Microsatellites are stretches of DNA abundantly interspersed in the genome of prokaryotes and eukaryotes ^[1][2][1,2], including humans, where they account for the 3% of the genome [3]. Their structure consists of short, tandemly repeated duplications of 1–6 base pairs, spanning between 20–100 bases [4] and primarily consisting of mono- and dinucleotides, although tri-, tetra-, penta-, and hexa-nucleotides microsatellite classes are present ^[4][5][4,5]. Their location is ubiquitous, occurring both in protein-coding as well as in non-coding DNA regions, with preference for the latter [6], and their functions range among several biological regulatory processes [7], including alternative splicing [8], transcription start/end site selection ^[9][10][9,10], nucleosome packaging [11], and methylation [12]. One of the most peculiar characteristics of microsatellites is their tendency to mutate. While 10⁻⁹ is the rate of mutations occurring in non-repetitive region of the DNA, microsatellite mutation rate ranges between 10⁻² and 10^{−6 [13][14]}[13,14]. Because of this, microsatellites are highly polymorphic, as the number of the repeats in a given locus is relatively unstable and frequently varies between individuals [15]. Due to this feature, deregulated microsatellite expansions are widely known to be the triggering cause of many neurological and neuromuscular diseases and, to date, more than 30 disorders are known to be caused by aberrant expansion of repetitive DNA sequences [16]. Although trinucleotides repeats, in particular GAA, CGG, CAG, and CTG, are commonly known to be responsible for the pathologic manifestation of nucleotide expansion disorders, also tetra- and penta-nucleotides expansions are disease causing [17], thus these diseases are collectively classified as disorders of DNA unstable repeat expansions [18]. Some peculiarities join this variegated group of pathologies: (i) The disease-causing expansion repeats are more unstable in the affected population, with a higher tendency to expand or to contract with respect to the polymorphic repeats of normal people, even if contraction events are rarer [16]. This also occurs as the repeats reach the threshold limit between a normal condition and the pathological state. Thus, even in unaffected families, de novo mutations can determine clinical manifestations [19]. (ii) The dynamic changes in the length of expansions are so marked that differences among patients, as well as in different tissues of the same affected proband, are common [16]. (iii) The more the expansions are transmitted from generation to generation, the earlier the disease symptoms appear in the newly affected individuals, a phenomenon known as clinical anticipation ^[20][21][22][20,21,22]. (iv) In most disorders, the length of repeat expansions influences the phenotype severity of affected individuals ^{[23][24][25][26][27][28][29]}[23,24,25,26,27,28,29]. Consequently, different defects for a given disease can be expressed among patients, including onset of symptoms and co-morbidities. For example, the incidence of diabetes in Friedreich’s ataxia (FA) ranges between 8–32% of patients ^[30][31][32][30,31,32], but the risk of developing is directly correlated to the number of GAA repeats in the frataxin (FXN) gene [33]. In the same way, in myotonic dystrophy (DM), patients carrying small CTG repeats (i.e., between 50–99) are asymptomatic or develop mild defects, such as cataracts, while a severe phenotype occurs in patients with 100–200 repeats [34].

Beside these common features, a high grade of diversity characterizes DNA expansion disorders, as the DNA expanded tracts can affect genes encoding proteins with different roles. Therefore, three different classes have been distinguished, by assembling disorders on the basis of which defects arise from expansions. The first class groups the diseases that are determined by a protein loss of function and are inherited by autosomal recessive or x-linked manner [19]. Typical examples are Fragile X (FXS) or FA, where CGG or GAA repeats determine the loss of the expression of the fragile X mental retardation protein (FMRP) and FXN, respectively ^[35][36][35,36]. The second group belongs to disorders characterized by an autosomal dominant inheritance and in which a protein gain of function occurs [18]. PolyQ diseases, for instance, are determined by CAG expansions in the coding region of 9 distinct genes that lead to the formation of glutamine residues in the final peptide of their encoded product [37]. PolyQ tracts exert a toxic effect mainly by causing aberrant nuclear and cytoplasmic protein aggregation and trapping transcription factors ^[38][39][38,39], chaperons, and proteins belonging to the ubiquitin–proteasome system (UPS) [40]. The third group of disorders are characterized by gain of function involving the transcribed RNA. DM1 and DM2 are respectively caused by the aberrant insertion of CTG and CCTG expansion repeats in the 3′ untranslated region (UTR) of dystrophin myotonic protein kinase (DMPK) [41] and in the first intron of zinc-finger protein 9 (ZNF9) [42]. Similarly, the CGG triplet expansion in FMR1 gene, ranging between 60–200 triplets, causes the fragile X–associated tremor ataxia syndrome (FXTAS) [43]. The pathogenic activity, both in DM and in FXTAS, lies on of the respective expanded mRNA molecules, which are able to sequestrate RNA binding proteins, such as muscleblind-like (MBNL) proteins in DM1, determining splicing alterations and impairments in protein expression ^{[44][45][46][47]}[44,45,46,47].

Oxidative stress has been widely reported to play a prominent role in neurodegenerative diseases ^[48][49][48,49], including disorders caused by DNA expansion repeats (Figure 1). Here, we report the most recent evidences connecting ROS imbalance and DNA expansion disorders, with particular emphasis to the pathway primarily involved in the regulation of cellular antioxidant response, the NF-E2 p45-related factor 2 (NRF2) signaling pathway.