4th UK spinal muscular atrophy (SMA) researchers network meeting

1. Introduction 

The fourth UK Spinal Muscular Atrophy Researchers Network Meeting was held at the Oxford Belfry Hotel on the 21st and 22nd November 2007. The meeting was focused on three main research areas: (1) SMN protein function, (2) Therapeutic strategies in SMA and (3) SMA animal models. Delegates representing academic research, industry and funding charities were in attendance and included: University of Oxford, University of London, University of Sheffield, University of Edinburgh, University of St. Andrews, Peninsula College of Medicine and Dentistry and Keele University, Summit Plc, the Jennifer Trust for SMA and the SMA Trust.

2. SMN protein function 

The opening presentation by Phil Young (Peninsula College of Medicine and Dentistry, Exeter) provided an update on the molecular analysis of the SMN protein. Dr. Young and his colleagues have developed a panel of expression constructs representing individual exon deletions spanning the entire coding sequence of SMN [1] and are using this series to illuminate the mechanisms whereby SMN is localised to Cajal bodies (CBs), sites of RNP storage and processing in the nucleus. Transient expression of each of these isoforms in HeLa cells and subsequent analysis of cell survival and SMN localisation provides strong evidence that a nine amino acid motif in exon 2b is necessary for correct CB targeting and further, that deletion of this motif has a dominant negative effect resulting in cell death. In addition his group have shown that this nine amino acid sequence contains the exon 2 self-association domain, and have also shown that sequences encoded by intact exons 3 and 6 are necessary for correct nuclear targeting [1]. Data from a second panel of GFP labeled constructs was also presented. These were used to express proteins encoded by individual exons 2b, 3 or 6, shortened proteins representing exons 1–5, 2a+2b, 2+3+6 or 6+7, or a double exon deletion isoform SMNΔ5Δ7 in HeLa cells and offered further evidence for specific roles for the previously identified functional motifs in self-association and sub-cellular targeting of the protein.

Next, Prof. Glenn Morris (RJAH, Oswestry) described his laboratory’s progress towards the production of novel monoclonal antibodies raised against components of the Gemin complex, a macromolecular SMN containing complex involved in the cytoplasmic assembly and nuclear import of snRNPs [2]. Prof. Morris presented work on the successful development and epitope mapping of multiple antibodies for each of the gemins 3–7. These were then used in conjunction with existing monoclonal antibodies for Gemin complex components SMN, unrip, fibrillarin, gemin 2 and profillin II to investigate sub-cellular distribution of complex components [3]. Data from sub-fractionation and immuno-histochemistry experiments were presented indicating that gemin 5 exhibits differential localisation patterns to the other major complex components leading to speculation that, although gemin 5 is necessary for snRNP formation in the cytoplasm, it is likely released from the SMN complex prior to import to the nucleus and CB.

The final presentation for this session was given by Dr. Judith Sleeman (University of St. Andrews, Fife). Dr. Sleeman provided data from on-going work analysing the relationship between SMN levels, nuclear gems and Cajal bodies (CBs) [4]. Her group’s work in the neural-like cell line SH-SY5Y has shown that CBs and gems associate more frequently when SH-SY5Y cells are differentiated through addition of both retinoic acid (RA) and brain-derived neurotrophic factor (BDNF). Data was also provided illustrating that over-expression of SMN alone is capable of increasing the association frequency of these two nuclear bodies in undifferentiated cells. SMN binds to the Sm proteins (core splicing snRNP components) and the CB marker, coilin. Symmetrical arginine dimethylation of coilin and Sm proteins is important for these interactions [5], [6], [7]. Interestingly, inhibition of methylation reduces the frequency of CB and gem co-localisation, with a more pronounced effect observed in differentiated cells. This suggests that methylation may be one mechanism involved in the increased association between gems and CBs observed in differentiated neural cell lines. Finally data was presented that depleted SMN levels achieved through siRNA-mediated knockdown can act to reduce the size and number of nuclear gems and also results in disruption of CBs.

3. SMN therapeutics 

Prof. Mimoun Azzouz (Sheffield University) began this session with an update on his work using lentivirus (LV) and adeno-associated virus (AAV) derived vectors to deliver therapeutic genes in neurodegenerative disorders. He has previously reported that rabies G pseudo-typed LV vectors encoding SMN are capable of delivering sufficient protein to partially rescue the disease phenotype in both SMA type I fibroblasts and a mouse model of SMA type I [8]. Continued refinement of this research strategy has lead Prof. Azzouz to develop a codon optimised construct for SMN, coSMN, which is designed to be translated at much higher efficiencies than the previously used human cDNA sequence. Data from trials using both LV and AAV based vectors to deliver the coSMN construct to SMA type I fibroblasts was then reported and shown to enhance SMN protein expression levels by a factor between two to three fold. On-going work aims to translate these improved SMN delivery methods to in vivo SMA models.

The second presentation in this session came from Summit Plc representative Natasha Thomas. Dr. Thomas gave an overview of how a fruit-fly model for SMA is being used as a reagent for high-throughput therapeutic screens of small molecules from the Summit Plc library. The major advantage of this screen over parallel strategies specifically focused on SMN up-regulation is the lack of a priori assumptions made on the mechanism of phenotypic rescue. To facilitate this process she described how the fly SMA model, which displays pathologically reduced dSmn levels resulting in motility defects and eventual death at late larval stages [9], has been adapted to allow large numbers of compounds to be screened efficiently. Several lead compounds have been identified through their ability to partially alleviate larval paralysis and survival using a food churning measure of activity. Phenotypic rescue conferred by these compounds was also assessed using larval body wall contraction rates and shown to have consistent effects. Data showing partial rescue of NMJ morphology in mutants treated with one of these compounds was also shown. Finally, plans were outlined to conduct further in-depth trials to confirm activity and establish the mechanism of action of these compounds and re-engineered analogues. These tests will include assays for phenotypic rescue in cell lines exhibiting siRNA-mediated knockdown of SMN, zebrafish injected with SMN knockdown morpolinos and eventually in genetically engineered mouse models of SMA.

The final presentation of this session was given by Haiyan Zhou (Hammersmith Hospital, Imperial College). Dr. Zhou gave an update on her work towards targeted oligonucleotide enhancer of splicing (TOES) based therapy for SMA aiming to correct SMN2 splicing and consequently increase full-length protein production [10]. The SMN2 locus exhibits a single coding nucleotide change in exon 7 with respect to SMN1. This change does not alter SMN protein sequence but does result in altered splicing dynamics causing the frequent skipping of exon 7 and the production of an unstable, truncated protein isoform, SMNΔ7, of reduced functional ability. Debate exists over the exact mechanism whereby this single base alteration confers this change in splicing dynamics. One model suggests that the consensus motif for an exonic splice enhancer (ESE) is disrupted by the substitution and thus prevents the binding of the splice recognition protein SF2/ASF. Dr. Zhou outlined the current status of the development and trail of bifunctional antisense oligoribonucleotides specifically designed to target exon 7 of SMN2 and, through the addition of a tail containing an SF2/ASF binding motif, aim to recruit SF2/ASF to the site. Data was presented showing that transfection of the most effective tailed antisense oligonucleotide in SMA fibroblast cell lines is capable of increasing full-length SMN at both RNA and protein level as measured by quantitative Real-Time PCR and Western blotting, respectively, and can also increase nuclear gem counts. On-going efforts towards resolving toxicity issues and increasing efficacy of tailed antisense oligonucleotide using chemical modification of the oligonucleotide were also outlined.

4. Oral poster session 

The second day of the meeting opened with an oral poster session. Delegates with static poster displays were invited to give a short oral presentation to outline and expand on their poster content. Seven exciting presentations were given. (1) Debra Shaw (Peninsula College of Medicine and Dentistry, Exeter) started with an update of her continued characterisation of the Ewing Sarcoma protein (EWS)/gemin protein complex in human cell lines. (2) Gary Todd (Peninsula College of Medicine and Dentistry, Exeter) then highlighted his current work analysing SMN-associated axonal transport complexes in cultured neuronal cell lines followed by (3) Nguyen thi Man (RJAH, Oswestry) presenting her latest data on the development of a two-site ELISA for measurement of SMN protein and its application to finding drugs for treatment of spinal muscular atrophy. Next, (4) Heidi Fuller (RJAH, Oswestry) gave an overview of her latest results describing the SMN interactome as detected through mass spectrometry followed by (5) Stuart Grice and Lin Lee (University of Oxford) presenting an update of their work researching SMN’s potential role in translational repression. (6) Birk Bäumer (University of Oxford) provided data to show that there is no evidence for direct genetic involvement of the feline or bovine SMA genes in human motor neuron disorders (7) and finally Nick Parkinson (University of Oxford) presented the latest data from his work into potential genetic interactions of SMN and SOD1.

5. Animal models 

The final session of the meeting was opened by Ji-Long Liu (University of Oxford). Dr. Liu has previously been involved in searching for the fruit fly orthologue of the Cajal body. His overview of this work described how this search led him to identifying the Cajal body and also to discover the histone locus body in Drosophila nuclei [11]. The former is involved in spliceosomal snRNP maturation, while the latter might be critical for histone mRNA processing. This research also led to the discovery of a cytoplasmic structure, the U body (for uridine-rich snRNP body), which he found to be present in numerous species from fly to human. His work on U bodies has shown that they contain both SMN and snRNPs [12]. Intriguingly, U bodies always associate with cytoplasmic processing bodies (P bodies) which contain mRNA degradation factors and RNAi machinery [12]. The interaction of U bodies and P bodies provides a fortuitous opportunity to investigate a potential role for SMN in regulatory networks of various small RNAs.

The next presentation was given by David Sattelle (University of Oxford) in which he provided an update on the development of the Caenorhabditis elegans (nematode) model of SMA. Prof. Sattelle gave a compelling overview of his SMA research that focuses on SMN-1, the worm orthologue of the human SMN protein. Knockdown of SMN-1 by RNA interference produces severe developmental and motility defects [13]. In addition, several new candidate SMN-interacting proteins have been identified in the worm [14]. Recently, the mutant worm line, smn-1(ok355), in which the smn-1 gene is almost entirely deleted, has been explored. This model expresses no endogenous SMN-1 protein but, as is also the case for the fly mutant [9], it can survive into late larval stages through the maternal contribution of functional SMN-1, thereby facilitating detailed studies of the progressive neuropathology associated with loss of the SMN protein. Striking defects in movement and pharyngeal pumping are detected. Prof. Sattelle and colleagues have used genetic ‘add-back’ experiments to deliver transgenically expressed SMN protein to neuronal and muscle tissues. First results suggest that neuronal, but not muscle, directed protein expression can partially alleviate the phenotype. In view of the extensive array of genetic, molecular and physiological tools available for this genetic platform, the C. elegans model represents an exciting resource for investigating the basic functions and interactions of SMN [15].

The penultimate talk of the session was delivered by Tom Gillingwater (University of Edinburgh). Dr. Gillingwater gave an update of his laboratory’s detailed characterisation studies of motor neuron pathology in both severe, SMN2 low [16], and intermediate, SMNΔ7 [17], mouse models of SMA. Dr. Gillingwater’s group have used a range of imaging techniques to visualise and quantify abnormalities in both axonal and neuromuscular junction (NMJ) architecture from levator auris longus, transversus abdominis, flexor digitorum brevis and deep lumbrical muscles at a range of disease stages in both SMA mouse lines [18]. He reported that significant pre-synaptic pathology was observed in all muscles from late-symptomatic mice of both strains, but that this loss was particularly acute in the postural slow-twitch muscle transversus abdominis in both strains, albeit much less severe in SMNΔ7 mice. Data was also presented highlighting the significant synaptic pathology present at both pre- and early symptomatic stages demonstrating that nerve loss is an early event in SMA. As well as confirming that NMJ pathology is not uniform between muscle groups his studies of the levator auris longus also showed that this heterogeneity can be apparent even within the same muscle. Detailed mapping of axonal inputs to various regions of this muscle suggested that relative susceptibility to pathology was linked to intrinsic differences in NMJ stability rather than muscle fibre type or motor unit size, with fast synapsing (FaSyn) NMJs appearing to be more at risk than their delayed synapsing (DeSyn) counterparts [18], [19].

In the final talk of the meeting Nick Parkinson (University of Oxford) gave an update on the characterisation of two novel mouse transgenic lines, SMNE134K and SMNΔ5. Dr. Parkinson began by describing the development of the SMNE134K transgenic mouse line in which the SMN1 single point mutation E134K, initially described in a severe type I SMA patient, has been over-expressed under the control of a shortened version of the human SMN1 promoter. This single amino acid change in the Tudor domain of the protein is predicted to disrupt Sm protein binding and hence prevent the formation of snRNPs prior to their import into the nucleus [20]. Dr. Parkinson showed that high levels of this protein isoform expressed throughout the mouse had no effect on a wild-type background and completely failed to extend survival when crossed onto the SMN2 low line, a reproducible model of type I SMA [16], indicating that the protein has no significant function. This finding was further backed up by parallel research over-expressing the same mutation in the zebrafish model system. Dr. Parkinson then went on to present the latest data from the SMNΔ5 transgenic mouse line that expresses a naturally occurring splice isoform of human SMN missing exon 5. This protein lacks the majority of the poly-proline repeat motif necessary for Profillin IIa binding and hence potentially implicates SMN in actin dynamics and cytoskeletal modeling [21]. Like the SMNE134K line this transgene has no effect on a wild-type background and is unable to rescue the Smn knockout mouse [22] suggesting that the SMNΔ5 protein is functionless. In contrast, when bred onto the SMN2 low background [16] the transgene is capable of rescuing the lethal phenotype from an average 5 days to over 90 days. These data suggest a functional role for the SMNΔ5 isoform but only in the context of other SMN species. On-going work aims to further define these interactions and assess the effect of this transgene on Profillin IIa localisation and expression levels.

Throughout the meeting and following the final presentation lively discussion took place identifying key goals for the UK SMA research community in the coming year. Additionally, drinks and an evening meal generously provided by the Jennifer Trust for SMA at the end of the first day afforded an excellent backdrop for the establishment of many new collaborative ventures between research groups from across the UK.

	Participants
	Oxford University Department of Physiology,
	Anatomy and Genetics
	Mara Almeida
	Dirk Bäumer
	Ruben Cauchi
	Sandrine Fraboulet
	Stuart Grice
	Will Motley
	Nick Parkinson
	David Sattelle
	Kevin Talbot
	Lin Lee
	Ji-Long Liu
	Institute of Biomedical and Clinical Science,
	Peninsula College of Medicine and Dentistry, Exeter
	Robert Morse
	Debra Shaw
	Gary Todd
	Philip Young
	Wolfson Centre for Inherited Neuromuscular Disease,
	Robert Jones & Agnes Hunt Orthopaedic
	& District Orthopaedic Hospital, Oswestry
	Heidi Fuller
	Glenn Morris
	Nguyen thi Man
	Centre for Integrative Physiology
	& Centre for Neuroscience Research,
	University of Edinburgh Old Medical School
	Tom Gillingwater
	Lyndsay Murray
	Jennifer Trust
	Richard Green
	Max Huxham
	Academic Unit of Neurology, Medical School,
	Sheffield University
	Mimoun Azzouz
	Dept of Gene Therapy, Division of
	Medicine, Imperial College
	Nicholas Mazarakis
	Dubowitz Neuromuscular Unit,
	Imperial College
	Haiyan Zhou
	University of St Andrews
	School of Biology, St Andrews
	Judith Sleeman
	Summit Plc, Oxford
	Natasha Thomas
	SMA Trust
	Sir James Gowans