Retroviral Vectors

To date, retroviral vectors (RVs) derived from Moloney murine leukemia virus (Markowitz et al. 1988; Riviere et al. 1995; Merten 2004), hereafter referred to as γRVs, are most frequently used in clinical trials for keratinocyte-based ex vivo gene therapy (Edelstein et al. 2007). They can accommodate up to ~9 kb of genetic material and stably transduce nearly 100% of clonogenic keratinocytes (Dellambra et al. 1998; Hirsch et al. 2017). The construction of γRVs involves the generation of packaging cell lines carrying the transgene construct, the RNA packaging signals (Ψ), cis-acting viral sequences—such as 5′- and 3′-long terminal repeats (LTRs) used as a promoter—and gag, pol, and env viral sequences (Pear et al. 1993; Chang and Yee 2012). Packaging cells then need large-scale production in compliance with good manufacturing practices (GMPs), which are required to conform to the guidelines recommended by agencies that control the authorization of the manufacture and sale of pharmaceutical products, including ATMPs. Among others, complete absence of replication-competent viruses (RCVs) is a stringent GMP requirement (Cornetta et al. 2008).

A major concern about the use of γRVs is the risk of insertional mutagenesis. Indeed, these vectors show a preferential integration pattern in actively transcribed genes, which can interfere with physiological gene expression, leading to potential deregulation of cellular processes (Scherdin et al. 1990; Wu et al. 2003; Mitchell et al. 2004; Lewinski et al. 2006; Osborn et al. 2013). Lymphoproliferative or myelodysplastic disorders, resulting from proviral insertion into proto-oncogene promoters of hematopoietic stem cells, were observed in γRV-based clinical trials for gene therapy of X-linked severe combined immunodeficiency (SCID) (Hacein-Bey-Abina et al. 2003, 2008; Aiuti et al. 2009), Wiskott–Aldrich syndrome (WAS) (Aiuti et al. 2013; Braun et al. 2014; Hacein-Bey Abina et al. 2015), and X-linked chronic granulomatous disease (X-CGD) (Stein et al. 2010). Of note, none of these adverse events were reported when γRV-transduced hematopoietic stem cells were used to successfully treat adenosine deaminase (ADA)-deficient SCID (Aiuti et al. 2009). In fact, γRV-transduced CD34⁺ cells were approved in 2016 in the European Union for the treatment of ADA-SCID (Strimvelis). Similarly, no insertional oncogenesis has been detected in γRV-based ex vivo gene therapy of recessively inherited junctional EB (JEB) and dystrophic EB (RDEB) (see below). Thus, to trigger tumor formation, insertional mutagenesis might require other oncogenic factors, which could be related to cell type, patient's individual genetic background, age, disease, transgene, or other mutations (Howe et al. 2008; Cavazza et al. 2013).

Nevertheless, the development of safer γRVs is desirable and many efforts are going to achieve this. Self-inactivating-γRVs (SIN-γRVs) are characterized by a deletion of the U3 region of the 3′LTR, which eliminates a strong viral enhancer, hence reducing the tendency of γRV to activate genes. An internal, often tissue-specific, promoter can be used to drive the transcription of the transgene (Yu et al. 1986; Kraunus et al. 2004; Schambach et al. 2006). During these studies, it has been clearly shown that insertional gene activation is determined by the characteristics of transcriptional regulatory elements carried by the vector and is largely independent from the vector type or design (Maruggi et al. 2009). Indeed, the LTR has strong enhancer activity in both γRV and lentiviral vectors (LVs), whereas an internal cellular promoter induces deregulation of gene expression less frequently and to a lesser extent in both vectors (Maruggi et al. 2009).

An alternative vector system, derived from the Avian sarcoma-leukosis α-retrovirus (ASLV), is currently available. In preclinical models, αRVs showed favorable safety features as compared with both γRVs and LVs, at least in hematopoietic cells (Suerth et al. 2012). Small-scale surveys of integration sites of αRVs in different cell types pointed to a more haphazard integration pattern as compared with other RVs, with no apparent preference for promoters or transcription start sites (Mitchell et al. 2004; Hu et al. 2007, 2008; Suerth et al. 2010, 2012, 2014). This promising outcome led to the development of SIN-αRV being able to efficiently transduce mammalian cells, attain long-term transgene expression in murine and human hematopoietic progenitors, and redress X-CGD in a mouse model (Kaufmann et al. 2013).

Lentiviral Vectors

Lentiviruses are a subclass of retroviruses, the best-known member being human immunodeficiency virus (HIV), sharing with its fellow γRVs the gag, pol, and env genes. LVs own a few additional regulatory genes, namely tat and rev, and some other accessory genes depending on the subtype (Malim et al. 1988; Miller and Sarver 1997). LVs have many attractive properties. In contrast to γRVs, LVs infect both dividing and nondividing cells; hence, they could transduce a broader variety of human cells (Bukrinskaya et al. 1996; Kaushik and Ratner 2004). LVs can be pseudotyped with vesicular stomatitis virus glycoprotein (VSV-G), thus allowing the binding to a widely expressed receptor on the plasma membrane of target cells, with a consequent increase of transduction efficiency. Although LVs might have a slightly safer integration profile than γRVs (Cattoglio et al. 2010), LVs and γRVs have similar LTR-enhancer activity (Maruggi et al. 2009), potentially increasing the expression of the gene where they integrate. Of note, down-regulation of gene expression was observed only in the context of LVs, which might impact on tumor-suppressor genes (Maruggi et al. 2009).

Similarly to γRV, the safety of LVs has been greatly improved by the generation of SIN-LVs (Zufferey et al. 1998). Many clinical trials have shown that SIN-LVs are quite safe for hematopoietic stem-cell-mediated gene therapy of X-SCID, WAS, and metachromatic leukodystrophy (Biffi et al. 2004, 2013; Galy et al. 2008; Zhou et al. 2010; De Ravin et al. 2016; Singh et al. 2017). Indeed, in the last few years, the use of LVs in gene therapy clinical trials has grown exponentially (Cavazzana et al. 2019; Naldini 2019) and will probably continue to do so in the future (The Journal of Gene Medicine, Gene Therapy Clinical Trials Worldwide, https://onlinelibrary.wiley.com/doi/full/10.1002/jgm.3015).

Despite the many advantages offered by LVs, RVs seem more manageable for ex vivo genetic modification of human keratinocytes. This is perhaps because Rev and the viral proteases of LVs are quite toxic for primary keratinocytes, leading to loss of a significant number of clonogenic cells during the transduction procedure (Haselhorst et al. 1998). Furthermore, although we have been able to limit the number of γRV integrations in keratinocytes to less than two per cell, we consistently detect more than 10 LV integrations in each clonogenic human keratinocyte, irrespective of which transduction protocol is used (Mavilio et al. 2006; Bauer et al. 2017; Hirsch et al. 2017; MC Latella et al., unpubl.). Of note, SIN-γRV showed a significantly lower bias for transcription start sites and promoters in human keratinocytes as compared with hematopoietic cells (Cavazza et al. 2013). Finally, preclinical studies aimed at combined cell and gene therapy of RDEB—a form of EB caused by mutations in COL7A1, a gene encoding for collagen VII (see below)—showed that virtually all clonogenic LV-transduced keratinocytes contained multiple rearranged, truncated, or fragmented collagen VII proteins (MC Latella et al., unpubl.). In contrast, only ~25% of γRV-transduced keratinocytes, analyzed at the clonal level, contained a single aberrant collagen VII, always in the presence of full-length protein (JW Bauer, M De Luca et al., unpubl.). As a practical consideration, RVs allow the generation of stable GMP-certified packaging cell lines that can be used for years, hence fostering reproducibility of the procedure and lowering the cost of each transduction process. Instead, LV proved to be more efficient and less toxic in transducing human primary fibroblasts (Woodley et al. 2003; Georgiadis et al. 2016; Lwin et al. 2019).

Herpes Simplex Viruses

Herpes simplex viruses (HSVs) are DNA viruses extensively used for the generation of vectors. Their broad cellular tropism, high-packaging capacity (almost 150 kb), and ability to transduce both dividing and quiescent cells have quickly attracted the interest of researchers for their use in in vivo gene therapy approaches, particularly for the treatment of neurological disorders (Spaete and Frenkel 1982; Artusi et al. 2018).

In contrast to RVs and LVs, HSVs and their related vectors remain episomally recircularized in the nucleus and are thought to replicate as a concatemer via a rolling circle-like mechanism (Subak-Sharpe and Dargan 1998; Whitley and Roizman 2001). To date, different types of HSV-1-derived vectors exist: (1) replication-defective rHSV-1 lacking essential replication genes, which are provided in trans by the cell line, (2) attenuated rHSV-1, containing deletions in nonessential genes that hinder their replication in vivo but not in vitro, and (3) HSV-1-derived amplicon vectors, which are the safest by virtue of their carrying just an origin of viral replication, a packaging signal, and the cassette of interest (Manservigi et al. 2010). Thus, replication, structural, and packaging proteins must be supplied in trans by a helper/packaging cell line (lacking packaging signal), achieving a lower chance of RCV generation, and hence a better safety profile. A historical issue that concerned HSVs has been the presence of contaminating helper viruses in vector stocks, which could lead to unwanted immune responses (Bowers et al. 2003). However, the most important pitfall related to the use of HSVs in genodermatoses is their episomal nature, which hampers a long-lasting therapeutic effect (Marconi et al. 2008).

In conclusion, genodermatoses can best be tackled through combined cell and gene therapy by means of γRV, most preferably SIN forms. It would be of great interest to see whether SIN-αRVs could attain similar performances, considering their safer integration profile.

Epidermolysis Bullosa

Inherited EB encompasses a heterogeneous group of rare, genetic disorders characterized by recurrent blistering of the skin and other stratified epithelia (Fig. 2). Blisters arise spontaneously or on minimal mechanical stress or trauma as a result of increased epithelial fragility, caused by molecular defects within genes encoding various structural proteins of the epidermal–dermal junction (Fig. 2; Fine et al. 2014). The incidence and prevalence of EB in the United States are ~1/53,000 and 1/125,000, respectively, and similar estimates have been obtained in some European countries (Fine 2010, 2016). EB affects individuals from all ethnic origins without gender predilection and displays either dominant or recessive patterns of inheritance. More than 1000 mutations in at least 14 structural genes (Fig. 2) have been documented causing distinct clinical manifestations, ranging from mild to severe, with local or generalized involvement and significant morbidity and mortality. This variety depends on several factors, both phenotypic (distribution and severity of the lesions, involvement of mucosae, age of onset) and molecular (targeted protein, type of mutation and degree of function loss, mode of inheritance, and genetic background) (Fine 2010; Intong and Murrell 2012).

Inherited EB encompasses three major forms, primarily based on the level of skin cleavage (and subsequent blistering formation), that is intraepidermal in EB simplex (EBS), within the lamina lucida in JEB and beneath the lamina densa in DEB (Fig. 2; Intong and Murrell 2012; Fine et al. 2014). Kindler syndrome (KS) is a fourth, more rare EB form caused by mutations in FERMT-1, the gene encoding Kindlin-1. In KS, blisters occur at multiple levels, intraepidermal, and within or beneath the basement membrane (Intong and Murrell 2012; Fine et al. 2014).

EB, especially in its generalized intermediate/severe forms, greatly impairs the quality of life of people suffering from this painful, chronic and visible condition, representing a heavy health care burden (Horn and Tidman 2002; Fine et al. 2005; Dufresne et al. 2015). There is no cure for EB. Currently available therapies are largely palliative, only partially alleviating the devastating clinical manifestations (Prodinger et al. 2019). Such therapies, either systemic or topical (Prodinger et al. 2019), are not sufficient to provide decisive relief from pain, symptoms, and mental stress and achieve satisfactory living standards for these patients. Long-lasting, curative therapies are urgently needed and several attempts have been made in this respect. As of today, molecular-based therapeutic strategies aiming at gene correction in combination with cell-based approaches focus on individually designed treatments, holding promise for more effective results. We will now focus on JEB and RDEB, the only forms tackled by combined cell and gene therapy approaches that have made it to the clinic.

Junctional Epidermolysis Bullosa

Generalized JEB is a chronic, devastating, and life-threatening condition caused by mutations in LAMA3, LAMB3, or LAMC2—encoding, respectively, α3, β3, and γ2 chains of laminin-332 (also known as laminin 5)—and in genes encoding collagen XVII and α6β4 integrins (Fig. 2; Fine et al. 2008, 2014). Hemidesmosomes are rudimentary or absent, hence blisters occur within the lamina lucida of the basement membrane (Fig. 2) and predominate at sites exposed to friction, trauma, or heat. In addition, hair, nail, and enamel defects are common, along with the involvement of the ocular surface and gastrointestinal and renal systems (Laimer et al. 2010; Yancey and Hintner 2010). Deleterious mutations in genes encoding laminin 332 are usually lethal early in life (Hammersen et al. 2016). Approximately 40% of patients with intermediate to severe generalized JEB die before adolescence and adults are at high risk of developing squamous cell carcinoma (Fine et al. 2008, 2014).

LAMB3-deficient generalized intermediate JEB was the first genetic skin disease successfully tackled by ex vivo gene therapy (Mavilio et al. 2006). Combined cell and gene therapy was tested in a phase I/II clinical trial and eventually developed into a long-lasting curative treatment for skin lesions of this EB subtype. Preliminary preclinical studies showed that holoclones could be stably transduced by γRVs carrying different transgenes, including the LAMB3 full-length cDNA (Morgan et al. 1987; Mathor et al. 1996; Dellambra et al. 1998). Transgenic autologous epidermal cultures were then grafted on patients’ selected skin areas, on surgical removal of diseased epidermis and proper preparation of the wound bed. This approach has proven effective in restoring patches of normal epidermis on large nonhealing epidermal lesions in two adults LAMB3-dependent JEB patients (Mavilio et al. 2006; De Rosa et al. 2014; Bauer et al. 2017).

More recently, ex vivo gene therapy proved to be lifesaving, as it succeeded in regenerating virtually the entire epidermis of a 7-year-old child suffering from a devastating form of LAMB3-dependent JEB with very poor prognosis (Hirsch et al. 2017). Through the entire 4-year follow-up, the transgenic epidermis remained robust and resistant to mechanical stress, free from blisters or erosions. It expressed physiological levels of laminin-332, had normal thickness and continuity of the basement membrane, normal numbers and morphology of hemidesmosomes, and, notably, normal wound healing on injury (Hirsch et al. 2017). Strikingly, the regenerated epidermis was entirely transgenic, as LAMB3 messenger RNA (mRNA) and laminin 332 were uniformly detected in all the analyzed skin sections. No immune response or inflammation were observed. Of note, these studies not only attest to a most likely permanent (4- to 14-year follow-up) functional restoration of the epidermal–dermal junction but also reveal a good safety profile.

Targeting LAMB3 into the genome of long-lived stem cells underpins the stability of the transgenic epidermis. In fact, these studies provided formal, direct evidence that holoclone-forming cells identify authentic epidermal stem cells. Using γRV integrations as clonal genetic marks, clonal tracing has formally shown that, in the long-term, human epidermis is sustained uniquely by a defined number of self-renewing holoclones, which therefore have all the hallmarks of stem cells (Hirsch et al. 2017). Clonal tracing has also shown that holoclones continuously generate pools of meroclones and paraclones, which behave as short-lived progenitors, persist for various periods of time, replenish differentiated cells, and play a crucial role both in the engraftment of epidermal cultures and in epidermal regeneration during wound healing (Hirsch et al. 2017). This latter notion is consistent with the observation that single clonogenic keratinocytes derived directly from a biopsy taken from normal, unwounded skin generate mostly holoclones, whereas keratinocytes cloned from wounded skin generate predominantly mero- and paraclones (Fig. 4). Taken together, these data shed light on the issue, raised by lineage tracing experiments in mice, whether mammalian epidermis is sustained by specific long-lived, self-renewing stem cells (Mascré et al. 2012; Hirsch et al. 2017) or equipotent progenitors (Clayton et al. 2007; Rompolas et al. 2016), the latter being also suggested by the mode of proliferation of cultured clonogenic human keratinocytes (Roshan et al. 2016).

Open in a separate window

Figure 4.

Clonal analysis from skin biopsies and primary cultures. Clonal analysis was performed as described in Figure 3B. Human skin samples were obtained as anonymized surgical waste, typically from abdominoplasty or mammoplasty. Skin samples were cut in three parts for (1) direct cloning from intact skin, (2) direct cloning from wounded skin, (3) preparation of primary cultures. (1) Clonal analysis performed directly from the skin biopsy unveiled that the majority of clonogenic cells of intact skin generate holoclones (stem cells). (2) The epidermis was mechanically removed from the center of the biopsy to mimic a wound and the biopsy was kept in the incubator in keratinocyte growth medium (Hirsch et al. 2017). Clonal analysis, performed 24 hours later from a skin section taken in between the two edges of the wound, showed that the majority of clonogenic cells during the wound-healing process generates meroclones and paraclones (transient amplyfing progenitors). A similar distribution of clonogenic cells was detected in primary cultures (3), confirming that such cultures recapitulate a wound-healing process.

Altogether, the three JEB patients treated received ~4 × 10⁸ transgenic clonogenic keratinocytes, none of which developed into a transformed clone, within a time frame that entailed >150 epidermal renewal cycles (Mavilio et al. 2006; De Rosa et al. 2014; Bauer et al. 2017; Hirsch et al. 2017). The notion that long-term maintenance of transgenic epidermis is in fact sustained only by holoclones, a small percentage of such clonogenic cells (Mavilio et al. 2006; De Rosa et al. 2014; Bauer et al. 2017; Hirsch et al. 2017), indicates that the number of independent genomic hits stably introduced into the patient with the graft is finite, contributing to lower the risk of oncogene activation.

Taking all these results together, combined cell and gene therapy for LAMB3-dependent JEB is probably the most advanced therapeutic gene therapy approach for genodermatoses, and the closest to obtaining final approval by regulatory authorities as an advanced therapy for JEB skin lesions (De Luca et al. 2019).

Dystrophic Epidermolysis Bullosa

DEB is caused by mutations in COL7A1, the gene encoding collagen VII, the major component of dermal anchoring fibrils, whose absence (or alteration) induces blister formation just beneath the lamina densa of the basement membrane (Fig. 2). Dominant DEB (DDEB) generally displays a mild phenotype, with blisters occurring on areas of trauma (with a predilection for the extremities), often resulting in scars, milia (keratin-filled cysts of sweat glands), and loss of nails (Das and Sahoo 2004). Instead, generalized RDEB is more severe. Blisters develop at birth and progressively result in extensive scarring, particularly in joints, and there is highly disabling pseudosyndactyly—fusion of web spaces between the digits—and involvement of the mucous membranes (Nystrom and Bruckner-Tuderman 2016). The quality of life of RDEB patients is regarded as the poorest among all EB forms (Nystrom and Bruckner-Tuderman 2016) and patients’ life expectancy of usually highly reduced, because they almost invariably develop aggressive squamous cell carcinomas as the disease progresses (Fine et al. 2009; Montaudie et al. 2016; Kim et al. 2018).

A variety of preclinical studies attempted to deliver collagen VII to RDEB skin, underlining some crucial issues for future gene therapy in humans. Experiments included direct intradermal injection of LVs containing COL7A1 (Woodley et al. 2004b) and either local or systemic delivery of recombinant collagen VII protein (Woodley et al. 2004a; Remington et al. 2009). The outcome of the latter strategy pointed to its potential ability to safely reverse RDEB keratinocytes phenotype, at least temporarily because the half-life of collagen VII was estimated to be ~1–2 months (Woodley et al. 2004a; Remington et al. 2009; Latella et al. 2017).

During the ex vivo gene therapy approaches pursued to tackle this ravaging disease, major hindrances were encountered in vector construction. The huge COL7A1 coding sequence (8833-nucleotides) makes its accommodation within γRVs a real challenge and negatively affects virus packaging, substantially reducing viral titer and limiting the efficiency of cell transduction (Christiano et al. 1994).

Marinkovich and colleagues recently reported early data from the first phase I/II ex vivo combined cell and gene therapy clinical trial for RDEB (Siprashvili et al. 2016). The trial involved transduction of cultured RDEB keratinocytes by means of γRVs bearing COL7A1, followed by generation of autologous transgenic epidermal grafts applied onto selected wounds of four patients. This approach appeared to be safe and well tolerated. In particular, no neutralizing immune response was observed to de novo synthesized collagen VII, no recombinant RV was detected in the bloodstream and no malignant transformation of transgenic keratinocytes occurred over a 12-month follow-up. To date, this study presented the most convincing evidence of deposition of functional collagen VII and restoration of anchoring fibril in the skin of RDEB patients. However, a variable, limited efficacy was reported, because at 12 months, ~50% of the transplanted areas were covered by transgenic, functional epidermis. Three patients showed a decline of collagen VII expression and blistering at transplant sites after 6 months, whereas the fourth patient showed a decrease in collagen VII and anchoring fibrils at 12 months postgrafting, suggesting that the benefit was tapering off (Siprashvili et al. 2016) (ClinicalTrials.gov identifier: NCT01263379).

We have initiated a similar γRV-based phase I/II trial on RDEB (ClinicalTrials.gov identifier: NCT02984085) by means of autologous epidermal cultures transduced with the same type of γRVs used for gene therapy of JEB. This trial is still ongoing, but initial outcomes confirm the aforementioned results reported by Siprashvili et al. In contrast to gene therapy in JEB, our initial data point to mosaic epidermal regeneration in RDEB patients. Patches of transgenic epidermis expressing collagen VII are indeed interspersed with patches of untransduced, mutant epidermis (JW Bauer and M De Luca et al., unpubl.).

Marinkovich and colleagues speculate that the partial restoration of functional collagen VII might be caused by an age-related decline of the regenerative potential of epidermal cultures (Siprashvili et al. 2016). In our opinion, this is unlikely; autologous epidermal and limbal cultures permanently (more than a 35-year follow-up) regenerate the epidermis and the corneal epithelium in massive skin and ocular burns irrespective of the age of patients (Pellegrini et al. 1999b, 2013; De Luca et al. 2006; Rama et al. 2010). The first two JEB patients grafted with autologous transgenic epidermal sheets were 37 and 50 years old, respectively (Mavilio et al. 2006; Bauer et al. 2017). The first patient now had a 14-year follow-up and the transgenic epidermis is functioning well. In principle, the aforementioned findings could be ascribed to an insufficient number of epidermal stem cells retrieved from the skin biopsy and/or targeted by γRVs. Indeed, holoclone analysis was not performed in Siprashvili's study; therefore, stem cells were not quantified as they were in our JEB studies. Nevertheless, although we did evaluate the number of transduced stem cells by holoclone formation in our RDEB transgenic grafts, our clinical results are similar to those of Siprashvili and colleagues.

Assuming that proper gene correction of epidermal stem cells has occurred, the partial restoration of collagen VII might relate to intrinsic differences between JEB and RDEB clonogenic keratinocytes and potential competition between COL7A1-transduced and -untransduced cells. JEB, but not RDEB, keratinocytes initiate colonies with low efficiency and show rapid clonal conversion leading to premature stem-cell depletion and replicative senescence (Mavilio et al. 2006; De Rosa et al. 2019). It has been recently shown that the epidermal stem-cell depletion and premature replicative senescence characteristic of JEB are caused by the perturbation of the YAP/TAZ signaling pathway (De Rosa et al. 2019). Such perturbation is not found in RDEB keratinocytes, whose clonogenicity and growth potential are similar to those of normal keratinocytes. Strikingly, gene therapy of JEB keratinocytes completely restores the YAP/TAZ pathway and, consequently, fully rescues epidermal stem-cell function and proliferative potential, both in vitro and in vivo after transplantation (De Rosa et al. 2019). It follows that, in contrast to RDEB, genetically corrected clonogenic JEB keratinocytes must hold a strong selective advantage over their gene corrected counterparts. Because the transduction efficiency of clonogenic RDEB keratinocytes was in the range of 50%–70% in both Siprashvili's and our study, it is tantalizing to speculate that cell competition between transduced and untransduced RDEB epidermal stem cells within the graft could explain the mosaic composition of the regenerated epidermis and the progressive decrease of collagen VII observed in RDEB patients treated with transgenic cultures.

It follows that a decisive improvement in combined ex vivo cell and gene therapy for RDEB could ensue from a precise control of the number of transduced holoclones contained in the graft, together with the development of new vectors and/or transduction procedures allowing genetic correction of virtually 100% of the entire stem-cell population of primary RDEB cultures. This should minimize the competition between transduced and untransduced RDEB cells and assure long-term regeneration of a fully functional transgenic epidermis. Such optimization procedures are underway. A hitherto unsolved issue might relate to potential competition between transgenic interfollicular epidermis and untransduced keratinocytes contained in the bulge or the isthmus of hair follicles (Snippert et al. 2010), which might contribute to interfollicular epidermis (Blanpain and Fuchs 2009). However, the notion that stem cells in the hair follicles contribute to wound repair but not to homeostasis of the epidermis (Claudinot et al. 2005; Ito et al. 2005), argues against this possibility, would the transgenic epidermis have a complete engraftment upon transplantation, as shown with JEB.

Last, but not least, hurdles also relate to intrinsic challenges in understanding the mechanisms underlying many rare diseases, which complicates the design of efficacious therapies. Particularly in RDEB, where skin cleavage is deeper than in JEB (Fig. 2), the clinical success of epidermal cultures strictly relies on proper selection of the patients including their immunological picture (Siprashvili et al. 2016). It also depends on control of inflammation and microenvironment of the receiving wound bed (which varies between JEB and RDEB and can hamper the engraftment of the culture), preparation of the wound bed, and postoperative procedures (De Luca et al. 2006), all of which are beyond the scope of this chapter.

Other gene therapy approaches for RDEB involve the use of LV- and HSV-derived vectors in autologous dermal fibroblasts, which can also produce to collagen VII (Marinkovich et al. 1993). A phase I clinical trial used SIN-LVs for correcting RDEB fibroblasts by an ex vivo strategy. This protocol (referred to as FCX-007) was used on five patients and envisaged local intradermal injection of transgenic fibroblasts (ClinicalTrials.gov identifier: NCT02810951) (Marinkovich et al. 2018). FCX-007 proved to be well-tolerated and safe in patients during the 52-week long follow-up, with no product-related adverse events or immunologic response being reported. However, data on the efficacy of FCX-007 are not yet available. A similar trial (ClinicalTrials.gov identifier: NCT0249381), but using codon-optimized COL7A1, reported comparable safety and early efficacy outcomes, pointing to a partial clinical synthesis and deposition of collagen VII, but without formation of mature anchoring fibrils (Lwin et al. 2019).

Another strategy to treat RDEB used KB103, an off-the-shelf topical gel delivering COL7A1 via an attenuated HSV-1 vector through an in vivo approach (ClinicalTrials.gov identifier: NCT03536143). KB103, targeting keratinocytes and fibroblasts, was developed for both severe RDEB and milder DDEB. However, its therapeutic effect is predicted to be temporary because of the episomal nature of HSV, which does not integrate within the patient's genome and is therefore eliminated on epidermal renewal. Finally, a proposed combined cell and gene therapy approach for RDEB envisages the use of iPS as these cells can be selected in culture after gene correction and induced to differentiate into keratinocytes. In principle, a single, very well-characterized iPS cell line can be used to produce large amounts of transgenic epidermis (Tolar et al. 2011; Sebastiano et al. 2014; Wenzel et al. 2014). This approach has not yet reached the stage of an approved clinical trial.

Recently, immunomyeloablation followed by allogeneic bone marrow (BM) transplantation has been proposed for the treatment of RDEB (Wagner et al. 2010). Variable clinical benefits were reported (improved healing of the wounds and reduced frequency of daily medication). But anchoring fibrils were not restored, despite increased collagen VII deposition at the epidermal–dermal junction (Wagner et al. 2010). Such an increased deposition might ensue from BM-derived subclinical inflammation resulting in heparin-binding EGF up-regulation and a consequent increase in mutated collagen VII synthesis in host cells, as observed with fibroblasts or mesenchymal stromal cells (Wong et al. 2008; Haniffa et al. 2009). In fact, it is not even clear which cell type contained in the BM aspirates and which mechanism could determine the limited reported clinical benefits, and whether such benefits could not rather be ascribed to a better management of the lesions attained during the long hospitalization of the patients. Indeed, the poorly defined trophic, anti-inflammatory or immune modulatory effects of BM-derived cells observed in other clinical settings is still a quite controversial issue (Bianco et al. 2013; De Luca et al. 2019). In summary, BM transplantation is not curative and in the absence of a curative effect, it is hard to accept 30% lethality (Wagner et al. 2010) in children whose life expectancy is of a few decades. Clinical trials envisaging less aggressive myeloablation are ongoing (NCT00881556).

Recessive Forms of EB

More than 650 different mutations causing RDEB have been identified within COL7A1. The majority of them cluster in distinct hotspots and are typical of certain ethnicities or groups of families (Intong and Murrell 2012; Fine et al. 2014). A proposed gene-editing strategy involves the eviction of a region carrying multiple genetic defects with the aim of simultaneously correcting a whole set of mutations (Escamez et al. 2010; van den Akker et al. 2011; Wertheim-Tysarowska et al. 2012). For instance, the c.6527insC mutation in COL7A1 was tackled by a combined delivery of template DNA and TALEN, by adenoassociated virus and adenoviral vectors, respectively. This strategy achieved correction of the mutation via HDR and additionally restored the reading frame via NHEJ (Chamorro et al. 2016).

An ex vivo exon-skipping approach was also developed to precisely excise COL7A1 exon 80 carrying the same c.6527insC mutation. This used a dual single-guide, RNA-guided Cas9 nuclease delivered as a ribonucleoprotein complex through electroporation and succeeded in deleting exon 80 and restoring COL7A1 reading frame in RDEB keratinocytes, which were able to produce functional collagen VII (Bonafont et al. 2019). Alternatively, the CRISPR/Cas9 system was delivered in RDEB-immortalized keratinocytes by plasmid transfection; however, genetic and functional HDR-based correction of the aforementioned mutation only occurred in a few selected clones (Hainzl et al. 2017).

CRISPR/Cas9-mediated homology-directed repair, delivered ex vivo in primary RDEB keratinocytes and fibroblasts through integration-deficient LVs, tackled a c.189delG (p.Leu64Trpfs*40) null mutation in exon 2 of COL7A1. Efficient COL7A1 editing was achieved, showing rescue of functional collagen VII and anchoring fibrils in vivo by means of transgenic skin xenografts (Izmiryan et al. 2018). An iPS-based approach was attempted to tackle a c.1837 C > T mutation on exon 14 of COL7A1. RDEB iPS cells were genetically corrected by means of TALEN-mediated HDR (Osborn et al. 2013). Finally, HDR-based gene editing restored laminin 332 expression in LAMB3-dependent JEB immortalized keratinocytes, but its efficiency was too low for clinical translation (Benati et al. 2018).

Dominant Forms of EB

Dominant forms of EB encompass most EBS and some DEB subtypes. Gene-editing techniques have been implemented for DDEB, covering TALEN nucleases and CRISPR/Cas9-mediated NHEJ approaches aimed at correcting c.8068_8084delinsGA mutation within COL7A1 (Shinkuma et al. 2016). A site-specific mutagenesis inducing NHEJ was performed on iPS generated from DDEB fibroblasts; gene-edited iPS cells produced keratinocytes and fibroblasts able to synthetize and secrete functional collagen VII (Shinkuma et al. 2016). A similar approach targeting RDEB was developed by combining CRISPR/Cas9 with iPS technology (Webber et al. 2016). A dominant hotspot mutation within the KRT14 gene, causing a generalized severe EBS was corrected using a double-nicking strategy (Kocher et al. 2017). This HDR-based approach involved CRISPR/Cas9 correction of a patient-derived immortalized EBS keratinocytes. This strategy resulted in a transfection efficiency of about 30% associated with a 16% pathogenic allele correction, assuming a wild-type/mutated allele ratio of 50/50 within treated cells (Kocher et al. 2017).

Other Genodermatoses

Epidermolytic ichthyosis (EI) is a rare autosomal genodermatosis, with an incidence of 1:100,000–400,000, characterized by blistering and subsequent reactive scaling of the skin (Müller et al. 2006). EI is inherited predominantly with a dominant pattern, although recessive forms have been reported. Cytoskeletal fragility of suprabasal epidermal cells is caused by dominant-negative mutations in either KRT1 or KRT10. As keratin 1 and keratin 10 are crucial cytoskeleton constituents, EI results in major impairment of intermediate filaments (Cheng et al. 1992; Chipev et al. 1992; Oji et al. 2010). To date, treating EI remains a challenge. Palliative treatments and medications aim at removing the excess of thickened skin layers. Recently, TALEN-based gene editing was performed with the aim of inactivating mutant KRT10 by introducing a frameshift mutation upstream of a premature termination codon in cultured EI keratinocytes. Despite phenotype restoration of selected clones, the therapeutic efficiency was low (3.7%) with high disruption rates on the wild-type allele (March et al. 2019).

In summary, proof-of-concept for the use of gene-editing technology in gene therapy of genodermatoses has been obtained, but several technical hurdles must be overcome before this technology can reach its full potential.

This work was supported by the Regione Emilia-Romagna, POR-FESR (E8IJ10000120007 and E92I16000220005).