Leading Edge: Emerging Drug, Cell, and Gene Therapies for Junctional Epidermolysis Bullosa

Abstract

Introduction:

Junctional epidermolysis bullosa (JEB) is a rare inherited genetic disorder with limited treatments beyond palliative care. A major hallmark of JEB is skin blistering caused by functional loss or complete absence of major structural proteins of the skin. Impaired wound healing in patients with JEB gives rise to chronic cutaneous ulcers that require daily care. Wound care and infection control are the current standard of care for this patient population.

Areas covered:

This review covers research and clinical implementation of emerging drug, cell, and gene therapies for JEB. Current clinical trials use topical drug delivery to manipulate the inflammation and re-epithelialization phases of wound healing or promote premature stop codon readthrough to accelerate chronic wound closure. Allogeneic cell therapies for JEB have been largely unsuccessful, with autologous skin grafting emerging as a reliable method of resolving the cutaneous manifestations of JEB. Genetic correction and transplant of autologous keratinocytes has demonstrated persistent amelioration of chronic wounds in a subset of patients.

Expert Opinion:

Emerging therapies address the cutaneous symptoms of JEB but are unable to attend to systemic manifestations of the disease. Investigations into the molecular mechanism(s) underpinning the failure of systemic allogeneic cell therapies are necessary to expand the range of effective JEB therapies.

1. Introduction

Epidermolysis bullosa (EB) is an inherited group of genetic disorders that primarily affect the skin. Of this group, junctional epidermolysis bullosa (JEB) is one of the most rare and severe types – affecting three in one million children [1]. JEB is caused by the absence or functional loss of heterotrimer laminin-332 (LM332), Type XVII collagen (C17), or integrin α6β4 [2-4]. Mutations in LM332 subunits that result in complete protein loss are characterized as “generalized severe” (formerly Herlitz-type), while mutations in C17 or reduced or dysfunctional LM322 result in “generalized intermediate” JEB (formerly non-Herlitz-type; Fig. 1B). Mutations affecting integrin α6β4 lead to JEB with pyloric atresia (JEB/PA; Fig. 1B) [5]. There is a notable lack of clear genotype/phenotype correlation in JEB, which has pushed the field to prioritize specific mutations and clinical descriptors over broad phenotypic classifications [6].

Figure 1. Skin and hemidesmosome structure.

A) Healthy skin is comprised of a multilayered, well-differentiated epidermis comprised of keratinocytes and an underlying dermis connected by the basement membrane zone. Hemidesmosomes (HDs) are the functional units that connect basal layer keratinocytes to anchoring fibrils in the dermis. B) Basement membrane zone (BMZ) deficiencies in JEB subtypes: generalized severe, generalized intermediate, and junctional epidermolysis bullosa (JEB) with pyloric atresia. Subtypes are not restricted to genetic origin, as different mutations in the same protein may give rise different molecular and clinical phenotypes. SCC, squamous cell carcinoma.

JEB is characterized by extreme skin fragility, mechanically induced blistering, and chronic inflammation, in addition to gastrointestinal, renal, and respiratory complications [7]. The current standard of care for JEB patients includes palliation of cutaneous disease manifestations with careful lancing of blisters, judicious wound care, infection management, and pain control, as well as supportive care for systemic manifestations. These include anemia of chronic disease and malnutrition/micronutrient deficiencies from increased metabolic demand and mucosal/dental barrier to adequate food intake [6-9]. To date, no standard curative therapies targeting the genetic etiologies of JEB have been established, with the exception of recent work reporting amelioration of chronic wounds following autologous transplant of gene-corrected keratinocytes [10-12].

Wound healing of the skin is a complex and dynamic process divided in three phases: inflammation, new tissue formation, and remodeling [13]. During the inflammation phase, chemokines interleukin (IL)-8, C-C motif chemokine ligand (CCL) 2, CCL3, and CCL5 are released to recruit neutrophils and macrophages, the phagocytic armament of the innate immune system, to the site of injury to prevent infection, clear cellular debris, and promote angiogenesis (Fig. 2A) [14,15]. New tissue is formed as leading-edge keratinocytes along the periphery of the wound acquire epithelial-to-mesenchymal transition characteristics to become migratory [16]. This process initiates proliferation and differentiation of adjacent keratinocytes, generating the cell mass necessary to seal the wound. These migratory keratinocytes deposit a scaffold of LM332 and fibronectin to aid in wound closure, while the underlying dermal fibroblasts support these keratinocytes through the formation and remodeling of the extracellular matrix (ECM; Fig. 2B) [16,17]. In the last phase of wound healing, the collagen that was deposited to initiate the process is degraded, while the cells playing transient roles in wound-healing response (e.g., neutrophils) undergo apoptosis as they do not contribute to long-term skin homeostasis (Fig. 2C) [18].

Figure 2. Phases of cutaneous wound healing.

A) Inflammation phase. Chemokines recruit immune system to prevent infection and clear cellular debris. Oleogel-S10 upregulates expression of pro-inflammatory chemokines to recruit neutrophils and macrophages. RGN-137 inhibits neutrophil chemotaxis to reduce neutrophil-associated inflammatory responses. B) New tissue formation phase. Leading-edge keratinocytes deposit LM332 and fibronectin to migrate toward the center of the wound; adjacent keratinocytes proliferate and differentiate. Oleogel-S10 reorganizes actin cytoskeleton in migratory leading-edge keratinocytes. RGN-137 stimulates angiogenesis. Increased deposition of LM332 by leading-edge keratinocytes is seen. C) Remodeling phase. Extracellular matrix is remodeled by matrix metalloproteins; endothelial and immune cells exit wound site or undergo apoptosis.

The basement membrane zone (BMZ) of the skin is studded with hemidesmosomes (HDs), which act as biological “rivets” to bind the upper epidermis and lower dermis. The major proteins that comprise HDs are LM332, C17, integrin α6β4, and Type VII collagen (C7) anchoring fibrils (Fig. 1A). Function-altering mutations in any of these components lead to various forms of EB. Adhesion proteins and receptors are concentrated at the HDs of the BMZ, allowing for crosstalk between the skin layers and imparting mechanical strength and regenerative cues that allow the skin to function as a cohesive structure [19]. LM332 is an ECM protein that interacts with integrin α6β4 on the basal surface of epidermal keratinocytes to convey external mechanical inputs, in addition to binding C7 to maintain adhesion to the underlying dermis. C17 interacts with extracellular LM332 and the intracellular domain of integrin β4 to stabilize the HDs [20]. The innate dysfunction of the ECM in JEB impacts both the skin barrier function and its resident immune population [21]. The ECM serves as a critical attachment site for immune cells, such as mast cells and developing thymocytes [21-24].

As JEB is a rare disorder with a relatively small patient population, expansive clinical trials and approved therapies are limited. Currently available commercial products aim to improve the wound-healing capacity of affected skin to limit chronic wounds and their associated clinical complications, like infection, dehydration, and pain. With the advent of powerful gene-editing technology, the focus of emerging translational research has shifted toward autologous gene correction and transplantation to support regeneration of a fully functional epidermis in patients with JEB. Understanding the underlying causes and mechanisms of this disorder is foundational to truly attend to the clinical manifestations inherent to the JEB patient population. In this review, we will discuss the latest research and clinical avenues that are propelling current trends in drug, cell, and gene therapies.

2. Drug Therapies

Pharmacotherapies for JEB seek to augment current standards of palliation, which include use of common local or systemic pain and pruritus agents, as well as micronutrient supplements. Emerging disease-specific therapies have targeted the wound-healing process. Two topical agents supporting tissue regeneration in JEB are currently under investigation in clinical trials: Oleogel-S10 (Amryt Pharma) and RGN-137 (RegeneRx Biopharmaceuticals, Inc.). Gentamicin, a broad-spectrum antibiotic, is being investigated in clinical trials for its use in resolving chronic wounds by promoting premature termination codon readthrough to restore function in affected HD components.

2.1. Oleogel-S10

Oleogel-S10 is a topical treatment of 10% triterpene extract (TE) derived from birch bark (containing 72-88% betulin) in refined sunflower oil. The active component of Oleogel-S10, TE, has been shown to influence two major axes of the wound-healing process – inflammation and new tissue formation (Fig. 2A, B). In primary keratinocytes, treatment with TE upregulated expression of the pro-inflammatory markers COX-2, IL-6, and IL-8 by stabilizing their mRNA in a p38 MAPK-dependent manner [25]. These factors are necessary to recruit neutrophils and macrophages as well as initiate downstream inflammatory pathways [26,27]. TE treatment induced formation of filopodia, lamellipodia, and stress fibers by activating the GTP-RhoA signaling pathway. This reorganization of the actin cytoskeleton promoted keratinocyte migration, re-epithelialization, and angiogenesis [25].

An international randomized, triple-masked Phase III clinical trial, EASE, is currently underway to assess the efficacy and safety of Oleogel-S10 in patients with dystrophic and junctional EB (clinicaltrails.gov: NCT03068780; clinicaltrialsregister.eu: EudraCT 2016-002066032). This trial will assess the ability of Oleogel-S10 to enhance wound closure compared to a refined sunflower oil placebo, both in combination with standard-of-care wound dressing [9,28]. Oleogel-S10 has shown promise in accelerating wound healing for superficial partial-thickness burns, for which scarring and infection are major concerns. In another Phase III trial, Oleogel-S10 was compared to concurrent application of standard-of-care antiseptic octenidine dihydrochloride gel on different regions of a burn on the same patient. Oleogel-S10 provided earlier burn re-epithelialization in 85.7% (p<0.0001) at a mean of one day (7.6 vs. 8.8 days, p<0.0001), although conclusions are limited by 16% subject drop-out and challenges with residual gel residue inhibiting masked photograph interpretation (EudraCT 2012-000362-38) [29]. While superficial partial-thickness burns are relatively easy to treat, more serious burns compel treatment with split-thickness skin grafts. Two Phase III trials using the same protocol were conducted to assess the efficacy of Oleogel-S10 compared to moist, non-pharmacological dressings on re-epithelialization of an intra-patient divided split-thickness skin graft donor site. Investigators found that the portion of the graft donor site treated with Oleogel-S10 closed a mean of 1.1 days faster than the control portion (15.3 vs. 16.5 days, p<0.0001; EurdaCT 2012-003390-26; EurdaCT 2012-000777-23) [30]. In a Phase II pilot trial, Oleogel-S10 was compared to a non-adhesive dressing in promoting re-epithelialization of dystrophic EB wounds, either both applied to separate sections of a larger wound or to separate smaller wound pairs (EurdaCT 2010-019945-24). Due to a small patient cohort (n=10, 12 wound pairs) and a short trial (14 days), a statistically significant conclusion was not reached, although wounds treated with Oleogel-S10 trended towards faster closure [31]. The currently available published outcomes of Oleogel-S10 treatment does not support a clinically significant improvement in wound closure with acceleration of only 1-2 days. The ongoing EASE trial will determine the utility of Oleogel-S10 in aiding wound closure in other EB subtypes with a larger patient cohort.

2.2. RGN-137

In a recent study, fluid collected from advanced blisters in JEB-affected skin demonstrated high levels of CXC chemokines, which specifically recruited CD16b+ neutrophils to JEB wounds [32]. While recruitment of neutrophils is important for initiating repair in acute wounds, the continuous presence of neutrophils in chronic wounds is associated with high levels of proteases (e.g., MMPs, elastase, cathepsin G) that degrade the BMZ and ECM [18]. RGN-137 is a topical gel containing thymosin β4 (Tβ4), a small, ubiquitous polypeptide that has a role in tissue regeneration [33]. Intracellularly, native Tβ4 primarily acts as a potent G-actin monomer-sequestering protein [34]. Extracellularly, the oxidized form of Tβ4 inhibits neutrophil chemotaxis in vitro and reduces neutrophil-associated inflammatory responses in vivo [35]. Tβ4 has also demonstrated efficacy in re-epithelialization and angiogenesis processes, both in the skin and cornea [36-38]. In the context of JEB, Tβ4 increases the deposition of LM332 independently of TGFβ [39].

The combined effect of Tβ4 on neutrophil control, keratinocyte migration, and LM332 deposition dynamics make RGN-137 an attractive treatment option for JEB patients with chronic wounds (Fig. 2A, B). To this end, a dose-escalation Phase II clinical trial investigating the effect of topical 0.01%, 0.03%, and 0.1% Tβ4 on chronic wounds in EB (junctional and dystrophic) was initiated in 2006 but was terminated in 2012 due to a lack of enrollment (NCT00311766). In a different chronic wound context, patients with stage III/IV full-thickness pressure ulcers or venous stasis ulcers were treated with RGN-137 (0.01%, 0.02-0.03%, or 0.1%). Patients whose ulcers healed after 12 weeks showed a trend toward faster wound closure when treated with 0.02-0.03% RGN-137 compared to 0.01% and 0.1% doses [40]. A new randomized, triple-masked, placebo-controlled Phase II clinical trial investigating the impact of RGN-137 on wound-healing in JEB (and dystrophic EB) began recruiting in mid-2019 (NCT03578029). Because of the inherent lack or dysfunction of LM332 in most generalized severe and some generalized intermediate JEB patients, it remains to be seen whether upregulation of LM332 deposition by RGN-137 will accelerate wound-healing.

2.3. Gentamicin

Gentamicin is a broad-spectrum aminoglycoside antibiotic approved for treatment of Gram-negative bacterial infections that has recently demonstrated efficacy in promoting readthrough of premature termination codons (PTCs) to restore full-length, functional proteins in EB [41-43]. Gentamicin promotes PTC readthrough by interfering with two adenosines, A1755 and A1756, on helix 44 of the 18S ribosomal RNA in the 40S subunit of eukaryotic ribosomes [44]. By flipping these two adenosines outward, gentamicin alters the decoding center that ensures accurate selection of aminoacyl-transfer RNAs (tRNAs) corresponding with mRNA codons. Gentamicin-ribosome interactions lead to decreased codon-anticodon recognition at the aminoacyl-tRNA acceptor site, which allows for incorporation of near-cognate aminoacyl-tRNAs at the location of PTCs [45]. Efficiency of gentamicin-induced PTC readthrough is determined by the type of stop codon and sequence (UAA, UAG, or UGA). A recent investigation demonstrated that UGA stop codons result in higher levels of readthrough than UAG and UAA stop codons, respectively [45]. Additionally, a cytosine at the +4 position and a uracil at the −1 position are significantly linked to enhanced gentamicin-induced readthrough response [45]. This evidence may narrow participant selection criteria for clinical trials focused on drug-induced PTC readthrough.

A double-blind, placebo-controlled Phase I/II clinical trial reported that both topical application and intradermal injection of gentamicin increased C7 expression in patients with recessive dystrophic EB (RDEB) harboring PTCs in C7 (NCT02698735) [42]. For each patient (n=5), two ulcerated sites were chosen for topical gentamicin (0.1%) and placebo comparison. For four of these five patients, two intact sites were chosen for two daily intradermal gentamicin injections (8 mg) or placebo comparison. Both methods of gentamicin delivery induced a significant increase in full-length C7 expression and formation of new anchoring fibrils compared to baseline assessments [42]. Topical gentamicin application enhanced wound closure and decreased new blister formation in four of five patients over the course of three months. Because gentamicin is known to have nephro- and ototoxic effects [41], this trial assessed kidney and auditory effects before and after gentamicin treatment. No systemic parameters changed significantly over the course of the trial for all patients [42].

Mutated LAMB3, which codes for the β3 subunit of heterotrimer LM332, accounts for 80% of generalized severe JEB cases [7,46]. Of those cases, approximately 95% of diseased alleles contain PTCs [46]. Gentamicin induces PTC readthrough of LAMB3 at efficiencies ranging from 2-27% in LAMB3-null keratinocytes, depending on the specific LAMB3 mutation [43]. Similar to previous evidence supporting the importance of sequence context and stop codon identity [45], LAMB3 mutations containing UGA stop codons or cytosines and uracils at the +4 and −1 locations, respectively, demonstrated the most robust response to gentamicin [43]. Compared to a control, gentamicin treatment resulted in rescued expression and localization of intact heterotrimer LM332 in HDs in in vitro organotypic culture [43]. A new open-label Phase I/II clinical trial assessing the efficacy of topical (0.5%) or intravenous (7.5 mg/kg) delivery of gentamicin on restoration of LM332 expression in β3-deficient JEB patients was launched in 2018 (NCT03526159). Given the encouraging in vitro data supporting gentamicin-induced PTC readthrough of LAMB3 and promising results in RDEB patients, this trial may establish an in vivo therapeutic benefit for a limited cohort of patients with generalized severe JEB.

Topical drugs targeting various phases of the healing process are a promising supplement to the daily wound care regimen of JEB patients. Both Oleogel-S10 and RGN-137 influence the immunological component driving persistent chronic wounds, albeit in different ways, in addition to promoting the keratinocyte migration necessary for effective wound closure. In a different approach, gentamicin modulates ribosome dynamics during translation to generate full-length proteins by reading through PTCs. Results from these ongoing, EB-specific clinical trials will offer a more definitive perspective on whether these drugs, in the absence of gene therapy, are able to improve the devastating skin manifestations of JEB.

3. Cell Therapy

Epidermal grafting has shown promise for JEB patients. However, other cellular therapies such as allogeneic hematopoietic cell transplantation and local or systemic treatment with stromal cells await adequate evidence to support routine use. Within the EB family of genetic disorders, therapy evaluation in RDEB is one of the most mature, likely due to its higher prevalence, improved patient survival, and the existence of preclinical animal models. However, cellular therapies which provide benefit by immune modulation would likely translate across EB subtypes.

3.1. Skin Grafting

The first cell therapy for JEB occurred in 1987, when three patients received epidermal autografts to chronically denuded areas [47]. Patient keratinocytes were harvested from intact skin by a suction blister technique, seeded onto Type I collagen sponges, cultured into multilayered epidermal sheets, and applied to wounds. All three patients demonstrated partial to complete re-epithelialization of wounded sites within ten months. More recently, full-thickness punch biopsies from unaffected sites have been used to resolve chronic ulcers in four LM332-deficient patients [48]. This method involved harvesting 3-4 mm of skin from the donor site with a biopsy punch, separating the epidermal and dermal layers, and then placing the epidermal portions 2-6 mm apart in the debrided ulcer bed. Over the course of ten years, 23 ulcers were grafted in the four patients. The donor sites healed without issue, and 70% of treated ulcers healed completely with an overall reoccurrence rate of 13% three months post-transplant [48].

Such successful autologous grafting relies on the availability of unaffected, healthy skin in patients with JEB (Fig. 3). Revertant mosaicism, or “natural gene therapy,” is a phenomenon in which the underlying cause of a genetic disease is corrected by somatic mutational events, usually during embryonic development [49]. One patient with generalized intermediate JEB (LM332-deficient) was treated with autologous transplant of punch biopsies from a revertant skin patch to multiple chronic ulcers [50]. Grafted sites did not form blisters 18 months post-transplant, and sequencing confirmed that the regenerated epidermis was comprised of revertant keratinocytes [50]. Although revertant skin grafts are capable of ameliorating blistering in the acceptor sites of LM332-deficient JEB patients, the utility of this approach does not seem to extend to other JEB etiologies. In one patient with generalized intermediate JEB (C17-deficient), a revertant skin patch was biopsied and cultured into epidermal sheets for subsequent transplant to an acceptor site prepared by epidermal stripping [51]. In contrast to LM332-deficient patients receiving revertant autografts, the graft site of the C17-deficient patient was fragile and blistered easily following mechanical manipulation. Notably, less than 3% of the graft was comprised of revertant keratinocytes [51]. In the absence of unaffected sites that give rise to viable grafts, genetic correction and subsequent transfer of autologous keratinocytes was successful in three patients [10-12].

Figure 3. Schematic of autologous epidermal grafting procedure.

Biopsies are taken from unblistered or intact skin to isolate keratinocyte cultures. In the case of revertant mosaicism, virally-mediated correction is avoided.

Engineering optimal skin grafts for JEB requires knowledge of the regenerative epidermal stem cell types, markers to identify and isolate them, and signals for self-renewal and proliferation. There are three distinct niches for skin stem cells: the basal layer of the epidermis, the hair follicle bulge, and the sebaceous gland [52-55]. Keratinocytes terminally differentiate as they rise from the basal layer of the epidermis to the stratum corneum, where they are eventually shed. Epidermal stem cells, or holoclones, self-renew and are capable of extensive proliferation. The progeny of stem cells, transient-amplifying cells or meroclones, are unable to self-renew but retain limited proliferative potential. Differentiated keratinocytes, or paraclones, are unable to self-renew or proliferate [56,57]. Holoclones can be identified in a mixed population of keratinocytes by their high expression of β1 integrin, the α isoform of transcription factor p63, and BMI-1 [56-58].

Isolating holoclones from JEB skin is technically difficult due to reduced numbers of that cell type compared to healthy skin [10,11]. Recent work has shown that intact HDs are necessary to maintain stem cell populations in the basal layer niche. In HDs, integrin α6β4 interacts with LM332, which transduces adhesion-dependent signals to activate Yes-associated protein (YAP) [59]. Active YAP translocates to the nucleus, where it acts as a transcriptional co-activator with its binding partner TAZ and DNA-binding transcription factors (i.e., TEADs) for genes necessary for cell proliferation, growth, and survival [60-62]. YAP activation through an integrin α6β4/LM332 interaction allows for preservation of epidermal stem cells, which are progressively lost in JEB patients [10,11]. Interestingly, forced YAP expression was able to preserve the clonogenic and proliferative capacity of LM332-deficient keratinocytes in vitro [59]. By culturing JEB patient-derived keratinocytes on LM332-coated culture vessels, holoclones can be preserved for grafting applications. Indeed, robust epidermal grafts that demonstrate long-lasting homeostasis and regenerative capacity are dependent on inclusion of a sufficient proportion of holoclones within the graft [63].

Ex vivo culture techniques are critical for the development of a successful graft. Culturing keratinocytes on a fibrin matrix instead of an uncoated plastic surface decreases graft contraction after enzymatic detachment, prolongs clonogenic ability of cells to three days, and allows for transport at room temperature (Fig. 3). Additionally, fibrin-supported grafts enable efficient engraftment after application to the wound bed [63,64]. Wound preparation prior to grafting is essential to ensure the performance of the graft. Removing diseased or necrotic tissue from the wound bed and managing any underlying infection greatly improves graft success (Fig. 3). Preparing the wound bed with cadaveric skin allografts helps reduce infection and provides a base for engraftment of autologous cultures following mechanical removal of the epidermal layer of the allograft [63,65]. In the absence of cadaveric allografts, collagen, hyaluronic acid-based substrates, or completely synthetic dermis structures containing fibroblasts have also been shown to mediate engraftment [63].

Proliferative signals aid in inflammation management and graft regeneration. High mobility group box 1 (HMGB1), a non-histone nuclear protein, is released by necrotic cells and has been characterized as a biomarker reflecting disease severity in RDEB [66,67]. In a mouse model of allogeneic bone marrow transplant, HMGB1 upregulated CXCR4 expression on bone marrow-derived platelet-derived growth factor receptor alpha-positive (PDGFRα+) mesenchymal stromal cells (MSCs) to recruit them to epidermal graft sites [68]. The presence of PDGFRα+ MSCs correlated with a reduction of local macrophage infiltrate and an increase in the anti-inflammatory cytokines TSG-6 and IL-10 [68]. This suggests that HMGB1 is an immune modulatory molecule that could be used to control local inflammation for grafting applications.

3.2. Allogeneic Hematopoietic Cell Transplant

The potential benefits of allogeneic hematopoietic cell transplant (alloHCT) in EB are many. Bone marrow cells have demonstrated the ability to differentiate in epithelial lineages and take residence in the skin [69-74], providing a potential source of healthy LM332, C17, or α6β4 integrin, although transdifferentiation of bone marrow-derived cells remains controversial [75]. AlloHCT establishes immune tolerance between a donor and JEB patient, permitting subsequent allogeneic epidermal grafting of wounds without need for somatic mosaicism of gene-corrected keratinocytes as a source for autologous grafts [76]. While no suitable preclinical model for JEB is available, alloHCT has shown benefit in a preclinical murine RDEB and C17-deficient generalized intermediate JEB model, as well as in clinical trials (NCT00478244) [74,77,78].

Unfortunately, at this time, clinical experience with alloHCT in JEB is limited and without evidence of benefit. In a female infant with generalized severe JEB (laminin β3-mutant), transplant of peripheral blood hematopoietic stem cells from her HLA-haploidentical father following myeloablative conditioning was performed with the goal of overcoming the immune obstacle of skin grafting [76]. No graft-versus-host disease was detected after alloHCT, and split-thickness skin grafts were harvested from the father and applied to wounded areas accounting for 80% of the patient’s body surface area. While the grafts healed well and had satisfactory biomechanical function, the patient succumbed to pulmonary failure and septicemia at day +96 after alloHCT [76].

This result was recapitulated in another infant with generalized severe JEB (laminin β3-mutant). The male infant was transplanted with a haploidentical, non-T-cell-depleted bone marrow graft from his mother following myeloablative conditioning [7]. Three weeks after transplant, the patient showed complete engraftment of hematopoietic donor cells and improvement of his chronic wounds. However, after a seven-week period of stability, as evidenced by stable body weight, decreased need for pain management, decreased albumin substitution, the patient demonstrated hematopoietic graft failure concurrent with the return of his severe phenotype. At day +129 after transplantation, the patient succumbed to infection [7]. It is inconclusive whether the temporary improvement of disease phenotype was from the transient anti-inflammatory effects of the pre-transplant conditioning regimen or from donor cells that were lost with shifting donor hematopoietic chimerism.

3.3. Stromal Cells

Stromal fibroblasts and bone marrow-derived MSCs secrete extracellular matrix proteins, like C7, which has augmented their utility in providing therapeutic benefit for RDEB patients with chronic wounds [79,80]. A randomized, double-blind, vehicle-controlled clinical trial reported that a single intradermal injection of allogeneic fibroblasts in the periphery of chronic erosions in RDEB patients reduced erosion area for +28 days post-injection, but not thereafter (isrctn.org: ISRCTN67757229) [79]. In another report, intradermal injection of allogeneic MSCs into the margins of chronic ulcers in two RDEB patients prompted re-epithelialization and C7 deposition at the BMZ compared to a vehicle control [76]. While allogeneic MSC injection did not produce acute adverse effects in either patient, the improvement in wound-healing was transient and diminished four months after administration [80].

Bone marrow-derived MSCs are posited to evade immune recognition, which was supported by the lack of acute adverse effects seen in allogeneic injection applications [80]. Compared to cutaneous injection, intravenous infusion of allogeneic MSCs holds potential to impart systemic relief from the wide-ranging manifestations of EB. In a small cohort of patients with RDEB (n=10), serial infusions of non-HLA-matched, bone marrow-derived MSCs were well-tolerated with improved wound healing extending to 4-6 months after treatment (EudraCT 2012-001394-87; ISRCTN46615946). While severe adverse events were reported following infusion, 78% of those events were determined to be unrelated to the MSCs [81]. In a randomized, double-blind study, utility of systemic infusion of bone marrow-derived MSCs for patients with RDEB was compared with (n=7) or without (n=7) concurrent treatment with cyclosporine [82]. In both groups, MSCs significantly improved rates of wound-healing (<0.001) and significantly decreased appearance of new blisters (p=0.003-0.004). Notably, there was no statistical difference in outcomes between patient groups treated with or without cyclosporine. No major severe adverse events were reported in either group. Finally, 10 adults with RDEB treated with systemic allogeneic bone marrow-derived MSCs demonstrated reduction in pruritus, as well as in serum levels of HGMB1, the single biomarker correlating with disease activity in RDEB (EudraCT 2014-004500-30) [83].

Unique subsets of dermal resident MSCs may provide superior beneficial anti-inflammatory benefits in EB. ATP-binding cassette subfamily B member 5 (ABCB5+) dermal MSCs demonstrate potent immunomodulatory effects in vitro, including skewing of monocyte differentiation to anti-inflammatory M2 macrophage fate, expression of programmed cell death protein 1 (PD-1) leading to suppression of conventional T cells, induction of regulatory T cells [84], and suppression of neutrophil extracellular trap formation [85]. Administration of ABCB5+ MSCs shortly after birth in a pre-clinical murine model of RDEB significantly prolonged survival, with treated mice showing decreased M1 macrophage skin infiltration [86]. Further investigations of ABCB5+ MSCs demonstrated the cells to mediate to anti-inflammatory M1-to-M2 shift by secreting IL-1 receptor antagonist. In a chronic iron overload murine model of chronic venous leg ulcers, intradermal injections of ABCB5+ MSCs promoted accelerated healing of excisional wounds [87]. These pre-clinical data supported development of several early phase clinical trials of ABCB5+ MSCs, including systemic administration for treatment of RDEB (NCT03529877; EudraCT 2018-001009-98).

Decidual stromal cells (DSCs) derived from full-term fetal membranes have been shown to have similar immunosuppressive effects and inhibit alloantigen-induced proliferation of peripheral blood mononuclear cells (PBMCs) in vitro [88]. In a recent study, a patient with generalized severe JEB was treated with serial, systemic infusions of allogeneic DSCs in combination with chronic blister treatment with allogeneic amniotic membranes [89]. The patient showed transient improvement in her skin blistering phenotype, but no amelioration of her upper airway complications. The patient also developed multi-specific, anti-HLA class I antibodies, indicating that she developed immunity to the allogeneic DSCs [89]. Interestingly, patient-derived PBMCs proliferated vigorously in response to DSCs compared to PBMCs from a healthy donor, suggesting that the acquired immunity was not the sole result of HLA-antigens. DSCs were shown to express LM332, but no anti-LM332 antibodies were detected in the patient. Instead, the patient had developed antibodies against bovine serum albumin, the main protein in the fetal calf serum used to supplement the DSC expansion medium [89]. The patient eventually succumbed to respiratory failure one year after treatment. This therapy was limited by the development of alloimmunization, which could potentially be avoided in the future following different DSC culture conditions [89]. In all, intravenous delivery of stromal cells to enhance wound-healing and prognosis in JEB patients remains an open avenue for further investigation.

With improvements in gene correction techniques, autologous epidermal grafting no longer requires revertant mosaicism for success. Epidermal progenitors can be isolated from affected skin, gene-corrected, and then cultured to regenerate healthy skin in JEB.

4. Gene Therapy

Gene therapy is the process of correcting genetic abnormalities to treat or prevent disease. In the case of JEB, this can range from correction of specific mutations to the total replacement of nonfunctional alleles. Loss-of-function mutations leading to generalized severe JEB are most commonly found within the β3 subunit of heterotrimer LM332. These are associated with some of the most severe clinical manifestations of the disease [90]. Much of the JEB-specific gene-editing research to date has focused on restoration of β3 subunit (LAMB3 gene) function. As such, corrective therapies have been limited to ex vivo genetic manipulation and subsequent autologous transplant [10-12]. A variety of viral and non-viral methods have been employed to aid in correction of JEB-causing gene mutations.

4.1. Viral Correction

In 2001, LAMB3 cDNA under control of a cytomegalovirus (CMV) promoter was delivered to patient-derived JEB keratinocytes by a retroviral vector [91]. Transgenic JEB keratinocytes containing wild-type LAMB3 cDNA were able to regenerate skin in vivo in a severe combined immunodeficient (SCID) mouse model, with correct distribution of the β3 subunit in the basal epidermal layer and evidence of HD formation between the dermis and epidermis. Additionally, β3 subunit restoration was able to rescue localization of several key BMZ components, including α6 and β4 integrins [91]. Gene expression signatures of matrix adhesion, membrane signaling, and growth regulation functions were significantly altered in uncorrected JEB keratinocytes compared to healthy control cells [92]. LAMB3 gene transfer was able to partially rescue these gene expression signatures to a greater degree than supplying extracellular LM332 to uncorrected JEB keratinocytes [92].

In similar experiments, β3-deficient patient keratinocytes were transduced with retrovirus containing LAMB3 cDNA under control of the Moloney leukemia virus long terminal repeat (MLV-LTR) promoter [10-12]. Keratinocytes demonstrating β3-subunit restoration were then subjected to autologous grafting in human patients. In one of these studies, the efficacy of graft establishment following cultivation of gene-corrected keratinocytes on plastic or fibrin-coated culture dishes was assessed for the first time in JEB wounds [11]. Both methods of culture allowed for full engraftment of epidermal sheets after ten days, and complete epidermal regeneration after one month [11]. As insertional mutagenesis is a concern in viral approaches, the retroviral insertion sites of the transgenic epidermis was profiled [10,11,93]. Approximately 5% of integrations were mapped to exons in one report, although the treated patient did not exhibit any tumor development or related adverse events [11]. Long-term follow-up in two treated patients demonstrated histologically normal epidermis at the grafting sites comprised solely of transgenic keratinocytes, which was resistant to blistering following mechanical stress [11,93].

Though less common in the JEB patient population, the α3 subunit (LAMA3 gene) of heterotrimer LM332 has also been a target of gene therapy interventions. In one report, a recombinant adenoassociated virus (rAAV) with a hybrid serotype (rAAV-DJ) [94], which is able to mediate homologous recombination (HR) at high efficiencies in human keratinocytes, was used to correct a homozygous point mutation in the splice acceptor site of exon 44 in LAMA3 [95]. Single-stranded DNA containing LAMA3 exons 43 and 44 flanked by inverse terminal repeats was delivered to primary JEB patient-derived keratinocytes. These were then enriched for HR correction events using a drug selection-free, adhesion-based assay [95]. The incidence of random integration for this approach was estimated to be around 10%, although the adhesion-based selection may have artificially rescued non-targeted cells as LM332 deposition is non-cell-autonomous [95]. A pooled population of corrected cells demonstrated restored deposition of LM332 in a monolayer, in organotypic culture, and when grafted onto nude mice [95].

4.2. Non-Viral Correction

While ectopic viral insertion of LAMB3 cDNA has proven successful in restoring normal function to diseased keratinocytes in vitro and in vivo, viruses raise biosafety, cost, and time concerns that may limit their use in clinical applications [96]. Non-viral methods have also been used to correct LAMB3-deficient keratinocytes. Sleeping Beauty (SB) is a plasmid-based, synthetic transposable element derived from fish that permanently integrates DNA cargo flanked by short inverted repeat sequences into a host genome using a “cut-and-paste” mechanism [97]. A transposon plasmid containing LAMB3 cDNA and a blasticidin resistance gene under the control of the CMV promoter was delivered to β3-deficient JEB keratinocytes with an SB transposase plasmid [98]. Following clonal analysis, SB-mediated transposition was shown to rescue β3-subunit expression in vitro and regenerate human skin with correctly polarized HDs on SCID mice in vivo. Additionally, transposon integration was localized to TA dinucleotide regions and were not found in intragenic sequences [98]. In another approach, Streptomyces phage φC31 integrase was used to stably insert CMV-driven LAMB3 cDNA and a blasticidin resistance gene into β3-deficient JEB keratinocytes [99]. φC31 integrase recognizes the 285-base pair attB targeting sequence on incoming DNA cargo and performs site-specific, unidirectional insertion at endogenous attP insertion sites in the mammalian genome [100]. Following analysis of ten clones, insertion of DNA cargo occurred at six previously identified pseudo-attP sites. Corrected JEB keratinocytes generated skin that demonstrated β3 protein rescue and intact HDs following transplant onto SCID mice [99].

4.3. CRISPR/Cas9-Based Correction

Clustered regularly interspaced short palindromic repeats (CRISPR) is a microbial adaptive immunity system that has proven its genome-editing utility in a wide variety of species. CRISPR-associated protein 9 (Cas9) is an endonuclease that can be targeted to specific locations in the genome by a guide RNA (gRNA), where it cleaves DNA to induce DNA repair pathways [101]. CRISPR/Cas9-initiated homology directed repair was recently used to correct β3-deficient JEB keratinocytes. This study employed a promoterless LAMB3 cDNA containing exons 3-23 [102]. This cDNA was engineered with a splice acceptor site before exon 3, conferring the cargo transcriptionally silent unless a recombination event allowed for correct splicing between endogenous exon 2 and exogenous exon 3 [102]. The cDNA, Cas9, and a gRNA targeting the intronic region between exons 2 and 3 were cloned into integration-defective lentiviral vectors for delivery into patient-derived keratinocytes [102]. This strategy did not utilize a drug selection step to enrich the edited population, instead relying on the restoration of LM332-mediated adhesion to select for corrected cells. Clones that incorporated one copy of the cDNA cargo were able to regenerate human skin with complete HDs and properly localized LM332 deposition in SCID mice [102]. Off-target indels are a major concern for therapeutic CRISPR/Cas9 approaches because of the DNA repair mechanisms initiated by Cas9-induced double-stranded breaks [103]. This study reported an indel frequency of 0.9% in one off-target analyzed while the remainder of analyzed off-targets were unaffected, although other high-fidelity Cas9 variants are available that may reduce this frequency [102,104]. This elegant approach combines the efficiency of viral delivery with the specificity of CRISPR/Cas9 to establish a template for standardized gene therapy for most documented LAMB3 mutations [105,106].

Progress has been made toward developing gene therapy approaches for patients with JEB. Currently, an open-label Phase I/II clinical trial is underway to assess the efficacy of a gamma-retroviral vector containing C17 cDNA to correct patient-derived epidermal stem cells for subsequent autologous transfer (NCT03490331). This trial, in combination with previous reports of successful correction of β3-deficient epidermal holoclones and autologous grafts that give rise to long-lived tissue homeostasis [10-12,93] is evidence that the field is expanding the horizons of what is possible in the clinic.

5. Expert Opinion

JEB is a complex disease, with manifestations ranging from acute cutaneous involvement (blistering, chronic wounds) to devastating systemic consequences (respiratory, gastrointestinal, oral) [6,90]. Despite the challenges in understanding, managing, and stabilizing the multi-faceted phenotype of JEB patients, researchers and clinicians advocating for the JEB patient population have expanded their repertoire of potential therapeutic strategies.

Interventional drug therapies, such as Oleogel-S10, RGN-137, and gentamicin, represent promising new directions in JEB treatment. Oleogel-S10 and RGN-137 target the phases of wound-healing to aid in the closure of persistent wounds in EB patients and improve outcomes of grafting procedures in burn patients, although a more conclusive study is required to determine the efficacy of these drugs for JEB patients. Oleogel-S10, in particular, has demonstrated an efficacy in improving outcomes in skin grafting procedures, although the clinical significance of reported outcomes is uncertain [30]. The potential combinatory effect of Oleogel-S10 with autologous, gene-corrected grafts may lead to accelerated wound-healing and an improved cosmetic appearance in treated JEB patients. Gentamicin has already demonstrated a wound-healing benefit for patients with RDEB by promoting PTC readthrough in C7, which led to a marked increase in full-length protein deposition [42]. Promising in vitro data suggests that gentamicin may have a similar effect in laminin-β3-deficient patients [43] and is currently under investigation in clinical trials. As the underlying genetic drivers contributing to chronic wounds in JEB are not corrected through interventional drug therapies, these approaches may only partially correct the clinical phenotype.

Exciting gains have been made in the implementation of gene correction and autologous grafting of JEB patient-derived keratinocytes, which circumvents the immune system obstacles encountered in allogeneic grafting approaches [10-12]. To date, gene correction of patient-derived keratinocytes used for autologous transfer in human patients has been retrovirally mediated, with replacement genes under the control of a non-native, constitutive promoter (MLV-LTR). These features pose potential barriers in large-scale therapeutic implementation due to the cancer risks associated with insertional mutagenesis through the random integration of viral vectors. Additionally, non-native promoters are unable to control gene expression of these replacement genes in response to the exogenous and endogenous signals central to skin homeostasis. Despite these limitations, virally mediated correction has proven to give rise to long-lasting, biomechanically sound grafts in a three patients with generalized severe JEB [10-12]. These positive results emphasize the necessity of crosstalk between the latest cell and gene therapy approaches by merging effective correction techniques with well-established grafting protocols. This system is a promising direction for relieving the hallmark cutaneous symptoms of EB. Indeed, as EB patients are at risk of developing aggressive squamous cell carcinomas, a suspected consequence of chronic wounds [107], the benefits of autologous transplant of virally mediated, gene-corrected keratinocytes may outweigh the inherent risks of this method.

The substantial advancements in JEB treatments have not been without difficulties. Allogeneic HCT, which has demonstrated partial amelioration of disease phenotype in patients with RDEB, has not been effective in patients with β3-deficient JEB and remains to be further explored in other JEB subtypes. Following alloHCT for β3-deficient JEB, patients consistently succumbed to infection and disease manifestations after an initial period of improvement and preliminary donor engraftment [7,76]. In particular, the relationship between the underlying genetic cause of JEB and the inability of patients to overcome infection post-transplant offers numerous investigative challenges. Such susceptibilities to overwhelming infection may reflect underlying immune deficiency compounded by the effect of alloHCT conditioning. A growing literature highlights LM332 in thymocyte development and raises need for future investigations of adaptive immunity in JEB [22,108-111]. Additional cellular therapies, including treatment with DSCs, have not been effective to date in relieving the long-term cutaneous symptoms of JEB due to alloimmunization against the donor cells [89]. However, different stromal cell populations, bone marrow- and dermal-derived MSCs, appear immunologically inert with genotype-independent immunomodulatory effects on clinical phenotype. Identification of increasingly potent MSC subtypes and current clinical trials in RDEB may translate to a successful therapeutic strategy for JEB patients in the future.

The long-term goal of researchers, clinicians, and the families of patients with JEB is to identify therapeutic strategies that ameliorate both the devastating cutaneous and systemic manifestations of this disease. Current therapies are limited in the sense that they target a narrow range of JEB phenotypes (cutaneous wound healing), while others remain unimproved (gastrointestinal and respiratory complications). AlloHCT and additional cellular therapies are appealing for their potential systemic benefit. However, greater investigation into the mechanism of disease amelioration realized in alloHCT for RDEB, but not in β3-deficient JEB, are required to optimize this therapy for use in all genetic EB etiologies. The field has made substantial advancements in its capacity to mediate gene-correction in primary JEB keratinocytes, which has already become a focal point in a clinical trial for C17-deficient cases of generalized intermediate JEB (NCT03490331). Regenerative medicine approaches to JEB offer the greatest hope for effective systemic individualized therapies for this heterogenous disease.

Article Highlights.

• Junctional epidermolysis bullosa is a rare, inherited genetic disease that has no cure.

• Drug therapies in clinical trials may provide therapeutic benefit by influencing phases in cutaneous wound-healing (Oleogel-S10 and RGN-137) or promoting readthrough of premature stop codons to generate full-length protein (gentamicin).

• Although systemic allogeneic cell therapies have demonstrated clinical benefits in patients with RDEB, these approaches have been largely unsuccessful in patients with β3-deficient JEB.

• Autologous skin grafting for closure of chronic wounds could be achieved if patients with JEB have areas of revertant mosaicism, in which somatic mutational events correct genetic defects, or by ex vivo genetic correction and subsequent autologous transplant.

• Current therapies are limited to the cutaneous symptoms of JEB, while systemic manifestations have demonstrated minimal improvement.

Abbreviations:

alloHCT

allogeneic hematopoietic cell transplant

BMZ

basement membrane zone

C7

Type VII collagen

C17

Type XVII collagen

EB

epidermolysis bullosa

ECM

extracellular matrix

HD

hemidesmosome

JEB

junctional epidermolysis bullosa

LM332

laminin-332

MSCs

mesenchymal stromal cells

PTC

premature termination codon

RDEB

recessive dystrophic epidermolysis bullosa