagen may lead to carcinoma migration and invasion (Pourreyron et al., 2014).

To circumvent these problems of redosing or controlling viral amount or expression levels, precise gene editing correction of RDEB could restore endogenous levels of type VII collagen and theoretically offer a one-time curative treatment as the corrected gene would be maintained into dividing cells. To this end, several studies (Hong et al., 2022, Naso et al., 2023, Osborn et al., 2020, Sheriff et al., 2022) have shown that base editing correction of RDEB-causing point mutations in patient-derived cells can efficiently restore type VII collagen protein expression, and skin grafting of corrected cells in mice can form functional AFs. While point mutations are the most common cause of RDEB, small deletion or insertion mutations cannot be directly corrected with base editing. However, as type VII collagen is a highly repetitive protein structure, removal of sections of the protein containing premature termination codon sequences can restore production of partially truncated but fully functional protein. To this end, CRISPRCas nucleases were used to excise exons containing these mutations, a technique known as exon skipping, and the truncated protein was found to be functional in restoring epidermis-dermis connection (Bonafont et al., 2019). Whereas RDEB can be treated by simply supplying the corrected gene, DDEB must be treated through the elimination of dominant negative mutations that impair the function of the wild-type protein. This can occur through either gene editing of the mutated gene (Shinkuma et al., 2016), gene silencing (Pendaries et al., 2012), or possibly exon skipping (NCT05529134). [Exon skipping is the basis of a few FDA-approved drugs and ongoing clinical trials for the neuromuscular genetic disease Duchenne muscular dystrophy, also caused by lack of a structural protein in muscle cells, suggesting a potential therapeutic avenue for RDEB.] One general limitation of gene therapy which persists for gene editing is the need to effect the change over the entire skin surface. Currently, there are no clinical trials yet for gene editing correction of dystrophic or junctional epidermolysis bullosa, but the success of pre-clinical approaches suggests that these efforts may be forthcoming.

Whereas EB can affect large body surface areas thereby requiring systemic approaches or therapies, some skin diseases have primarily focal manifestations that may be more amenable to local administration of nucleic acid therapeutics. Patients with pachyonychia congenita predominantly have the eponymous thickened nails as well as painful focal plantar and palmar hyperkeratosis that greatly limit their ability to walk unaided. Treatment is currently limited to physical trimming of calluses and nails or use of topical keratinolytics. Pachyonychia congenita is caused by dominant-negative mutations in any one of the five genes (KRT6A, KRT6B, KRT6C, KRT16, or KRT17) encoding keratin proteins that results in a dysfunctional keratin filament network and fragile keratinocytes (O’Toole et al., 2024). Preclinical studies showed that siRNA-based knockdown of the mutant transcripts could restore the keratin filament network (Leachman et al., 2008, Hickerson et al, 2008), and this work was extended to a single-patient Phase Ib trial (Leachman et al., 2010). The siRNA, named TD101, was injected twice weekly into the superficial dermis at the center of a plantar callus with increasing dose and volume as tolerated. Subjective patient assessment of the callus as well as the size of the callus was significantly improved after two months of treatment compared to a vehicle-control-treated contralateral plantar callus of similar size, but this improvement was just as rapidly lost following discontinuation of the drug. While systemic side effects were not observed, intradermal delivery was technically challenging and hampered by significant pain requiring benzodiazepines, pain medications, and nerve blocks. Like pachyonychia congenita, additional skin diseases like Epidermolysis Bullosa variants, Congenital Bullous Ichthyosiform erythyroderma, Xeroderma pigmentosum, Darier’s Disease, Hailey-Hailey, Tuberous Sclerosis, and Neurofibromatosis, all show variability of penetrance in the affected skin surface. Thus, focusing on treating the most severely affected areas might decrease potential side effects and facilitate the development and testing of gene therapies.

While nucleic acid therapeutics are ideal for treatment or correction of genodermatoses, they also have great potential for wound healing, inflammatory skin diseases, or skin cancers. Use of programmable siRNAs allows for more narrow and specific regulation of target molecules, which may diminish systemic side effects. Ruxolitinib and other JAK inhibitors are widely used for inflammatory or autoimmune skin conditions including psoriasis, atopic dermatitis, alopecia areata, and vitiligo, but they bind to a kinase domain that is conserved across multiple JAK family inhibitors which can lead to unintended side effects (Papp et al., 2021). Local injections of siRNAs specifically targeting JAK1 transcripts or upstream IFN-γ receptors not only reduce levels of the target transcripts and downstream JAK1-dependent IFN-γ-inducible chemokines, but also reduced inflammatory responses in ex vivo human skin and in mouse models (Tang et al., 2023, Tang et al., 2022). In other cases, siRNAs can uniquely target genes for which no current small molecule or antibody drug has been developed. Topical application of siRNA against Fidgetin-like 2 (FL2), a microtubule-severing enzyme that is a negative regular of cell migration, was shown to significantly enhance the speed of wound closure and the quality of re-epithelization and collagen organization (Charafeddine et al., 2015, O’Rourke et al., 2019).

Additional indirect mechanisms to treat skin disease are also either available or in development. Specific examples include Givosiran for porphyria, which could be extended to treat cutaneous porphyrias (Balwani et al., 2020). Alternatively, strategies deployed to allow for the expression of proteins that treat inflammatory skin diseases are in development. For example, one unique strategy is to indirectly deliver the encoded protein to the skin. In a preclinical study, Cheng and Farbiak and colleagues utilized the innate tendency for LNPs to transfect the liver but incorporated a signal peptide sequence for secretion into the packaged mRNA to allow efficient extracellular secretion (Cheng et al., 2023). In this way, the liver produces high amounts of an encoded protein, but rather than having expression constrained to a transfected hepatocyte, would allow the body’s endogenous mechanism of protein trafficking to systemically deliver a desired protein. Addition of a signal peptide into anti-hTNFα (etanercept) resulted in its expression in systemic circulation, providing therapeutic benefit in an imiquimod-induced psoriasis mouse model.

CONCLUSION

In this review, we highlighted several types of nucleic acid therapeutic approaches including gene editing, gene replacement, and gene silencing. Both viral vectors and non-viral vectors have been developed and are being investigated for their translation into the clinic. While there is only one topical gene therapy for local wounds currently used in clinics, many pre-clinical studies and ongoing clinical trials bring hope for the future integration of nucleic acid therapies to the dermatology clinic.