- Top of page
- MATERIALS AND METHODS
- Author contributions
- For more information
- Supporting Information
Spinal muscular atrophy (SMA) is a neurodegenerative disorder that presents as progressive muscle wasting and loss of motor function. It is caused by the degeneration of motor neurons, specifically the anterior horn cells of the spinal cord and is one of the leading heritable causes of infant mortality worldwide (Crawford & Pardo, 1996; McAndrew et al, 1997; Pearn, 1978). SMA is caused by a deficiency of the SMN protein. There are two nearly identical SMN genes, the telomeric SMN1 and the centromeric SMN2 (Boda et al, 2004; Echaniz-Laguna et al, 1999; Monani et al, 1999). While the protein coding capacity of SMN2 is identical to that of SMN1 (Jablonka et al, 2000a), there is a translationally silent nucleotide variation in exon 7 of SMN2 (Lorson et al, 1999; Monani et al, 1999). This C to T transition results in alternative splicing of SMN2 and exclusion of exon 7. From SMN1, >95% of the transcripts include exon 7 and express the full-length SMN protein. From the SMN2 mRNA, ∼85% of the messages lack exon 7 (Gavrilov et al, 1998; Gennarelli et al, 1995; Lorson et al, 1999; Monani et al, 1999) and express a truncated form of the protein (SMNΔ7). The SMNΔ7 protein is inactive and cannot fully compensate for the loss of SMN1 (Burnett et al, 2009; Lorson & Androphy, 2000; Lorson et al, 1998).
SMN2 is a potent disease modifier for SMA, and there is an inverse relationship between the number of copies of SMN2 and clinical severity. Most cases of SMA harbour homozygous deletions of the SMN1 gene but retain at least one copy of SMN2 (Brahe et al, 1996; Campbell et al, 1997; Hahnen et al, 1995-1997; Jablonka & Sendtner, 2003; Melki, 1997; Talbot et al, 1997; van der Steege et al, 1996; Velasco et al, 1996). The relationship between SMN2 copy number and disease severity has been confirmed in SMA mouse models (Hsieh-Li et al, 2000; Michaud et al, 2010; Monani et al, 2000). Homozygous deletion of the single copy of the mouse Smn gene is embryonic lethal (Schrank et al, 1997). Introduction of two copies of the human SMN2 transgene supports viability but these animals have motor function defects and an average life span of 4–6 days. Increasing the number of SMN2 copies decreases disease severity and increases life span. High copy number SMN2 transgenic mice were phenotypically ‘normal’ (Monani et al, 2000).
Because SMA carriers with only one copy of SMN1 are clinically asymptomatic, 50% of normal SMN levels should protect from disease. If SMN2 can be stimulated to express more full length SMN mRNAs, synthesis would be directed towards increased amounts of the active SMN protein (Cherry & Androphy, 2012). Although the threshold level of SMN necessary to maintain motor neurons is not known, only 10–15% of SMN2 transcripts contain exon 7 and express functional SMN, so doubling or tripling the amount of full length SMN2 mRNA should be clinically significant (Meyer et al, 2009).
There is no treatment for SMA. Therapeutic modalities for treatment of SMA that are being actively pursued include oligonucleotides to restore SMN2 exon 7 inclusion, gene transfer using viral vectors, and cell replacement with motor neuron differentiated stem cells (Corti et al, 2008, 2010; DiDonato et al, 2003; Dominguez et al, 2011; Foust et al, 2010; Hua et al, 2010, 2011; Passini et al, 2010, 2011; Porensky et al, 2012; Valori et al, 2010; Williams et al, 2009). Several laboratories have undertaken screens for drug-like compounds that increase cellular levels of the SMN protein from the SMN2 gene. Compounds that have been shown to increase SMN2 expression include various histone deacetylase (HDAC) inhibitors, aclarubicin, indoprofen, splicing modifiers, a DcpS inhibitor, anti-terminators, proteasome inhibitors and inhibitors of multiple signalling pathways (Andreassi et al, 2001; Avila et al, 2007; Bowerman et al, 2010, 2012; Burnett et al, 2009; Chen et al, 2012; Farooq et al, 2009; Garbes et al, 2009; Hahnen et al, 2006; Hastings et al, 2009; Heier & DiDonato, 2009; Jarecki et al, 2005; Kernochan et al, 2005; Kwon et al, 2011; Lunn et al, 2004; Makhortova et al, 2011; Narver et al, 2008; Singh et al, 2008; Wolstencroft et al, 2005; Zhang et al, 2001, 2011). Because many of these activators are non-specific and can have off-target effects, their long-term safety remains to be determined. Compounds that have advanced into clinical trials have demonstrated mixed results. There is clearly need for additional drug candidates (Darras & Kang, 2007; Sproule & Kaufmann, 2010; Sumner, 2006).
We previously reported the development of an SMN2-luciferase reporter assay to identify compounds that increase SMN expression from the SMN2 gene (Cherry et al, 2012). This assay has been used at two screening centres to screen over 300,000 compounds (Cherry et al, 2012; Xiao et al, 2011). From a screen of 115,000 compounds at the Laboratory for Drug Discovery in Neurodegeneration (LDDN), 462 hits were identified and 19 ‘high’ priority compounds were selected on the basis of their activity, potency, specificity, lack of overtly toxic functional groups, and potential tractability for chemical modification. Here we report the selection and further characterization of two compounds as potential leads for new SMA therapies.
- Top of page
- MATERIALS AND METHODS
- Author contributions
- For more information
- Supporting Information
SMA is primarily caused by the loss or mutation of both copies of the SMN1 gene. The SMN gene is required for viability (Schrank et al, 1997) and the copy number of SMN2 inversely correlates to disease severity (Feldkotter et al, 2002; Harada et al, 2002; Monani et al, 2000; Wirth et al, 1999). The potential to express functional full-length SMN protein and the presence of at least one copy in nearly all SMA patients makes SMN2 an attractive therapeutic target for the treatment for SMA. Although the threshold level of SMN protein necessary to maintain motor neurons has not been adequately determined, it is reasonable to predict that doubling or tripling the amount of full length SMN protein should be sufficient to prevent or diminish clinical progression. Motor neurons appear intact at birth in SMA murine models, so there appears to be a temporal window for restoration of SMN and survival of motor neurons (McGovern et al, 2008). Consistent with this, an inducible mouse model of SMA demonstrated that whole-body restoration of SMN decreases disease severity even if induction occurs after the onset of symptoms (Lutz et al, 2011).
As a potent disease modifier for SMA, SMN2 has become a high priority target for SMA therapeutics. Many of the compounds in development are either general neuroprotective compounds or compounds re-purposed from other indications. HDAC inhibitors including sodium butyrate, SAHA, phenyl butyrate, trichostatin A (TSA), LBH589 and M344, are the most commonly studied class of SMN inducers and have all been reported to increase SMN transcription (Andreassi et al, 2004; Avila et al, 2007; Brichta et al, 2003; Chang et al, 2001; Garbes et al, 2009; Hahnen et al, 2006; Riessland et al, 2006; Sumner et al, 2003). Sodium butyrate and SAHA have also been shown to increase exon 7 inclusion in SMN2 transcripts. The HDAC inhibitor VPA has also been characterized as a splicing enhancer for SMN2 (Brichta et al, 2003).
SMN2 based reporters have been used to identify novel compounds for the treatment of SMA. Our original splicing cassette was used in low-throughput mode to identify the phosphatase inhibitor sodium orthovanadate (Zhang et al, 2001), aclarubicin (Andreassi et al, 2001) and indoprofen (Lunn et al, 2004). These compounds have been shown to increase full-length SMN protein expression from the SMN2 gene. Sodium orthovanadate, a phosphatase inhibitor, might affect SR (serine/arginine rich) protein phosphorylation state and thereby modulate SMN splicing (Zhang et al, 2001). Interestingly, both aclarubicin and sodium orthovanadate activities are enhanced in the presence of the transcription factor Stat5 (signal transducers and activators of transcription 5), while Stat5 knockout abrogates their effect (Ting et al, 2007). The hormone prolactin can activate Stat5, which in turn increases SMN expression (Farooq et al, 2011). In the same study, prolactin treatment in the SMNΔ7 mice improved gross motor function and increased survival. This suggests that SMN expression may be regulated by a signal transduction pathway in response to cytokines or growth factors. Indoprofen, a non-steroidal anti-inflammatory drug, was also identified using our first generation reporter assay. The mechanism of action for indoprofen has not been determined but recent evidence suggests that it has anti-terminator activity. Aminoglycosides also have anti-terminator activity and have been shown to stabilize the SMNΔ7 protein and presumably increase its functionality (Heier & DiDonato, 2009; Mattis et al, 2008). An independent SMN2 promoter screen was used to identify the 2,4 diaminoquinazoline series of compounds (Jarecki et al, 2005). The quinazoline compounds may act by binding to and inhibiting the scavenger decapping enzyme, DcpS (Singh et al, 2008). Other modes of action being explored include proteasome inhibition, SMN stabilization, inhibition or activation of signal transduction, and targeted regulation of exon 7 inclusion and are currently at different stages of development.
We previously reported a new screen that combined the benefits of the splicing and transcriptional assays for SMN2 expression (Cherry et al, 2012). In that report, we described three novel compounds that increased SMN expression in the reporter cells and were confirmed in 3813 SMA derived fibroblasts. Early medicinal chemistry efforts revealed that these compounds lacked the characteristics desired for further development. Here we characterize two new scaffolds, LDN-75654 and LDN-76070, which induce expression of full-length SMN from SMN2. Both compounds increase SMN-luciferase expression in the reporter assay and levels of endogenous SMN protein in 3813 SMA derived primary fibroblasts with low micromolar EC50s. Neither of the compounds displayed inhibition of HDAC 3, 6 or 8 activity in vitro. These compounds also displayed selectivity for SMN2-luciferase expression when tested in two additional cell lines for specificity, SMN1-luciferase and the SV40min-luciferase. These compounds have been tested in other high throughput screens without being identified as hits, suggesting the utility of the multi-faceted screening platform.
RT-PCR analysis confirmed that LDN-76070 and LDN-109657 increased the amount of total SMN-luciferase fusion transcripts, with a concomitant increase in the amount of full-length transcripts (Fig 2B). LDN-76070 does not share structural similarity with other compounds shown to increase SMN expression. The pathway and targets involved in its activity are still unknown. LDN-75654 also increased the amount of detectable SMN-luciferase fusion protein, but it promoted little change at the mRNA level. This suggests that LDN-75654 functions post-transcriptionally, perhaps by increasing translation efficiency or decreasing SMN protein turnover. Post-transcriptional regulation of SMN expression is not without precedent. The aminoglycosides are anti-terminators that act post-transcriptionally by allowing read through of the stop codon in exon 8 of the SMNΔ7 transcript (Mattis et al, 2006, 2009; Wolstencroft et al, 2005). The additional amino acids produced by this read-through stabilize the SMNΔ7 protein product. Due to the design of our screen, activity of anti-terminators like the aminoglycosides would not be detected. LDN-75654 is not structurally similar to aminoglycosides or other compounds reported to regulate SMN expression. However, the oxazole carboxamide backbone of LDN-75654 is similar to that of leflunomide, a pyrimidine synthesis inhibitor used to treat rheumatoid arthritis (Bartlett et al, 1991; Ruckemann et al, 1998). Leflunomide was also active in our reporter screen and was identified independently using our SMN2-luciferase reporter cells (PubChem CID = 3899). We are currently examining the nature of this similarity.
Investigation of the combinatorial effects of these scaffolds could provide further insight into their general mechanism of action and might enhance their therapeutic efficacy. Combining LDN-75654 with the transcriptional activators LND-76070, LDN-109657, or SAHA produced a greater than additive stimulation of the SMN2-luciferase reporter. This effect was most apparent when LDN-75654 was combined with SAHA, which has been shown to increase both SMN transcription and exon 7 inclusion. We propose that the complementation seen with these compounds is a confirmation that the compounds are working through separate and distinct mechanisms to induce SMN2 expression. The effect of LDN-76070 was partially masked by the addition of SAHA. SAHA appears to overwhelm the transcriptional machinery and blunt the efficacy of LDN-76070, suggesting that these compounds stimulate SMN2-luciferase expression through similar or overlapping pathways. The observation that these two compounds still produce a greater than additive effect in the constant ratio experiments may be due to the secondary effect that SAHA has on the splicing efficiency and inclusion of exon 7. The dramatic increase in SMN2-luciferase activation observed with the combination of SAHA with LDN-75654 suggests that LDN-75654 augments SMN expression in addition to the splicing and transcriptional increases produced by SAHA, possibly through effects on protein stability or translation.
Systematic delivery of LDN-76070 promoted greater than threefold increases in SMN protein levels in the brain and spinal cord, suggesting the ability to penetrate the blood brain barrier. It dramatically increased lifespan and gross motor function in these animals as evidenced by the improvements in TTR tests. LDN-76070 displayed excellent stability in mouse liver microsomes, half-life of 40 min, but has low aqueous solubility. While LDN-75654 promoted increases in SMN expression in both the reporter assay and SMA derived fibroblasts, treatments with this compound in pilot experiments with the SMNΔ7 mice resulted in weak and inconsistent increases in SMN protein levels. LDN-75654 had a relatively short half-life of 15 min in in vitro mouse liver microsome assays. We assume that this lack of metabolic stability accounts for its poor activity in the animals.
There has been an unprecedented increase in the therapeutic pipeline for SMA (Cherry & Androphy, 2012; Lorson & Lorson, 2012). Unlike early studies focused on neuroprotective agents including carnitine, riluzole and more recently olesoxmine and ceftriaxone, the compounds describe in this manuscript are designed to target SMN2 expression. The largest class of compounds that is currently being studied for their ability to increase SMN expression is the HDAC inhibitors. These have shown modest extension of lifespan in SMA model mice (Chang et al, 2001; Riessland et al, 2010), however, VPA and phenylbutyrate failed to exhibit significant clinical efficacy in human SMA protocols. Other drug-like small molecules that have increased lifespan of SMA mice include an orally bioavailable quinazoline analog, the aminoglycoside TC007, and prolactin (Butchbach et al, 2010; Farooq et al, 2011; Mattis et al, 2009).
LDN-76070 induces SMN2 transcription, while LDN-75654 acts post-transcriptionally. Neither inhibited HDAC activity in vitro. LDN-76070 treated animals lived an average 11 days longer than DMSO treated littermates, representing an increase of lifespan of more than 180%. Despite the apparent DMSO toxicity, LDN-76070 treatment still increased the lifespan of these animals by 48% over that of untreated animals. The chemical scaffolds described here, derived from a high-throughput screen of a chemical diversity library to increase SMN protein, show promise as these compounds exhibited the predicted activities in mice. Further work is necessary to improve their pharmacological properties and determine their precise mode of action.