Mitochondrial pathophysiology in Friedreich's ataxia

Authors

  • Pilar González-Cabo,

    1. Program in Rare and Genetic Diseases, Centro de Investigación Príncipe Felipe, Valencia, Spain
    2. Instituto de Biomedicina de Valencia, CSIC, Valencia, Spain
    3. CIBER de Enfermedades Raras (CIBERER), Valencia, Spain
    Search for more papers by this author
  • Francesc Palau

    Corresponding author
    1. Instituto de Biomedicina de Valencia, CSIC, Valencia, Spain
    2. CIBER de Enfermedades Raras (CIBERER), Valencia, Spain
    3. Facultad de Medicina de Ciudad Real, Universidad de Castilla-La Mancha, Ciudad Real, Spain
    • Program in Rare and Genetic Diseases, Centro de Investigación Príncipe Felipe, Valencia, Spain
    Search for more papers by this author

Address correspondence and reprint requests to Dr. Francesc Palau, Centro de Investigación Príncipe Felipe, Valencia, Spain. E-mail: fpalau@cipf.es

Abstract

Neurological examination indicates that Friedreich's ataxia corresponds to a mixed sensory and cerebellar ataxia, which affects the proprioceptive pathways. Neuropathology and pathophysiology of Friedreich's ataxia involves the peripheral sensory nerves, dorsal root ganglia, posterior columns, the spinocerebellar, and corticospinal tracts of the spinal cord, gracile and cuneate nuclei, dorsal nuclei of Clarke, and the dentate nucleus. Involvement of the myocardium and pancreatic islets of Langerhans indicates that it is also a systemic disease. The pathophysiology of the disease is the consequence of frataxin deficiency in the mitochondria and cells. Some of the biological consequences are currently recognized such as the effects on iron–sulfur cluster biogenesis or the oxidative status, but others deserve to be studied in depth. Among physiological aspects of mitochondria that have been associated with neurodegeneration and may be interesting to investigate in Friedreich's ataxia we can include mitochondrial dynamics and movement, communication with other organelles especially the endoplasmic reticulum, calcium homeostasis, apoptosis, and mitochondrial biogenesis and quality control. Changes in the mitochondrial physiology and transport in peripheral and central axons and mitochondrial metabolic functions such as bioenergetics and energy delivery in the synapses are also relevant functions to be considered. Thus, to understand the general pathophysiology of the disease and fundamental pathogenic mechanisms such as dying-back axonopathy, and determine molecular, cellular and tissue therapeutic targets, we need to discover the effect of frataxin depletion on mitochondrial properties and on specific cell susceptibility in the nervous system and other affected organs.

Abbreviations used
DN

dentate nucleus

DRG

dorsal root ganglia

FRDA

Friedreich's ataxia

OXPHOS

oxidative phosphorylation

PGC-1

peroxisome proliferation activator receptor ϒ-coactivator 1 alpha

ROS

reactive oxygen species

Friedreich's ataxia (FRDA) is caused by mutations in the FXN gene that encodes frataxin, a 210-amino acid small protein of 23 kDa. The mature protein of 14.2 kDa retains amino acids 81–210 and is located in the mitochondrial matrix in association with the inner mitochondrial membrane. The GAA trinucleotide expansion (Campuzano et al. 1996) associated with epigenetic changes (Festenstein 2006; Ku et al. 2010; Sandi et al. 2011) in the first intron of the FXN gene affects its correct transcription and reduces the amount of frataxin within the cell and mitochondria. It is widely accepted that the pathogenic mechanisms associated with the disease is because of frataxin deficiency (Campuzano et al. 1996; Schmucker and Puccio 2010; Martelli et al. 2012). Thus, it is essential to know how the lack of frataxin in the mitochondrial matrix produces wide and systemic changes in the pathology, clinical picture, and natural history of FRDA patients, which involve peripheral and central nervous systems, heart myocardial fibers, and the pancreatic islets of Langerhans. Furthermore, it is important to understand the mechanisms that target the effects of this disease to neural cells and connections. Therefore, herein we will review current knowledge on the effect of frataxin depletion on mitochondria and cells, and how modifications in mitochondrial physiology may affect specific neural and other cell structures and functions that underlie the pathophysiology of the disease. In addition, we review other mitochondrial functions have been linked to the neurodegeneration but which has not been paid enough attention in Friedreich's ataxia.

Mitochondrial and cellular consequences of frataxin deficiency

There is increasing consensus that neurodegenerative diseases are the result of neuronal dysfunction and not cell death per se. Deficits in mitochondrial function play a relevant role in neurodegenerative disorders and have been linked to Parkinson's disease, Huntington's disease, hereditary spastic paraplegia, Friedreich's ataxia, and Charcot-Marie-Tooth disease (DiMauro and Schon 2008; Tatsuta and Langer 2008). Mitochondria are important organelles present in every cell type, but are especially important in the nervous system. Mitochondrial functions are essential for neuronal development and maintenance, where they participate in energy production, calcium homeostasis, maintenance of membrane potential, folding of proteins by chaperones, axonal and dendritic transport, and the release and reuse of neurotransmitters in the synapses. Given the enormous variety of functions mediated by mitochondria, it is not a surprise that mitochondrial dysfunction has severe consequences at the cellular level, which is closely related to aging and human neurological diseases. Classically, mitochondrial disorders have been related to mitochondrial respiratory failure. However, involvement of the respiratory chain and oxidative phosphorylation (OXPHOS) is not always present in every case. In the last years, a major role of mitochondria in the pathogenesis of neurodegenerative disorders has become increasingly accepted. Mitochondrial dysfunction can be caused by several pathways: (i) alteration in the mitochondrial energy production and oxidative stress pathway; (ii) defects in mitochondrial quality control; (iii) modification in the mitochondrial network dynamics; and (iv) disruption in the communication of mitochondria with other organelles. In addition, dysregulation of cell and mitochondrial iron metabolism is a pathogenic mechanism in neurological disease, which is especially relevant in Friedreich's ataxia (Richardson et al. 2010; Huang et al. 2011). We are now focusing on how these mitochondrial functions are affected or might be involved in the pathogenic mechanisms of FRDA. Whereas some of the pathways and functions have been studied in depth (e.g., bioenergetics and redox status, and iron homeostasis), there are other pathways for which data are scarce in FRDA (e.g., mitochondrial biogenesis and quality control, mitochondrial dynamics, and calcium homeostasis).

Iron metabolism

Iron is vital to cell survival, but its levels must be regulated and constantly chaperoned because of its redox activity and reactive oxygen species (ROS) generation that have deleterious consequences for the cell. Although not fully understood, mitochondria have an important role in maintaining cellular iron homeostasis. There is some evidence pointing toward the relevance of communication between both mitochondrial and cytosolic compartments in the modulation of iron metabolism (Richardson et al. 2010). In mitochondria, iron is used to generate Fe-S clusters (ISC) and synthesize heme groups. These prosthetic cofactors are responsible for the activity of several enzymes involved in a number of metabolic reactions (Lill and Kispal 2000).

Iron–sulfur clusters are essential components of respiratory electron transfer complexes as well as enzymes of the tricarboxylic acid cycle. Frataxin participates in ISC biogenesis (Adinolfi et al. 2009; Lill 2009; Schmucker et al. 2011). The ISC biogenesis hypothesis is currently the most accepted regarding the biochemical function of frataxin. The consequence of frataxin deficiency is the abnormal function of enzymatic activities and therefore, uncoupling of the electron transport chain and electron transmission (Duby et al. 2002; Gerber et al. 2003; Muhlenhoff et al. 2002; Ramazzotti et al. 2004; Shan and Cortopassi 2012; Shan et al. 2007; Yoon and Cowan 2003). Correct ISC in mitochondria is intimately linked to cellular iron homeostasis (Huang et al. 2011). Therefore, failure to assemble mitochondrial Fe-S proteins results in increased cellular iron acquisition and eventually, mitochondrial iron overload (Lill et al. 2012). Lack of frataxin causes dysregulation of iron metabolism that sometimes induces iron deposits, which has been observed in several FRDA patients (Ramirez et al. 2012) as well as in different models of frataxin deficiency such as yeast (Babcock et al. 1997), C. elegans (Gonzalez-Cabo et al. 2011), and mouse (Whitnall et al. 2012).

Heme synthesis occurs partly in the mitochondria and partly in the cytoplasm. Hemoproteins are involved in a broad spectrum of crucial biological functions including oxygen binding (hemoglobin), oxygen metabolism (oxidases, peroxidases, catalases and hydroxylases), and electron transfer (cytochromes). Several studies have demonstrated a physical interaction between frataxin and ferrochelatase (He et al. 2004; Yoon and Cowan 2004). Frataxin could have a role as a mitochondrial iron donor involved in heme metabolism. Moreover, knockdown of frataxin in mammalian cells decreases mitochondrial heme metabolite levels (Schoenfeld et al. 2005) and increases heme biosynthesis (Lu and Cortopassi 2007), suggesting that frataxin may have a role in the biogenesis of heme-containing proteins.

Alteration in mitochondrial energy production and oxidative stress

The main function of mitochondria is energy production through oxidative phosphorylation that maintains cellular activity. Bioenergetics is divided into two steps, the electronic transport chain that generates a gradient of protons that is then used to produce ATP by oxidative phosphorylation. Disruption of the respiratory chain and OXPHOS activities may change the mitochondrial membrane potential and decrease ATP production. In addition, OXPHOS is the primary source of endogenous ROS, which are produced as toxic products of respiration, namely the superoxide anion (O2) and H2O2. During the transfer of electrons to the molecular oxygen in the respiratory chain, there is a leak of electrons (1 to 5%), which mainly participate in O2 generation. Factors that decrease the efficiency in the coupling of the electron transport chain increase superoxide production.

Increased intracellular ROS levels modify the physiological redox balance, forcing the cell toward an oxidative stress status. Lack of frataxin may favor such a redox imbalance. The exact sequence of events that occur in FRDA cells has not been clarified yet. In most cellular and animal models of frataxin deficiency, increases in ROS have been observed, in some cases associated with mitochondrial iron deposits. These models include yeast (Babcock et al. 1997; Foury and Cazzalini 1997; Koutnikova et al. 1997; Wilson and Roof 1997; Foury 1999; Radisky et al. 1999), Drosophila (Anderson et al. 2005; Llorens et al. 2007), Caenorhabditis elegans (Vazquez-Manrique et al. 2006), and mouse (Puccio et al. 2001; Simon et al. 2004; Al-Mahdawi et al. 2006). These changes have also been detected in endocardic biopsies (Rotig et al. 1997), cerebellum (Waldvogel et al. 1999), and fibroblasts (Garcia-Gimenez et al. 2011) from FRDA patients. Evidence for iron deposits has been demonstrated in heart necropsy studies (Lamarche et al. 1980, 1993). Biochemical, functional, and genetic experiments showing an interaction between frataxin and complex II subunits in yeast, human cells (Gonzalez-Cabo et al. 2005) and Caenorhabditis elegans (Vazquez-Manrique et al. 2006) suggest that there is a direct role of frataxin in the electron transport chain and mitochondrial respiration. Thus, frataxin deficiency could entail a disruption of the electron transport chain.

Defects in mitochondrial quality control

When additional power supply by a specific tissue is needed, cells respond by inducing mitochondrial biogenesis and increasing the number of new mitochondria. Factors that affect mitochondrial biogenesis are aging, fusion and fission of mitochondrial network, oxidative stress, and gene regulation by several transcription factors (Lopez-Lluch et al. 2008). The peroxisome proliferation activator receptor ϒ-coactivator 1 alpha (PGC-1α) is a transcriptional co-activator that has emerged as a master regulator transducing different physiological stimuli in metabolic programs, often by stimulating mitochondrial biogenesis (Lin et al. 2005). It has been identified as a target pathway in FRDA (Coppola et al. 2009). Oxidative stress and mitochondrial biogenesis have been linked in FRDA fibroblasts (Garcia-Gimenez et al. 2011). Superoxide anion levels and the expression of PGC-1α and the mitochondrial transcription factor mtTFA were seen to be increased in the fibroblasts of two patients with late onset disease. Moreover, levels of the upstream signals p38 MAPK and AMPK kinases were also increased. These findings suggest a relationship between oxidative stress and mitochondrial biogenesis. Interestingly, PGC-1α and mTFA expression was down-regulated by the presence of the antioxidant idebenone in cell culture. On the contrary, such an increase in PGC-1α and mTFA expression was not observed in fibroblasts from an early onset patient. Discrepancies in the regulation of biogenesis may be explained by age at onset and the natural history of the disease. These data suggest an inverse correlation between disease severity and the capacity to respond to cellular energy needs by mitochondria. The less frataxin there is (which is associated with a more severe phenotype), the less capacity the cell may have to activate the program of mitochondrial biogenesis and generate an energy response. The full implications of these findings on mitochondrial pathophysiology and disease evolution in FRDA require additional studies.

In most types of cells, autophagy is an adaptive mechanism of starvation, acting by degrading intracellular proteins to obtain amino acids under conditions of limited nutrients. By means of continuous regeneration of cytoplasmic material and structures, autophagy plays an important role in neuronal survival. Changes in autophagy have been linked to neurodegenerative disorders such as Huntington's disease, Parkinson's disease, and Alzheimer's disease (Cheung and Ip 2011). Autophagy works as a mechanism of defense against injuries that cause oxidative stress as well. Oxidative stress radicals activate macroautophagy to facilitate the elimination of damaged organelles (e.g., depolarized mitochondria). In an inducible knock-out mouse model of FRDA, cell bodies of sensory neurons of the dorsal root ganglia (DRG) showed the typical pattern of autophagy, characterized by the presence of vacuoles (Simon et al. 2004). The YG8R humanized mouse expressing the abnormal GAA expansion (two expansions of 190 and 90 GAA repeats) in a knock-out genome background (Al-Mahdawi et al. 2006) showed a large number of vesicular structures similar to autophagosomes and autolisosomes, as well as lipofuscin deposits in DRG cells. This humanized rescued mouse displayed a degeneration of large sensory neurons of the DRG and a mild neurological picture with involvement of motor behavior and mild ataxia that partially resembled the clinical picture of FRDA patients. Recently, (Schiavi et al. 2013) observed that reduction in frataxin protein induced autophagy in C. elegans and lymphoblasts derived from FRDA patients, suggesting that the induction of autophagy by frataxin suppression may be preserved in evolution. There is, however, no report specifically addressing the role of mitophagy in frataxin deficiency and FRDA pathology.

After continuous oxidative injuries, the increase in oxygen radicals may damage lysosomal membranes, which may induce abnormal changes in cell structure and function. In the case that autophagy mechanisms do not avoid ROS damage and do not support cell needs for survival, other pathways may be activated, which include neuronal death by apoptosis. Lack of frataxin affects cell viability and proliferation. Complete inactivation of the Fxn gene in a knock-out mouse induced early embryonic lethality associated with cellular death by apoptosis and necrosis (Cossee et al. 2000). This complete knock-out mouse also exhibited necrosis in several cells. Cell death studies in frataxin-deficient cell lines have given variable results. In a study of several neuronal cell lines, the authors concluded that only Schwann cells showed alteration of the cell cycle before dying by apoptosis (Lu et al. 2009). On the other hand, silencing frataxin in human neuroblastoma SH-SY5Y cells induced large-scale apoptosis in neuron-like differentiated cells (Palomo et al. 2011); this apoptotic cell death occurred through p53 and activation of caspase 3. Recent studies in pancreatic β cells in both clonal and primary rat cultures demonstrated sensitization to apoptosis when depleting frataxin, with an endoplasmic reticulum stress-mediated mechanism being proposed to be involved (Cnop et al. 2012). It seems that the response of frataxin-depleted cells to apoptotic stimuli is variable and may depend on the type of cell. In that sense, one could predict that there is a connection between the quality control of mitochondria and cells, mechanisms of cell death, cell specificity and neurological pathways involved in the pathophysiology of FRDA.

Mitochondrial network dynamics

Over the last years, the relationship between energy production and the organization of the mitochondrial network has become more apparent. Mitochondrial movement and the fusion and fission pathways of mitochondrial dynamics with rapid changes between fragmented and tubular states are emerging in the field of axonal biology and in the pathogenesis of peripheral neuropathies associated with mitofusin 2 (Baloh et al. 2007; Misko et al. 2010) or GDAP1 (Niemann et al. 2005; Pedrola et al. 2008), in addition to spastic paraplegias (Ferreirinha et al. 2004; Joshi and Bakowska 2011). The Parkinson's disease-related genes PINK1 and PARKIN have also become relevant in mitochondrial dynamics and mitochondrial quality control (Vives-Bauza et al. 2010).

Human fibroblasts carrying genetic changes in the subunits of the respiratory chain complexes show a fragmented mitochondrial network (Koopman et al. 2007). However, few data on the mitochondrial network structure in frataxin-deficient cells are available. In the yeast frataxin-defective model, yfh1∆, mitochondria showed a fission pattern that has been attributed to oxidative stress (Lefevre et al. 2012). However, the authors of this report did not observe such a fragmentation of mitochondria in FRDA fibroblasts, indicating that the mitochondrial response to different stresses or injuries may depend on the cell type physiology, cell adaptation capacity and biological microenvironment.

Failure in the communication with other organelles

Interaction and communication with other organelles are important factors for regulating mitochondrial function in cells. These organelles and cellular structures are the nucleus, endoplasmic reticulum (ER), Golgi, and cytoskeleton. Molecular interaction with cytoskeletal microtubules and neurofilaments is important for the proper anterograde and retrograde transport of mitochondria across axons and dendrites. Mitochondria play a relevant role in calcium homeostasis, especially via communication with the ER (de Brito and Scorrano 2010). Abnormal changes in the cell metabolism of calcium have been observed in several neurological diseases including peripheral neuropathies such as Charcot-Marie-Tooth 2A (CMT2A) disease caused by MFN2 mutations (de Brito and Scorrano 2008) and GDAP1-related CMT neuropathies (Pla-Martin et al. 2013), which may be associated with dysfunction of the mitochondrial-associated membranes (MAMs), calcium trafficking and store-operated Ca2+ entry (SOCE).

In FRDA fibroblasts, limited organization of the cytoskeleton is related to the increase in merged glutathione levels in cytoskeletal proteins (Pastore et al. 2003). The authors of this finding suggested that oxidative stress in frataxin-defective fibroblasts could act as the trigger for destabilization of the cytoskeletal structure. Later on, this same group observed similar results in necropsy material from the spinal cord of FRDA patients, where they found increased amounts of dynamic tubulin compared to the stable form, implying an alteration in microtubule polymerization and therefore, the cytoskeleton network in FRDA patients (Sparaco et al. 2009). Despite increasing interest in mitochondrial communication with other cell structures, there are no reports addressing the abovementioned aspects of mitochondrial biology in FRDA pathophysiology. Few data relating frataxin and Ca2+ metabolism are available. Over-expression of frataxin in mammalian adipocytes induces Ca2+ uptake into isolated mitochondria, which up-regulates tricarboxylic acid cycle flux and respiration, and increases cellular ATP content (Ristow et al. 2000). On the contrary, chelation of intracellular Ca2+ with desferoxamine rescues FRDA fibroblasts from oxidant-induced death (e.g., H2O2) (Wong et al. 1999). It seems that calcium homeostasis may have a role in the biology of frataxin and the study of Ca2+ physiology in FRDA deserves more attention.

Frataxin is an evolutionary relevant small protein for which definite functions in mitochondria and cell physiology have not been fully established. Most of the knowledge has been revealed by examining cell and animal models lacking frataxin. Current experimental work points toward biochemical and signaling pathways that mainly involve the biosynthesis of ISC and heme prosthetic factors, oxidative stress status, and iron homeostasis. However, these biochemical events that are usually observed in most of these models are not universal: oxygen radical production and increased activity of antioxidant enzymes is not always evident (Chantrel-Groussard et al. 2001; Seznec et al. 2005; Shidara and Hollenbeck 2010), impairment of ISC-containing enzymes is not always a consequence of frataxin loss of function (Rotig et al. 1997; Gonzalez-Cabo et al. 2010; Hick et al. 2013), and iron accumulation in mitochondria is not invariably an early event (Cossee et al. 2000; Puccio et al. 2001). In an attempt to explain the whole biological phenomenon in one hypothesis, Bayot et al. (2011) suggested that frataxin deficiency induced an increased basal sensitivity to ROS as opposed to an absolute increase in oxygen radicals, that is, defective cells display hypersensitivity to an oxidative insult. The whole picture of the disease pathogenesis is even heterogeneous at the histological level. The affected pathways and functions show a differential pattern in different organs and tissues (Fig. 1). For instance, late iron deposits are observed in myocardial fibers, but not in nervous system structures such as the DRG. Whether these biochemical and pathological differences are related to distinct cell susceptibility to the damage associated with frataxin depletion needs further studies in disease models, either cellular or animal. Nevertheless, whatever the cellular mechanisms involved, frataxin deficiency has consequences on the physiology of the targeted organs. In the following section, we will address disease pathophysiology.

Figure 1.

Schematic representation of the biological pathways altered in Friedreich's ataxia (FRDA) as consequences of frataxin deficiency in mitochondria. There is a representation of the three major organs involved in the pathogenesis of FRDA, that is, nervous system, heart and pancreas. Data are obtained from human or mouse specific tissues or the analysis of cultured cells derived from patients (i.e., fibroblasts) or from cell lines (i.e., human neuroblastoma). Specific biological mechanisms that are altered in all reports are indicated by a continuous arrow. In contrast, a dashed arrow indicates that the altered mechanism in the specific organ/tissue has not always been reported. All the four pathogenic mechanisms have been observed in the myocardium, but this is not the rule for the nervous system and pancreas. Iron deposits have been detected only in the heart in both patients' necropsies (Lamarche et al. 1980, 1993) and in the YG8R (Al-Mahdawi et al. 2006) and the conditioned MCK-Frda knock-out (Puccio et al. 2001) mouse models. The most frequently studied enzymes related to Fe-S clusters as prosthetic factors are the respiratory chain complexes and aconitase. Aconitase activity is altered in patients' heart biopsies (Rotig et al. 1997) and the mouse models YG8R (Al-Mahdawi et al. 2006) and conditioned knock-out mice (Puccio et al. 2001). However, evidence for changes in the respiratory chain complexes has been obtained only in the MCK-Frda mouse (Puccio et al. 2001). Data from FRDA patients heart necropsies are not conclusive as reduced activity has been found by some researchers (Bradley et al. 2000) but not by others (Rotig et al. 1997). A similar disparity of findings has been observed in the nervous system of patients, where a reduced activity of aconitase has been detected in the dorsal root ganglia. In contrast, no changes in the respiratory chain complexes activities have been observed in either patients (Bradley et al. 2000), conditioned knock-out mice (Puccio et al. 2001) or neurons derived from reprogrammed iPS cells (Hick et al. 2013). Oxidative stress has been described in every organ or cells derived from several disease models, which include cardiac tissue in the YG8R mouse (Al-Mahdawi et al. 2006) and the MCK-Frda knock-out mouse (Al-Mahdawi et al. 2006; Wagner et al. 2012), several nervous tissues in the YG8R mouse (Al-Mahdawi et al. 2006), and pancreatic β-cells of a conditioned knock-out mouse (Ristow et al. 2003; Cnop et al. 2012); however, no oxidative sensitizing has been observed in frataxin-silenced β-cells (Cnop et al. 2012). Apoptosis has been observed in specific tissues of mouse models, particularly of pancreatic β cells (Ristow et al. 2003) and nervous system (Cossee et al. 2000), and in some cell lines of β cells islets (Cnop et al. 2012). However, in neuronal cell lines results are contradictory as apoptosis has been observed in some of them (Lu et al. 2009; Palomo et al. 2011) and not in others (Lu et al. 2009). In FRDA patients there are some data suggesting that apoptosis may occur: the number of dorsal root ganglia (DRG) propioceptive neurons from necropsies is decreased (Koeppen et al. 2009; Koeppen 2011), something that also has been detected in post-mortem pancreatic islets β cells (Cnop et al. 2012), which could be because of an apoptotic process. On the other hand, tissue degeneration associated with vacuoles and lipofuscin accumulation have been reported in YG8R mouse (Al-Mahdawi et al. 2006) and conditioned Prp-Frda mouse (Simon et al. 2004); this finding suggests that autophagy changes may be either a protective biological response to frataxin deficiency or part of pathogenic mechanisms related to the disease. This figure was produced using ServierMedical Art (http://www.servier.com/serviermedical-art/powerpoint-image-bank).

Friedreich's ataxia: pathophysiology of frataxin deficiency

FRDA pathophysiology is a consequence of the partial deficiency of frataxin in patients homozygous for the GAA expansion. A minority of patients, however, are compound heterozygotes for an expanded GAA repeat allele and a point or small mutation (missense, nonsense, splicing, frame-shift, or ins/del nucleotide change) that induces a loss-of-function protein mechanism (Cossee et al. 1999; De Castro et al. 2000).

Clinical symptoms and neurological examination (Harding 1981; Durr et al. 1996; Koenig and Dürr 2000; Pandolfo 2009) indicate that FRDA corresponds to a mixed sensory and cerebellar ataxia, which affect the proprioceptive pathways in the peripheral nervous system, spinal cord, and nuclei of the cerebellum. Later on, the pyramidal tracts, related to muscle weakness and spasticity, are also involved. Most of these neural structures are linked by long axons connecting Pacini's and Ruffini's corpuscles and unencapsulated nerve endings with sensory neurons localized in the DRG and Clarke's columns. These structures are also connected with the gracile and cuneate nuclei and the dorsal nucleus of Clarke in the lower medulla, medial lemniscus, and output projections of the cerebellum. Thus, major structures of the nervous system involved in the neuropathology and pathophysiology of FRDA include peripheral sensory nerves, the DRG, posterior columns, spinocerebellar and corticospinal tracts of the spinal cord, gracile and cuneate nuclei, dorsal nuclei of Clarke, and the dentate nucleus (DN) in the deep cerebellar nuclei.

The pathology of FRDA was recently summarized by Koeppen (Koeppen 2011; Koeppen and Mazurkiewicz 2013). This author and his colleagues have studied in depth the neuropathology of FRDA in a series of sequential papers related to the DRG (Koeppen et al. 2009), the sensory nerves (Morral et al. 2010), and the cerebellum (Koeppen et al. 2011). The DRG are entirely affected, with an important reduction in large neurons and a lack of large myelinated fibers in the central axons and dorsal root nerves; however, the fine unmyelinated nerves fibers are conserved (Hughes et al. 1968; Inoue et al. 1997; Koeppen et al. 2009; Koeppen 2011). Satellite cells participate in the generation of residual nodules (misnamed Nageotte nodules) that are composed of clusters of nuclei from residual neurons and a thickened rim of ferritin-positive cells. Both DRG neurons and satellite cells exhibit changes in the ferritin and ferroportin proteins, indicating abnormal iron metabolism (Koeppen et al. 2009; Koeppen 2011). It seems that both types of cells, sensory neurons and satellite cells, are affected in FRDA (Koeppen 2011; Koeppen and Mazurkiewicz 2013). There is no evidence as to where the primary defect occurs or whether the cell pathophysiology may be the consequence of abnormal satellite cell-neuron interaction. In addition, the involvement of the communication between neuronal bodies and dorsal spinal root axons is another point that requires further investigation. In any case, the role of frataxin deficiency in primary cellular defects and nerve structure interactions remains elusive.

The outstanding lesion of sensory nerves consists of a lack of myelination of axons from large sensory neurons in the DRG and axonal degeneration (Dyck et al. 1968); this axonopathy has been associated with a dying-back mechanism (Barbeau 1980; Said et al. 1986). Dying-back axonopathy should affect sensory nerves as well as central spinocerebellar and corticospinal motor tracts; however, such a mechanism has only been intensively investigated in the peripheral sural nerves, showing distal predominance of fiber loss (Hughes et al. 1968; Said et al. 1986; Jitpimolmard et al. 1993), with data from central axons being scarce (Fig. 2). Moreover, recent pathological studies in sural nerve biopsies suggest that Schwann cells may play a role in the pathogenesis of hypomyelination and that the disease process may affect small neurons too (Morral et al. 2010). The Schwann cell role in nerve pathology and architecture in FRDA patients come from immunohistochemical studies of the cytoplasmic protein S100α and the basal lamina protein laminin 2. S100α immunoreactivity in Schwann cells is reduced in the patients' nerves. In contrast, laminin 2 expression is maintained, but the honeycomb pattern observed in normal nerves is modified in FRDA nerves, which is consistent with the depletion of large myelinated fibers (Morral et al. 2010). It seems that Schwann cell-axon communication is defective as well. However, no clear evidence on the primary involvement of Schwann cells exists, but hypomyelination because of abnormal communication between Schwann cell and axon, together with slow axonal degeneration, may explain the pathogenic mechanisms in FRDA sensory neuropathy (Morral et al. 2010).

Figure 2.

Schematic drawing of axonal pathology in peripheral and central nervous systems. Dying-back sensory axonal neuropathy is represented by loss of large myelinated axons (dashed black line) and abnormal myelinization (loss of the black line in the myelin), in association with loss of large proprioceptive neurons in the dorsal root ganglia (DRG). Involvement of central axons is indicated by dashed black lines in the spinal cord that represent abnormal medial lemniscus and dendrite connections. Degenerated fasciculi and tracts are indicated in white. Motor neurons, interneurons and motor nerves axons and myelin are normal. This figure was produced using ServierMedical Art (http://www.servier.com/serviermedical-art/powerpoint-image-bank).

The cerebellar cortex projects connections toward the thalamus, having an intermediate relay station at the deep cerebellar nuclei, which represent outputs of the cerebellar cortex: the dentate nucleus (the largest), the emboliform nucleus, the globose nucleus and the fastigial nucleus. The deep nuclei axons exit the cerebellum via the superior cerebellar peduncle and ascend to the contralateral thalamus before projecting directly to the primary motor and pre-motor association cortices. Such a pathway offers the cerebellum access to the corticospinal projections that participate in the organization and coordination of voluntary complex movements. Cerebellar histological structures have been less studied than those of spinal cord and nerves in FRDA. The analysis of autopsy specimens from 24 patients performed by Koeppen and collaborators (Koeppen et al. 2011) shed some light on our knowledge of the role of cerebellum in the pathophysiology and clinical picture of FRDA, especially on the cause of central disturbances such as dysmetria, dysarthria, and dysphagia. In FRDA, the dentate nucleus size is small and has high iron concentration that is not properly removed from the affected brain tissue (Koeppen 2011). The role of iron metabolism and other metals such as copper and zinc in the cerebellum has recently been investigated (Koeppen et al. 2012); however, further studies that explain the role of metals in the pathology of DN and the cerebellar component of ataxia are needed. Histological and immunohistochemical studies have revealed loss of large neuron-specific enolase (NSE)-reactive neurons, with preservation of small neurons, loss of large inhibitory GABA-ergic terminals and formation of grumose degeneration. In contrast, Purkinje cells somata and dendrites in the cerebellar cortex remain intact. In addition, the DN displays depletion of the terminals positive for vesicular glutamate transporters 1 and 2 (VGluT1 and VGluT2), which arise from climbing and mossy fibers. These results suggest that the circuitry between Purkinje cells and the dentate nucleus in the cerebellar cortex is altered, whereas processing of GABA-mediated (inhibitory) and glutamate-mediated (excitatory) impulses remain normal in the cerebellar cortex, it is defective in the DN large neurons.

In such a situation, what is the role of mitochondrial defects caused by the lack of frataxin in both axonal neuropathy and neurons? What are the pathological changes affecting mitochondrial distribution and function in axons? Is there any relationship between frataxin deficiency and the biochemical changes in mitochondria and bioenergetics, mitochondrial trafficking and distribution, mitochondrial-endoplasmic reticulum communication, calcium homeostasis or mitochondrial quality control, as suggested in other neurodegenerative disorders? (Schon and Przedborski 2011). Whereas pathological lesions of the nervous system and several pathogenic hypotheses in mitochondrial biochemistry (see above) have been well described, few data on the defects of mitochondrial function in axon biology have been reported. In a knock-down Drosophila model, DfhIR, frataxin depletion induced a dying-back neuropathy early in development, which was associated with a reduced mitochondrial membrane potential and depolarized organelles in the cell bodies of ventral ganglion (apparent at the late larval development), axons and neuromuscular junctions of segmental nerves. This was followed by a failure of retrograde axonal transport and abnormal synaptic distribution (Shidara and Hollenbeck 2010). These data in an invertebrate animal lead us to postulate defects in the mitochondrial transport through the axon and distribution in specific nerve structures in FRDA.

FRDA pathophysiology may not be (at least not only) the consequence of a neurodegenerative process. There are two other possibilities that are not contradictory to each other (Morral et al. 2010; Koeppen 2011): (i) neurodegeneration may affect neurons, satellite cells and Schwann cells, the glia-neuron interaction being a relevant point in the pathogenic mechanisms causing the disease; and (ii) as DRG neurons, satellite cells, Schwann cells, and the axons of sensory nerves and dorsal roots have a common origin in the neural crest, a major developmental defect may be part of the disease process. In any case, there is a selective neurological and pathological pattern in FRDA that involves specific structures in the nervous system with specific susceptibility to becoming defective: large (but not only) sensory neurons in the DRG, satellite cells, neurons of the dorsal Clarke's column, large neurons of the DN, peripheral and central large axons communicating with specific neurons, and axonal connections between such neurons and Purkinje cells. Discovery of such a specific susceptibility to frataxin deficiency may help us to define therapeutic approaches to FRDA.

Heart disease in FRDA is present is more than 90 percent of patients and may sometimes precede neurological signs (Hewer 1969; Harding 1983; Durr et al. 1996). The cardiomyopathy and cardiopulmonary consequences are usual causes of death occurring in the late third and fourth decades of life. The pathological changes show left myocardial hypertrophy that used to be concentric or septum hypertrophy, hypertrophied myocardial fibers with pleomorphic nuclei, interstitial fibrosis with excessive endomysial connective tissue, and degenerating and necrotic muscle fibers (Koeppen 2011). Intra-cytoplasmic granular deposits of iron are observed in the histological studies of necropsic hearts (Lamarche et al. 1980, 1993). It has been proposed that such deposits may contain cytosolic and mitochondrial ferritin (Michael et al. 2006; Koeppen 2011). The electron-dense inclusions in mitochondria suggest that the iron accumulation in ferritin is mitochondrial. However, recent studies suggest that iron overload and iron-mediated damage to cardiomyocytes and myocardial scarring not only occur in cardiac mitochondria, but also in the cytosol (Ramirez et al. 2012). Iron trafficking between the mitochondria and cytosol is becoming relevant in iron metabolism (Richardson et al. 2010; Huang et al. 2011). Such a communication between both cellular compartments seems to be affected by frataxin depletion, as observed in the muscle creatine kinase (MCK) conditional frataxin knock-out mouse. In the heart of these mice, there are changes in iron trafficking, which lead to cytosolic iron deficiency (Whitnall et al. 2012). It will be interesting to investigate a possible mechanistic relationship between abnormal regulation of iron metabolism in frataxin deficiency and the iron deposits observed in the hearts of FRDA patients. Very recently, lipid metabolism was associated with frataxin deficiency in fly (Navarro et al. 2010) and mouse models (Wagner et al. 2012). Whereas glial cells were affected in the fly, the study of both neural and cardiac conditioned knock-out mice suggested a role of lipid pathways in heart disease in FRDA. Β-oxidation of fatty acids occurs in the mitochondria and fat is the major source of carbon fuel in the heart. Wagner et al. (2012) observed hyperacetylation of cardiac mitochondrial proteins from knock-out mice, and evidence exists that it is caused by the inhibition of the mitochondrial NAD+-dependent deacetylase sirtuin-3 (SIRT3). Since SIRT3 is an important regulator of fatty acid β-oxidation, lipid homeostasis is a new target to investigate the pathogenesis of FRDA.

The incidence of diabetes mellitus and glucose intolerance is estimated to be around 10 and 30% in FRDA patients, respectively (Harding 1981; Durr et al. 1996). Diabetes is caused by a progressive loss of the pancreatic β-cells. As the islets of Langerhans become smaller, they undergo invaginations of normal exocrine pancreatic tissue. Bioenergetic deficiency may be the major cause of β-cell death, but other mitochondrial functions and their relationship with frataxin depletion remain to be investigated in depth in FRDA.

Conclusion

To understand the role of mitochondria in FRDA pathophysiology, along with the pathogenic changes at the cellular level, we have to consider the effect of frataxin deficiency on mitochondrial functions in cell and tissue physiology. These functions include bioenergetics, mitochondrial trafficking, mitochondrial relationship with other cellular organelles especially the endoplasmic reticulum, calcium homeostasis, and mitochondrial quality control (Schon and Przedborski 2011). Definition of redox status in FRDA is relevant, especially because abnormal oxidative stress is a therapeutic target. However, it is important to know the relative balance between excessive generation of ROS because of abnormal respiratory function and an impaired ability to recruit antioxidant defenses against endogenous oxidative stress (Santos et al. 2010). Oxidative stress may affect the cytoskeleton (Pastore et al. 2003; Sparaco et al. 2009) and impair axonal transport (De Vos et al. 2008) and mitochondrial transport through the axon (Shidara and Hollenbeck 2010), which may be related to the dying-back mechanism of axonopathies, not only in sensory nerves, but also in the sensory (medial lemniscus) and pyramidal tracts of the spinal cord and cerebellar projections (Fig. 2). Furthermore, the effect of mitochondrion compartmentalization within the nervous system and differences in tissue susceptibility to frataxin deficiency (Dubinsky 2009) are becoming increasingly relevant in understanding the pathophysiology of FRDA, as has been postulated for axonal degeneration (Court and Coleman 2012) and other neurodegenerative disorders (Schon and Przedborski 2011). To develop new therapeutic approaches for FRDA, we need to understand how the abovementioned functions in cells and structures of the nervous system, heart and pancreas are more sensitive to frataxin depletion than other body structures and cell types.

Acknowledgements

The research work in our laboratory is funded by grants from the Spanish Ministry of Economy and Competitiveness, the Instituto de Salud Carlos III (ISCIII), the European Community's Seventh Framework Program FP7/2007-2013 under the grant agreement no. 242193 EFACTS, the Collaborative Joint Project TREAT-CMT awarded by IRDiRC and funded by ISCIII grant IR11, the Generalitat Valenciana (Prometeo program), the Fundació Marató TV3, the Fundación Alicia Koplowitz, and the Fundación Isabel Gemio. CIBERER is an initiative developed by the Instituto de Salud Carlos III in cooperative and translational research on rare diseases.

Conflicts of interest

Both P. González-Cabo and F. Palau have no conflict of interest.

Ancillary