1. Introduction
RNA-based therapies constitute elegant options to restore gene function in various genetic diseases. Therapeutic intervention at the posttranscriptional level during the course of naturally occurring mRNA maturation in the cell represents a viable opportunity to manipulate the mRNA sequence prior to protein translation. Examples of such strategies include splice-modulating therapies, such as antisense oligonucleotide (ASO)-mediated exon skipping and spliceosome-mediated RNA
trans-splicing (SMaRT), both of which are currently under development for the therapy of several monogenetic diseases [
1,
2,
3].
Others [
4,
5] and we ourselves [
6,
7] have applied SMaRT as an emerging RNA editing tool for the specific correction of recessively and dominantly inherited mutations. So far, SMaRT has successfully been implemented in RNA repair studies for genetic disorders such as epidermolysis bullosa (EB) [
8,
9], muscular dystrophy [
4,
10], Alzheimer’s disease [
11], and Huntington’s disease [
12]. SMaRT uses the cell’s own splicing machinery to recombine a rationally designed RNA
trans-splicing molecule (RTM) with the target pre-mRNA of interest, generating a new chimeric gene product. This is in contrast to group I intron-based (also referred to as ribozyme-based)
trans-splicing RNAs that act without the help of the endogenous spliceosomal machinery. Group I intron ribozymes are mainly found in protists, bacteria, or bacteriophages [
13], and similar to SMaRT they can be used for RNA repair via the replacement of the 3′ part of the mRNA with the 3′ exonic sequence provided by the ribozymes [
14,
15] or by splicing a suicide gene into the target RNA to kill tumor cells [
16,
17]. In general, however, the efficiency of both these RNA technologies needs to be significantly improved for a prospective clinical application.
The efficiency and specificity of SMaRT mainly relies on the RTM’s composition and binding affinity to the respective target locus of interest. The RTM carries the wild-type coding region to be inserted into the target pre-mRNA, splicing elements for efficient splicing, and a binding domain (BD) that specifically binds an intron of the target pre-mRNA and is critical for initiating RNA recombination. SMaRT is suitable for the specific exchange of 5′-, 3′-, or internal parts of a given transcript, referred to as 5′, 3′, or double RNA
trans-splicing [
18,
19]. Thereby, any pathogenic mutation within the respective gene region can be replaced with a single type of RTM. For EB, a severe skin blistering disease [
20], different
trans-splicing strategies have been described to correct EB-associated mutations in
COL7A1 [
7,
19],
COL17A1 [
9],
KRT14 [
6], and
PLEC [
8]. Although important progress in RTM design has been achieved, particularly in the selection of the BD sequence [
21,
22], the low efficiency of the technology remains an obstacle to its broad in vivo application. However, our first ex vivo and in vivo studies in recessive dystrophic epidermolysis bullosa (RDEB) mouse models showed that a low RNA reprogramming efficiency can be sufficient to restore protein function [
7,
23]. Via transduction of a
COL7A1-specific RTM into RDEB keratinocytes using a lentiviral SIN vector, we achieved an RNA repair efficiency of ~2% in an isolated single cell clone, which resulted in restoration of normal levels of type VII collagen [
7,
24]. As type VII collagen has a reported half-life of two months, at least in murine skin [
25], it can be assumed that a low percentage of
COL7A1 repair at the RNA level may be sufficient to restore a wild-type phenotype in DEB (dystrophic EB) skin.
In general, a higher
trans-splicing efficiency is probably needed for dominant negative mutations which also contribute to the etiology of other EB subtypes, as well as other genodermatoses such as epidermolytic ichthyosis [
26,
27]. In this regard, Wally et al. demonstrated that RNA
trans-splicing-mediated repair of a dominant negative
KRT14 mutation in EB simplex (EBS) keratinocytes resulted only in a partial reversion of the phenotype [
6]. Thus, a higher reprogramming efficiency is still required for the treatment of dominantly inherited diseases and, certainly for any
trans-splicing-based in vivo RNA repair strategy, which can be achieved for instance via splicing modulation using short antisense molecules [
4,
28].
ASOs can be designed to bind mutation-harboring exons during the pre-mRNA splicing process in order to promote their skipping [
1,
29,
30]. In the optimal case, the resulting mRNA, while shorter, is translated into a fully functional protein. The feasibility of this strategy was recently shown in a mouse model, wherein skipping of the mutation-harboring exon 80 resulted in restoration of
COL7A1 function [
31]. In this study, however, the outcome of ASO application we explored was not the skipping of a disease-associated exon, but rather the enhancement of
trans-splicing efficiency due to inhibition of competitive
cis-splicing events within the target region. Studies in spinal muscular atrophy showed that co-expression of a therapeutic RTM engineered for
SMN2 correction, together with an antisense RNA molecule that specifically blocked the competing
cis-splice-site, led to an improvement in
trans-splicing efficiencies both in vitro and in vivo [
4,
32]. Furthermore, our group recently demonstrated the possibility of enhancing the
trans-splicing efficiency of a given RTM via antisense RNAs (asRNAs) in the dystrophic subtype of epidermolysis bullosa (DEB) using a GFP-split model system [
28].
The aim of this study was to improve the RNA
trans-splicing efficiency of a previously described therapeutic
KRT14-RTM [
6] by the co-administration of randomly generated asRNAs expressed via a plasmid and rationally designed short ASOs, applied as oligonucleotides, that bind potential splicing elements within the
KRT14 target region. Dominant mutations in
KRT14 are responsible for the generalized severe form of epidermolysis bullosa simplex (EBS-gen sev). Incorporation of the mutant K14 into the intermediate filament (IF) network compromises its integrity, leading to its collapse into protein aggregates under conditions of stress [
6,
33,
34]. Using our established fluorescence-based screening system [
21,
22] we investigated the impact of
KRT14-specific antisense molecules on the
trans-splicing efficiency of our
KRT14-RTM and on the general
cis-splicing pattern within the
KRT14 target region. We achieved increased RNA repair levels after the addition of specific antisense molecules of varying lengths in our screening system. Thus, we were able to identify potential splicing modulators that may further improve the RNA
trans-splicing technology to a level sufficient to support its implementation in a clinical setting.
3. Discussion
In recent years, RNA
trans-splicing has been predominantly applied in RNA repair studies for various genetic diseases or for the RTM-mediated delivery of “suicide genes” into tumour cells [
35,
36]. Reprogramming at the pre-mRNA level represents an elegant approach for the editing of aberrant RNA molecules that maintains endogenous regulation of the target transcript, making the technology potentially safe and well suited for the therapy of certain genetic conditions. These include disorders caused by mutations affecting large genes or dominantly inherited mutations, difficult to correct via conventional full-length cDNA replacement strategies. In this respect, we have already demonstrated promising results with a
COL7A1-targeting RTM capable of correcting ~40% of all RDEB-associated mutations encoded within the ~3300-nucleotide long target pre-mRNA [
7]. In the case of dominant negative mutations in EBS-gen sev, Cao et al. showed with an inducible EBS mtK14 mouse model, that dominance of the mutation is dependent on the ratio of wild-type to mutated alleles. A reduction in expression of the mutated allele by ~50% appeared sufficient to overcome the EBS phenotype [
37]. In this respect, RNA
trans-splicing-mediated correction of the mutant RNA is expected to not only increase the amount of wild-type
KRT14 transcripts, but also simultaneously decrease the level of mutant transcripts by competing with the normal
cis-splicing events. Indeed, we could achieve partial reversion of the phenotype of EBS keratinocytes associated with a dominant-negative mutation in
KRT14 with an appropriate
KRT14-targeting RTM [
6].
With the recent advances in the identification and selection of the BD for the RTM [
18,
21,
22], the main obstacle that remains is the low efficiency of the technique, which limits its potential for future in vivo or clinical application. Therefore, we have developed a fluorescence-based screening system aimed at identifying factors with the potential to increase the
trans-splicing efficiency, such as antisense RNA molecules [
28]. In the course of this study, we could identify antisense RNA molecules capable of enhancing the
trans-splicing efficiency induced by the
KRT14-targeting RTM by up to 20-fold. The system proved practical for the screening of an unbiased library of randomly generated sequences of varying lengths and binding positions along the target pre-mRNA, as well as short, site-specific ASOs rationally designed to block splice sites and splicing enhancer elements. Our results revealed both antisense strategies to be attractive for future RNA repair studies in patient cells, but also highlight several points for consideration.
In the described screening system, the
KRT14-scRTM induced a
trans-splicing reaction between the 5′ GFP coding sequence on the RTM and the 3′ GFP sequence located immediately after intron 7 on the target pre-mRNA molecule. Notably, some of the best-acting antisense molecules identified (asRNA34 and ASO9) were not expected to directly inhibit the competing
cis-splicing of intron 7. ASO9 was expected to bind to the 5′ region of exon 7, whereas asRNA34 bound a 324-nt region spanning all of exon 5, intron 5, and half of exon 6. As such, asRNA34 was rather expected to interfere with
cis-splicing of intron 5, whereas ASO9 may or may not interfere with
cis-splicing of intron 6. The mechanisms by which these molecules enhanced
trans-splicing efficiency are not yet completely understood. However, blocking potential splicing enhancer and silencer sequences located within the
KRT14-scMG is likely to impact both types of splicing reactions (
cis and
trans). asRNA34, in particular, would block many predicted enhancer and silencer sequences, and thereby affect the general spliceosome-mediated splicing efficiency (
Figure S3) in an unpredictable manner. In this respect, increasing amounts of asRNA34 added to the cells led to a significant enhancement of both
cis- and
trans-splicing reactions. Besides blocking potential splicing motifs within the MG sequence, asRNA34 may also exert a general stabilizing effect on the target pre-mRNA that promotes both
cis- and
trans-splicing reactions. Thus, while we achieved increased
trans-splicing rates by the addition of asRNA34, the concurrent upregulation of
cis-splicing in the target pre-mRNA would be expected to negate any of its associated benefits, as no alteration in the ratios of wild-type (corrected) transcripts to mutant transcripts can be expected.
In contrast to long asRNA sequences, the mechanisms of action by which short ASOs exert their effects are more clearly defined, as they are designed to inhibit splicing either by directly blocking the splice donor/acceptor sites located within the exon–intron junctions (as in the case of ASO7), or by blocking potential splicing enhancer motifs adjacent to the intron–exon junction (ASO9). Both ASO7 and ASO9 resulted in enhanced trans-splicing of the RTM without any detectable increase in cis-splicing events within the target-pre-mRNA, suggesting that these shorter ASOs are better suited for our purposes. The most auspicious ASO7, targeting the exon 7–intron 7 boundary, additionally exhibited an inhibitory effect on the competitive cis-splicing of intron 7 as expected. The observed 2-fold inhibition in cis-splicing of intron 7, combined with the 4-fold increase in trans-splicing will not only reduce the level of mutated transcripts in the patient cells but also shift the ratio of wild-type to mutant transcripts further towards wild-type. This is expected to decrease the dominant negative effect of the disease-causing mutation to a level that may be sufficient to fully restore the wild-type phenotype in patient cells.
This study describes a straightforward and practical protocol that can be adapted to the screening of any type of splicing modulator, including pharmacological compounds. Our results also demonstrate that there are different mechanisms by which antisense technologies can impact splicing events and further underscore the need to comprehensively evaluate the effects of such molecules on different processes and outcomes including, for example, RNA stability. To this end, future studies will not only focus on evaluation of this therapeutic strategy in EBS patient-derived keratinocytes, but also on the investigation of mechanisms of action and precise quantification of corrected and mutant transcripts using next-generation sequencing platforms.
4. Materials and Methods
4.1. Construction of asRNA Library and Antisense Oligonucleotides
An asRNA library was constructed by the fragmentation of the PCR-amplified
KRT14 target region spanning from exon 5 to intron 7 by sonication according to an already established protocol [
21,
28]. PCR amplification was performed using genomic DNA from a healthy donor and the GoTaq DNA polymerase (Promega, Madison, MI, USA). The resulting fragments were treated with a DNA Terminator End Repair Kit (Lucigen, Middleton, WI, USA), cloned into a pcDNA 4.0 plasmid (Invitrogen, Carlsbad, CA, USA) using the restriction site for EcoRV, and analyzed for their correct orientation (complementary to the target region) by sequence analysis. PCR products were digested with the corresponding restriction enzymes for 1 h at 37 °C, purified with a Illustra GFX PCR DNA and Gel Band Purification Kit (GE Healthcare, Chalfont St. Giles, UK) and ligated with T4 DNA ligase (Thermo Fisher Scientific, Waltham, MA, USA), all according to the manufacturer’s protocols. The ligation was then transformed into chemically competent SoloPack
® Gold Competent Cells (Agilent Technologies, Santa Clara, CA, USA) and plasmid preparations were carried out using a Plasmid Mini Prep Kit (Sigma-Aldrich, Taufkirchen, Germany), according to the manufacturer’s protocol. Sequence analysis of all plasmids and PCR products was performed using a 3500 ABI automated sequence analyzer and ABI PRISM dye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA, USA).
4.2. Screening Constructs
Screening constructs used for detection of
cis-splicing and correct
trans-splicing were designed and cloned according to the protocol of Bauer et al. [
21]. Concisely said, the
KRT14-scRTM carries a 5′ split-GFP part consisting of the first nucleotides (nt: 1–336) from a full length acGFP (pacGFP vector, Clontech, Mountain View, CA, USA) and a BD sequence that specifically targets Intron 7 of
KRT14 cloned into the pcDNA3.1D/V5-HIS (Invitrogen, Carlsbad, CA, USA) backbone. The
KRT14-scMG spanning exon 5 to intron 7 was generated using a
KRT14 exon 5 forward primer (5′-GATCAAGCTTCACCACAGAGGAGCTGAACCGCGAGGTGGC-3′) and a
KRT14 intron 7 reverse primer (5′-GATCGGATCCGGGGAAGAGGTGGGAAGAGGACGTTACC-3′) for amplification via GoTaq DNA polymerase (Promega, Madison, MI, USA). Afterwards, the amplified PCR product was cloned into the screening vector backbone downstream to the CMV-promoter (pcDNA3.1, Invitrogen, Carlsbad, CA, USA) using HindIII and BamHI restriction sites [
21,
28]. The
KRT14-scMG also carries the 3′ split-GFP portion consisting of the rearward part of acGFP (nt: 337–720) to enable a reconstitution of the full length acGFP upon accurate
trans-splicing.
4.3. Cell Culture and Plasmid Transfection
For triple-transfection studies, the human embryonic kidney cell line HEK293 (Stratagene, La Jolla, CA, USA) was used and grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 100 U/mL penicillin–streptomycin (Biochrom, Berlin, Germany) at 37 °C and 5% CO2 in a humidified incubator. Passaging of the cells was carried out with Trypsin (0.05%)-EDTA (0.02%) (Biochrom, Berlin, Germany) treatment followed by Trypsin-EDTA inactivation with Solution A (7.25 g HEPES; 1.8 g glucose; 0.22 g NaCl; 0.27 g Na2HPO4; 1 mL phenol red; 900 mL H2O; pH 7.4) supplemented with 10% FBS. Cells were centrifuged at 350× g for 5 min and seeded in fresh tissue culture 6-well plates or 60 mm plates.
Transient transfections of screening plasmids were performed the next day using the jetPEI reagent (Polyplus-transfection SA, Illkirch, France) according to the manufacturer’s protocol. The transfected DNA amounts varied in the different experiments:
Section 2.2: 0.5 µg
KRT14-MG + 0.5 µg RTM + 3 µg asRNA were transfected into HEK293 cells cultivated in 6-well plates;
Section 2.3: 0.5 µg
KRT14-MG + 0.5 µg RTM + 70 nM ASO were transfected into HEK293 cells cultivated in 6-well plates.
4.4. Flow Cytometric Analysis
The amount of trans-splicing events, manifested in the expression of GFP, was measured 48–72 h after transfection using a Beckman Coulter FC-500 FACS analyzer (Beckman Coulter, Vienna, Austria). The transfected HEK293 cells were washed once with phosphate buffer saline (Dulbecco’s PBS) and detached from the cell culture plates using Trypsin-EDTA. Trypsin-EDTA was inactivated with Solution A (7.25 g HEPES; 1.8 g glucose; 0.22 g NaCl; 0.27 g NaHPO4; 1 mL phenol red; 900 mL H2O; pH 7.4) supplemented with 10% FBS. After centrifugation for 5 min at 350× g, 10,000 cells were analyzed according to their GFP expression and geometric mean with the Kaluza 1.3 software (Beckman Coulter, Brea, CA, USA).
4.5. RNA Isolation and cDNA Synthesis
HEK293 cells were harvested 48–72 h post transfection and RNA was isolated using the Innuprep RNA Mini Kit (Analytik Jena, Jena, Germany) according to the manufacturer’s protocol. Afterwards cDNA synthesis was performed with 3 µg RNA using the iScript™ cDNA Synthesis Kit (BioRad, Hercules, CA, USA) according to the manufacturer’s protocol.
4.6. sqRT-PCR Analysis of Cis- and Trans-Splicing Levels
sqRT-PCR was performed to detect fused GFP transcripts in treated HEK293 cells. For sqRT-PCR analysis, a 5′ GFP-specific forward primer (5′-GCTGACCCTGAAGTTCATCTG-3′), a 3′ GFP-specific reverse primer (5′-CGCCGATGGGGGTATTCTGCTGG-3′), cDNA of treated cells and GoTaq® qPCR Master Mix (Promega, Madison, WI, USA) was used. The PCR was performed using a Bio-Rad CFX™ system (BioRad, Hercules, CA, USA) with the following conditions: 95 °C for 2 min, and 50 cycles of 20 s at 95 °C, 20 s at 62 °C, and 25 s at 72 °C. Experiments were carried out in duplicates and repeated two times. Correct PCR products were verified by direct sequence analysis.
To analyze differences in cis-splicing events after transfection of HEK293 cells exon–exon junction-specific forward (Exon5/6: 5′-AGCTCAGCATGAAAGCATCCCT-3′, Exon6/7: 5′-AGGACGCCCACCTCTCCTCC-3′) and reverse (Exon6/7 5′-GGAGGAGAGGTGGGCGTCCT-3′, GFP: 5′-GGTCAGCTCGATGCGATTCACC-3′) primers, cDNA of treated cells and GoTaq® qPCR Master Mix (Promega, Madison, WI, USA) was used. The PCR was performed using a Bio-Rad CFX™ system with the following conditions: 95 °C for 2 min, and 50 cycles of 20 s at 95 °C, 20 s at 64 °C, and 25 s at 72 °C. Experiments were carried out in duplicates and repeated two times. Correct PCR products were verified by direct sequence analysis.
4.7. Western Blot Analysis
Enhancement of asRNA/ASO-induced trans-splicing manifests in an increase in full-length GFP expression, which was quantified at the protein level via Western blot analysis 48–96 h post transfection. HEK293 cells were washed with PBS, harvested, and resuspended in radioimmunoprecipitation assay buffer RIPA (Santa Cruz Biotechnology, Heidelberg, Germany). Samples were denatured for 5 min at 95 °C in 4× SDS loading buffer (0.25 M Tris-HCl; 8% SDS; 30% glycerol; 0.02% bromphenol blue; 0.3 M β-mercaptoethanol; pH 6.8). SDS-gel-electrophoresis was performed at 140 V for a maximum of 2 h using a NuPAGE 12% Bis-Tris gel (Invitrogen, Carlsbad, CA, USA) and NuPAGE MOPS running buffer. Proteins were electro-blotted onto a nitrocellulose membrane (Amersham Hybond-ECL, Amersham, Buckinghamshire, UK) at 0.25 A for 75 min. The membrane was blocked with Western Blocking Reagent (Roche, Basel, Switzerland) in Tris buffered saline (TBS) + 0.2% Tween (TBS-T) for 1 h at RT and incubated with a primary anti-GFP rabbit IgG antibody (MBL-598, MBL International, Woburn, MA, USA) in a dilution of 1:750 in TBS-T at 6 °C overnight. After three washing steps (3 × 10 min) with TBS-T the membrane was incubated with a secondary HRP Envision+ labelled anti rabbit antibody (1:100 in TBS-T, Agilent Technologies, Santa Clara, CA, USA) and incubated for 90 min at room temperature. Finally, GFP expression was visualized using the Luminata™ Forte Western HRP Substrate (Millipore, Burlington, MA, USA). α-actinin (H-2) mouse monoclonal IgG antibody (1:1000 in TBS-T; Santa Cruz. Dallas, TX, USA) and a secondary HRP Envision+ labelled anti mouse antibody (1:100 in TBS-T Tween, Agilent Technologies, Santa Clara, CA, USA) were used to detect α-actinin which served as protein loading control. The relative quantification of GFP expression was measured and calculated using the Image Lab 5.2.1 (Bio-Rad) software.
4.8. Statistical Analysis of SqRT-PCR Data
For statistical analysis, an unpaired Student’s t-test (two-tailed) was performed with the GraphPad Prism 5.03 software (GraphPad Software, San Diego, CA, USA) to prove statistical significance between the values of controls (pcDNA 4.0 & scrASO) and samples (asRNA34, ASO7, ASO9). A sample size of at least n = 4 was analyzed for mean ± SEM to provide a p-value.
4.9. Analysis of Splicing Enhancer and Silencer