An RNA-targeted therapy for dystrophic epidermolysis bullosa

Abstract

Functional impairment or complete loss of type VII collagen, caused by mutations within COL7A1, lead to the severe recessive form of the skin blistering disease dystrophic epidermolysis bullosa (RDEB). Here, we successfully demonstrate RNA trans-splicing as an auspicious repair option for mutations located in a wide range of exons by fully converting an RDEB phenotype in an ex vivo pre-clinical mouse model based on xenotransplantation. Via a self-inactivating (SIN) lentiviral vector a 3' RNA trans-splicing molecule, capable of replacing COL7A1 exons 65-118, was delivered into type VII collagen deficient patient keratinocytes, carrying a homozygous mutation in exon 80 (c.6527insC). Following vector integration, protein analysis of an isolated corrected single cell clone showed secretion of the corrected type VII collagen at similar levels compared to normal keratinocytes. To confirm full phenotypic and long-term correction in vivo, patches of skin equivalents expanded from the corrected cell clone were grafted onto immunodeficient mice. Immunolabelling of 12 weeks old skin specimens showed strong expression of human type VII collagen restricted to the basement membrane zone. We demonstrate that the RNA trans-splicing technology combined with a SIN lentiviral vector is suitable for an ex vivo molecular therapy approach and thus adaptable for clinical application.

Figure 1.

LV-RTM-S6m transduced RDEB patient keratinocytes showed trans-splicing corrected type VII collagen expression. (A) The bidirectional SIN lentiviral vector contained the coding sequence of COL7A1 exons 65–118 and the 224 bp BD in antisense direction expressed by a spleen focus forming virus promoter (SF-Pro). The GFP and puromycin cassette (Puro) under the control of a mnCMV promoter was inserted in sense direction. (B) Stable transduction of LV-RTM-S6m into an RDEB keratinocyte cell line induced accurate trans-splicing into the endogenous COL7A1 pre-mRNA, leading to the replacement of endogenous COL7A1 exons 65–118 by a wild-type copy provided by the RTM. Asterisk = silent mutations. (C) PCR analysis on genomic DNA of LV-RTM-S6m and LV-RTMm-woBD transduced RDEB keratinocytes showed full-length genomic RTM integration (3.6 kb). Positive control: LV-RTM-S6m plasmid. (D) Immunofluoresence staining showed a specific signal indicating type VII collagen expression (red) in LV-RTM-S6m transduced RDEB keratinocytes, similar to healthy wild-type keratinocytes (normal kc) and hardly visible in negative controls (untransduced and LV-RTMm-woBD). Cell nuclei: 4′,6-diamidin-2-phenylindol (DAPI, blue). Scale bar (SB) = 20 μm.

Figure 2.

Single cell clone C47 showed strong type VII collagen expression. (A) Immunofluorescence analysis showed strong type VII collagen expression (red) in C47. Vector integration in transduced keratinocytes was confirmed by GFP expression (green). The cell nuclei were stained with DAPI (blue). SB = 50 μm. (B) Immunoblotting of extracted cell lysates and supernatants revealed expression as well as secretion of corrected type VII collagen in C47 (arrows), barely visible in untransduced and LV-RTMm-woBD transduced RDEB keratinocytes. Positive control: normal keratinocytes (kc). α-actinin and Ponceau staining confirmed equal loading.

Figure 3.

Trans-splicing corrected type VII collagen expression is functional. C47 showed an increased binding ability to the ECM proteins collagen IV and fibronectin compared to untransduced and LV-RTM-woBD transduced RDEB keratinocytes. Mean of four values for collagen IV and five values for fibronectin and error bars (SEM) including one-way ANOVA analysis with a Tukey's multiple comparison test are given. ns = not significant; *P-value < 0.05; **P-value < 0.01; ***P-value < 0.001.

Figure 4.

Analysis of the trans-spliced COL7A1 mRNA in the LV-RTM-S6m corrected RDEB single cell clone C47. (A) Via sqRT-PCR using primers specifically hybridizing to endogenous COL7A1 exon 62/63 and the introduced silent mutation in exon 65 on the RTM, we detected a product of 216 bp, corresponding to the trans-spliced COL7A1 mRNA in LV-RTM-S6m transduced RDEB cell pool and single cell clone C47. LV-RTM-woBD transduced cells served as negative control. (B) Sequence analysis of trans-spliced COL7A1 mRNA shows the correct exon 64/65 junction as well as the amplified silent mutations. (C) SqRT-PCR analysis showed that 2.113% of COL7A1 mRNA in C47 was correctly trans-spliced. Mean of four individual experiments and error bars (SEM). ***P-value < 0.001.

Figure 5.

Full-length integration of the provirus into the C47 genome. (A) Full-length RTM integration was detected by Southern Blot analysis using HindIII, flanking the RTM, and a 327 bp probe spanning from COL7A1 exon 70 to parts of the spacer sequence of LV-RTM-S6m, expecting a predicted band size of 4441 bp upon RTM integration. (B) Southern blot analysis showed full-length RTM integration in LV-RTM-S6m and LV-RTM-woBD transduced keratinocyte pool, as well as in single cell clone C47 by detecting a band at 4.4 kb.

Figure 6.

Grafting of skin equivalents derived from C47 revealed long-term correction of the RDEB phenotype. (A–D) H&E staining of 12 weeks old skin grafts derived from C47 (C), normal keratinocytes (B), and untransduced keratinocytes (A) showed normal differentiation into human skin. As a control we included a murine tissue sample (D). (E–H) Immunofluorescence staining of type VII collagen in 12 weeks old skin grafts derived from C47 (G) showed protein expression at the BMZ (red) at comparable levels as detected in grafts from normal human keratinocytes (F). Grafts expanded from untransduced RDEB keratinocytes (E) showed only weak type VII collagen expression at the BMZ. As negative control, we included murine wild-type tissue (H). The cell nuclei were stained with DAPI (blue). e = epidermis; d = dermis. C7 = type VII collagen. SB = 50 μm.