Spinal Muscular Atrophy – Is Newborn Screening Too Late for Children with Two SMN2 Copies?

ABBREVIATIONS

INTRODUCTION

Spinal muscular atrophy (SMA) is the most common neurodegenerative disease in childhood. In more than 95% of cases it is caused by a homozygous deletion of exon 7 of the SMN1 gene, encoding the survival motor neuron (SMN) protein, which is involved in a variety of cellular processes (RNA biogenesis and splicing, translation, transcription, apoptosis, endocytosis, DNA damage repair, etc.) [1–3]. When there is a critical deficiency of SMN protein, there is a rapid degeneration of bulbar and spinal motor neurons [4].

SMA-specific therapies, that can correct SMN deficiency, have been available for a few years. Considering the irreversibility of motor neuron damage, there is wide consensus concerning the need for newborn screening to allow treatment in the first weeks of life. The dramatic benefit of NBS on the motor outcome of most patients has already been clearly demonstrated [5–8].

As a species-specific attribute, humans have a second, almost identical copy of the SMN1 gene, called SMN2. SMN2 differs from SMN1 only by the exchange of five nucleobases. A critical C to T transition in exon 7 leads to an altered splicing of SMN2 and thus to a significantly reduced production of functional SMN protein from the SMN2 gene. The amount of SMN-protein, which is produced by SMN2, is usually insufficient to prevent disease onset in SMA patients. Nevertheless, increasing copies of SMN2 is generally associated with a milder phenotype of SMA [9, 10]. Complete absence of the SMN1 and SMN2 gene leads to embryonic lethality and is incompatible with life. The gene dose of SMN2 currently represents the major disease-modifying factor, apart from other genetic modifiers that have been previously described [1]. Patients with one SMN2 copy are extremely rare. They show a severe clinical picture called SMA type 0 with respiratory failure at birth, and potentially, arthrogryposis. The majority of infants with two SMN2 copies are expected to develop Werdnig-Hoffmann’s disease (SMA type 1). Early initiation of disease specific therapy is critical. Previously, our group has reported that a relevant proportion of SMA patients with two SMN2 copies must be considered “early symptomatic,” (i.e., already showing an active disease process with a decline of motor neurons evidenced shortly after birth). This is best explained by the time-dependent, pivotal role of the SMN protein in the development of the peripheral nervous system in the last months of pregnancy and in the first months of life [11, 12].

SMA screening was implemented as part of the national newborn screening program in Germany in October 2021. SMA newborn screening has been launched in the majority of states in the U.S., and is expected to become a standard nationwide procedure in other European countries in the near future. The aim of this work, is to characterize the prognosis of the extremely vulnerable group of SMA patients with two SMN2 copies identified through newborn screening and provided with timely initiation of treatment.

PATIENTS AND METHODS

The description of the NBS pilot projects, the method for screening for SMA, the confirmation of homozygous deletion of SMN1 and SMN2 quantification at neuromuscular referral centers were previously published [5, 6, 13, 14].

Twenty-one children born between January 2018 and January 2021 were included in this study. Inclusion criteria were detection by NBS, two copies of the SMN2 gene, treatment start with Nusinersen or Onasemnogene abeparvovec within the first 6 weeks of life (range 10–39 days, median 17 days, mean 20.3 days), and a minimum of 9 months of age at last examination. Data cut was in November 2021. The follow-up period of individual patients ranged from 10 months to 3.5 years of age. Data collection was performed as part of a prospective cohort study [6]. The local ethics committee of the participating universities (Project No. 18–269) approved the study.

Twenty of 21 children were initially treated with the antisense oligonucleotide Nusinersen since gene replacement therapy with Onasemnogene abeparvovec, which has been approved for early treatment since summer 2020, was not yet available. In 9 of 20 initially Nusinersen-treated patients, therapy was switched to Onasemnogene abeparvovec during the course (age 5–28 months, mean 14.7 months, median 13 months). The indication for gene replacement therapy was mostly the parents’ wish to avoid the need for sedation associated with Nusinersen treatment. None of the children had therapy complications that may have had an impact on motor development. During the observation phase of this study, no child had switched treatment from Nusinersen to Risdiplam, which is a small molecule that is available orally and was approved by the EMA in March 2021 for SMA patients aged two months an older.

The CHOP INTEND, a reliable measure for patients with SMA, albeit with a strong “ceiling effect”, comprises 16 items for evaluation of motor skills [15] and was regularly performed on all children. The first examination consistently took place before the start of treatment, the latest time being at 5 weeks of age in patient 10 (Fig. 1A). For infants with repeated maximum CHOP INTEND scores of 64, the test was discontinued. The Hammersmith Infant Neurological Examination Section 2 (HINE-2) is an additional assessment tool for evaluating motor milestones in infants with SMA. It consists of eight sections [16] and was also administered until the maximum score of 26 was repeatedly reached. Depending on the motor skills, endurance of patients, and the patient’s understanding of the tasks, the Gross Motor Scale of the Bayley Scales of Infant and Toddler Development [17] was performed at age 12–24 months.

Measurement of ulnar CMAPs was done according to standard procedures. Ulnar CMAPs were performed in the first weeks of life and at different intervals and are available from all patients.

RESULTS

Motor outcome

CHOP INTEND

Seventeen of 21 patients (81%) achieved the maximum score of 64 points. Four patients slowly increased their CHOP INTEND values to a score of 42–60 points at an age of > 15 months. Two children with low CHOP INTEND scores (9 and 27 points) at baseline, finally improved to 42–51 points but were not able to sit unassisted during their entire follow-up period. Two children with initial scores of 35 and 33 finally reached 64 points and learned to walk independently. The two children with a decline from scores above 40 to scores below 35 within the first 2–3 weeks of life never reached a maximum score in the CHOP INTEND and currently exhibit an SMA type 3 phenotype. Patient 32 started with 47 points, presenting with a very slow increase during the first months. He deteriorated at age 10 months despite continuous SMA therapy when he suffered a severe, first pneumonia and attained an upper limit of 42 points only. All children with CHOP values above 50 had no deterioration and all reached 64 points (Fig. 1A).

Fig. 1

(A) Course of CHOP INTEND scores, (B) Course of HINE-2 scores. X-axis: Age in weeks, y-axis: scoring results. (C) Bayley gross mwallowingw score and (D) Bayley gross motor age scales. X-axis: Age in months, y-axis: scoring results. The more intensely colored area represents the norm range of healthy children. ^*Data from HINE-2 and CHOP INTEND from patients 1-43 have already been published in our previous manuscripts [5, 6].

Motor WHO-milestones

Fifteen of 21 children (71%) within the first 9 months of life eventually reached the WHO milestone,, sitting unassisted“. Four more patients (10%) presented with a delay in acquiring this motor milestone at the age of 10–17 months. Two children (9.5%) have not yet learned to sit alone (age at last examination 16 months and 30 months, respectively).

Twenty of 21 children were at least 15 months old at the last visit (or had already reached this milestone earlier) and could be assessed for their ability to walk with assistance. Twelve of 20 (60%) reached this motor milestone in an age-appropriate time frame, while five children (25%) were delayed and ambulated with assistance at age 17–23 months. Three children (15%) have so far not achieved this milestone (18, 30 and 26 months at last examination).

Nineteen of 21 children were at least 18 months old or could already walk independently at the time of last examination, so the WHO milestone “walking alone” could be assessed. Ten of 19 children (53%) reached this milestone within the normal age range of 18 months. Three more patients gained independent walking at the age of 20–23 months. Six children are not yet able to walk unassisted despite being 18–30 months of age. Three of these children have progressed to ambulation with assistance (Fig. 2).

Fig. 2

Age at achievement of WHO-Milestones. Green: Sitting without support, yellow: walking with assistance, red: walking independently. Sign “syringe”: Start of therapy with Nusinersen. Sign “double helix”: Treatment with Onasemnogene abeparvovec. ^*Data from patients 1-43 have already been published in our previous manuscripts [5, 6].

HINE-2 score

HINE-2 testing was performed in all children to describe motor functions not included in the WHO motor milestones such as head control and motor skills in prone position. Thirteen of 21 children (62%) showed normal and age appropriate motor development. Six children (38%) did not reach the maximum score, displaying heterogeneous underperformance in the HINE-2 score (7–25 points), with four of them (19%) continuously approaching the maximum and three children (14%) far behind (Fig. 1B). The two mildly affected children with HINE scores of 24 and 25 at > 2 years of age have proximal weakness only and cannot walk or stand independently. All children with HINE scores of 23 or lower at the corresponding age have proximal and axial weakness with difficulties also in a prone position.

Gross motor Bayley scores

In fourteen patients, a Bayley motor test was administered at least once. Seven children (50%) who attained all motor milestones within normal limits, had normal Bayley test scores. Two children (14%) started with below-average values but improved to test results consistent with their age; both achieved the ability to walk independently (one with an insignificant delay). Five of 14 children (36%) significantly underperformed on Bayley testing, all of them with delayed attainment of milestones (Fig. 1C/D).

DISCUSSION

In this study, we investigated the outcome of children diagnosed with SMA in NBS, focussing on children with only two copies of the SMN2 gene and who received SMN targeted treatment within the first 6 weeks of life. In summary, 57% of the patients developed without any signs of a motor neuron disease. This implies the achievement of motor milestones within the WHO predefined age ranges, the absence of dysphagia or respiratory problems as well as normal results on the Bayley motor tests. Another 14% with initial delay, progressed to a near normal or normal development. Despite early diagnosis and treatment, 29% of patients developed manifest proximal weakness, half of which were type 2 or type 3 SMA, respectively. One patient (4.8%) developed respiratory problems. 14% have mild chewing problems and two patients (9%) are fed via gastrotube. All patients with respiratory or feeding problems had additional muscle weakness.

A key question to be answered is whether clinical or neurophysiological findings at the first examination following NBS, usually at ages 7–10 days, are predictive of children’s outcomes. These data would facilitate adequate counselling of families, especially regarding treatment expectations.

We found that the CHOP INTEND might give some indicative information at rather low or rather high scores: in this study cohort, values below 30 were related with a less favourable motor outcome whereas values above 50 were associated with a good prognosis. Values between 30 and 50 were not associated with outcome. Some children in this range may deteriorate later despite medical treatment. However, predictions should be treated with caution because the number of cases is too small for in-depth statistics.

Absence or presence of deep tendon reflexes are ineffectual as clinical prognostic markers. Ulnar CMAP amplitudes values above 3 mV might indicate a favourable course. Low CMAPs may predict a low SMN2 copy number at initial presentation and suggest urgency of therapy, but a low CMAP amplitude does not necessarily predict poor motor outcome.

It should be noted, that children from this study with poorer outcomes did not have a longer duration to initial treatment onset than patients with better outcomes.

This study supports the hypothesis [5] of an already intrauterine onset of the disease for the potentially worse prognosis of children with 2 SMN2 copies compared with children with more copies. CMAPs are frequently used as a surrogate parameter for the stratification of SMA patients in clinical studies. Decrease of CMAPs precedes loss of motor function [19]. In this study, patients with 2 SMN2 copies showed a median value of 1.2 mV on ulnar motor studies at first presentation. This was far below the values which were found in patients with 3 or more SMN2 copies, indicating that, in the group of patients with 2 SMN2 copies, some degree of denervation already occurred during pregnancy.

Our results are also consistent with the NURTURE and SPR1NT studies [8, 20] respectively. In the Nurture study, which investigated early Nusinersen treatment in presymptomatic children, a cut-off of the ulnar CMAP of > 1 mV was used [21]. In the SPR1NT study, which investigated presymptomatic children, undergoing early gene replacement therapy, a cut-off of 2 mV was used.

In both studies, a smaller proportion of patients with two SMN2 copies were found to have relevant signs of motor neuron loss. This supports our conclusion that an unremarkable clinical impression at initial presentation is no guarantee of later absence of symptoms. Notwithstanding the fact that ulnar CMAPs above 2 mV were used as an inclusion criterion in the SPR1NT study, a small proportion of children with 2 SMN2 copies became symptomatic with signs of motor developmental delay or dysphagia.

This work addresses the question of outcome of children with two SMN2 copies who received a diagnosis of SMA through newborn screening and were treated either with SMN2-targeted treatment (Nusinersen) or gene replacement therapy (Onsamnogene abeparvovec) at 10–39 days of age. In this cohort, 20 of 21 children received Nusinersen as first therapy due to the fact that gene replacement therapy with Onasemnogene abeparvovec, was only approved in Germany in the summer of 2020 and the children presented in this study were at least 10 months old at the time.

This study produced comparable results to NURTURE and SPR1NT both of which were conducted in children diagnosed shortly after birth. As the outcomes reported in these studies are not significantly dependent on drug choice, we assume that the data generated from this patient cohort are independent of the initial type of SMN-targeted therapy [8, 20].

In conclusion, although NBS has significantly improved outcome, parents of children with SMA and 2 SMN2 copies detected by NBS should be counselled that there is no guarantee for a cure.

A minority of 15% may still be at risk of significant physical disability despite early treatment and up to 30% of cases present with delayed motor development, which may correspond to mild proximal weakness along the child’s developmental trajectory. CHOP-INTEND-scores and CMAP might facilitate outcome prediction, but a larger number of cases is needed to confirm this.

LIMITATIONS

This is a descriptive study in which the limited number of patients prevents a more in-depth statistical analysis, or renders such an analysis meaningless. The small numbers are a limitation as well as that only one child was treated with Onasemnogen abeparvovec.

DECLARATIONS

DISCLOSURES

All authors have indicated that they don’t have any non-financial conflicts of interest.

Funding Sources

Not applicable.

AUTHOR’S CONTRIBUTION

Oliver Schwartz conceptualized and designed the clinical study, collected clinical and electrophyisological data and co-drafted the manuscript.

Heike Kölbel conceptualized and designed the clinical study, collected clinical and electrophyisological data and reviewed the manuscript.

Astrid Blaschek and Ulrike Schara collected clinical and electrophyisological data and reviewed the manuscript.

Dieter Gläser performed the genetic confirmation and the SMN2 copy number determination and reviewed the manuscript.

Siegfried Burggraf developed the screening method of the newborn screening from dried blood spots on behalf of the respective project owner.

Wulf Röschinger performed the newborn screening from dried blood spots on behalf of the respective project owner.

Wolfgang Müller-Felber conceptualized and designed the clinical study, collected clinical and electrophyisological data, and reviewed and revised the manuscript.

Katharina Vill conceptualized and designed the clinical study, collected clinical and electrophyisological data, co-drafted the initial manuscript, and reviewed and revised the manuscript.

All authors approved the final manuscript as submitted and agreed to be accountable for all aspects of the work.