Gene-based Therapy for Spinal Muscular Atrophy

1. Introduction

Spinal muscular atrophy (SMA) is an autosomal recessive disorder that progressively disables motor neuron units in the spinal cord (Crawford & Pardo 1996). Motor neurons are specialised nerve cells that transmit electrical impulses from the spinal cord to the muscles that control movement in the limbs (e.g. for sitting, standing and walking) and in the body trunk (e.g. for swallowing and breathing). The dysfunction of these motor neurons results in progressive muscular weakness and atrophy, loss of movement capability and eventually, in severe forms of the disease, to respiratory failure.

The disorder was originally identified by Werdnig in 1891 and for the next hundred years treatment was limited to orthopedic aids to extend mobility and to nutritional and respiratory support to prolong life (Sumner & Crawford 2018). However, without mechanical ventilation, survival rarely goes beyond two years of age. SMA occurs in 1 in 6,000–10,000 live births and is the most common inherited cause of infant mortality (Van de Ploeg 2017).

The molecular mechanism that triggers SMA was identified by Lefebre et al. (1995) and that insight has opened the way to gene-based therapies that can moderate the progress of the disease. Recent developments have focused on gene-editing coupled with the generation of new motor neurons (motoneurogenesis) as a treatment that could even reverse the effects of SMA. This report examines the potential of gene-based therapies in the treatment of SMA and considers the specific question:

“should gene-editing be used with motoneurogenesis in the treatment and/or prevention of SMA?”

1. Spinal muscular atrophy (SMA)
   1. Types of SMA

SMA categories are based on age at onset and severity of symptoms (Zerres et al. 1997):

• Type I (Werdnig–Hoffmann disease):onset in the first few months of life with rapid loss of motor tone before 6 months of age; inability to sit unassisted; poor head control, but cognition unaffected; feeding support required by age one; tongue and pharynx muscles affected, putting these infants at risk of respiratory failure; death typically between ages one and two. This is the most common form of the disease affecting about 60% of cases (Van de Ploeg 2017).
• Type II(Dubowitz disease): onset after 6 months of age; ability to sit unassisted, but not to walk; survival beyond 2 years but less than 30 years (Kolb & Kissel 2015).
• Type III (Kugelberg–Welander disease): onset after 18 months of age; ability to walk unaided, although this may be lost requiring wheelchair assistance in later childhood (Kramer & Gitler 2017); little respiratory muscle weakness, no effect on life expectancy (Kolb & Kissel 2015).
• Type IV (adult form of the disease): onset in the second or third decade life; neuromuscular decline in later decades; normal life span (Porensky & Burghes 2013).
  1. Pathogenesis of the disease

All cells in the body require the survival motor neuron protein (SMN); complete elimination of the protein leads to cellular and progressively organism death. It is decreased levels of SMN within spinal cord neurons that triggers the neuromuscular pathology of SMA (Porensky & Burghes 2013). In infants with SMA1, irreversible loss of motor neurons begins during the first 3 months of life, and 95% of motor neurons are lost before age 6 months.

SMN is produced primarily by the SMN1 gene, but this gene is deleted or non-functional in about 96% of patients with SMA (Beroud 2003). Humans have a variable number of copies of a nearly identical gene, survival motor neuron 2 (SMN2), that can also produce SMN (see Fig1).

Fig 1: schematic of SMN gene (Kolb & Kissel 2015, Fig 4, p. 837).

However, a C to T substitution in an exonic splicing enhancer results in the exclusion of exon 7 during transcription of SMN2 means that most of the resulting protein is truncated and unstable; only about 15% of typical SMN1 levels of full-length SMN are produced by SMN2 and these are not sufficient for normal cell maintenance (Porensky & Burghes 2013). Higher number of copies of SMN2 per genome has been shown to correlate with decreased level of severity of the disease. Type I SMA is characterized by only two copies of SMN2; Type II by three or four copies; Type III & IV by five or more copies (Sumner & Crawford 2018).

The identification of the underlying genetic cause of SMA has prompted the investigation of gene-based therapies to induce increased SMN protein levels.

2.3.  Established gene-based treatments of SMA

Gene-based technologies allow targeted manipulation of the human genome for therapeutic purposes. Modifications can replace defective DNA by: 1) the addition of beneficial exogenous genes to specific sites in the genome; 2) the removal of damaging genes; and 3) the specific editing of DNA strands to correct mutations within the gene structure (Maeder & Gersbach 2016).

Fig 2:  Strategies for therapeutic genome editing (Maeder & Gerbach 2016, p. 435, Fig 3).

The application of gene therapy can be in vivo or ex vivo (see Fig 2). In an in vivo application, a carrier vector delivers either functional genetic material or a modification-inducing agent (e.g. a nuclease protein) that is targeted to a specific gene. In an ex vivo application, cells are modified and propagated in vitro and then transplanted into the patient.

Two established gene-based therapies for the treatment of SMA have used the in vivo mechanisms described by Chipman et al. (2012) to produce: 1) SMN2 enhancement; and 2) SMN1 replacement.

SMN2 enhancement by ASO

Antisense oligonucleotides (ASOs) can manipulate in vivo the pre-messenger RNA sequences that promote or inhibit exon splicing during RNA processing (Sumner & Crawford 2018). The ASO Nusinersen has been shown to enhance the effectiveness of existing SMN2 genes by promoting the inclusion of exon 7 into SMN2 transcripts during splicing of SMN2premRNA, which leads to increased encoding of normal full-length SMN protein. ASOs disperse widely when injected into the cerebrospinal fluid, without requiring a carrier; but they are not persistent and cannot cross the blood-brain barrier, so they must be delivered to the central nervous system by intermittent intrathecal injection.

Finkel et al. (2017) assessed the effectiveness of Nusinersen in a trial involving infants 80 ASO-treated infants (and 42 untreated infants), over a period of two years. At the completion of the trial, only 39% of ASO-treated infants failed to survive or required permanent assisted ventilation compared with 68% of the untreated infants; 51% showed markedly improved motor ability.

Nusinersen was approved for the treatment of SMA in USA in 2016. It is now subsidised in Australia through the Pharmaceutical Benefits Scheme.

SMN1 replacement by scAAV9-SMN

The second therapy uses Avexis-101 (scAAV9-SMN), an adeno-associated virus (AAV) carrying exogenous SMN1 cDNA, to deliver a complete working copy of the SMN1 gene to the central nervous system (Mendell 2017). This new copy of SMN1 sits inside motor neuron cells, but does not incorporate into the child’s DNA; it’s essentially a supplemental gene that induces SMN expression in vivo within the cells and peripheral tissues. AAV vectors are persistent and can cross the blood-brain barrier; they can be delivered to the central nervous system in a one-time postnatal intravenous injection.

In a study of 12 SMA infants treated with Avexis-101, Mendell (2017) found that all survived beyond the age of 20 months without requiring permanent assisted ventilation; all but 1 achieved levels of motor-functionality in feeding, sitting and talking, beyond those typical for SMA infants.

Although the patients in these trials were still producing SMN three years after treatment, it is not yet established that the single dose of this gene therapy will last throughout life. However, Avexis-101 is expected to be approved for the treatment of SMA in USA in 2019.

Patient screening

Both ASO and scAAV9-SMN therapies have been shown to deliver significant improvements in motor function as well as improvements in survival (Sumner & Crawford 2018). However, the precipitous drop in motor neurons in the early course of SMA means that, by the time that clinical diagnosis is made, a significant portion of motor neurons are most likely are already lost. A diagnosis of SMA in pre-symptomatic patients through screening would allow treatment of the disease before the onset of motor neuron loss or development of clinical symptoms (Chien 2017). Earlier treatment has been shown to deliver better outcomes – more patients respond and the response is larger (Sumner & Crawford 2018).

Non-invasive prenatal diagnosis of SMA has been shown to be feasible through the genetic analysis of fetal cells circulating in maternal blood (Beroud 2003) and more recently, a protocol for newborn genetic screening has been developed to predict SMA severity (Sumner & Crawford 2018). In the United States, the American Advisory Committee on Heritable Disorders in Newborns and Children has recently recommended nationwide newborn screening for SMA (Sumner & Crawford 2018). In Australia, newborn babies are now being routinely screened by testing drops of blood taken from the baby's heel (Guthrie test).

1. Gene-editing with motoneurogenesis in the treatment of SMA

Recent studies have investigated the potential of an ex vivo process that combines in vitro gene-editing to ‘correct’ pluripotent stem cells with motoneurogenesis to supply fully-functional motor neurons to the SMA patient.

3.1.  Motoneurogenesis

Chipman et al. (2012) describe three techniques to generate motor neuron (MN) cells ex vivo for transplantation to the patient (Fig 3):

Fig 3: Techniques for the derivation of motor neurons (Chipman et al. 2012, p322, Fig 1).

• Embryonic stem (ES) cells are propagated in vitro from cells extracted from pre-implantation blastocysts, i.e. in humans, the egg stage about 5 days after fertilisation (Klimanskaya et al. 2014). ES cells proliferate indefinitely in culture and can be manipulated to produce many cells of many types of cells in the body. However, their usage implies termination of tissue that has the potential to generate life.
• Induced Pluripotent stem (iPS) cells are propagated in vitro from sample somatic cells of the patient (Yu & Thomson 2014). iPS cells are reprogrammed from skin fibroblasts, expand robustly, retain the capacity to generate motor neurons rapidly and efficiently ( and can be gene-edited in vitro. iPS cells are patient-specific but remain proliferative and can become tumorigenic. Their usage does not require termination of the source and so provide a more ethically acceptable alternative to ES cells (Chipman et al. 2012).
• Induced motor neurons (iMN) are not based on stem cells; instead, sample somatic cells in skin fibroblasts taken from the patient are reprogrammed to convert them directly into induced motor neuron (iMN) cells. iMNs generated this way are patient-specific and are anatomically and physiologically equivalent to endogenous motor neurons; however, they cannot be edited before transplantation (Chipman et al. 2012).
  1.     Gene-editing

A number of studies (e.g. Ebert 2009; Hester et al. 2011) have demonstrated the potential of ex vivo gene-editing to ‘correct’ pluripotent stem cells and use them as the source of fully-functioning motor neuron cells.

Gene-editing utilises the body’s endogenous cellular repair machinery to incorporate changes into the DNA strand sequence (Maeder & Gersbach 2016). A double-strand break (DSB) in the DNA is repaired either by direct religation of the cut ends (nonhomologous end-joining, NHEJ) or by mapping to the template provided by the parallel chromosome (homology-directed repair, HDR). Breaking the DNA strand at specific target points and introducing an exogenous template stimulates HDR repair mode; in that way, any sequence differences in the provided template are incorporated into the repaired site.

A number of platforms exist to induce DSBs at specific target sites in DNA while avoiding or minimizing collateral damage elsewhere in the genome (Maeder & Gersbach 2016).  Three of these (zinc finger nucleases, transcription activator-like effector-nucleases, and meganucleases) require specific proteins for each target DNA site. A more recently developed platform is Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR associated protein 9 (CRISPR/Cas9) nucleases. CRISPR technology emulates an adaptive immune system found in bacteria that interferes with an invading virus by cutting its DNA at a specific site; CRISPR-associated (Cas9) proteins specify the DNA site to be cut by the Cas9 “scissors”. The encoding of the target site in only a short region of gRNA that can be easily altered makes the CRISPR/Cas9 technology a flexible and re-usable tool in gene-editing.

3.3.  Application of gene-editing with motoneurogenesis to SMA

Corti et al. (2012) applied gene-edited motoneurognesis to SMA subjects in a mouse model. iPS cells were generated from skin fibroblasts and genetically modified in vitro. Single-stranded oligonucleotides were used to direct the edit of a T to C at position +6 of exon 7 in the survival motor neuron 2 (SMN2) gene. The inclusion of exon 7 in transcription converted the SMN2 gene into an SMN1-like gene that produces normal amounts of full-length SMN. Corrected motor neurons derived from the edited iPS cells were then transplanted back into the subjects. Treated subjects showed reduced SMA symptoms and extended life span compared with the control group of untreated subjects. The genetic modification was shown to be permanent and heritable (Corti et al. 2012).

Although the therapeutic outcomes of their study were positive, Corti et al. (2012) reported an issue that would limit the application of their technique, i.e. the oligonucleotides vectors used can produce mutations that interfere with normal cell function and might lead to tumorigenesis. This issue was addressed by Zhou et al. (2018) in a human-based trial that used CRISPR technology but using Cpf1 nuclease (instead of Cas9) as the targeting mechanism. Cpf1 was shown to be a more accurate and more efficient genome editing system for seamless genetic conversion of SMN2 to an SMN1-like gene. SMN expression was restored in the edited iPS cells and their derived motor neurons, without cell aberrations, i.e. with no exogenous sequences, normal karyotype and absence of tumorigenesis.

The positive results achieved by Zhou et al. (2018) were for an adult patient with Type 3 SMA. Further trials involving early-onset SMA cases are still needed to gain regulatory approval of this approach. However, the results show that the combination of gene-editing with motoneurogenesis is a safe, feasible treatment of SMA that can be used to effectively reverse the pathogenic effects of SMN deficit and provide a permanent, heritable cure for the disease.

1. Benefits and risks

The combination of gene-editing with motoneurogenesis targets the underlying genetic mechanisms that cause the disease. It can prevent further motor neuron loss/damage and, most importantly, can also replace motor neurons already lost at the time of diagnosis.

4.1.       Major benefits of this approach are:

• For the patient – an effective one-time treatment that can
  • improve quality of life by restoring full muscular mobility and control
  • extend life expectancy to normal levels
  • remove the SMA-carrier risk in later parenthood.
• For society –
  • a cost-effective treatment; in comparison, the cost to the Australian Government of Nusinersen is approximately A$370,000 per year of life for an estimated 160 patients (Australian Government Department of Health 2019).
  • a universal treatment that does not require wide-scale screening programs; although screening for SMA is now being included in standard early post-natal programs in Australia, these programs are not available in many other countries.
  • a heritable cure that will eliminate the disease over time (Maeder & Gerbasch 2016).
  • a process model that can be reused; the techniques to isolate, edit, propagate and transplant cells back to the patient can be readily adapted to target other similar genetic conditions (Maeder & Gerbasch 2016).
  1.     Key risks associated with this approach are:

• Ethical concerns – the use of ES cells has raised the issues of consent and termination of life. Both of these issues are now resolved by the use of iPS cells derived from the patient (Chipman et al. 2012).
• Patient safety – early trials involving in vitro gene-editing of pluripotent stem cells identified potential collateral cell damage that raised the risk of cancer.

The recently developed techniques for more accurate site-targeting using CRISPR/Cpf1 have mitigated these risks (Zhou et al. 2018).

• Non-optimal return on investment – a prevention approach might be more cost-effective. Genetic screening of intending parents can identify SMA carriers before conception. This would allow for in vitro fertilization (IVF) with pre-implantation genetic testing so that only those fertilised embryos without the disease are implanted (Cooper 2018). A government-sponsored pilot test of pre-pregnancy screening for SMA is currently underway in Australia (Scott & Armitage 2018).

1. Conclusion

Spinal muscular atrophy (SMA) is a neurodegenerative disorder affecting the motor neurons in the spinal cord, leading to reduced mobility and early death. The condition is caused by a deficit in the survival motor neuron protein (SMN) as a result of genetic inheritance of a missing or dysfunctional survival motor neuron (SMN1) gene. Current gene-based therapies for SMA are designed to increase SMN levels, either by introducing exogenous SMN1or by enhancing the effectiveness SMN2, a parallel gene can also produce SMN but only in limited amounts. Both approaches have been shown to be effective in improving mobility and extending life span in SMA patients, but neither is a cure as they do not re-generate motor neurons.

Gene-editing with motoneurogenesis does not rely on any existing motor neurons. In vitro gene-editing of induced pluripotent stem cells to amend SMN2 to an SMN1-like gene can provide new fully-functional motor neurons for transplant back into the patient. The process employs established techniques for cell harvesting and transplantation; the CRISPR/Cpf1 technology has been shown to be a safe and efficient tool to achieve accurate gene-editing. The corrected motor neurons have been shown to be effective as a treatment of SMA in mouse models and more recently in human trials.

Gene-editing with motoneurogenesis will provide feasible, low-risk treatment of SMA, with positive therapeutic outcomes for the individual patient; for society, the genetic mechanisms of this approach offer a heritable cure not only for SMA but possibly also for other similar disorders. This analysis concludes that gene-editing coupled with motoneurogenesis should be used in the treatment and/or prevention of SMA.

• References
  
  ​​​​​
•   Australian Government Department of Health 2019, ‘The Australian Government listed nusinersen on the Pharmaceutical Benefits Scheme (PBS) from 1 June 2018 for the treatment of Type 1, Type 2 and Type 3a spinal muscular atrophy (SMA)’, viewed 10.5.2019 at:  http://www.health.gov.au/internet/main/publishing.nsf/Content/MC17-021776-SMA.

• Beroud, C 2003, ‘Prenatal diagnosis of spinal muscular atrophy by genetic analysis of circulating fetal cells’, Lancet, 361, pp. 1013–14.
• Chien, Y-H 2017, ‘Presymptomatic Diagnosis of Spinal Muscular Atrophy Through Newborn Screening’, The Journal of Pediatrics, 190, pp. 124-129.
• Chipman, P, Toma, J & Rafuse, V 2012, ‘Generation of motor neurons from pluripotent stem cells’, Progress in brain research, vol. 201, pp. 313-331, http://dx.doi.org/10.1016/B978-0-444-59544-7.00015-9.
• Cooper, L 2018, ‘Genetic testing plus IVF can sidestep genetic disease and reduce the need for high-priced therapies’, downloaded 9/5/2019 from: https://www.statnews.com/2018/12/14/ivf-plus-genetic-diagnosis-sidestep-genetic-disease/.
• Corti, S, Nizzardo, M, Simone, C, Falcone, M, Nardini, M, Ronchi, D, Donadoni, C, Salani, S, Riboldi, G, Magri, F, Menozzi, G, Bonaglia, C, Rizzo, F, Bresolin, N, Comi, G 2012, ‘Genetic correction of human induced pluripotent stem cells from patients with spinalmuscular atrophy’, Science Translational Medicine, 4, 165ra162, pp. 1-15.
• Crawford, T & Pardo, C 1996, ‘The neurobiology of childhood spinal muscular atrophy’, Neurobiol Dis, 3, pp. 97–110.
• Ebert, A, Yu, J, Rose, F, Mattis, V, Lorson, C, Thomson, J & Svendsen, C 2009, ‘Induced pluripotent stem cells from a spinal muscular atrophy patient’, Nature, vol. 457, pp. 277-281.
• Finkel, R, Mercuri, E, Darras, B, Connolly, A, Kuntz, N, Kirschner, J, Chiriboga, C, Saito, K, Servais, L, Tizzano, E, Topaloglu, H, Tulinius, M, Montes, J, Glanzman, A, Bishop, K, Zhong, Z, Gheuens, S, Bennett, C, Schneider, E, Farwell, W & De Vivo, D 2017, ‘Nusinersen versus Sham Control in Infantile-Onset Spinal Muscular Atrophy’, The New England Journal of Medicine, 377, pp. 1723-1732.
• Hester, M, Murtha, M, Song, S, Rao, M, Miranda, C, Meyer, K, Tian, J, Boulting, G, Schaffer, D, Zhu, M, Pfaff, S, Gage, F & Kaspar, B 2011, ‘Rapid and Efficient Generation of Functional Motor Neurons From Human Pluripotent Stem Cells Using Gene Delivered Transcription Factor Codes’, Molecular Therapy, vol. 19, no. 10, pp. 1905-1912.
• Klimanskaya, I, Kimbrel, E & Lanza, R 2014, ‘Embryonic stem cells’, in R Lanza, R Langer & J  Vacanti (eds), Principles of Tissue Engineering, Fourth Edition, pp. 565-579.
• Kolb, S & Kissel, J 2015, ‘Spinal Muscular Atrophy’, Neurol Clin, 33, pp. 831-846.
• Kramer, N & Gitler, A 2017, ‘Raise the Roof: Boosting the Efficacy of a Spinal Muscular Atrophy Therapy’, Neuron, 93, pp. 3-5.
• Lefebvre, S et al. 1995, ‘Identification and characterization of a spinal muscular atrophy determining gene’, Cell, 80(1), pp. 155–165.
• Maeder, M & Gersbach, C 2016, ‘Genome-editing Technologies for Gene and Cell Therapy’, Molecular Therapy, vol. 24 no. 3, pp. 430-446.
• Mendell, S 2017, ‘Single-Dose Gene-Replacement Therapy for Spinal Muscular Atrophy’, The New England Journal of Medicine, 377, pp. 1713-22.
• Porensky, P and Burghes, A 2013, ‘Antisense Oligonucleotides for the Treatment of Spinal Muscular Atrophy’, Human Gene Therapy, 24, pp. 489-498.
• Scott, S & Armitage, R 2018, ‘Genetic testing: SMA added to newborn heel prick in profound change in medical screening’, downloaded 9/5/2019 from: https://www.abc.net.au/news/2018-10-14/genetic-testing-sma-added-to-newborn-heel-prick-test/10359622.
• Sumner, C & Crawford, T 2018, ‘Two breakthrough gene-targeted treatments for spinal muscular atrophy: challenges remain’, The Journal of Clinical Investigation, 128, no. 8, pp. 3219-3227.
• Van de Ploeg, A 2017, ‘The Dilemma of Two Innovative Therapies for Spinal Muscular Atrophy’, The New England Journal of Medicine, 377, p. 18.
• Yu, J & Thomson, J 2014,  ‘Induced Pluripotent Stem Cells’, in R Lanza, R Langer & J  Vacanti (eds), Principles of Tissue Engineering, Fourth Edition, pp. 581-592.
• Zerres, K, Wirth, B & Rudnik-Schoneborn, S 1997, ‘Spinal muscular atrophy - Clinical and genetic correlations’, Neuromuscular Disorders, 7, pp. 202–207.
• Zhou, M,  Hu, Z, Qiu, L, Zhou, T, Feng, M, Hu, Q, Zeng, B, Li, Z, Sun, Q, Wu, Y, Liu, X, Wu, L & Liang, D  2018, ‘Seamless Genetic Conversion of SMN2 to SMN1 via CRISPR/Cpf1 and Single-Stranded Oligodeoxynucleotides in Spinal Muscular Atrophy Patient-Specific Induced Pluripotent Stem Cells’, Human Gene Therapy, 29 (11), pp. 1252-1263.