Cutaneous gene therapy

Abstract

Possibilities of using the skin for somatic gene therapy have been investigated for more than 20 years. Strategies have included both direct gene transfer into the skin and indirect gene transfer utilizing cultured cells as an intermediate step for gene manipulation. Viral as well as nonviral vectors have been used, and both gene addition and gene editing have been performed. Although cutaneous gene therapy has now begun translating into clinical medicine (as seen by the first clinical gene therapy project of an inherited skin disorder) further developments are still required.

• Epidermis
• gene editing
• inherited skin diseases
• keratinocytes
• stem cells
• telomerase

Introduction

The skin was one of the first tissues considered for somatic gene therapy—initially as a general vehicle for the supply of gene products. Thus in 1987 publications described cutaneous expression of three different proteins: human growth hormone, factor IX and adenosine deaminase 1–3. Since these pioneering studies numerous experiments have been performed, exploring possibilities for gene therapy of diseases such as inherited skin diseases and wounds (reviewed in 4–6). Several factors are inhibiting the translation of these projects into the clinic including loss of gene expression, immune reactions and limited blood supply 7–10.

Epidermis is a natural source of secreted proteins with both local and systemic effects, and after gene transfer numerous medically relevant proteins can be secreted from genetically modified epidermal tissue into the blood 11. These proteins include hormones (such as leptin, erythropoietin, and human growth hormone), the coagulation factors factor VIII and factor IX, apolipoprotein E, α1‐antitrypsin, various cytokines, monoclonal antibodies, and antimicrobial peptides 8,12–19. Most studies used constitutive expression of these proteins, but regulated gene expression can also be achieved after both viral and nonviral gene transfer using both the tetracycline system and synthetic ligand‐driven intracellular heterodimerization 20–23. It is possible to select in vivo for the gene‐targeted cells 24. Pfützner et al. showed that colchicine treatment increased the percentage of multidrug resistance (MDR) expressing keratinocytes to almost 100% on mice 25. The transduced keratinocytes were able to proliferate normally and capable of forming a stratified differentiated epidermis.

In tissue culture, gene transfer using optimized murine retroviral vectors can lead to transduction frequencies approaching 100%, making it unnecessary to select for transduced cells 26. The transgene expression and copy number in the keratinocytes can be even further increased by employing a centrifugation technique for the transduction 27. Fibronectin can also be used to enhance the transduction frequency 28.

Key messages

• After more than 20 years of research cutaneous gene therapy has now reached the clinical phase.

• Future developments will include improved gene delivery methods.

• Cutaneous gene therapy, initially relevant for rare inherited skin diseases, may eventually be used for systemic disorders as well.

Stem cells

Epidermal stem cells have been extensively studied and are obviously important for skin gene therapy. In general, stem cells can be defined as cells with the capacity for unlimited self‐renewal and the ability to generate daughter cells that undergo differentiation. In human interfollicular epidermis, basal cells thought to be one class of epidermal stem cells, are found in clusters located at the tip of the dermal papilla 29. These cells are multipotent and can be induced to differentiate into hair and cells of the sebaceous gland (reviewed in 30). Stem cells in the bulge area of the hair follicle are probably more multipotent and can give rise not only to the hair lineages but also to sebocytes and interfollicular epidermis 31. One prevailing model is that, under steady state conditions, stem cells are unipotent (i.e. producing only follicular or only interfollicular epidermis), whereas during regeneration after tissue damage, they display multipotency. Stem cells in the hair follicle bulge do not normally contribute cells to the epidermis, but after epidermal injury they are recruited into the epidermis. Although the bulge‐derived cells acquire an epidermal phenotype, most are eliminated from the epidermis over several weeks 32. Thus, epidermal and hair follicle stem cells are considered as distinct populations. Plucked hair follicle‐derived cells have been used in preparing skin equivalents, and can form epithelium after implantation in deep wounds 33,34. Thus the hair follicle bulge area is considered as an abundant, easily accessible source of actively growing, pluripotent adult stem cells 35. Interestingly, hair follicle stem cells have been shown to be able to differentiate into neurons, glial cells, keratinocytes and smooth muscle cells, and in animal models they enhanced the rate of nerve regeneration and the restoration of nerve function 36.

Telomerase in stem cells

Telomerase is active in the germ line, immortal, and tumor cells, and in cells from continuously renewing normal tissues, and is thought to be correlated with an indefinite life span (reviewed in 37. In general, it is hypothesized that stem cells do not have sufficient telomerase to prevent telomere shortening associated with continuous tissue renewal. However, telomerase cannot be used as a skin stem cell marker, since an enriched population of epidermal stem cells exhibited less telomerase activity than the more actively proliferating transit amplifying cells 38. Telomerase and telomeres are important for the normal function of epidermal stem cells. Shortening of telomeres inhibits mobilization of stem cells, impairs hair growth, and suppresses their proliferative capacity. In contrast, overexpression of Tert, the protein component of telomerase, promoted stem cell mobilization, hair growth, and stem cell proliferation 39,40. This function does not require the telomerase RNA component, which encodes the template for telomere addition, and therefore operates through a novel mechanism independent of its activity in synthesizing telomere repeats.

Ectopic expression of telomerase has been used to immortalize various human cell types, and ectopic expression of the telomerase catalytic subunit in combination with two oncogenes results in direct tumorigenic conversion of normal human epithelial and fibroblast cells 41,42. We reported the establishment of epidermolysis bullosa simplex (EBS) keratinocytes with an extended life span by telomerase expression 43. After immortalization the cells retain normal growth and differentiation characteristics such as the ability to generate an epidermis with normal morphology in organ cultures and also a normal karyotype.

Direct transfer of genes into the skin

Genetic material—in the form of both viral and nonviral vectors—can be transferred directly to the skin. Naked DNA applied topically onto the skin leads to gene expression, but the expression is transient and at low levels 44. Plasmid DNA in various liposomal spray formulations sprayed onto mouse or human skin likewise leads to gene expression, but at low levels 45. A novel gene transfer method based on biphasic lipid vesicles is reported to enhance DNA transfer through intact stratum corneum 46. Higher levels of gene expression after direct delivery can be obtained using intracutaneous injection of DNA, liposomes, a gene gun or electroporation (reviewed in 27).

For prolonged gene expression in epidermis, integration into the genome of stem cells is required. Both retro‐ and lentiviral vectors are efficient for integration into cultured cells, including stem cells, and after direct injection both vectors result in long‐term gene expression in vivo13,47. Vectors based on adeno‐associated virus (AAV) have also been evaluated for in vivo application. Subcutaneous injection of AAV leads to long‐term expression of human factor IX, erythropoietin, and interferon 48, and intracutaneous injection of an AAV vector leads to gene expression for more than 6 weeks with expression in keratinocytes, both in the hair follicle and in eccrine sweat glands 49.

Viral and nonviral gene transfer methods may be combined. Retroviral vectors were produced in tissue‐cultured primary human keratinocytes after DNA‐mediated gene transfer followed by transduction of cocultured keratinocytes and prolonged reporter gene expression 50. This method was evaluated in a pig model using selection for transduced keratinocytes (employing MDR as selective marker) (unpublished data).

The ‘metabolic sink’ approach

Several studies investigated the capacity for genetically modified keratinocytes or skin fibroblasts to metabolize toxic compounds accumulating in various inborn errors of metabolism (reviewed in 51). Both fibroblasts and keratinocytes were able to produce significant amounts of adenosine deaminase (ADA) and metabolize deoxyadenosine and adenosine (accumulating in ADA deficiency), but it was calculated that blood supply would be rate‐limiting for clinical applications 2,10. Keratinocytes from patients with ornithine aminotransferase deficiency (OAT deficiency) overexpressing OAT have a large capacity for protein expression, and ornithine uptake and metabolism was increased to supranormal levels 52,53. We used similar strategies to investigate the possibility of using the skin for gene therapy of phenylketonuria (PKU) caused by deficiency of phenylalanine hydroxylase (PAH) and increased levels of phenylalanine 54. The PAH enzyme requires a cofactor (BH₄), and the rate‐limiting step in the synthesis of BH₄ is GTP cyclohydrolase I (GTP‐CH). In tissue culture we showed that primary human keratinocytes and patient fibroblasts overexpressing PAH and GTP‐CH were able to metabolize phenylalanine. Transgenic mice expressing these two enzymes under control of skin‐specific promoters were cross‐bred to PKU mice to obtain mice only expressing the enzymes in the skin 55. Homozygous PKU mice having expression of the enzymes in the basal layer of epidermis (driven by the K14 promoter) had statistically significant decreased levels of serum phenylalanine compared to control homozygous mice without the transgenes. However, the decrease was only minor showing that the system should be developed further. Additional steps in the clearance pathway—both transport of phenylalanine into the cells and downstream processes—were subsequently manipulated in tissue culture in order to enhance the clearance of phenylalanine 56. These results suggested that enzymes other than GTP‐CH involved in BH₄ synthesis or regeneration are rate‐limiting, such as 6‐pyruvoel‐tetrahydrobiopterin synthase (PTPS) or 4a‐carbinolamine dehydratase (PCD). Interestingly, in the muscle tissue, recent data show that expression of PTPS, in addition to PAH and GTP‐CH, leads to very high levels of phenylalanine clearance (Beat Thony, personal communication).

Wound healing

Non‐healing wounds represent a major clinical problem. Several strategies have been investigated with the purpose of promoting wound healing by use of gene transfer, including stimulation of the granulation process, the vascularization, the reepithelialization or the scar quality 57,58. Recent progress in this field includes transient expression of human host defense peptides, and gene transfer of multiple genes to improve the wound healing process 59–61. One of the major regulators of tissue repair in the skin is platelet‐derived growth factor (PDGF) 62. This protein is produced at the injury site, promotes neovascularization, and stimulates dermal fibroblasts to proliferate, migrate, and synthesize extracellular matrix components. Normally the protein has no direct effect on keratinocytes since these cells do not have PDGF receptors. We showed that expression of human PDGF receptors can be achieved in epidermal keratinocytes by retroviral transduction, and that ligand activation of such gene‐modified skin equivalent enhances cell proliferation 63. Gene transfer resulting in the production of both PDGF A and B has been described to enhance the wound healing process, and, thus, it may be of advantage to make adjacent epidermis sensitive to PDGF stimuli.

Gene editing

Specific repair of disease‐causing mutations can be achieved using several methods. One strategy that has been used in the skin is based on oligonucleotides that stimulate the correction of single‐base mutations in mammalian cells 64,65. In cultured melanocytes a point mutation in the tyrosinase gene was corrected resulting in restoration of tyrosinase enzymatic activity, melanin synthesis, and pigmentation 66. Likewise, topical application and intradermal injection of the oligonucleotides to albino mouse skin resulted in dark pigmentation of several hairs in a localized area 67. Gene repair in primary human keratinocytes may be more difficult by these methods 68. Thus, a better understanding of the underlying mechanisms of the gene conversion process may help to increase the targeting frequency, and the promising repair technique based on artificial zinc finger proteins is worth trying for cutaneous applications 69,70. In epidermolysis bullosa simplex, caused by keratin mutants expressed in basal cells, gene‐corrected keratinocytes from the basal layer will have a growth advantage 71. Thus strategies aiming at correcting keratin mutations in keratinocytes are indeed worth pursuing.

Gene therapy of inherited skin diseases

Numerous experiments aiming at the development of gene therapy of inherited skin diseases have been described, and a clinical trial was initiated in Italy involving ex vivo retroviral mediated transfer of the β3 gene of laminin‐5 to patients with junctional epidermolysis bullosa (JEB) 72,73. This disease is caused by defects in proteins important in the adhesion of basal layer cells to the dermis, including laminin‐5, a major basal lamina component formed by three distinct polypeptides, α3, β3, and γ2. Stable transduction of JEB patient stem cells, long‐term expression in vivo, and complete phenotypic correction of mutated keratinocytes, combined with the facts that corrected cells are more adherent to the basal lamina, and thus may have a selective advantage in vivo, led to trial.

Several research groups have focused on the development of gene therapy of recessive dystrophic epidermolysis bullosa (RDEB). This disease is caused by mutations in the COL7A1 gene that codes for the epidermal adhesion protein collagen VII. PhiC31 bacteriophage integrase was used to integrate COL7A1 cDNA into keratinocytes from RDEB patients and normalize the disease phenotype 74. Intradermal injection of RDEB fibroblasts overexpressing type VII collagen into intact RDEB skin is another strategy that stably restored correctly localized type VII collagen expression in vivo75. Finally, injection of lentiviral vectors expressing human type VII collagen transduced dermal cells, which synthesized and exported type VII collagen into the extracellular space and reversed the RDEB phenotype 76.

Xeroderma pigmentosum (XP) is an inherited disease associated with increased risk for the development of skin tumors in sun‐exposed areas. Impaired nucleotide excision repair is deficient leading to ultraviolet‐induced hypermutagenesis and, thus, predisposition to cancer. In tissue culture, the phenotype of XP keratinocytes was normalized after retrovirus‐mediated transfer of the adequate wild‐type XP gene 77, but even more impressing, recombinant adenovirus carrying a human XPA gene was used for in vivo gene therapy in UVB‐irradiated skin of XP‐mutant mice, and here injected virus prevented development of squamous cell carcinoma 78. Gene delivery to the skin is therefore a promising tool for reconstitution of specific DNA repair defects in XP patients.

Hair follicle gene therapy

As mentioned previously, hair follicle stem cells are highly pluripotent and can form neurons, glial cells, and other cell types, and it was suggested that hair follicle stem cells may serve as general gene therapy targets for regenerative medicine 79.

Retroviral vectors encoding reporter genes have been used to transduce cultured hair follicle stem cells. After skin reconstitution, transgene expression was observed in all skin epithelial compartments including the hair follicle epithelium, sebaceous gland, and epidermis. In addition, transgene expression was observed for at least 6 months 80.

The protein Sonic hedgehog (Shh) is involved in the initiation of hair growth. After adenoviral mediated expression of the protein in the skin after direct injection, it was found that the hair follicle size increased and that there was a marked acceleration of the onset of new hair growth in the region of adenovirus administration 81. It was suggested that the strategy may be beneficial in the treatment of some forms of alopecia associated with chemotherapy 82.

Immune reactions

The skin is a very immunologically active tissue, and gene transfer to skin has been shown to induce transgene‐specific immune responses (reviewed in 83). Ex vivo gene transfer approaches, where keratinocytes are transduced in culture and transplanted back to patients, may avoid signals provided to the immune system by in vivo administration of vectors 84. It was shown that activation of transgene‐specific immune responses in ex vivo gene transfer targeted to keratinocytes requires cross‐presentation of transgene product to antigen‐presenting cells. This process is amenable to immune modulation, and thus immunological reactions may be circumvented using this strategy.

Gene transfer into the skin has elicited effective immune responses against a large variety of different infectious agents and tumor antigens in animal models (reviewed in 85. DNA transfer into the skin can lead to tumor regression by enhancing the tumor‐directed immune response, and cytokine genes may be delivered directly to a tumor to stimulate destruction by the host immune system 86–90. Several cell types that are present in the skin microenvironment, including keratinocytes in epidermis and fibroblasts in dermis, have the capacity to secrete cytokines that initiate and control immune responses 91. Furthermore, the skin has resident epidermal and dermal dendritic cells, a very potent (‘professional’) antigen‐presenting cell. Transfected dendritic cells have been shown to present transgenic peptides through both MHC class I‐ and II‐restricted pathways 92,93. In addition, this cell type is able to take up proteins from transfected neighboring cells in the case of secreted proteins or by phagocytosis of necrotic or apoptotic cells 94,95. Both of these mechanisms can induce humoral and cellular immune responses.

Conclusions

For gene therapy applications, the skin, the largest organ in the body, offers several advantages compared to other organs including easy accessibility, well understood biology, and safety parameters. Important factors arguing for continued focus on this field include controlled gene expression, 100% targeting of cultured cells in vitro including stem cells, prolonged gene expression even after in vivo applications, and excellent transplantation methods. Thus cutaneous gene therapy will remain a potential approach not only for local diseases but also in the future most likely for systemic diseases. Development of vectors allowing topical gene transfer, efficient targeting of long‐lived cells and regulated gene expression will pave the way for the treatment of diseases requiring short‐term or long‐term effects.

Figure 1. Number of publications in PubMed about skin gene therapy.

Figure 2. Combined viral and nonviral gene transfer into human keratinocytes performed as previously described50. a Monolayer cultures of human keratinocytes. The cells were transfected with a retroviral construct containing the LDL receptor gene (left panel) or cotransfected with the same construct and the pPAM3 plasmid containing gag pol and env genes (right panel). The LDL receptor protein was visualized using immunostaining at the indicated time points after transfection. (Green = FITC‐labeled LDL receptor protein; red = propidium iodine‐staining of nuclei) for colours please see (50). After cotransfection the cells produce retroviral vectors that subsequently transduce 100% of the cells. b Raft cultures of human keratinocytes cultured in the liquid‐air interphase. The cells were transfected in monolayer with a retroviral construct containing a GFP gene and the pPAM3 plasmid containing gag pol and env genes and subsequently induced to differentiate to a multilayered tissue. GFP protein was visualized using immunostaining as described 50. The anti‐GFP antibody was visualized using peroxidase‐staining leading to the brown color. GFP‐positive cells are found in all layers of the tissue. In control cultures transfected with the GFP construct alone no GFP‐positive cells were found at this time point (13 days after transfection) (data not shown).