Onasemnogene‐abeparvovec administration to premature infants with spinal muscular atrophy

Abstract

Twin girls born at 30 weeks' gestation with spinal muscular atrophy (SMA) received onsasemnogene‐abeparvovec (OA) at 3.5 weeks of life. They had no treatment‐related adverse events, normal acquisition of motor milestones, and normal neurological examination at 19 months. Genotyping revealed 0 copies of SMN1 and a single, hybrid SMN2 gene containing the positive genetic modifier c.835‐44A>G. This was associated with full‐length SMN2 blood mRNA expression levels similar to a 2 copy SMA infant. The observed favorable outcomes are likely due to the genetic modifier combined with early drug administration enabled by prematurity.

Introduction

The neuromuscular disease spinal muscular atrophy (SMA) is caused by biallelic mutations of the survival motor neuron 1 gene (SMN1). ¹ Patients retain at least one copy of a paralogous gene SMN2, which is alternatively spliced generating reduced levels of full‐length SMN protein. SMA disease severity correlates inversely with SMN2 copy number. Patients with the most common type I SMA most often harbor two copies of SMN2 and die by age 2 without treatment. SMA type 0 is a rare disease form in which infants typically have a single copy of SMN2 causing congenital onset weakness and limited survival after birth. ²

Onsasemnogene‐abeparvovec (OA), an adeno‐associated virus 9 (AAV9) gene therapy, is one of three available SMN‐inducing treatments for SMA. ¹ , ³ , ⁴ Each of these drugs can improve motor function, but initiation of treatment as quickly after birth as possible is critical for optimal outcome. ¹ , ⁵ There is limited information regarding the safety and efficacy of treatment provided to premature infants. ⁶ , ⁷ Furthermore, because of the predicted severity of disease in infants with a single copy of SMN2, current guidelines leave treatment decisions to physician discretion in this patient group. ⁸ Here, we report our experience with SMA twins with one copy of SMN2, who received OA at 33 weeks' gestational age (GA).

Methods

Institutional review board approvals (University of Iowa) were obtained for data and human sample collection. Written informed consent was obtained from all participants. Clinical history was abstracted from the medical record. Blood was collected from the twins, their mother, and two age‐similar SMA patients (Table 2).

Table 2.

Blood sample identification, SMN gene copy numbers, and treatment information of participants.

Case ID/Description	SMN1 copy #	SMN2 copy #	Initial SMA treatment age (type)	Age at OA treatment	Age at blood draw
Parent	2 (cis)	1 (hybrid)	–	–	–
Twin B	0	1 (hybrid)	3.5 weeks (OA)	3.5 weeks	8 months
Child X	0	2	4 weeks (OA)	4 weeks	42 months
Child Y	0	4	2.5 weeks (nusinersen)	7.5 months	12 months

Genetic analysis

Blood samples were analyzed with the P021 Multiplex Ligand‐dependent Probe Amplification (MLPA) kit from MRC Holland^® to determine SMN1 and SMN2 copy number. The SMN1 and SMN2 genes were sequenced by NGS following a previously established protocol. ⁹

RNA extraction and RT‐qPCR

RNA was extracted and purified using the PAXgene^® Blood RNA Kit (Qiagen/BD), and RNA clean up was performed with RNeasy kit (Qiagen). RT‐qPCR was performed as previously described on a 7900HT Real‐Time PCR System using multiple housekeeping genes. ¹⁰

Case report

Monochorionic, monoamniotic twin girls were born by emergency cesarean section at 30 weeks and 2 days gestation due to fetal decelerations in Twin B. Birth weights were 1560 grams (Twin A) and 1590 grams (Twin B). Apgar scores at 1 and 5 minutes were 7 and 8 (Twin A) and 3 and 7 (Twin B). Both twins were intubated at birth with Twin A extubated at day of life (DOL) 28 and Twin B at 27. Nasogastric (NG) tube feeds were initiated DOL 8 for both girls. Routine head ultrasounds were normal in both twins on DOL 7 and 28. Fetal echocardiogram of Twin B showed a pericardial effusion, and postnatal echocardiogram showed a trace pericardial effusion; no cardiac abnormalities were observed in Twin A.

Newborn screening suggested spinal muscular atrophy (SMA). Confirmatory genetic testing (performed at Athena Diagnostics and repeated at Invitae Diagnostics) showed 0 SMN1 copies and 1 SMN2 copy. Neurologic examination on DOL 18 was normal for gestational age; reflexes could be elicited, and no tongue fasciculations were noted. Both twins were treated with 1 mg/kg prednisolone followed by OA, 1.1 × 10¹⁴ vector genomes (vg) per kg, at 33 weeks and 5 days gestational age (DOL 24). Troponin I, AST, ALT, and platelets remained normal. Neurological examinations showed normal tone and reflexes throughout the hospitalization. They were discharged on DOL 51 (GA 38 weeks) on a steroid taper, which was completed at 3 months of age. At the time of discharge, both twins had zone 3, stage 0 retinopathy of prematurity.

Twin B required NG tube feeds until DOL 78 as swallow studies showed aspiration with thin liquids and poor oral strength. Both had normal swallow studies at 3 months of age, but continued to cough with thin liquids, which was managed with thickening. Swallow studies at 2 years showed mild–moderate oral‐pharyngeal dysfunction. In early infancy, both girls were hospitalized briefly for various upper respiratory viral infections such as the rhino/enterovirus, parainfluenza virus, and the respiratory syncytial virus/entero bronchiolitis. They have not required hospitalization since age 16 months (Twin A) or 6 months (Twin B).

Serial neurological examinations have shown normal motor tone, bulk and strength, normal reflexes, and absent tongue fasciculations through 2 years of age. At 30 months, formal evaluation showed speech and language delay in both girls (Table 1).

Table 1.

Developmental milestones of Twins A and B.

Age at visit in months (corrected for prematurity)	Motor skills	Social/cognitive skills
8 (6)	Sitting without support
12 (10)	Crawling, pulling to stand	Single words
15 (13)	Walking independently
19 (17)	Running, using utensils	1–4 words, stranger anxiety
30 (28)	Running, climbing	Speech and language delay

DNA and RNA analyses

DNA sequencing revealed that both twins have a single SMN2 copy that includes two PSVs (paralogous sequence variants) from SMN1: one located in intron 6 (position Chr5:69372304) and the other in exon 8 (Chr5:69373081) making this a SMN2‐SMN1 hybrid gene (Fig. 1A). ⁹ Importantly, the PSV in intron 6 is a known positive modifier variant c.835‐44A>G. ¹¹

Figure 1.

Schematic representation of the SMN2‐SMN1 hybrid gene present in this family and RT‐qPCR analysis of SMN mRNA levels in peripheral blood. (A) The hybrid structure is comprised of almost all paralogous sequence variants (PSVs) from SMN2 (represented in purple), except for two specific PSVs from SMN1 (in orange) located in intron 6 (c.835‐44A>G) and exon 8 (c.*239A>G). The T in position c.840 defines this gene as SMN2. PSV8 is not present in this scheme as it is no longer considered a PSV. ¹⁵ (B) SMN1 transcript is detected in the mother of the twins (parent, black bars) but not in Twin B or in children with SMA with 2 or 4 copies of SMN2. (C) SMN2‐FL and (D) SMN2‐Δ7 expression of one twin were comparable to an SMA patient with 2 copies of SMN2, but less than an SMA patient with 4 copies of SMN2. (⁺2 copies SMN1 in cis, *SMN1/2 hybrid gene).

Their mother also harbored this hybrid gene. SMN1 and SMN2 mRNA levels were analyzed in whole blood samples from mother, Twin B (insufficient RNA obtained from Twin A) and 2 other SMA children (SMN2 copy numbers 2 and 4), who had been treated with OA neonatally (Table 2). Blood samples were obtained several months after OA treatment. Given the short lifespan of blood cells, ¹² there is predicted to be little OA still present in these cells at the time of sampling. As anticipated, SMN1 transcript was only detected in the mother (Fig. 1B). In contrast, all SMA subjects had SMN2 transcripts that contained exon 7 (FL) as well as those that lacked exon 7 (Δ7). Twin B had approximately the same level of SMN2 transcripts as a SMA child with 2 copies of SMN2 suggesting that the positive modifier c.835‐44A>G improves exon 7 inclusion as previously shown. ¹¹ Twin B and the infant with two copies of SMN2 had approximately half of both the FL and Δ7 transcripts compared to a SMA child with four SMN2 copies (Fig. 1C,D).

Discussion

We describe twin girls with SMA treated with OA at 33 weeks' gestational age, who have normal motor function at 2 years old. These cases illustrate the challenges in predicting phenotype based on standard SMN2 copy number testing and demonstrate that delivery of OA in prematurity can be both safe and highly effective.

The single SMN2 copy detected in these twins predicted a very severe disease phenotype, which would be minimally responsive to treatment. However, further post hoc analysis revealed that this single SMN2 gene is a SMN2‐SMN1 hybrid. Hybrid genes may ameliorate the SMA phenotype and have been detected in 5–10% of patients, ¹³ , ¹⁴ but present data indicate a much higher proportion. ¹⁵ Even more importantly, the detected SMN2 gene contains the positive modifier variant c.835‐44A>G. ⁹ , ¹¹ The variant is localized to a region of intron 6 that functions as a HuR‐dependent splice silencer. In the presence of this modifier, HuR binding is decreased and exon 7 inclusion can increase by approximately 30% as demonstrated in minigene studies. ¹¹ Consistent with the splice modifying effect of this variant, FL‐SMN2 mRNA levels were approximately equal in the blood of one of our twins compared to another SMA patient with two copies of SMN2, but well below those seen in a patient with four SMN2 copies. These results suggest that without treatment the twins would have had a phenotype similar to two copy SMA infants. Importantly, the twins reported here were asymptomatic at the time of treatment. In the SPRINT trial, presymptomatic treatment of two copy SMA infants with OA resulted in good outcomes, although some patients showed deficient acquisition of some motor milestones. ⁴

Like the other two SMA therapeutics, early (presymptomatic) OA treatment has the greatest treatment effect, likely related to preventing rapid irreversible motor unit loss during the perinatal period in severe patients. ¹⁶ It has also been shown that SMN protein levels normally fall during the last trimester and first few months after birth suggesting it is most needed during early developmental periods. ¹⁰ In the case of a premature birth, there is an opportunity to deliver a SMA therapeutic at an earlier stage of motor neuron development and potentially more effectively suppress early degeneration. One previously published case described an infant born at 36 weeks' gestation with SMA type 0 and a single copy of SMN2, who received nusinersen starting on DOL 14 followed by OA at 114 days. After treatment, the infant had improvement in motor function with a 30‐point improvement in CHOP‐INTEND score, but remains profoundly weak. ¹⁷ In our twins, OA was safely administered to infants with SMA at 33.5 weeks gestational age, with excellent motor outcome at 2 years.

The twins reported here have mild language delay at 30 months. There are several possible contributing factors including prematurity ¹⁸ and twin gestation. ¹⁹ There is evolving evidence that language development is delayed in some infants with SMA ²⁰ , ²¹ although the data are mixed. ²² , ²³ , ²⁴ The twins received prednisone to prevent known OA immune‐mediated side effects including acute liver injury, thrombocytopenia, thrombotic microangiopathy, and elevated troponin I. ²⁵ , ²⁶ Although postnatal steroids have a long history of use in premature infants, primarily in the context of prevention or treatment of bronchopulmonary dysplasia, they have also been associated with increased risk of neurodevelopmental disorders. ²⁷ This risk varies by type of corticosteroid and duration of treatment. ²⁷ , ²⁸ The optimal immunosuppression for use with OA in premature infants is unknown. Finally, for all children, social and genetic factors can affect language development.

Each of the three approved therapies for infants with SMA has risks and benefits that must be weighed for the individual patient. In addition to the immune‐mediated risks, necrotizing enterocolitis has been reported in term infants after OA administration. ²⁹ In the two infants reported here with a single copy of SMN2, the combination of the genetic modifier and early presymptomatic treatment resulted in excellent motor outcome. The clinician should consider the clinical status of the infant, SMN2 copy number, and positive modifier variants in making early treatment decisions in SMA.

Author Contributions

SMB: Writing, editing and RNA analysis. ASA: Original draft, data collection and editing. DR: Data collection and editing. EFT: DNA analysis, writing and editing. LBP and MCR: Data analysis and editing. CJS: RNA analysis, concept development, writing and editing. KDM: initial conceptualization, oversight and writing and editing.

Conflict of Interest

KDM receives clinical trial support from Biogen, Scholar Rock, and Biohaven. CJS has or is receiving grant support from Biogen, Actio Biosciences, and Roche. She has also received financial compensation for consulting and/or speaking from Novartis, Roche, Genentech, and Biogen. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Supporting information

Table S1.

Funding Statement

This work was funded by NINDS grant R35 NS122306; National Institute of Neurological Disorders and Stroke of the National Institutes of Health grant P50NS053672.

Data Availability Statement

The data that support the findings of this study are available in the main body and supplementary material of this article. Additional data that are not publicly available due to privacy or ethical restrictions are available on request from the corresponding author.