Gene Therapy and the DMD Heart
The promise of gene therapy is to restore expression of a normal protein in a cell containing a genetically deficient or defective version. In that sense, gene therapy raises the possibility of a cure for DMD, not only for skeletal muscle but also for the heart. Strictly speaking, gene therapy is use of nucleic acid, either DNA or RNA (or both) to alter the function of a cell. A variety of approaches have been tried to carry nucleic acid into a cell, including direct injection of DNA (so-called “naked” DNA), DNA transfected into a cell by compacting it with a glycoprotein/polycation scaffolding, encapsulated in liposomes that fuse with the cell membrane, or most commonly packed in a virus shell. Viruses have evolved to efficiently enter cells and transmit DNA to the nucleus, and many can be “gutted” of many or all of their normal genes to make room for transferring in new genes. The new gene is referred to as a “transgene,” virus-mediated entry into a cell is termed “transduction,” and the virus carrying the transgene is called a “vector.” While many different viruses have been used for transduction, a small member of the parvovirus family, adeno-associated virus (AAV), is most commonly used due to its safety profile (it is not associated with any disease), efficiency of transduction, widespread biodistribution following intravenous delivery, and ease of manipulation. Whereas the normal “wild type” adeno-associated virus tends to integrate its DNA into a specific locus on chromosome 19, gutted versions containing a transgene persist in the nucleus as circular episomes, separate from the cellular chromosome. This lack of integration markedly reduces the “genotoxicity,” meaning the chance for causing genetic mutations in cellular genes, but it also increases the chance the new gene could be “lost,” since it is not permanently incorporated into the cellular genome.
Critical to the success of gene therapy is ensuring adequate expression in the diseased cell, meaning skeletal and heart muscle cells in patients with DMD. There are two major determinants of adequate expression, the AAV serotype and promoter. The serotype (a word originally used to define related viruses distinct enough not to show cross reactivity with immune serum) determines the amino acid sequence of the outer shell of the virus, which contains the “key” to enter a cell; different serotypes bind to different cell surface receptors to gain entry. While most but not all naturally occurring AAV serotypes enter skeletal muscle cells, only a subset also enter heart muscle cells. Each serotype has its own unique biodistribution, as do many new synthetic serotypes being created and explored to preferentially transduce specific tissues. A seemingly infinite variety of capsid variations are possible, and finding those best suited for DMD is underway 4. No virus has yet been discovered or created that is completely specific for a given cell type, and most viruses transduce liver cells after given intravenously due to features of circulation and liver uptake, in addition to other tissues. Once the virus binds and enters a cell, the second determinant regarding expression is the activity of the enhancer/promoter in that cell. There are many well-known naturally occurring strong enhancer/promoters, such as from a cytomegalovirus gene, the elongation factor 1a gene, and chicken β-actin, as well as many synthetic and hybrid versions, that are capable of activating gene expression across a wide variety of cell types. While these promoter/enhancers might work well in muscle, they may also express the transgene, in this case a small form of dystrophin, in cells that don’t normally express it, which could cause harm. Thus, gene therapy vectors for DMD use enhancers/promoters derived from muscle-specific genes such as muscle creatine kinase, desmin, and myosin heavy chain, in order to restrict expression to muscle cells. Taken together, in order for gene therapy to work in the DMD heart, it must be comprised of a vector serotype that transduces cardiac muscle cells and an enhancer/promoter that is active or “on” in those cells.
The choice and structure of the transgene is also an important consideration, especially in DMD. Because the full-length dystrophin protein is 3,685 amino acids, it requires 11,595 DNA nucleotide bases to encode it (3 bases per amino acid). AAV shells are quite small, about 70 nanometers, and they can only package approximately 4,700 bases of DNA. From this vantage point, some viruses with large packaging capacities such as herpes simplex virus might seem more attractive, but many such larger virus types do not transduce muscle cells well. The main solution employed so far to solve this size dilemma has been to use a shortened version of dystrophin. Much of the dystrophin protein within the so-called rod domain is comprised of numerous repeats, which can be deleted to fit into AAV. While these so-called mini-dystrophin (akin to dystrophin in Becker patients) or micro-dystrophin (with nearly all repeats deleted) proteins are not normal, they are sufficient to restore significant function to muscle cells. Such an approach markedly improved cardiomyopathy in aged mdx mice, resulting in improvements on electrocardiography, hemodynamic measurements, and prevented dobutamine-stress induced cardiac death 5.
Ongoing clinical development of gene therapy for DMD is currently being spear-headed by three main companies. Sarepta is using an AAV serotype derived from rhesus monkeys called rh74. An advantage of this serotype due to it being a monkey virus is that, ostensibly, fewer people have pre-existing immunity to AAVrh74 compared to human-derived AAV serotypes. Their micro-dystrophin gene is driven by an engineered hybrid promoter, MHCK7, derived by combining sequences from the creatine kinase and alpha-myosin heavy chain complex promoters and shown to have high expression in both cardiac and skeletal muscle cells6. Thus, it is anticipated this virus will be effective in treating the heart. Similarly, Solid Biosciences uses an AAV9 serotype which enters skeletal and cardiac cells as well as a variety of other tissues, and a muscle creatine kinase 8 promoter, which also shows robust cardiac expression in preclinical studies. In contrast, while the serotype being used by Pfizer is also an AAV9 serotype, their mini-dystrophin gene is driven by the human creatine kinase promoter that for which some data suggest it is not particularly active in cardiac cells. This construct may do well in skeletal muscle, but whether it significantly ameliorates DMD-associated cardiomyopathy is unclear.
As an alternative to replacing the defective dystrophin gene with a shortened version, another approach is to express a gene to help the disease muscle cell in other ways. The GALGT2 gene encodes glycosyltransferase that increases expression of a variety of proteins that can improve muscle cell function such as utrophin, plectin1, agrin, several laminins and integrin. Indeed, AAV-mediated gene transfer of GALGT2 improves not only skeletal muscle pathology in the mdx mouse but also is very effective at improving cardiac function 7and is currently in clinical trials (https://clinicaltrials.gov/ct2/show/NCT03333590).
Some dystrophin mutations are amenable to exon skipping; that is, their alterations or deletions are contained within a single exon that essentially derails the rest of the protein by changing the reading frame and resulting in a premature stop codon. In such cases, antisense oligonucleotides designed to specifically interfere with splicing of that particular exon can cause it to be excluded from the final mRNA during splicing, restoring the reading frame and thus the remainder of the protein. In fact, a chemically stabilized antisense oligonucleotide (so-called “morpholino”) specific for exon 51 is FDA approved for patients with mutations in that exon, and other antisense molecules are being developed for other exons. The major limitation of this approach is achieving sufficient quantities to enough muscle nuclei to have a broad and high impact on restoring dystrophin expression, though conjugation of the antisense with peptides may improve delivery. Another approach to overcome this limitation is to place the antisense sequence in a U7 small nuclear RNA expressed in an AAV, again leveraging the virus biology to achieve high levels in the muscle nucleus. Interestingly, the combination of both the peptide-conjugated morpholino together with AAV-U7 appears even better than either approach alone8.
Due to recent biotechnological advances, over the past few years the definition of gene therapy has expanded from one that is primarily about “therapy using nuclei acids or genes,” such as AAV-insertion of a functional gene into the nucleus or anti-sense oligonucleotides to skip an exon as described above, to one that encompasses “therapy OF genes.” By leveraging and adopting a bacterial system of proteins and template or “guide” RNAs, we can now precisely edit genes in situ ; that is, we can make deletions or insertions into the cellular genome at specific locations based on the unique DNA sequence. Again a nuclei acid is used as a therapeutic, as the hybridization to DNA, or the “matching” of the guide RNA sequence to the DNA, determines where in the DNA to cut and/or insert. Critical to the process is the bacterial protein as well, known as the clustered regularly interspaced short palindromic repeats (CRISPR)-associated protein (Cas). There are several Cas family members, with Cas9 the most commonly one in use. This new capability is rapidly transforming our vision of what diseases might someday be treatable with gene therapy, including DMD. In fact, encouraging proof-of-principle studies in mice have been published. For example, in the mdx mouse that contains a mutation in exon 23 of dystrophin, CRISPR/Cas9 was successful used to delete the mutant exon, restore the reading frame, and partially restore dystrophin expression, resulting in improved function of both skeletal and cardiac muscle9,10. While promising, there remain a number of barriers to clinical use of CRISPR/Cas9 technology. The follow-up study of those mice showed persistence, long-term expression one year later, but detailed analysis also reviewed immunity to the bacterial Cas9 protein and unintended “off target” gene editing 11. The latter has been a concern in the field since its inception, and exactly how precisely we can edit the genome without inducing other mutations that either could be damaging to the cell or cause other diseases such as cancer is unknown. Another potential barrier to CRISPR/Cas9 use in humans was recently uncovered: pre-existing immunity to bacterial Cas9, given humans are colonized with a microbiome, which in a mouse model destroyed cells edited with CRISPR/Cas9 due to anti-Cas9 immunity 12.
Respiratory Outcome Measures for Trials of Emerging Therapies in Duchenne Muscular Dystrophy
While the goal of gene modifying and molecular therapies for DMD is to improve survival, the duration of prospective clinical trials are too limited to be capable of demonstrating an impact on this outcome. As such, surrogate measures of respiratory function which correlate with survival and quality of life (QoL) are necessary. Measures must also be sufficiently sensitive to change within the population for whom the therapeutic effects can be the most meaningful, those younger individuals with DMD.
In order to design clinical trials of novel therapies and assess their impacts on respiratory function, it is necessary to appreciate the natural history of the time course and evolution of respiratory muscle and pulmonary function in DMD as well as the impact that standard therapies may have on that evolution. Furthermore, and perhaps more importantly, multiple individuals with identical genetic features may have significantly different trajectories for cardiopulmonary function. As such, it will be necessary to stratify randomization by phenotypic behavior in order to more accurately evaluate the effect of emerging therapies.