Allie Mazzella 

In honor and memory of Katherine Santiago

Until 1995, Spinal Muscular Atrophy (SMA) research was primarily concerned with treating symptoms rather than the disease itself—easing patient care and family living. But since the discovery of the genes and DNA structures associated with the disease, SMA research has skyrocketed, and tremendous strides have been made especially in the past ten years. This review will highlight the major discoveries in SMA research, with regards to both what causes the disease as well as treatment options. By emphasis on the DNA sequences and structures, the review will demonstrate how the root of the problem is also the root for the solution, and how the more DNA-centric experimental treatments have been the most effective in the search for the cure.

Spinal Muscular Atrophy (SMA) is a genetic neuromuscular disorder characterized by motor neuron loss leading to muscle weakness and atrophy. It is the number one genetic killer of infants, and was virtually unheard of until the 1980s. Today there are four main types of SMA identified, with many different subtypes and many experimental treatments in the works. However, there is no cure. SMA is an ideal disease to study because it has a large impact in society, as well as follows basic biological phenomena. It is autosomal recessive and it only involves a few key genes. The genes involved are a part of an inverted duplication and are also subject to copy number variations; the genes directly demonstrate how and why the disease occurs, and what treatments options are feasible. This review will cover the highlights of the history of SMA research, as well as an advanced discussion of the DNA structure of the genes. Truly comprehending these structures will lead to an understanding of the disease as a whole. 

History of the Disease: 1990s

Two physicians, Guido Werdnig and Johan Hoffman, first discovered SMA in 1890 (1). It was described as a disease in which the α-motor neurons, or survival motor neurons (SMNs), are absent from the spinal cord, and therefore the communication between brain and muscle movement is severed. As a result, muscles that receive no messages are never used, and atrophy. For the next one hundred years, little else was known about the disease aside from this basic characterization as well as its death toll (1). Still, nothing was known about the disease’s cause or development.

This all changed in the 1990s. Firstly, the entire SMA gene was discovered and sequenced, found on a region of chromosome 5q (2). Then in 1995, French scientist Suzie Lefebvre and her team at L’Institut National de la Santé discovered the exact gene and mutation on that was demonstrated to cause of 95% of SMA in patients (3). SMA was also demonstrated to exhibit a completely autosomal recessive form of inheritance. Finally, this paper was able to categorize SMA into different types based on level of severity and onset, as well as the genetics. Type I became known as Werdnig-Hoffman disease in honor of the scientists who first discovered it, and is the most severe type. Types II, III, and IV are progressively less severe, and occur in infants over the age of 1, juveniles, and adults respectively (3).

Although it is just one gene that actually causes the disease in the majority of cases, a few years later it was discovered that the disease is dictated by two genes in conjunction—one that actually causes the disease, and one that dictates the severity. These genes are Survivor Motor Neurons 1 and 2 (SMN1 and 2 respectively). SMN1 codes for the proteins that encompass and manufacture the survival motor neurons described above. Therefore, in order for SMA to occur, SMN1 must suffer a loss-of-function mutation in a 140 base-pair critical region, or simply be deleted (4). As the disease is autosomal recessive, homozygous loss of SMN1 causes SMA. Once SMN1 is rendered nonfunctional, SMN2 determines severity because it is almost identical to SMN1. When SMN1 is turned off, SMN2 is the only possible producer of the main SMN protein. However, as this is not SMN2’s main purpose, it is not nearly as effective as SMN1 and therefore cannot completely compensate (Figure 1). Sometimes people have two, three, or four copies of SMN2, and more copies of SMN2 lead to better compensation of SMN1 loss-of-function (4). Therefore, people with types II, III, and IV SMA have more copies of SMN2 and are therefore able to promote later onset. It is hypothesized that SMN2 is the result of a large duplication which includes SMN1 and evolved as a compensation mechanism directly for SMA. SMN1 and SMN2 are revealed to be almost completely genetically identical. What distinguishes SMN2 from SMN1 is a single base-pair change, but why this is so crucial remained unknown for a few years.

 History of the Disease: 2000s

With the full sequencing of the human genome in 2000, more about SMN1 and SMN2 is discovered. With the discovery of RNA splicing, both SMN genes are partitioned into 8 exons, with exon 2 being split into 2a and 2b. In this process, the distinction between SMN1 and SMN2 becomes clear. The single base-pair change that differentiates SMN2 from SMN1 is a C to T transition mutation in exon 7, which as a result causes SMN2 to splice out exon 7 during mRNA processing (5). As a result, the usually 294 amino acid SMN protein is shortened and much less functional without exon 7. This demonstrates both the importance of every single base pair, as well as explains why SMN2 cannot replace SMN1.

Armed with the knowledge of what causes the disease and how it does so, the last ten years of research have been about finding the cure. Many early treatments attempted to replace the SMN proteins, or find a new way in which to encode them. But this has not proven effective, and the most successful approaches have been the DNA-centric ones.

Two major discoveries have come about through the DNA-centric approach. Firstly, while 95% of SMA cases are caused due to the SMN1/SMN2 mechanism detailed above, the other 5% of cases occur due to a variety of other mutations and types of SMA (6). An autosomal dominant form of SMA was discovered in 2005, with key gene BICD2 at the center. An X-linked form of SMA was also discovered in the same year. Secondly, in looking for model organisms to perform SMA experiments, it has been revealed that the SMN gene is highly evolutionarily conserved. Model organisms for SMA research have ranged from yeast to zebrafish and mice, and both of the SMN genes, their functions, and SMA have been proven to exist for hundreds of thousands of years (6).

 

Molecular Basics of SMA

The SMN genes are about 20 kilobase pairs long. SMN1 is located from base-pair 70,924,940 to base pair 70,953,014, while SMN2 is located from base-pair 70,049,522 to 70,077,594 (2). Both genes are on chromosome 5, on the long arm q at position 13.2. SMN1 is located on the telomeric portion of the chromosome, whereas SMN2 is centromeric in location on the chromosome (3). But it is only mutations in the telomeric copy of the gene, SMN1, that cause SMA. Mutations in the centromeric gene, SMN2, can be modifiers of SMA, but cannot actually cause the disease.

Both genes are part of a five hundred kilobase inverted duplication on chromosome 5q13, much bigger than either of the genes (2). Because this region of the chromosome was duplicated and therefore has many repetitive elements, the region has an increased probability of rearrangements and deletions, resulting in copy number variation. This is why SMA occurs randomly at such a high frequency.

Due to SMA research, a second function of the SMN1 gene was revealed. SMN1 is also critical in RNA-processing. SMN1, along with SMN-interacting-protein-1 (SIP1) form a complex with several spliceosome-associated snRNP proteins (7). Additionally, nuclear SMN1 DNA has been found in Gemeni of the coiled bodies (GEMS), which are also alleged to have a role in RNA processing. It has been discovered that SMN1 exon 2 shares homology with several nucleic acid binding factors, particularly high mobility group (HMG) proteins. This homology allows SMN1 to participate in RNA, singled-stranded DNA and double-stranded DNA binding, which makes SMN1 ideal for a role in RNA splicing. SMA patients, demonstrating the predicted correlation between the protein sequence and function, have exhibited reduced RNA-binding activity (7). Additionally, SMA patients show dramatic reorganization of spliceosomes in response to the lack of functional SMN protein (7). Because of this dual-role of SMN, SMA can be investigated as both a motor neuron disease as well as a general splicing disease.

The SMN protein that SMN1 is known to code for has three domains—Gemin-2 binding, YG-box, and Tudor, which are mainly all involved in RNA processing. The SMN protein has a secondary structure mainly composed beta sheets and turns, with 1 large alpha helix on the end (7). Exon 7 encodes the C-terminus of this protein and this is the portion that SMN2 encoded protein is lacking

The SMN2 gene encodes for protein products that are also involved in RNA splicing. But these proteins has different responsibilities than SMN1 proteins, such as being involved in the biogenesis of snRNPs, rather than aiding in the functions of snRNPs like SMN1 (7). Through the study of SMN2, it has been determined that SMN2 splicing can produce four distinct RNA transcripts, known as isoforms. Only one of them, isoform d, can function as an SMN1 replacement and encodes for a fully functional SMN1 protein (5). But these other RNA transcripts encode the majority of SMN2 proteins, which perform other duties in the cell. Alternative RNA splicing allows for these different transcripts to be possible.

Although much is now known about SMN1 and SMN2 and their involvement with SMA, the actual process of how the lack of SMN1 causes the death of motor neurons is not entirely understood (4). The current step in SMA research is figuring out this pathway.

Alternative Forms of SMA

As mentioned above, BICD2 is the gene whose mutation causes the autosomal-dominant form of SMA. BICD2 is located on chromosome 9, and is usually involved with dynein-mediated transport around the cell for vesicles. However, whenever the gene suffers a heterozygous loss-of-function mutation, the patient presents with symptoms of SMA. This form of SMA is called SMALED2 (8). Although transport is important in every cell, it is particularly important in the spinal cord, with the highest levels of BICD2 expression detected in motor neuron cells (8).

Two independent papers published in 2013 with a mutation to the BICD2 gene gave two hypotheses for why BICD2 mutations present phenotypically similar to SMN1 mutations. The first two papers describe a C to T base pair substitution causing an amino acid change from serine to leucine at position 107. This mutation was found to cause enhanced dynein binding. The paper published by Neveling et al. hypothesized that the enhanced binding causing BICD2 accumulation in centrosomes and the Golgi apparatus, 2 places with high microtubule transport activity. They propose that the vesicle trafficking between the endoplasmic reticulum and Golgi slowed motor neuron cells’ G1 normal tasks so much so that they died and never even reached M phase (8). The second paper published by Oates et al. proposed an alternative idea with regards to enhanced dynein biding. They hypothesized that enhanced dynein binding decreased motor neuron’s precursor cell’s ability to divide, resulting in a block of motor neuron outgrowth and impaired motor neuron cells in embryonic development (9). Closely monitoring of SMALED2 in an embryo would confirm or disprove this hypothesis, but both technology and research is not up to human models yet.

As more information about autosomal-dominant SMA as well as other alternative forms of SMA comes to light, there is also controversy with regards to the definition of the disease. Should SMA be defined by its symptoms, or its genetics? For over 100 years, SMA was solely characterized by the absence of motor neurons in the spinal cord and subsequent muscular atrophy. However, genetics has revealed that there are many mechanisms by which these characteristics may come about, and while phenotypes may be similar, the mechanisms behind are very different. As treatments have begun to target DNA mechanisms as opposed to proteins mechanisms, the different forms of SMA can no longer be treated the same. Some scientists argue that the alternative forms of SMA should be categorized as their own diseases. Others believe that while alternative forms of SMA may require alternative treatments, the diagnosis should remain the same, and all forms of SMA impact and educate how to treat the other forms.

Treatment: The mTOR pathway

On the surface, the cure for SMA appears simple—inject the missing SMN1 protein into the patients. But this fails for a few reasons. Although the survival motor neuron gene and protein is highly evolutionarily conserved, using other hosts to manufacture the protein and inject it has proven ineffective. This process has failed due to such a high specificity in motor neurons for each individual; the body treats foreign SMN1 as an infection, and attacks it almost instantly.

Majority of SMA treatments today use one of two alternative approaches. Treatments either aim to engineer SMN2 to “replace” SMN1 when it has been knocked out, or seek to stop SMN1 from being knocked out in the first place (1). In choosing the second approach, one needs to determine the exact pathway by which the lack of SMN1 causes motor neuron death, which is where research is now.

Very recently, such a pathway may have been discovered. In December of 2014, a paper published in Human Molecular Genetics detailed the miR-183/mTOR pathway, by which SMN is said to regulate neural axon growth (10). The proposed pathway states that decreased SMN protein levels changes micro-RNA expression in motor neurons. The specific micro-RNA impacted, miRNA-183, is in charge of regulating the protein mTOR at the post-translational level. mTOR is the known protein involved with regulating axon outgrowth out of the cell body of a neuron. Therefore, when SMN decreases, miRNA-183 becomes deregulated, mTOR becomes deregulated, and axons are unable to grow out of neurons, thereby leading to motor neuron cell death (10). A very recent mouse model in which micro-RNA183 is inhibited, shows recovery of some motor neuron cells, now able to function without SMN1 (10).

Alternative pathways have not yet been proposed, but more evidence and experimentation will be needed in order to verify this one, such as other models of the disease like zebrafish, or using alternative SMN1 mutations in the mice to test for similar results. Although very early, if correct, the mTOR pathway provides a mechanism by which treatments can operate in order to cure SMA.

 

Diagnosing

The biggest advancement in SMA research is the production of an easy and simple carrier test (11). With a sensitivity of 100% and a specificity of 96.2%, the blood test compares the levels of SMN1 and SMN2 DNA in the body and tests for mutations, and if SMN1 is knocked out, the number of copies of SMN2 one has. The test relies on qRT-PCR, and then sequencing. It is then compared to a library of SMN1/SMN2 mutations, to determine if there is a mutation, and if so, what to expect (11). Both parents are tested to determine if they carry a mutation in SMN1. As long as one parent is homozygous for the normal allele, there is very little risk of a child having SMA. The biggest problem with SMA used to be detection—as 95% of the cases are an autosomal recessive disease, no one would undergo the invasive and expensive test, and then parents would be unprepared. Now as all parents can easily be tested, there is less risk and more awareness with regards to caring for children with SMA.

 

Conclusion

In the past twenty-five years, research on Spinal Muscular Atrophy has come a long way. The disease itself has been recognized, categorized, and is now relatively understood with regards to DNA sequences, structures, and mechanisms. And because of this, the research towards finding a cure for SMA has also been able to make tremendous strides. However, the biggest advancement may be increased awareness about SMA in the general public. Because the disease was so poorly understood, very few people knew about it or were tested for it. Today, people are very active towards the fight for a cure, with many organizations dedicated to helping families and funding research. In particular Boston Children’s Hospital is recognized as one of the top SMA treatment and research facilities in the country. They receive major grants for SMA research, most recently in Febuary 2015 for $140000 for Dr. Mustafa Sahin’s work on “mTOR and Protein Synthesis in SMA.” In summary, as the knowledge, awareness, and research continues to grow towards fighting SMA, the cure grows closer and closer.

References

1) Van Meerbeke J.P. Progress and Promise: The Current Status of Spinal Muscular Atrophy Therapeutics. 2011. Discovery Medicine 065: 62-78.

2) Brzustowicz L.M, Lehner T, Castilla L.H, Penchaszadeh G.K, Wilhelmsen K.C, Daniels R, Davies K.E, Leppert M, Ziter F, Wood D, Dubowitz V, Zerres K, Hausmanowa-Petrusewicz I, Ott J, Munsat T.L, Gilliam T.C. Genetic mapping of chronic childhood-onset spinal muscular atrophy to chromosome 5q11.2-13.3. 1990. Nature 344:540–541.

3) Lefebvre S, Burglen L, Reboullet S, et al. Identification and characterization of a spinal muscular atrophy-determining gene. 1995. Cell 80:155-165.

4) Farrar MA, Kiernan MC. The Genetics of Spinal Muscular Atrophy: Progress and Challenges. 2014. Neurotheraputics 29: 290-301.

5) Kashima T., Manley J. A negative element in SMN2 exon 7 inhibits splicing in spinal muscular atrophy. 2003. Nature Genetics Letters 1-4.

6) Wirth B. An Update of the Mutation Spectrum of the Survival Motor Neuron Gene (SMN1) in Autosomal Recessive Spinal Muscular Atrophy (SMA). 2000. Human Mutation 15: 228-237.

7) Boda, B., Mas, C., Giudicelli, C., Nepote, V., Guimiot, F., Levacher, B., Zvara, A., Santha, M., LeGall, I., Simonneau, M. 2004. Survival motor neuron SMN1 and SMN2 gene promoters: identical sequences and differential expression in neurons and non-neuronal cells. Europ. J. Human Genetics 12: 729-737.

8) Neveling K, Martinez-Carrera LA, Holker I, et al. Mutations in BICD2, which encodes a golgin and important motor adaptor, cause congenital autosomal-dominant spinal muscular atrophy. 2013. Human Genetics 92:946-954.

9) Oates, E. C., Reddel, S., Rodriguez, M. L., Gandolfo, L. C., Bahlo, M., Hawke, S. H., Lamande, S. R., Clarke, N. F., North, K. N. Autosomal dominant congenital spinal muscular atrophy: a true form of spinal muscular atrophy caused by early loss of anterior horn cells. 2013. Brain 135: 1714-1723.

10) Kye M.J., Niederst E.D., Wertz M.H. et al. SMN regulates axonal local translation via miR-183/mTOR pathway. 2014. Human Molecular Genetics 1;23(23): 6318-6331.

11) Feldkotter, M., Schwarzer, V., Wirth, R., Wienker, T. F., Wirth, B. 2002. Quantitative analyses of SMN1 and SMN2 based on real-time lightCycler PCR: fast and highly reliable carrier testing and prediction of severity of spinal muscular atrophy. 2002. Human Genetics 70: 358-368.