Spinal muscular atrophy (SMA) is an autosomal recessive neuromuscular disorder that manifests across a wide range of severity. The cardinal clinical feature of SMA – diffuse skeletal muscle weakness – is a consequence of dysfunction or loss of alpha motor neurons. SMA is caused by loss-of-function mutations or deletions of the gene SMN1 (Gene ID
6606). A wide range of disease severity can be partially attributed to the presence of variable copy number of a neighboring near-identical gene, SMN2 (Gene ID
. A single base pair difference between these two genes greatly reduces the efficiency of exon 7 inclusion into mature transcripts from the SMN2 gene, but the coded protein sequence from full length SMN2 derived transcripts is unaffected by this change. SMN2 thus produces only a fraction of the functional full length protein compared to SMN1 
. The SMN gene is constitutively expressed in all eukaryotic cells and necessary to cell viability 
. Motor neurons are particularly vulnerable to reduced SMN protein levels for reasons unknown. The invariable presence of the partially functional disease-ameliorating SMN2 gene in individuals with SMA offers an attractive target for development of therapeutics 
. Proof-of-concept studies with small molecule, antisense oligonucleotide or gene therapy approaches targeting various mechanisms to upregulate expression of SMN2 have extended survival and improved motor function in SMN-deficient animal models 
. Clinical trials of available putative SMA treatments that non-specifically act on SMN expression have, so far, failed to demonstrate efficacy 
One impediment to efficient trial design is that the typical individual with SMA old enough to cooperate with motor function testing declines very slowly 
. Many affected children settle into a plateau phase with stable function for years even though they may have manifested progressive weakness in the first year or two of life 
. The consequence of this course is that a meaningful attenuation of the rate of degeneration would necessarily take a long time to demonstrate. Availability of a valid clinical measure of SMN expression could thus accelerate clinical trials of an SMN-enhancing therapeutic, particularly in the early, dose-finding, phase of development. Given that most new treatments currently under development intend to increase SMN synthesis in the nervous system, measurement of the abundance of SMN transcript or protein in tissues accessible to clinical sampling could be an immediate and plausible biomarker that fulfills the need for an early read-out of target engagement.
The continuous spectrum of SMA phenotype severity is generally divided into three “Types” of SMA based upon the history of specific gross motor abilities achieved before the disease curtailed further developmental progress 
. “Type I” defines those who never sat independently, “Type II” those who sat but never walked, and “Type III” those who were able to achieve independent ambulation. Over time, an individual having a milder type of SMA may decline to a level that overlaps in function with that manifested at an early point in the course of individuals affected by more severe types. The genomic number of SMN2 copies correlates with SMA types 
. Severity of motor impairment in SMA is likely multifactorial, although disabling mutation of SMN is the essential first step. The potential broad range of SMA phenotype that is predicted by loss of SMN1 is focused to some extent by characterization of SMN2 copy number, but within each specific SMN2 genotype there is broad spectrum of motor function, and over time the severity of motor impairment can vary even more.
The BforSMA study was designed to explore potential biomarkers of disease severity in accessible peripheral tissues – blood and urine – that may be of value in the execution of clinical trials. The cohort was selectively recruited to represent subjects with a wide range of SMA type, motor function, and age, who were then characterized clinically in a manner that would best power two independent projects. Results of the first project, an unbiased biomarker discovery effort, are presented in a companion paper (see companion paper, Finkel et al. 
). Here, we present the second project, a targeted analysis of relationships between SMN transcript and protein levels to SMN genotype, SMA clinical type, present motor function, and age. The project had two major goals. The first was to determine whether a previously identified association between SMN2 copy number and disease severity could be confirmed in an SMA cohort prospectively recruited to include individuals having all three types of SMA manifesting a broad range of motor impairment. A second, more focused, goal was to evaluate a possible relationship between SMN transcripts and protein levels in blood samples from SMA subjects compared to controls, and within the range of SMA clinical severity. Relationships between any of these measures of SMN in blood samples and clinical features may, with proper further development, prove invaluable for clinical trials of SMN-enhancing therapeutics.
Important to this project was the recognition that age is an important potential confounder to be included in the statistical analysis. The power of this is enhanced by targeted recruitment of a clinical cohort in which SMA type and age are not highly correlated. At the outset of the study, the extent to which this broad distribution could be accomplished was unknown, given early childhood mortality of subjects with SMA Type I. Site investigators were encouraged to keep this goal in mind during subject recruitment.
To date, studies of SMN transcript and protein levels from amniocytes, skin fibroblasts, lymphocytes or peripheral blood mononuclear cells (PBMCs) from SMA subjects have reported variable results, but a general correlation of transcript or protein quantity with phenotype is identified only in individuals at the lower end of the SMA functional continuum 
. Protein levels in tissues from SMA Type I infants were noticeably lower than those of healthy controls, while the differences in protein levels between Types II, III, carriers and control subjects were not significant. Existing studies have not controlled for present level of motor function or subject age, and the possibility that these factors undermine or enhance the relationship of blood assessments of SMN to clinical variables is unexplored. In addition, newly developed SMN transcript measurements by absolute quantification and improved ELISA-based protein assays have the advantage of increased sensitivity and reliability 
, and may prove helpful in characterizing the relationships between protein and disease severity in various tissues. It is thus not yet clear if blood-derived SMN transcript or protein levels can be used as a sensitive and meaningful marker of disease severity.