S-Nitrosylation of MEF2: a common mechanism underlying neurological disorders?

KEYWORDS:

• adult neurogenesis
• Alzheimer’s disease
• autism
• drug screening
• nitric oxide
• Parkinson’ disease
• reactive nitrogen species
• redox signaling
• S-nitrosylation
• stroke

“…pharmacological intervention to normalize or preserve MEF2 function may prove to be a promising approach for a variety of neurological disorders.”

Figure 1.  Context- and disease-dependent pathways regulated by SNO-MEF2.

Aberrant S-nitrosylation of MEF2 (forming SNO-MEF2) controls neuronal differentiation and survival in context- and disease-dependent manners [8,9]. The SNO-MEF2C/PGC1α pathway is involved in death of A9 dopaminergic neurons in Parkinson’s disease [8]. The SNO-MEF2C/Bcl-xL pathway contributes to cerebrocortical neuron death in stroke [9]. The SNO-MEF2A/TLX cascade negatively regulates adult neurogenesis in Alzheimer’s disease [9].

Physiological levels of reactive nitrogen species (RNS), such as nitric oxide (NO), act as signaling mediators, creating versatile redox signaling networks. Post-translational oxidative modification of protein thiol by RNS, termed S-nitrosylation (or chemically, S-nitrosation), has emerged as one of the essential mechanisms for modulating protein function; subsequent or alternative oxidation by reactive oxygen species can also contribute to this form of regulation [1]. Under certain pathological conditions, including exposure to environmental neurotoxins, excessive NO is produced, generating nitrosative stress with consequent aberrant S-nitrosylation of protein thiols that would otherwise not be affected in this manner [1]. We have observed this type of aberrant S-nitrosylation in a variety of neurodegenerative diseases, including Parkinson’s disease (PD), Alzheimer’s disease (AD), amyotrophic lateral sclerosis, autism spectrum disorders (ASDs), vascular dementia and stroke [1]. In some instances, the aberrant nitrosylation reaction can mimic or enhance the effect of mutations of the gene encoding the S-nitrosylated protein, and thus sporadic cases of disease can simulate rare familial forms [1]. Here, we discuss our recent findings regarding S-nitrosylation of the myocyte enhancer factor 2 family of transcription factors (MEF2A–D), which we have shown are essential regulators of neuronal differentiation, maturation and survival [2–4]. The resulting formation of SNO-MEF2 contributes to the pathogenesis of various neurological disorders.

The cause of sporadic PD is not completely elucidated. However, epidemiological studies have implicated environmental and genetic interactions as contributing factors to sporadic PD. In particular, co-exposure to the pesticides paraquat (PQ) and maneb (MB) has been implicated as a significant risk factor for the development of PD [5], although this evidence has remained contentious. Interestingly, PQ/MB exposure exacerbates PD-like phenotypes displayed by transgenic mice overexpressing a mutant form of the α-synuclein (SNCA) gene; this mutation, A53T α-synuclein, causes a dominantly familial form of PD in humans [6]. However, given that none of the existing PD animal models mimics the human disease perfectly, and the fact that differences in genetic background often complicate data interpretation in transgenic rodents, it is of critical importance to generate better PD models in human context.

We and our colleagues recently addressed these issues by generating an isogenic human induced pluripotent stem cell (hiPSC) model of PD in which the disease and the control samples are genetically identical except for the SNCA locus, one allele of which carries the A53T mutation in the diseased cells but not in the genetically corrected cells [7,8]. We used this robust system to probe the influence of genetic factors and pesticide exposure on PD pathogenesis. We differentiated A53T α-synuclein mutant hiPSCs and control corrected hiPSCs into A9-type dopaminergic (DA) neurons, the neuronal type most susceptible in the early phases of PD. We found that the A53T neurons exhibited increased oxidative/nitrosative stress even under basal conditions. Moreover, acute exposure to PQ/MB mimicked these observations in the normal A9 DA neurons. Importantly, low-dose PQ/MB exposure that did not affect normal A9 DA neurons led to increased nitrosative stress and cell death in A53T A9 DA neurons, showing that the α-synuclein mutant neurons were more susceptible to pesticides.

Global gene expression studies revealed that targets of the transcription factor MEF2C were among the genes whose expression was most affected by the presence of mutant A53T α-synuclein. Mechanistically, we found that increased RNS caused S-nitrosylation of MEF2C protein (forming SNO-MEF2C) at cysteine residue 39, thus reducing its DNA binding and transcriptional activation. As a result, this process inhibited transcription of MEF2C target genes, in particular the critical protein PGC1α, a guardian of mitochondrial integrity and bioenergetics [8]. Aberrant regulation of the MEF2C-PGC1α pathway due to S-nitrosylation led to mitochondrial dysfunction, which in turn contributed to A9 DA neuronal apoptosis. Remarkably, a non-nitrosylatable form of MEF2C restored the activity of this pathway and protected A9 neurons from pesticide exposure (Figure 1). Thus, this approach allowed us to elucidate a novel molecular pathway that links GxE (genes by environment) to mitochondrial dysfunction, oxidative stress and neuronal cell death in PD [8].

The pathogenic involvement of SNO-MEF2 is not limited to PD. We have also observed aberrant SNO-MEF2 in a mouse model of stroke (middle cerebral artery occlusion [MCAO]) and in transgenic AD mouse models as well as human patient brain [9]. As previously shown, excess NO promotes neuronal cell death in stroke [9]. Using primary cortical neurons and Mef2c knockout mice, we found that MEF2C controls expression of Bcl-xL, an antiapoptotic member of Bcl2 family proteins, in vitro and in vivo [9]. Formation SNO-MEF2C reduces Bcl-xL expression, leading to NO-mediated neuronal cell death [9]. This defined MEF2C-Bcl-xL cascade encountered in cerebrocortical neurons after stroke is different from the MEF2C-PGC1α cascade found in PD hiPSC-derived A9-type DA neurons [8], demonstrating the existence of disparate context- and disease-dependent prosurvival pathways triggered by MEF2 (Figure 1). In addition, expression of non-nitrosylatable MEF2C diminished neuronal cell death after MCAO stroke in mice, consistent with the notion that SNO-MEF2C contributes to neuronal death in vivo in stroke models [9]. Importantly, this was the first report to demonstrate that S-nitrosylation of MEF2 mediates neurodegeneration in vivo [9].

Additionally, our recent work has shown that SNO-MEF2 also modulates adult neurogenesis, and this effect can contribute to disease states [9]. Adult neurogenesis occurs throughout life in the dentate gyrus of the hippocampal formation, even in humans. A recent study assessing integration of nuclear-bomb-test-derived ¹⁴C in genomic DNA estimated that 700 new neurons are generated in the hippocampus per day, and approximately one-third of hippocampal neurons are reconstituted during adult life in humans [10]. This form of adult neurogenesis has been found to be compromised in several neurological disorders, including AD [11]. Additionally, NO has been reported to negatively regulate adult neurogenesis [12]. Our laboratory recently found that MEF2A is a major MEF2 isoform in adult neural progenitors [9], and Mef2a knockout mice have decreased adult neurogenesis in the hippocampus [9]. Experimentally, we observed that expression of non-nitrosylatable MEF2 in vivo ameliorates the defect in adult neurogenesis in the TG2576 mouse model of AD [9]. Coupled with the observation of aberrantly S-nitrosylated MEF2 in human AD brains, this finding is consistent with the notion that SNO-MEF2 inhibits adult neurogenesis in AD.

In light of the recent development of CRISPR/Cas9 gene editing technology, introduction of non-nitrosylatable MEF2 is now achievable at the genome level [13]. Hence, in the future CRISPR/Cas9 technology should allow further validation of the importance of S-nitrosylated MEF2 in the pathogenesis of neurodegenerative diseases. For example, it will be feasible to engineer an MEF2(C39A) mutant into hiPSCs from AD and PD patients and to analyze its effect on pathological phenotypes. In addition, it will be intriguing to generate knock-in mice with the MEF2(C39A) mutation and cross them with AD or PD mouse models. These ‘disease-in-dish’ and ‘in vivo knock-in mice’ models will allow us to further test the effect of aberrant MEF2 nitrosylation in various neurodegenerative diseases.

The pathogenic contribution of SNO-MEF2 in downregulating normal transcriptional activity in both hiPSC and mouse models of disease indicates that MEF2 is a potential target in the treatment of neurodegenerative diseases. Exploiting this premise and aiming at developing a therapeutic, we recently developed a high-throughput screening assay for small molecules capable of modulating MEF2 activity in human neural cells [8]. We performed high-throughput screening in a human context (using human neural progenitor-derived cells) on several different chemical library collections, including pharmacologically active compounds (LOPAC 1280), known kinase/phosphatase inhibitors, and pharmacologically active and structurally diverse small molecules related to stem cells (StemSelect). One of the top hits was isoxazole [8]. Notably, isoxazole stimulates MEF2 transcriptional activity and rescues PD patient hiPSC-derived neurons from mitochondrial toxins [8]. In addition, Petrik et al. recently reported that administration of isoxazole increases adult hippocampal neurogenesis and improves memory in mice in an MEF2-dependent manner [14]. MEF2 proteins dimerize and bind to DNA [15], so small molecules that affect this dimerization may also modulate MEF2 transcriptional activity. These findings indicate that modulation of MEF2 activity is a feasible approach for developing novel therapeutics for neurodegenerative diseases.

We have also begun to investigate an alternative strategy for modulating MEF2 activity involving molecules that inhibit or reverse the aberrant S-nitrosylation of MEF2, since S-nitrosylation disrupts MEF2 binding to DNA and thus transcriptional activity [9]. Our structural analysis of the MEF2/DNA complex revealed a pocket in close proximity to the nitrosylated cysteine [9]. Small compounds that fit in the pocket and prevent the S-nitrosylation may preserve MEF2 transcriptional, neurogenic and prosurvival actions in the face of nitrosative stress in neurodegenerative diseases.

Interestingly, oxidative and nitrosative stress are also associated with neurodevelopmental disorders such as autism and intellectual disabilities [16,17]. Our group and Eric Olson’s laboratory reported that Mef2c knockout mice display autistic phenotypes or learning and memory deficits, respectively [18,19]. Following these reports, an ever increasing number of clinical reports have found deletions or mutations in the MEF2C gene in patients with ASD and intellectual disability [4]. In addition, recent integrative functional genomic analyses identified MEF2C as one of the hub transcription factors controlling risk genes for ASD [20]. Based on some very preliminary data that we have obtained on human brains, we speculate that not only genetic mutations but also possibly S-nitrosylation of MEF2C by environmental factors may contribute to the development of ASD and other intellectual disabilities. Thus, small molecules that inhibit aberrant S-nitrosylation of MEF2 may one day be considered for the treatment of neurodevelopmental disorders such as ASD.

Excessive RNS are produced in a number of neurodevelopmental and neurodegenerative disorders. Our recent work demonstrates that aberrant formation of SNO-MEF2 is a potential unifying mechanism underlying the pathogenesis of several neurodevelopmental and neurodegenerative disorders. Hence, pharmacological intervention to normalize or preserve MEF2 function may prove to be a promising approach for a variety of neurological disorders.

Acknowledgements

The authors thank members of the Lipton laboratory, including Sarah Moore, Rameez Zaidi and Scott R McKercher, for their discussions and assistance with this manuscript.

Financial & competing interests disclosure

S-i Okamoto was supported by a Shiley-Marcos Alzheimer’s Disease Research Center (UCSD) Pilot Award and NIH grant R21 MH102672. R Ambasudhan was supported by a United Mitochondrial Disease Foundation grant (13-102). SA Lipton and work reported herein was supported in part by NIH grants R01 NS086890, P01 HD29587, P01 ES016738 and P30 NS076411, and awards from the Michael J Fox Foundation, the California Institute for Regenerative Medicine (CIRM TR4-06788) and the Brain and Behavior Research Foundation. The authors disclose that SA Lipton and S-i Okamoto are the named inventors on an issued patent describing methods to modulate MEF2 for neuroprotection, and this patent is assigned to their institution, the Sanford-Burnham Medical Research Institute. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.