An antisense oligonucleotide efficiently suppresses splicing of an alternative exon in vascular smooth muscle in vivo

Abstract

Targeting alternative exons for therapeutic gain has been achieved in a few instances and potentially could be applied more broadly. The myosin phosphatase (MP) enzyme is a critical hub upon which signals converge to regulate vessel tone. Alternative exon 24 of myosin phosphatase regulatory subunit (Mypt1 E24) is an ideal target as toggling between the two isoforms sets smooth muscle sensitivity to vasodilators such as nitric oxide (NO). This study aimed to develop a gene-based therapy to suppress splicing of Mypt1 E24 thereby switching MP enzyme to the NO-responsive isoform. CRISPR/Cas9 constructs were effective at editing of Mypt1 E24 in vitro; however, targeting of vascular smooth muscle in vivo with AAV9 was inefficient. In contrast, an octo-guanidine conjugated antisense oligonucleotide targeting the 5′ splice site of Mypt1 E24 was highly efficient in vivo. It reduced the percent splicing inclusion of Mypt1 E24 from 80% to 10% in mesenteric arteries. The maximal and half-maximal effects occurred at 12.5 and 6.25 mg/kg, respectively. The effect persisted for at least 1 mo without toxicity. This highly effective splice-blocking antisense oligonucleotide could be developed as a novel therapy to reverse vascular dysfunction common to diseases such as hypertension and heart failure.

NEW & NOTEWORTHY Alternative exon usage is a major driver of phenotypic diversity in all cell types including smooth muscle. However, the functional significance of most of the hundreds of thousands of alternative exons has not been defined, nor in most cases even tested. If their importance to vascular function were known these alternative exons could represent novel therapeutic targets. Here, we present injection of Vivo-morpholino splice-blocking antisense oligonucleotides as a simple, efficient, and cost-effective method for suppression of alternative exon usage in vascular smooth muscle in vivo.

INTRODUCTION

Alternative exon usage (AEU) is present in most genes (1) and is a major driver of phenotypic diversity in all cell types, including smooth muscle (2–4). The thousands of alternative exons provide an attractive target for gene therapies that, by suppressing or activating the alternative exon, could shift the expression of naturally occurring protein isoforms for therapeutic gain. This approach is used in the treatment of rare genetic diseases such as Spino-Muscular Atrophy (5). A variation of this approach is used to induce skipping of a mutated constitutive exon in Duchenne’s Muscular Dystrophy, thereby restoring the reading frame, as described [reviewed in Crooke et al. (6)].

Myosin phosphatase is a key nexus where signals converge to regulate vascular smooth muscle tone, thus influencing blood flow and pressure. A 31-nt alternative exon (E24) within the myosin phosphatase regulatory subunit (Mypt1 = PPP1R12a), by variably coding for a COOH-terminal leucine zipper motif (LZ), determines the enzyme’s ability to bind cGMP-dependent protein kinase 1α (cGK1α) and its activation by nitric oxide vasodilator signaling [reviewed in Brozovich et al. (7) and Fisher (2, 8)] (Fig. 1). Previous studies using germline (Cre-Lox) modified mice have shown that suppression of splicing of Mypt1 E24 sensitizes vascular smooth muscle to NO/cGMP vasodilator signaling and has a favorable effect on blood pressure (10, 11). This provides a foundation for testing novel approaches targeting E24 in vascular smooth muscle that could be translated into new therapies for humans with hypertension and heart failure.

Figure 1.

Alternative exon 24 of myosin phosphatase regulatory subunit (Mypt1) sets vasodilator sensitivity. Isoforms of Mypt1 are generated by the alternative splicing of the 31-nt exon 24 (9). Skipping of E24 generates an mRNA that codes for the COOH-terminal LZ motif (LZ⁺). Inclusion of E24 shifts the reading frame and codes for a distinct COOH-terminus lacking the LZ motif, as well as generating a premature termination codon (LZ⁻). The Mypt1 LZ motif is required for LZ-mediated heterodimerization with protein kinase-GIa and activation of the enzyme by vasodilator signals such as nitric oxide and atrial natriuretic peptides that use cGMP as a 2nd messenger. The goal of this study was to develop gene therapy approaches for suppression of splicing of Mypt1 E24 thereby forcing a switch to the E24⁻/LZ⁺ isoform of Mypt1 (arrow). NO, nitric oxide; ANP, atrial natriuretic peptide; PKG1, cGMP-dependent protein kinase 1; MP, myosin phosphatase; PP1cat; catalytic subunit of type-1 protein phosphatase; LZ, leucine zipper.

The Achilles’ heel of gene therapies has been inefficiency of delivery to the target cell type in vivo. While vectors and approaches have been tested and optimized for delivery to cardiac and skeletal muscle, liver, and other cell types (12, 13), little attention has been paid to smooth muscle cells (SMCs) dispersed throughout the different organ systems of the body. The few reports that have included SMCs in systematic and quantitative studies of cell tropism of gene delivery have generally reported low efficiency in SMCs [as seen in Westhaus et al. (14)]. A strategy targeting Mypt1 E24 to improve vascular function would have to target the SMCs of the microcirculation, of which there has been little study. In this study, we tested two approaches targeting Mypt1 E24 for potential therapeutic applications: 1) clustered regularly interspaced palindromic repeats (CRISPR)/Cas9 editing in vitro followed by viral gene delivery in vivo and 2) antisense oligonucleotide (ASO) injections in vivo.

METHODS

Animals

The animal protocols used in this study were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Maryland School of Medicine and adhere to National Institutes of Health guidelines. C57BL6J mice were purchased from the Jackson Laboratory (JAX Stock No. 000664). Mypt1 exon 24 floxed (E24^fl/fl) mice were previously described (11) and were maintained in a C57BL/6J genetic background. Mice were housed at the University of Maryland animal facilities at 22°C room temperature with 45% humidity under 12-h:12-h light/dark cycles. Mice were euthanized via CO₂ inhalation followed by exsanguination. Tissues were quickly dissected and either placed into RNAlater (Invitrogen) for later purification of RNA, frozen on dry ice for purification of DNA, or frozen by immersion in 2-isomethylbutane at −80°C and stored at −80°C for protein analysis, or placed in ice-cold PBS for imaging of fresh intact tissues.

CRISPR/Cas9

The design of small guide RNAs (sgRNAs) was performed on the Benchling.com platform using the Design CRISPR Guides function. The target region on the human PPP1R12A gene was set from 155,881 to 156,420 bp (gene sequence ENSG00000058272), allowing the screening of a 540-bp long region flanking PPP1R12A E24 for potential sgRNA targets. Guides were subsequently evaluated for use in mice by aligning the sequence of each sgRNA with the PPP1R12A murine sequences using Molecular Evolutionary Genetics Analysis (MEGA) 7.0.21 software. Guides with predicted efficiency and specificity scores exceeding 10 and 60%, respectively, based on the human genome, and additionally showing >65% sequence conservation between the human and the murine sequences, and immediately upstream of an NNGRRT staphylococcal protospacer adjacent motif (PAM) sequence, were selected for in vitro evaluation.

The px601-AAV-CMV::NLS-SaCas9-3xHA-bGHpA;U6::BsaI-sgRNA vector (hereafter referred to as “px601”) (Addgene Plasmid No. 61591) constructed on a pAAV backbone and harboring a cytomegalovirus (CMV) promoter-driven Staphylococcus aureus (Sa)Cas9 and a customizable U6-driven sgRNA scaffold (15), was used as the vector for sgRNA delivery. Oligonucleotides containing sgRNA sequences were ordered from Integrated DNA Technologies with 5′-CACC and 5′-AAAC overhangs to the sense and antisense oligonucleotides, respectively. A guanine (G) was added to the 5′-end of the sgRNA sequence, if one was not present, to enhance transcription from the U6 promoter. sgRNAs were annealed and cloned into px601 according to the protocol of the Zhang Laboratory (available online). The sgRNA-px601 and control empty vector constructs were prepared by bacterial transformation and culture and purified using ChargeSwitch-Pro Plasmid Miniprep kit (Thermo Fisher Scientific) according to the manufacturer’s protocols. Plasmids were sequenced to confirm presence of the sgRNA.

Transfection Experiments and Analysis of Editing of Mypt1 E24

HEK293T cells were seeded at a density of 10⁵ cells/well in 24-well plates. After 24 h at ∼80% density, cells were transfected in duplicate with 1 µg of plasmid using Lipofectamine 3000 Reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. Cells were harvested 72 h later, and genomic DNA was purified using PureLink Genomic DNA Mini Kit (Thermo Fisher Scientific) per the manufacturer’s protocol.

Adeno-Associated Virus Serotype 9 Constructs

The px601-sgRNA plasmids were expanded by bacterial culture and purified using the EndoFree Plasmid Maxi Kit (Qiagen) according to the manufacturer’s instructions. A minimum of 150 mg of purified DNA per plasmid was shipped on dry ice to the Boston Children’s Hospital viral core for custom adeno-associated virus serotype 9 (AAV9) production and purification. The control AAV9-Green Fluorescent Protein (GFP) vector was also purchased from the Boston Children’s Hospital viral core to assess AAV9 transduction efficiency in target tissues. A different AAV9 construct, pAAV.CMV.HI.eGFP-Cre.WPRE.SV40 (hereafter referred to as AAV9GFP-Cre), in which the CMV promoter and enhancer drives expression of enhanced green fluorescent protein (EGFP) and Cre recombinase, was obtained from Addgene (Plasmid No. 105545).

Animal Studies

Adult male and female C57B/6J mice, 12 wk of age, were used for AAV9-sgRNA injections. Each animal received a single injection of 10¹¹ gc of AAV9-sgRNA, AAV9-GFP, or vehicle (saline) via the tail vein. Animals were euthanized 3–6 wk post-injection. Female E24^fl/fl mice were injected intraperitoneally with AAV9GFP-Cre at a dose of 2.67 × 10¹³ gc/kg or empty vector at postnatal days 8 to 12 (PND8-12) and assayed 2–3 wk post-injection.

Antisense Oligonucleotides

A 25-mer antisense oligonucleotide (ASO) hybridizing to the 5′ splice site of Mypt1 E24 was synthesized by GeneTools (Fig. 5A). These ASOs, described as Vivo Morpholinos, are conjugated to octoguanidine and have a phosphorodiamidate backbone to improve cell uptake and stability, respectively (16). A validated inactive control ASO with the same chemistry was also supplied by GeneTools. ASOs were received as a powder and dissolved in water as a stock at a concentration of 0.5 mM and stored at room temperature. Before injection, the ASOs were diluted in saline to give the appropriate dose and injected intraperitoneally. Control mice received either the control ASO or vehicle. For dose-response assays, adult male and female mice received three doses every other day of ASO at 0.2, 1.0, 6.25, or 12.5 mg/kg and were euthanized the next day. For time-course analysis, male and female mice were treated with 12.5 mg/kg ASO every other day for three doses and euthanized on days 1, 4, 7, 14, or 28 after the last dose. Small mesenteric arteries were collected for analysis of mRNA and protein.

Figure 5.

An antisense oligonucleotide targeting the 5′ splice site causes effective and long-lasting suppression of Mypt1 E24 splicing. A: sequence of the ASO and its target sequence at the Mypt1 5′ splice site (ss) are shown. SBASOE24, or scrambled ASO or vehicle (saline) were injected intraperitoneally every other day times three doses at the doses shown. B: effect on splicing of Mypt1 E24 in mesenteric arteries was measured by RT-PCR with primers flanking the alternative exon followed by separation of PCR products by 2.5% agarose gel electrophoresis. Percent splicing inclusion was quantified by dividing the E24⁺ band by the total of the E24⁺ and E24⁻ bands. Topmost band is artifactual (ns) and was excluded from analysis. A dose of 12.5 mg/kg nearly completely suppressed splicing of Mypt1 E24 with an EC₅₀ of 6.25 mg/kg. Suppression of Mypt1 E24 splicing in mesenteric arteries was maintained for 28 days after 3 IP injections every other day of SBASOE24 at 12.5 mg/kg. C: mesenteric arteries from ASO-treated mice as described above were homogenized and 12 μg of protein assayed by Western blot as described in methods. A single membrane was probed with a rabbit polyclonal antibody specific for Mypt1LZ⁻ followed by a mouse monoclonal antibody recognizing all Mypt1 isoforms followed by IRdye-conjugated secondary antibodies (800 CW and 680 RD, respectively). Mypt1 LZ⁻ and total Mypt1 signals on each membrane were quantified with a Licor Odyssey gel scanner to provide on-blot normalization of the LZ⁻ signal, which is reported as Mypt1 LZ⁻/total Mypt1 normalized to the control sample. Representative blot image is shown. Note that the reduced signal for total Mypt1 in this blot at days 1 and 28 is due to variation in loading as it is not present in other samples on other blots. Low protein yield from these small arteries precluded rerunning of gels to adjust loading amounts. Merged image gives a visual representation of LZ⁻ vs. total Mypt1 expression, with the switch from yellow to orange indicating a reduction in LZ⁻:total Mypt1. A reduction of the Mypt1 LZ⁻:total Mypt1 ratio is present at day 1 and maintained out to 28 days. All data were analyzed by ANOVA, followed by Dunnett`s correction for multiple comparisons; n = 3–5/group, except 0.2 mg/kg dose ASO (n = 2); *P < 0.05 vs. control. SBASO, splice-blocking antisense oligonucleotide; ss, splice site; RT, reverse transcriptase.

DNA Analysis

Genomic DNA was purified from tissues using the PureLink kit (Invitrogen). Deletion of Mypt1 E24 and flanking intronic sequences was detected using PCR with primers 5′- ATGTTTAGGCATGCCGATGT-3′ (intronic region, 155,894 to 155,913 bp on PPP1R12A ENSG00000058272) and 5′- GCTTTGACTTTCTGGGAAGATG-3′ (intronic region, 156,626 to 156,647 bp on PPP1R12A ENSG00000058272) under standard cycling conditions. PCR products were resolved by 2% agarose gel electrophoresis. This was complemented by the Surveyor nuclease assay using the Surveyor Mutation Detection Kit (IDT) as per the manufacturer’s protocol. The Surveyor Mutation Detection Kit uses an endonuclease that cleaves DNA with high specificity at sites of mismatches and other distortions, allowing the detection of mutated DNA. Purified genomic DNA and PCR bands separated and purified from gel electrophoresis were sequenced by the Sanger technique to confirm CRISPR/Cas9-induced mutations.

RNA Analysis

RNA analysis was performed as previously described (11, 17). In brief, dissected tissues were lysed by homogenization, and total RNA was purified with RNeasy or PureLink columns (Invitrogen) with on-column DNase treatment as per the manufacturer’s instructions. RNA yield was quantified by optical absorbance (NanoDrop). RNA was reverse transcribed with Superscript IV reverse transcriptase enzyme and oligo dT primers. Mypt1 E24 splice variants were amplified by conventional PCR in a single PCR using the following primers: 5′- TGCAGTTGGAAAAGGCTACC-3′ (forward) and 5′- TCAAGGCTCCATTTTCATCC-3′ (reverse). Smtn 3′ alternative exon E20 splice variants were amplified in a separate reaction using primers 5′- TTCTTCCCTGAGGCTTTTGA-3′ and 5′- ATCATGTCCTCCACCTCCAC-3′. Mypt1 Exon24 and SMTN E20 splice variant PCR products were separated with 2.5% agarose gel electrophoresis, visualized, and quantified with SYBR Safe staining and a Li-Cor Odyssey digital imager, and data were reported as %Mypt1 E24 and %Smtn E20 splicing inclusion (PSI). Smooth muscle myosin heavy chain E6 splice variants were measured by qPCR with predesigned Taqman probes as previously described (17) and also reported as PSI. As a negative control, RT enzyme was omitted from the reverse-transcription reaction mix.

Protein Analysis by Western Blot Analysis

Protein analysis by Western blot was performed as previously described (18). In brief, tissues were homogenized using a pellet pestle motor in a 1:10 volume of 10× RIPA buffer (EMD 20–188) plus 1% protease inhibitor cocktail and centrifuged at 14,000 rpm for 5 min. The pellets were resuspended in a lysis buffer containing 125 mM Tris·HCl, 20% sucrose, 10% SDS, and 1% protease inhibitor cocktail, clarified by centrifugation at 14,000 rpm for 5 min, and protein concentration was determined (Pierce BCA Protein Assay Kit, Thermo Fisher Scientific). Lysates were diluted in sample buffer, heated to 95°C for 5 min, loaded onto a 4–15% Tris-glycine gel (Mini-Protean TGX; Bio-Rad), separated at 200 V for 35 min, and transferred to a PVDF or nitrocellulose membrane at 100 V for 1.5 h. The Mypt1 LZ⁻ isoform was detected with a rabbit polyclonal antibody specific for this isoform at 1:8,000, followed by a secondary IRDye conjugated anti-rabbit antibody (800 CW). Membranes were reprobed with a mouse monoclonal antibody raised against rat Mypt1 residues 723–840 (BD Biosciences No. 612165) and detected with an IRDye conjugated secondary antibody (680RD). Signals were detected by scanning on a Li-Cor Odyssey infrared scanner. After background subtraction, LZ⁻ signal was divided by total Mypt1 signal for on-blot normalization and reported as fold change versus control.

Imaging

Freshly dissected tissues in PBS were imaged with a Leica FL3 fluorescent stereomicroscope, and images were captured using a SPOT RT camera and software v3.1 (Diagnostic Instruments, Sterling Heights, MI). Subsequently, tissues were then fixed in optimal cutting temperature (OCT) and processed for cryosectioning. Cryosections were observed using a fluorescent microscope (Leica DM LB), and images were captured with the SPOT RT camera.

Statistics

Data were analyzed by ANOVA followed by multiple comparisons using Dunnett’s method and SigmaPlot version 15.0 software. For serum assays, paired data were analyzed by t test or Mann–Whitney U statistic, comparing control versus SBASO. Differences with a P value < 0.05 were considered significant.

RESULTS

Efficient CRISPR/Cas9 Editing of Mypt1 E24 in Vitro

Thirteen sgRNAs were identified by Benchling for potential use to target human Mypt1 E24 and intronic flanking sequence. Four sgRNAs were eliminated due to low-predicted efficiency scores, and four were eliminated due to mismatches between human and murine DNA sequences or the absence of PAM sequence in the murine sequences. The remaining five sgRNAs were selected for further evaluation, individually and in combinations, the latter to achieve total excision of E24 (Fig. 2A). The sgRNAs were cloned into px601 plasmid, transfected into HEK293 cells, and genomic DNA purified for analysis as described in methods. Edits induced by single sgRNAs were not resolvable by agarose gel electrophoresis (not shown). Edits induced by single sgRNAs resulted in two bands in the surveyor endonuclease assay, while only a single band was observed in the control DNA (Fig. 2B). Sequencing of DNA edited by single sgRNAs confirmed 1–3-bp deletions and insertions (indels) at the site targeted for editing (not shown). Transfection of cells with px601-sgRNA duets 6+ (11, 9 or 12), 5+ (11, 9 or 12) resulted in deletion of the intervening DNA resulting in PCR bands of sizes (bp) 500, 600, 470, 450, 570, and 420, respectively (Fig. 2C). The efficiency of editing ranged from 10% (sgRNAs 6 + 9, 5 + 9) to 40% (sgRNAs 6 + 11, 5 + 11). This represents an underestimate of approximately one-half in the editing efficiency since only ∼50% of the cells in the culture were transfected, as indicated by a fluorescent reporter. Sequencing of excised bands was consistent with CRISPR/Cas9 editing upstream of the PAM sequences as expected (Fig. 2A).

Figure 2.

Targeting alternative exon 24 of myosin phosphatase regulatory subunit (Mypt1) with CRISPR/Cas9. A: sequence of the target region of human Mypt1 E24 (boxed) and flanking intronic sequence (start-end positions 155881–156420). Five sgRNAs (5, 6, 9, 11, 12) selected for evaluation based on criteria described in the text are represented by gray pentagons positioned over their target sequence. Nucleotides are color coded. sgRNAs were cloned into a px601 expression vector and transfected into HEK293 cells. B: successful editing of Mypt1 E24 by single sgRNAs is shown with the Surveyor nuclease assay. The unedited gene product appears on this 2% agarose gel as a band of 754 nt. Successful editing results in 2 bands running below the control band which together add up to the size of the parent band. C: sgRNAs tested in tandem via dual plasmid transfections as shown yielded genomic edits of Mypt1 E24 with sizes of the PCR amplified recombined product as shown. Thus, the size of the deletion is equal to the size of the parent band (∼750 bp) minus the size of the recombined band. sgRNA, small guide RNA; C, control plasmid; MW, molecular weight; bp, base pairs.

Because splicing of Mypt1 E24 is not present in cultured cells, including cultured SMCs, the effect of CRISPR/Cas9 genomic editing on splicing of E24 could not be assessed in vitro. Therefore, plasmids containing the most effective sgRNAs (6 and 11) were used to generate recombinant AAV serotype 9. These AAVs, when injected into mice, did not induce editing of Mypt1 E24 in the target SMCs in any tissue within the mouse (data not shown). Thus, a series of experiments was performed to test recombinant AAV targeting of SMCs and the E24 target.

AAV Targeting of Mypt1 E24 in SMCs in Vivo

AAV9-GFP was intravenously injected into mature male and female mice. Imaging of tissues 3–6 wk after injection showed robust green fluorescence in the liver in sections and in the heart in whole mounts (Fig. 3, A and B). Imaging of the mouse intestine also revealed bright green fluorescence in the fat surrounding the intestine, but not in the intestine itself (Fig. 3C). The fat in the mesentery surrounding mesenteric arteries exhibited GFP positivity (Fig. 3D). With the fat stripped away, the underlying artery appeared negative for GFP fluorescence (Fig. 3E). At the same exposures, background fluorescence was not observed in control tissues from vehicle (saline)-injected mice (Fig. 3F).

Figure 3.

AAV9GFP injection results in high reporter protein expression in liver and heart and mesenteric fat but not intestine nor mesenteric arteries. Mature mice were injected intravenously with AAV9GFP (10¹¹ gc) and tissues imaged fresh in whole mount (B–F) or in section after cryofixation (A). Liver (A), heart (B), and mesenteric fat surrounding the small intestine (C) and mesenteric artery (D) show high GFP fluorescence, while the mesenteric artery after the fat is stripped away shows minimal GFP fluorescence. E: control tissue shows no fluorescence at the same exposure. Scale bars are 500 µm in B and C and 200 µm in A and D–F.

The use of reporter proteins as a surrogate for successful AAV transduction may result in undercounting due to transient reporter protein expression and/or expression of the reporter protein below the threshold of detection (21). To overcome these limitations, AAV9-GFPCre was injected into E24 flox mice. Assay of genomic DNA from tissue lysates for Cre-mediated recombination at the floxed Mypt1 locus serves as a sensitive, quantitative, and permanent marker of viral transduction and transgene expression. The results corroborate those obtained with the fluorescent reporter. Genomic DNA isolated from the liver had the highest level of recombination at ∼80% (Fig. 4, A and B) indicating successful transduction in the great majority of liver cells. The heart exhibited the second highest level of recombination at ∼40%, while the target mesenteric arteries and other smooth muscle tissues including aorta, bladder, and small intestine showed recombination rates of 20% or less. Given that the target of these gene therapies, splicing of Mypt1 E24 into the mature mRNA, only occurs in fully differentiated SMCs (18, 22, 23), it is not surprising that AAV9GFPCre injected into Mypt1 E24 flox mice had no effect on the inclusion of E24 into the mature Mypt1 mRNA in these tissues (Fig. 4C).

Figure 4.

AAV9GFPCre induces recombination at the E24 flox locus with differing tissue efficiencies in vivo AAV9GFPCre was injected at a dose of 2.67 × 10¹³ gc/kg into female Mypt1 E24 flox mice. A: 3 wk later recombination at the floxed locus was determined by a PCR assay of purified genomic DNA. B: percent recombination for each tissue was calculated by dividing the lower molecular weight band by the sum of the 2 bands. Liver had the highest percent recombination followed by the heart. Smooth muscle containing tissues all had recombination of 20% or less; n = 3; data are means ± SD. *P < 0.05, **P < 0.01 by ANOVA followed by Dunnett’s multiple comparison test. C: mRNA isolated from the different tissues was assayed for Mypt1 E24 inclusion in an RT-PCR assay using primers that flanked the alternative exon followed by 2.5% agarose gel separation of PCR products. As expected from the genomic assay E24 is present in mRNA in tissues from AAV9GFPCre-injected mice. ThAo, thoracic aorta; AbAo, abdominal aorta; BLA, bladder; MA, mesenteric artery.

Antisense Oligonucleotides Efficiently Block Splicing of Mypt1 E24

Splice-blocking antisense oligonucleotides (SBASOs) represent an alternative strategy to CRISPR/Cas9 gene editing for shifting isoform expression for therapeutic purposes. A Vivo morpholino designed to hybridize with the 5′ splice site of Mypt1 E24 (Fig. 5A) was synthesized by GeneTools (referred to here as the splice-blocking antisense oligonucleotide against E24, SBASOE24). When injected intraperitoneally at a dose of 12.5 mg/kg every other day for three doses, SBASOE24 nearly completely suppressed splicing of Mypt1 E24, as evidenced by a switch to the E24⁻ splice variant in the mesenteric arteries (Fig. 5B). A dose-response relationship was observed, with an EC₅₀ at 6.25 mg/kg and no effect observed at a dose of 1 mg/kg or lower. The effect of SBASOE24 on splicing of Mypt1 E24 was prolonged. After three injections of 12.5 mg/kg ip every other day times three doses, the effect of SBASOE24 on splicing of Mypt1 E24 in mesenteric arteries persisted for 28 days after the last injection (Fig. 5B). Longer time points have not been examined. Mypt1 protein isoform expression followed suit. By 6 days after the first injection of the SBASO, the ratio of Mypt1LZ⁻ to total Mypt1 was reduced to ∼40% of control values (Fig. 5C). This reduction in Mypt1 LZ⁻ protein was maintained for up to 28 days after the last injection. In limited analysis, the effect of SBASOE24 was specific, as there was no effect on the splicing of smooth muscle myosin heavy chain (Myh11) alternative exon 6 (PSI: control, 88% ± 3; ASO, 85% ± 4, n = 3–4; P > 0.05) nor smoothelin (SMTN) alternative exon 20 (PSI: control, 27% ± 6; ASO, 29% ± 6; n = 3–4, P > 0.05).

There was no evidence of toxicity of the SBASO at the dose used. After 2 wk of injection at 12.5 mg/kg three times per week in mature male and female mice, there was no significant increase in serum liver enzymes indicating liver injury and nor was there an increase in serum creatinine indicating kidney dysfunction (Table 1). Additionally, there was no increase in serum cytokines indicating immune activation, nor any indication of tissue damage or inflammatory infiltrates on histological examination of the liver and kidney (not shown).

Table 1.

Biochemical analysis of serum collected after 2 wk of SBASO 12.5 mg/kg treatment

	ALT, U/L		Creatinine, mg/dL		IL-1β, pg/mL	IL-6, pg/mL		TNF-α, pg/mL
		P value		P value			P value
Female
Control	23.67 ± 9.07		0.33 ± 0.02		(23.7, 30.0, 945.9, ND)	32.29 ± 44.95		(7.3, 9.6, 233.3, ND)
SBASO	33.33 ± 6.51	0.21	0.32 ± 0.13	0.83	(ND, 46.6, 15.9)	19.31 ± 5.93	0.629	(ND, ND, 7.0)
Male
Control	16.00 ± 1.00		0.50 ± 0.18		ND	(1.2, ND, 14.9)		ND
SBASO	19.33 ± 5.86	0.39	0.34 ± 0.03	0.20	ND	(16.2, 9.8, ND)		ND

Values are means + SD; n = 3–4 mice/group. ALT, alanine transaminase; IL, interleukin; TNF, tumor necrosis factor; ND, not detected; SBASO, splice-blocking antisense oligonucleotide. Parenthesis indicates values for each mouse. P > 0.05 for all groups.

DISCUSSION

The recent success of the COVID mRNA vaccine and other gene-based therapies have reignited tremendous enthusiasm in this field (9, 24). A major hurdle, dating to the early days of this field, is still being addressed: the efficiency of transduction of target cell types in vivo. SMCs dispersed throughout different organ systems play a central role in many common chronic human diseases, including hypertension, heart failure, atherosclerosis, asthma, gastrointestinal dysmotility, and premature labor, among others. Yet, studies evaluating the efficacy of gene delivery for gene therapy purposes have predominately focused on delivery to cardiac and skeletal muscle, as well as non-muscle cell types such as liver and brain.

Here we demonstrate using complementary techniques that AAV9 is highly effective for transgene expression in the liver, moderately effective in the heart, and relatively ineffective in mesenteric arteries and other SMC-containing tissues such as the aorta and bladder. This finding is consistent with a prior high-throughput study that tested 30 different AAV serotypes for transduction and transgene expression in vitro and in vivo (14). They observed that, of the eight tissues tested, smooth muscle had the lowest average vector copy number per diploid genome and concluded that smooth muscle was resistant to AAV transduction. In pilot studies, we tested AAV with serotypes 1, 2, and 5 but did not observe significant expression in smooth muscle with these vectors either. Older studies achieved AAV transduction of SMCs in vitro, or in balloon-injured or occluded, locally perfused vessels in vivo or ex vivo (19, 25, 26). However, the applicability of these findings to systemic targeting of AAV to SMCs in vivo, dispersed throughout different organ systems, is limited, particularly in vitro studies given SMC phenotypic modulation in vitro. The older studies are also limited by cruder assays of transgene expression in which efficiency was not well quantified. A study by Zincarelli and coworkers systematically examined the tropism of AAV1-9 after tail vein injection and observed that AAV9 had the broadest tropism and transgene expression (13). As such, AAV9 has become the most widely used AAV serotype for transgene delivery in experimental settings. Many factors besides serotype play a role in successful viral gene therapy, including the promoter driving transgene expression (27). The mechanisms behind poor transduction of and transgene expression in SMCs remain largely undefined and require further study. A variety of sophisticated techniques such as selected evolution are being used to develop novel AAV capsids (28). A very recent report using this method identified a promising variant of AAV9 with tropism for vascular smooth muscle cells and pericytes in vivo (29).

The splice-blocking ASO tested in this study was highly efficient at its target, specifically reducing splicing of Mypt1 E24 to near background levels. Because alternative E24 is only included in Mypt1 mRNA in fully differentiated SMCs [reviewed in Fisher (2, 3)], the suppression of its splicing indicates that the octo-guanidine conjugated ASO was avidly taken up into this cell type. This study focused on the efficacy of the ASO against Mypt1 E24 in small arteries, the target in hypertension and heart failure. The pattern of splicing of Mypt1 E24 is tissue-specific in rodents and humans, with nearly 100% inclusion in phasic smooth muscle such as portal vein, bladder, and intestine, predominance of E24 inclusion in the small arteries, and predominance of E24 skipping in tonic smooth muscle of the large arteries and veins (11, 20, 22, 30–35). Whether this ASO will have similar efficacy against E24 in SMCs in other vessels or organs systems in vivo requires further study. The initial reporting of the efficacy of the octo-guanidine-conjugated ASO, tested by injection into a GFP splicing reporter mouse (36), showed varying degrees of activity in different tissues and was particularly effective in the small intestine, liver, and kidney (blood vessels were not assayed) (16). Subsequently, this chemistry has been successfully used to target a variety of tissues in vivo, mostly in mice but in other model organisms as well (37–39). The tissue tropism of this formulation of ASO requires further study.

The dose response and kinetics of the effect of SBASOE24 are favorable. The dose of 12.5 mg/kg is within the range of dosing of oligonucleotides used therapeutically in humans. The pharmacodynamic experiment demonstrated a prolonged effect after SBASOE24 administration of at least 1 mo. This is consistent with studies in humans, in which, for example, an RNAi oligonucleotide targeting PCSK9 mRNA in the liver is effective at lowering serum cholesterol when injected subcutaneously twice per year (40), and the previously mentioned Spinraza is effective when injected intrathecally four times per year (5). The duration of the effect is a function of the turnover of the oligonucleotide and the turnover of its target, in the case, Mypt1 mRNA and protein. We did not measure the tissue nor serum concentrations of the ASO as this is technically challenging (41). The prolonged effect is consistent with a long tissue dwell time for RNaseH-independent (non-degrading) ASO, as reported by others (reviewed in (6, 42). The rate of turnover of Mypt1 mRNA and protein is not defined, but in general, cytoskeletal proteins tend to have longer half-lives, on the order of weeks in SMCs in vivo (43). That the full effect of ASO on mRNA and protein is observed ∼1 wk after the first injection suggests a turnover on the order of several days. However, we cannot rule out the possibility that the ASO speeds the turnover of Mypt1 mRNA and protein, or that the mRNA and protein isoforms have different rates of turnover. For example, the E24⁺ mRNA could have faster turnover because of the premature termination codon in the E24⁻ variant triggering nonsense-mediated decay. The LZ⁻ protein isoform contains a unique COOH-terminal ubiquitylation motif that could trigger its degradation. Regarding the pharmacodynamics, the ASO was injected intraperitoneally. This represents a third space. Whether the same efficacy and pharmacodynamics will be observed with other routes of administration as are used clinically, for example, subcutaneously, requires further study. While we showed here the specificity of the ASO against Mypt1 E24 by assaying two other SMC-specific alternative exon splice variants, more comprehensive study of the potential off-target effects of the ASO is required.

The near-complete efficacy of SBASOE24 against its target Mypt1 E24 in vivo likely relates to this being a weak alternative exon for which the default is exon skipping. A number of cis- and trans-regulatory factors are required for inclusion of the exon (20, 44) likely making splicing of the exon susceptible to suppression by the SBASO, especially in comparison to the relatively weak effect of SBASOs against constitutively spliced exons. We have shown a similar near-complete switch to the E24⁻ isoform of Mypt1 in various rodent models of disease (34, 45–47), further illustrating the susceptibility of suppression of E24 splicing [reviewed in Fisher (2)]. Switching to the Mypt1 E24⁻/LZ⁺ isoform increases smooth muscle sensitivity to NO/cGMP-mediated relaxation, as the COOH-terminal LZ motif is required for LZ-mediated heterodimerization with cGMP-dependent protein kinase and its activation of myosin phosphatase (11, 22, 48). Deletion of E24 in genetically modified (Cre-Lox) mice suppressed the hypertensive response in the chronic slow pressor ANG-II model of hypertension (10). These data, combined with the efficacy of SBASOE24 in achieving the same switch to Mypt1 E24⁻/LZ⁺, provide a strong foundation for developing SBASOs against this target as novel therapies to lower vascular resistance and blood pressure and improve blood flow in diseases where defective vasodilator signaling and vascular dysfunction play a prominent role such as hypertension and heart failure. However, whether this specific formulation of ASO could serve in this capacity is uncertain. A prior study described acute toxicity at standard Vivo-morpholino dosing delivered to mice by tail vein injection manifest as blood clotting and death (49). We have also observed Vivo-morpholino toxicity at supratherapeutic dosing. Modifications of ASOs to improve delivery and reduce toxicity are an active area of investigation and includes chemical modifications, peptide and antibody conjugates, and others (24, 50). Further study is indicated to identify the best formulation of ASO for therapeutic targeting of mRNAs within smooth muscle cells of the vascular and other organ systems.

GRANTS

This research was supported in part by the University of Maryland Scholars Program (an initiative of the University of Maryland: MPowering the State); National Heart, Lung, and Blood Institute Grants R01-HL066171 and R01-HL142971-01A1 (to S.A.F.); U.S. Department of Veterans Affairs Grant BX004443 (to S.A.F.); and a state of Maryland TEDCO Mil award (to S.A.F).

DISCLOSURES

S.A.F. has Patent No. 11,607,461 and a patent pending related to this work. None of the other authors has any conflicts of interest, financial or otherwise, to disclose.

AUTHOR CONTRIBUTIONS

C.D.d.A., M.M., and S.A.F. conceived and designed research; C.D.d.A., M.M., and N.C. performed experiments; C.D.d.A., M.M., N.C., and S.A.F. analyzed data; C.D.d.A., M.M., N.C., and S.A.F. interpreted results of experiments; C.D.d.A., M.M., and N.C. prepared figures; S.A.F. drafted manuscript; C.D.d.A., M.M., N.C., and S.A.F. edited and revised manuscript; C.D.d.A., M.M., N.C., and S.A.F. approved final version of manuscript.