Combination molecular therapies for spinal muscular atrophy: How much is enough?

In this issue, Harada and colleagues describe their experience with five patients with 5q-linked spinal muscular atrophy (SMA) type I who were exposed to both nusinersen and onasemnogene abeparvovec-xioi (onasemnogene). Four received gene replacement therapy 14 to 19 mo after initiating treatment with nusinersen. Three patients continued on nusinersen after onasemnogene. One patient received onasemnogene first followed by nusinersen 3 mo later. Patients ranged from 6.5 to 23 mo of age at gene replacement therapy and were 17–29 mo old at the time of last evaluation. Lee al reported last year two infants who were treated with nusinersen at age 5.5 and 5 mo, respectively, followed by onasemnogene at age 9 months. Both were improving at the time of writing and showed no apparent adverse effects. The neuromuscular field rejoices in the availability of now three treatments for this devastating disease, all of which clearly alter the natural history. The groundwork for this revolutionary treatment was laid with collaboration by clinicians, scientists, and advocacy groups. Elegant natural history studies and development of novel outcome measures sensitive to change in weak infants pushed the field forward. The first therapy to be approved for SMA was nusinersen, an antisense oligonucleotide (AS0) designed to promote inclusion of exon 7 in the SMN2 transcript. It is delivered intrathecally every 4 mo after initial loading doses. It was approved in December 2016 and was shown in clinical trials to allow long-term survival without full time ventilation of infants with SMA type I. This treatment revolutionized the outlook for newly diagnosed families and offered hope of improvement or stabilization for older children. In 2019, a second medication, onasemnogene, was approved by the U.S. Food and Drug Administration (FDA) for children with SMA under age 2 y. This gene replacement therapy is administered one time as an intravenous infusion. Finally, risdiplam, the first oral/per gastrostomy drug for SMA, received FDA approval in 2020. Risdiplam is a small molecule that promotes exon 7 inclusion, a mechanism similar to that of nusinersen. While each treatment has been shown to be effective individually, there are situations in which combination therapy is being considered by families and healthcare providers. For example, an infant might not be eligible for onasemnogene based on abnormal liver enzymes or antiAAV9 antibodies at diagnosis, so treatment is initiated with nusinersen. If laboratory studies return to normal, the family might wish to receive onasemnogene (sequential dosing). Another situation may occur when there is an apparent lack of response to the initial treatment. If a child is continuing to lose motor milestones or is not improving as expected with the first drug, then either switching to a different treatment or adding a drug with a different mechanism of action might be contemplated. The option of several therapies raises new questions, particularly about the potential benefits and toxicities of combining them. All three medications increase the amount of full length (fl) SMN protein in surviving anterior horn cells, although through different mechanisms. Information about both the benefits and the adverse effects of these therapies comes from clinical trials and limited real-world data from the nusinersen experience. We do not know how much fl SMN protein is sufficient to treat the disease or to prevent ongoing loss of motor neurons. Carriers of the SMA mutation are clinically healthy despite having only one functional copy of SMN, suggesting a threshold effect. We do not know exactly how much fl SMN protein is produced by any of the current therapies. The clinical effect of nusinersen increases over time with repeated dosing. The question of how much fl SMN is “enough” is critical when one considers combining these medications. In a postmortem study of SMA patients treated with nusinersen, the number of cells in the spinal cord expressing fl SMN protein was equal to or greater than that seen in age-matched controls, and the levels of protein increased proportionately with increasing number of doses. If each of the current treatments achieves sufficient fl SMN levels in the spinal cord, then combinations of such drugs that act by increasing fl SMN might not offer added benefit. A second question relates to the age of the child and timing of treatment. Elegant work by Sumner's group suggests that, in both healthy controls and children with SMA, fl SMN protein and transcript are highest in prenatal life and fall rapidly by age 3 mo. These data suggest there is a developmental period from 3 mo pre-term to 3 mo post-term when the action of fl SMN protein is most important. Perhaps there is a window of opportunity during which there might be particular benefit from combining therapies that increase fl SMN expression to achieve more rapid response, while a single therapy might be sufficient after this window closes. The optimal therapeutic window might even include the pre-natal period. A third question relates to what tissues and organs require fl SMN and have dysfunction that contributes to the disease phenotype. Increasing evidence points to a role of fl SMN in multiple tissues, including skeletal muscle (and neuromuscular junction) and peripheral nerve. Therapies with different routes of administration and different tissue targeting might be used synergistically to ensure that all critical tissues are treated. More information about the role of fl SMN Received: 15 October 2020 Revised: 23 October 2020 Accepted: 25 October 2020