Most neurodegenerative diseases, including Alzheimers, Parkinson and Amyotrophic lateral sclerosis, have a common underlying etiology in their pathogenesis; a protein(s) misfolds and accumulates in an intracellular or extra cellular environment 
. Concomitantly there is a series of poorly understood events that lead to degeneration of a selective subset of cells resulting in the loss of function. In ALS, several aberrant regulatory processes have been reported, such as, mitochondrial dysfunction 
, vacuolization 
, glutamate cytotoxicity 
, axonal retrograde transport 
and oxidative stress 
. Some of these regulatory pathways have been linked to one another in a sequential manner 
, while there are many that are obviously not connected. In spite of these complexities, there are many common features of disease pathogenesis between sporadic and familial, SOD1-dependent, ALS 
suggesting that common, but unknown, molecular mechanisms are at work. Identifying and deciphering these connections is thus vital to understanding the disease and for effectively developing therapies.
In the present work, this issue has been approached by starting from a known attribute; that misfolded mutant SOD1 is linked to familial amyotrophic lateral sclerosis and utilizing a functional genomic approach to identify cellular proteins involved in regulation of this attribute. A cell-based, β-galactosidase assay that monitors SOD1 solubility/folding in cells, 
was adapted for the whole genome RNAi screening. Earlier studies 
had demonstrated that blockage of the proteasome by MG132, leads to the elevation of the assay signal due to accumulation of SOD1. However, inhibitors were not suitable for optimizing the assay for an RNAi screen, because of the differences in the treatment procedures. Thus, siRNA targeting S4 (PSMC1), a component of 26 S proteasome was selected and used for assay optimizations (). Screening for modulators of SOD1 solubility/folding was carried out utilizing a whole genome RNAi library containing siRNA pools targeting 21,125 targets, thus providing a near comprehensive scan of genomic proteins. The screen results were processed using two independent statistical methods (MAD score and Z score) minimizing the influence of any anomalies introduced by a particular approach (). The screen results underrepresented the hits that lowered the level of SOD1 below the typical hit selection threshold of −3SD due to a threshold effect. The range of MAD score values reflects this phenomenon (13 MAD to ~−6 MAD). Thus, a relaxed hit selection criterion of −2SD was selected to compensate. When the hits that belonged to the selection criteria were classified in to functional categories, it revealed that proteins from the G-protein coupled receptor super family were involved in both up and down-regulation of soluble SOD1 levels (), demonstrating the complexity of the GPCR signaling network. Further, the occurrence of transcriptional and translational regulatory proteins among the top classes of protein targets involved in up or down regulation of SOD1 was consistent with their role in controlling protein expression levels. In addition, the leading biological processes including GPCR signaling and transcriptional regulation from the RNAi screen, represented many of the cellular modulators of superoxide dismutase 1 identified from an earlier expression cloning screen 
. Human counterparts of 6 of the 10 murine proteins 
that were identified in the earlier screen (Table S2
) had MAD scores indicating decreased SOD1 expression upon knockdown. Thus, levels of these targets correlate with levels of SOD1; increased expression leads to increased SOD1 and decreased expression, reduces SOD1.
Pathway analyses of the hits from the RNAi screen revealed many protein interaction networks. The network named “Skeletal and Muscular System Development and Function, Tissue Morphology, Inflammatory Response" with the highest score contained over 40 proteins that met the hit criteria (). A protein of particular interest in the network because of its relationship to the ALS was the TAR DNA binding protein or TDP-43. TDP-43 is a 414 amino acid, 43 kDa protein that regulates gene expression at the level of transcription, splicing and mRNA stability 
. Aggregates of TDP-43, including phosphorylated and truncated forms are found in sporadic and many familial forms of ALS 
. The molecular mechanism(s) by which misfolding of TDP-43 or formation of its aggregates lead to neurodegeneration is poorly understood. A linkage between TDP-43 and SOD1 is suggested not only by the related pathology, but also because of apparent reciprocal relationship of the two proteins in inclusions 
. TDP-43 aggregates are not observed in familial SOD1-associated ALS patients that harbor SOD1 inclusions. However, it should be noted that TDP-43 has been found to be mis-located to the cytosol 
in some of these patients. It has also been hypothesized that SOD1 dependent and independent ALS (sporadic and other forms of familial ALS) arise from independent mechanisms 
In addition to TDP-43 protein, other TDP-43 associated gene products in the top network also significantly affected SOD1 levels upon knockdown, highlighting the relevance of the network in SOD1 regulation (, Table S1
). In addition SETDB1, a histone methyl transferase linked to TDP-43 in the network based on its interaction identified in a mass-spectrometry based proteomics study 
significantly affected SOD1 levels (MAD score 5.3) upon its knockdown. SETDB1 knockdown increased the soluble SOD1 levels both in the screen ( and Table S1
) and in independent experiments (data not shown). Further, RBMX, a hetero-nuclear Ribonucleo protein (hnRNP) identified among the hits in the top network with a MAD score of 9.9, belongs to the same family of RNP's that has been reported to interact with TDP-43 
. These observations further support the involvement of TDP-43 in the regulation of SOD1.
Validation experiments confirm that the action of TDP-43 on SOD1 observed in the RNAi screen. The lowering of the TDP-43 levels with siRNA treatment resulted in the concomitant increase in SOD1 levels (). A converse experiment showing increased TDP-43 expression resulted in a suppression of SOD1 expression as measured by β-gal activity (). The inverse correlation between increased TDP-43 and decreased SOD1 expression suggests a repressor role for TDP-43. Such a repressor function of TDP-43 on select gene targets has been observed by others 
. The validation of 5 out of 11 target genes from the TDP-43 linked protein interaction network as regulators of SOD1 expression (Figure S4
A, B) adds further credence to the importance of the network (). The current finding that TDP-43 regulates SOD1 is also consistent with a recent micro array study, where a modest increase in SOD1 mRNA levels was observed upon TDP knockdown 
. The increase in SOD1 reporter mRNA levels as observed by qPCR analysis () upon TDP-43 knockdown could result because of alterations in transcription or mRNA stability, brought about by TDP-43. However the effect could also be mediated indirectly through an unknown factor which may or may not involve splicing, another known function of TDP-43. As the above experiments were carried out using artificial cDNA SOD1 fusion constructs, such a mechanism would not be observed. Thus, future work will be required to address these possibilities and their importance in the details of the mechanism of TDP-43 repression.
The data presented here () revealed an unexpected regulatory link associating TDP-43 and SOD1. Interesting questions for future work will be to determine whether mutant TDP-43 or the aggregated forms of wildtype TDP-43 lose or gain function and how these processes might relate to SOD1 and the dominant toxic-gain-of-function phenotype.