Cognitive dysfunction in Duchenne muscular dystrophy: a possible role for neuromodulatory immune molecules

Abstract

Duchenne muscular dystrophy (DMD) is an X chromosome-linked disease characterized by progressive physical disability, immobility, and premature death in affected boys. Underlying the devastating symptoms of DMD is the loss of dystrophin, a structural protein that connects the extracellular matrix to the cell cytoskeleton and provides protection against contraction-induced damage in muscle cells, leading to chronic peripheral inflammation. However, dystrophin is also expressed in neurons within specific brain regions, including the hippocampus, a structure associated with learning and memory formation. Linked to this, a subset of boys with DMD exhibit nonprogressing cognitive dysfunction, with deficits in verbal, short-term, and working memory. Furthermore, in the genetically comparable dystrophin-deficient mdx mouse model of DMD, some, but not all, types of learning and memory are deficient, and specific deficits in synaptogenesis and channel clustering at synapses has been noted. Little consideration has been devoted to the cognitive deficits associated with DMD compared with the research conducted into the peripheral effects of dystrophin deficiency. Therefore, this review focuses on what is known about the role of full-length dystrophin (Dp427) in hippocampal neurons. The importance of dystrophin in learning and memory is assessed, and the potential importance that inflammatory mediators, which are chronically elevated in dystrophinopathies, may have on hippocampal function is also evaluated.

the fatal, x chromosome-linked disease Duchenne muscular dystrophy (DMD) is caused by the loss of the structural protein dystrophin (Bulfield et al. 1984). Interactions between dystrophin and the dystrophin-associated protein complex (DAPC), which includes the membrane-spanning β-dystroglycan component, serves to link the subsarcolemmal cytoskeletal actin to the extracellular matrix in skeletal muscle. In such contracting cells, dystrophin, through interactions with the DAPC, protects the sarcolemma against the mechanical stresses of repeated contractions. Thus loss of functional dystrophin, which comprises 0.1% of the total human genome (Barbujani et al. 1990; Koenig et al. 1987), causes muscle fibers to become more susceptible to contraction-induced damage. This, in turn, results in muscle inflammation and myophagocytosis (Anderson et al. 1987), leading to progressive physical disability, eventual immobility, and finally, premature death in affected boys. There is currently no cure for DMD, with patients experiencing an average life expectancy of just over 20 years (Yiu and Kornberg 2008) and most eventually succumbing to death via cardiorespiratory failure.

Chronic inflammation is a key symptom and contributory factor in the pathogenesis of DMD, due in part to the secretion of proinflammatory immune mediators from damaged dystrophin-deficient muscle fibers (De Paepe and De Bleecker 2013; Porter et al. 2002). Proinflammatory cytokines, including tumor necrosis factor (TNF)α (Kuru et al. 2003; Porreca et al. 1999), interleukin (IL)-1β (Evans et al. 2009), IL-6 (Messina et al. 2011; Rufo et al. 2011), and the promotor of ongoing inflammation, IL-17 (De Pasquale et al. 2012), are elevated in muscle biopsies from patients with DMD.

Although dystrophinopathies have principally been studied in the context of skeletal muscle dysfunction, DMD itself is actually a multisystem disorder, because dystrophin is also expressed in cardiac and smooth muscle, endocrine glands, and neurons. In healthy individuals, neurons in the central nervous system (CNS) express full-length dystrophin (Dp427). Specifically, it is expressed in the hippocampus, cerebellum, cerebral cortex, and amygdala (Bies et al. 1992; Knuesel et al. 2000; Lidov et al. 1990, 1993; Sekiguchi et al. 2009). Consistent with the functional importance of dystrophin in the CNS, boys with DMD often exhibit varying degrees of nonprogressing cognitive impairment (Anderson et al. 2002; Bresolin et al. 1994), with their intelligence quotients (IQ) shifted downward one standard deviation below the normal range (Felisari et al. 2000; Nardes et al. 2012). Moreover, boys with DMD have difficulty communicating. They exhibit social behavior problems and have poor facial affect recognition (Hinton et al. 2006, 2007). Interestingly, expression of Dp71, a protein product produced through a mutation in the dystrophin gene and only expressed in the CNS, has been linked to intellectual disability without an attendant muscular dystrophy phenotype (de Brouwer et al. 2014). In the context of the frontal cortex and hippocampus, both of which facilitate learning and memory processes, patients with DMD exhibit deficits in verbal, short-term, and working memory (Hinton et al. 2000, 2001; Snow et al. 2013). Because it is now recognized that activation of the immune system in the periphery greatly influences the normal function of the CNS (Marin and Kipnis 2013), this review will assess the possible role of inflammatory mediators in DMD-associated cognitive dysfunction, focusing on the hippocampus.

Potential Role of Inflammatory Molecules in DMD-Associated Cognitive Dysfunction

Chronic inflammation is a key aspect of DMD pathophysiology. Indeed, corticosteroids with potent anti-inflammatory effects are the most effective treatment for delaying the onset and progression of the disease (Gloss et al. 2016). Skeletal muscle biopsies exhibit necrotic and degenerating fibers surrounded by macrophages and CD4+ lymphocytes in the early stages of the disease, when new myocytes are still being produced. However, the capacity to regenerate new myofibers becomes exhausted and muscle fibers are eventually replaced by connective and adipose tissue. There is also evidence of immune cell infiltration into muscle tissue and activation of the complement system (Haslett et al. 2002), which is an important effector arm of both the innate and adaptive divisions of the immune system. The inflammatory response is due, in part, to the secretion of cytokines from damaged dystrophin-deficient muscle fibers (De Paepe and De Bleecker 2013; Porter et al. 2002). Consistent with this, DMD muscle biopsies display altered cytokine profiles with elevated levels of proinflammatory mediators such as TNFα (Kuru et al. 2003; Porreca et al. 1999), IL-1β (Evans et al. 2009), IL-6 (Messina et al. 2011; Rufo et al. 2011), and IL-17 (De Pasquale et al. 2012).

Whereas cytokines have a well-characterized role in the response of tissues to infection, these signaling molecules also have neuromodulatory actions. In addition to immune cells, neurons, glia and the endothelial cells of the microvasculature in the CNS secrete cytokines and also express receptors for these immune mediators. There is significant cross talk between the immune and nervous systems, since cytokines can have indirect effects on neuronal activity by stimulating secretion of neuromodulatory molecules from glia or endothelial cells (Allan and Rothwell 2001; Montgomery and Bowers 2012). Moreover, they can bind directly to receptors on neurons, where their neuromodulatory actions can subsequently influence cognitive function. The hippocampus, a structure on which the acquisition of new declarative memories is absolutely dependent, expresses receptors for IL-1β (Gardoni et al. 2011), TNFα (Sairanen et al. 2001), and IL-6 (Schobitz et al. 1993).

Cytokines can promote neurite outgrowth, neurogenesis, and neuronal survival, in addition to being able to regulate synaptic transmission and synaptic plasticity (Marin and Kipnis 2013). However, chronically elevated levels of inflammatory cytokines can result in neuronal dysfunction. For example, IL-1 (Terrando et al. 2010) and IL-6 (Sparkman et al. 2006) have been implicated in lipopolysaccharide (LPS)-evoked cognitive dysfunction, and TNF receptor type 1 facilitates memory deficits associated with sepsis (Calsavara et al. 2015; Haslett et al. 2002). IL-17 is also a central regulator of CNS inflammatory responses and works synergistically with TNF and IL-1 under inflammatory conditions (Gaffen 2009). It is upregulated in hippocampal neurons following stimulation with LPS, indicating a role for it in neuroinflammation and the associated cognitive impairment (Sun et al. 2015). Disease states associated with such pathophysiology include epilepsy, stress-related disorders, and neurodegenerative diseases (Allan et al. 2005; Glass et al. 2010; Griffin 2006; Nguyen et al. 1998; Vezzani et al. 2011). Because neuroimmune interactions promote homeostasis of the nervous system (Marin and Kipnis 2013), chronic inflammation may also be a contributory factor in cognitive dysfunction exhibited by dystrophin-deficient patients with DMD.

Expression of Dystrophin in the CNS

The dystrophin gene has many independent, and tissue-specific, promoters for brain and skeletal muscle cells (Blake et al. 2002). Within the brain, dystrophin expression is only one-tenth of that found in muscle. However, brain tissue exhibits much greater variability in the protein products that are generated from the dystrophin gene. Included are the full-length protein (Dp427), on which this review will focus, and shorter proteins (Dp71, Dp260, Dp140, and Dp116; Gorecki and Barnard 1995). Quantitatively, Dp71 is the main dystrophin gene product in the brain, and its location around perivascular astrocyte endfeet (Haenggi et al. 2004; Tadayoni et al. 2012) suggests a role in blood-brain barrier (BBB) function (Amiry-Moghaddam et al. 2004), which may be a consideration in the ability of peripheral inflammatory molecules gaining access to the brain.

In the CNS, Dp427 is only found in neurons, and only within specific regions of the brain such as the hippocampus, amygdala, cerebellar Purkinje cells, and neocortex (Anderson et al. 2002; Bies et al. 1992; Chamberlain et al. 1988; Chelly et al. 1990; Comim et al. 2011; Cyrulnik and Hinton 2008; Knuesel et al. 2000; Lidov 1996; Sekiguchi et al. 2009). Similar to its function in striated muscle, dystrophin in neurons and glia associates with the DAPC of membrane-spanning proteins that link the intracellular cytoskeleton to the extracellular matrix. However, the DAPC in the brain is unlikely to act as a mechanotransducer as it does in skeletal muscle (Hendriksen et al. 2015). Furthermore, several variants of brain DAPC exist due to both the variety of CNS dystrophin protein products and the fact that they can also associate with DAPC components that are absent from skeletal muscle cells, such as β-dystrobrevin, ε-sarcoglycan, and the γ-syntrophins (Waite et al. 2009, 2012). It is likely, however, that this interaction of Dp427 with the DAPC in neurons does play a critical role in the formation and maintenance of new synaptic connections (Rodius et al. 1997).

Changes in the Structure and Function of Dystrophic Brains

Considerable debate continues as to whether loss of Dp427 alters the architecture of DMD brains, with some investigators noting relatively minor changes in DMD patient brains (al-Qudah et al. 1990; Bresolin et al. 1994; Dubowitz and Crome 1969; Rae et al. 1998) and others reporting a range of brain abnormalities including neuronal loss, neurofibrillary tangles, dendritic changes, and cortical atrophy (Itoh et al. 1999; Jagadha and Becker 1988; Lv et al. 2011; Rosman 1970; Rosman and Kakulas 1966; Septien et al. 1991; Yoshioka et al. 1980). Indeed, no clear correlations have been made between the different types of brain abnormalities observed in patients with DMD and the degree of intellectual impairment suffered. Moreover, investigations into possible biochemical and neuromodulatory mechanisms underlying the dystrophin-associated cognitive deficits in human DMD are relatively limited. The brains of boys with DMD have been reported to be hypometabolic in their use of glucose (Bresolin et al. 1994; Lee et al. 2002), a phenomenon that occurs in other conditions with known associated cognitive deficits, and may be related to reduced synaptic activity (Jueptner and Weiller 1995).

The functional role of specific proteins within the CNS is usually made much easier with the development of transgenic knockin or knockout mouse models. Therefore, the fortuitous generation of the mdx mouse (Bulfield et al. 1984), which is deficient in Dp427 (Blake and Kroger 2000) due to a nonsense mutation occurring in exon 23 of the dystrophin gene (Arechavala-Gomeza et al. 2010), has provided an opportunity for more functional studies on DMD to be carried out. Mdx mice are genetically comparable to human DMD and recapitulate striated muscle dysfunction, where skeletal muscle fibers progress through cycles of degeneration and regeneration until functional myofibers are replaced by collagen and adipocytes. Degeneration and regeneration of myofibers is associated with muscle inflammation (Grounds et al. 2008), but fibrosis and loss of limb muscle function is not as severe in mdx mice as in humans (Coulton et al. 1988; Dangain and Vrbova 1984), possibly due to compensatory mechanisms (Manning and O'Malley 2015). Nonetheless, consistent with human studies (al-Qudah et al. 1990; Bresolin et al. 1994; Dubowitz and Crome 1969; Rae et al. 1998), most work carried out on mdx mice has failed to note any gross abnormalities in brain structure (Bulfield et al. 1984; Miranda et al. 2009; Torres and Duchen 1987; Yoshihara et al. 2003). However, changes in cell number, size and/or shape have been noted in regions of the cerebral cortex and brain stem of mdx mice (Carretta et al. 2001, 2003, 2004; Minciacchi et al. 2010; Sbriccoli et al. 1995). Furthermore, enlargement of the lateral ventricles, possibly due to gray matter atrophy, has also been reported (Xu et al. 2015).

Similar to findings in patients with DMD, muscle levels of IL-6, TNFα, and IL-1β are also elevated in mdx mice (Huang et al. 2009; Huynh et al. 2013; Kurek et al. 1996), but alterations in circulating or CNS levels of inflammatory mediators in this mouse model have not yet been reported.

Hippocampal Dysfunction in mdx Mice

The neurobehavioral profile of mdx mice is characterized by deficits in cognitive function (Chaussenot et al. 2015; Muntoni et al. 1991; Perronnet et al. 2012; Vaillend et al. 1995, 2004), behaviors that are linked to dysregulation of hippocampal and amygdalar function. For example, the mdx mouse exhibits deficits in its capacity to learn and store spatial memories relative to controls (Vaillend et al. 1995, 2004). It also displays deficiencies in associative learning as well as in general processes of memory consolidation that are dependent on both the hippocampus and the amygdala (Chaussenot et al. 2015). Changes in long- and short-term memory formation and consolidation have been associated with altered synapse morphology and plasticity linked to a loss of Dp427 (Miranda et al. 2009; Vaillend et al. 1995, 2004). Interestingly, other in vivo and in vitro hippocampal-dependent functions, such as spatial learning, CA1 NMDA-dependent long-term potentiation (LTP), and its converse, long-term depression (LTD), have been found to be either intact (Sesay et al. 1996) or actually enhanced relative to controls (Dallerac et al. 2011; Vaillend and Billard 2002; Vaillend et al. 1998, 1999, 2004).

Dysregulated Synaptic Receptor Clustering in mdx Mice

GABA_A receptors.

Within the hippocampus, alterations in the number and localization of several receptor types have been noted in the absence of Dp427. Dp427 is exclusively expressed postsynaptically in dense puncta on neuronal cell membranes of hippocampal inhibitory synapses (Brunig et al. 2002; Kim et al. 1992; Knuesel et al. 1999, 2000; Levi et al. 2002; Sakamoto et al. 2008). Here, through its interaction with components of the DAPC, it makes a crucial contribution to synapse structure and function (for reviews see Haenggi and Fritschy 2006; Hendriksen et al. 2015; Perronnet and Vaillend 2010; Waite et al. 2012). This is likely to be facilitated through interactions between the syntrophin-dystrobrevin subcomplex (Compton et al. 2005; Matsumura et al. 1992), β-dystroglycan, and the synaptic cell adhesion molecules neurexin and neuroligin (Craig and Kang 2007; Graf et al. 2004; Kang et al. 2008; Sudhof 2008; Sugita et al. 2001; Waite et al. 2012; Zhang et al. 2010), all of which may be critically involved in anchoring GABAergic (for review see Fritschy et al. 2012) and other neurotransmitter receptors or channels in place at specific points in the postsynaptic membrane (Krasowska et al. 2014). Thus loss of dystrophin leads to the generation of malformed synapses (Knuesel et al. 1999, 2001; Kueh et al. 2008; Miranda et al. 2009, 2011), which will significantly alter how affected neurons respond to afferent stimuli (Graciotti et al. 2008). This will inevitably disrupt the precise and coordinated spatiotemporal and stratified neural network activity that is critical for normal hippocampal function and cognition (for review see Cohen et al. 2015).

This feature of dystrophin function has been particularly well described for inhibitory α2-subunit-containing GABA_A receptors in CA1 and CA3 regions of the hippocampus (Knuesel et al. 1999, 2001). Indeed, the fact that GABA_A receptor clustering in dendritic and pyramidal layers of mdx mice can be restored to normal levels by administration of specific antisense oligonucleotides established for the first time a direct link between receptor clustering and Dp427 expression (Vaillend et al. 2010). Furthermore, the proposed cellular correlate of new memory formation, NMDA-dependent LTP, can be restored in mdx mice using the same treatment. This evidence supports the suggestion that clustering of GABAergic neurotransmitter receptors at inhibitory synapses is essential for normal neuronal excitability and synaptic transmission (Dallerac et al. 2011).

More recently, the Wnt intracellular signaling pathway has emerged as an important player in the development and maintenance of normal neuronal structures within the CNS, as well as in modulating synaptic plasticity and cognitive function (for reviews see Inestrosa and Arenas 2010; Oliva et al. 2013). Significantly, Fuenzalida et al. (2016) demonstrated that the aforementioned deficiency in GABAergic signaling within hippocampal CA1 neurons of mdx mice can be restored by stimulating the noncanonical Wnt-5a pathway, which serves to increase the number of inhibitory synapses and GABA_A receptors in CA1 without any attendant increase in GABA release probability or release sites.

Nicotinic acetylcholine receptors.

Alterations in the expression, distribution, and function of other receptor subtypes have also been noted in dystrophic brains. For example, behavioral studies reporting a reduced response to nicotine in a passive avoidance memory task suggested altered nicotinic acetylcholine receptor (nAChR) expression or responsivity (Coccurello et al. 2002), work that has been supported by in vitro research demonstrating a reduction of hippocampal α3-nAChR mRNA (Wallis et al. 2004) and α7-nAChR binding sites (Ghedini et al. 2012). However, these alterations, which would be expected to lead to decreased nicotinic cholinergic synaptic signaling, appear to be compensated for by an upregulation in nAChR-mediated cholinergic transmission in the same brain region (Parames et al. 2014). However, the fact that the expression of hippocampal muscarinic receptors is unchanged in dystrophic mice (Yoshihara et al. 2003) indicates that dystrophin does not exert a generalized effect on the expression of all cholinergic receptor subtypes. Nonetheless, a reduction in acetylcholinesterase activity in the hippocampus of mdx mice may be illustrative of a non-receptor-dependent form of compensation for an overall reduction in cholinergic signaling in the CNS of dystrophic mice (Cohen et al. 2015; Comim et al. 2011).

Kainic acid/AMPA receptors.

Yoshihara et al (2003) demonstrated that kainic acid/AMPA receptor density was significantly reduced in several brain regions, including the hippocampus, in the mdx mice relative to wild-type animals. However, this may reflect another homeostatic compensatory mechanism, this time for increased glutamatergic input (Miranda et al. 2009, 2011).

In the context of this review, it is interesting to note that changes in neuromodulatory cytokines such as IL-6, IL-1β, and TNFα are known to affect both presynaptic and postsynaptic channels and receptors (Vezzani and Viviani 2015). Indeed, as previously mentioned, circulating levels of these proinflammatory cytokines are elevated in patients with DMD (Chahbouni et al. 2010), but so far there is no evidence that brain levels of these neuromodulatory cytokines are altered.

Changes in Hippocampal Synaptic Function in mdx Mice

Consistent with a reduction in the size and number, but not function (Kueh et al. 2008), of postsynaptically localized neuronal GABA_A receptor clusters in mdx mice (Anderson et al. 2002; Knuesel et al. 1999, 2001; Kueh et al. 2008, 2011; Vaillend et al. 2010), one might expect reduced inhibitory input to the neuron to be manifested as a measurable increase in overall neuronal excitability in affected regions of the brain. However, the picture emerging from electrophysiological studies is not quite so straightforward.

Whereas NMDA-induced Schaffer collateral-CA1 LTP in mdx hippocampus is abnormally enhanced in a bicuculline-sensitive manner, which would tally with reduced inhibitory input onto the CA1 neurons (but see Sesay et al. 1996; Vaillend et al. 1998, 1999, 2004), a significant increase in the frequency of spontaneous miniature inhibitory postsynaptic currents (mIPSCs) has been recorded from mdx CA1 hippocampal neurons. This is likely to be caused by increased neurotransmitter release probability from somatic inhibitory synapses contacting CA1 pyramidal neurons (Graciotti et al. 2008), due to an increased number of inhibitory synapses innervating the mdx CA1 neurons (Miranda et al. 2009; Perronnet and Vaillend 2010). This suggestion was supported by a contemporaneous study by Del Tongo et al. (2009) in which a significant increase in the number of parvalbumin-expressing GABAergic interneurons in the dorsal hippocampus of the mdx mouse was observed.

The increase in hippocampal GABAergic innervation associated with an absence of dystrophin may have arisen as a systemic compensation for impaired GABAergic synaptic clustering and function but still appears to be insufficient to fully restore GABA concentrations within the hippocampus (Xu et al. 2015). Graciotti et al. (2008) speculated that an increase in inhibitory transmitter release specifically onto CA1 neurons may also occur via a retrograde signaling system, possibly involving presynaptically expressed scaffolding proteins such as neurexins (Craig and Kang 2007; Kang et al. 2008; Pilgram et al. 2010; Sudhof 2008; Sugita et al. 2001; Zhang et al. 2010). Not only do neurexins form direct structural links with the DAPC, which, in turn, facilitates the postsynaptic clustering of GABA_A receptors (Fritschy et al. 2012; Graf et al. 2004; Krasowska et al. 2014; Kueh et al. 2008), but they also modulate presynaptic release of neurotransmitters and vesicle cycling (Futai et al. 2007; Olsen et al. 2006; Sugita et al. 2001; Tanaka et al. 2000). Evidence for such a retrograde signaling mechanism has emerged from studies on Drosophila, where dystrophin isoforms retrogradely regulate presynaptic transmitter release from both inhibitory and excitatory synapses (Bogdanik et al. 2008; Fradkin et al. 2008; van der Plas et al. 2006; Wairkar et al. 2008). However, conclusive evidence for the existence of such a retrograde signaling mechanism occurring specifically at mammalian inhibitory synapses, where Dp427 is exclusively expressed, remains elusive. Tantalizingly, Vaillend and colleagues have demonstrated that dystrophin loss affects presynaptic ultrastructural organization in hippocampal excitatory glutamatergic synapses by increasing not only the size of their postsynaptic densities (Miranda et al. 2009) but also the density of docked vesicles (Miranda et al. 2011). Such a change in the density of docked vesicles, possibly in conjunction with reduced dendritic inhibition (Vaillend and Billard 2002; Vaillend et al. 1999), is likely to alter excitatory synapse physiology and may account for the facilitated induction of NMDA-dependent LTP observed at Schaffer collateral-CA1 synapses in mdx mice (Vaillend and Billard 2002). Moreover, such a potentiation of excitatory output could, at least in part, account for the increase in mIPSC frequency observed in the CA1 region of mdx mice (Graciotti et al. 2008), because it is well established that hippocampal inhibitory interneurons are directly innervated by glutamatergic excitatory projections emanating from the pyramidal cell layer (Gulyas et al. 1993; McBain and Dingledine 1993). It is still unclear how a cytoskeletal protein expressed exclusively at inhibitory CA1 synapses can exert such an effect on excitatory synapses, but dysregulation of calcium homeostasis (Hopf and Steinhardt 1992; Tuckett et al. 2015), defective brain vascular permeability (Goodnough et al. 2014; Nico et al. 2004), and/or dysbindin relocalization (Miranda et al. 2011) have been posited as possible mechanisms for this phenomenon.

A Role for Immune Molecules in Synaptic Transmission

Although not yet investigated in dystrophin-deficient models, it is interesting to note that receptors for immune molecules in the CNS have expression patterns that bear similarities to the synaptic receptors discussed previously. The major histocompatibility complex class I (MHC I) is a complex molecule that provides the cellular signature that allows the immune system to differentiate between self and non-self. MHC I is expressed in neurons (Corriveau et al. 1998) of the thalamus, hippocampus, cortex, and cerebellum (Marin and Kipnis 2013), with the latter three showing overlap with brain regions expressing DP427 (Anderson et al. 2002; Bies et al. 1992; Chamberlain et al. 1988; Chelly et al. 1990; Comim et al. 2011; Cyrulnik and Hinton 2008; Knuesel et al. 2000; Lidov 1996; Sekiguchi et al. 2009). MHC I is upregulated in skeletal muscle from muscular dystrophy patients (Nagappa et al. 2013), but it is not yet known if there are changes in expression of this immune molecule in the CNS of patients with DMD. However, MHC I has been detected clustered at the postsynaptic density of synapses of hippocampal neurons (Goddard et al. 2007), the site of expression of postsynaptic receptors that are altered in dystrophic hippocampal neurons. Furthermore, hippocampal neurons from a MHC I knockout mouse display enhanced synaptic plasticity and potentiated LTP (Goddard et al. 2007; Huh et al. 2000), features that bear a striking similarity to those reported in mdx mice (Dallerac et al. 2011; Vaillend and Billard 2002; Vaillend et al. 1998, 1999, 2004).

Cytokines represent another class of immune molecules that are crucial to the homeostatic function of the nervous system, and given the chronic elevations in their secretion in dystrophinopathies, they are discussed below in relation to their potential role in the changes in hippocampal structure and function in dystrophic neurons.

Role of Cytokines in Hippocampal Function

IL-1β.

IL-1β is crucial for the generation of NMDA-dependent LTP in hippocampal CA1 neurons (Schneider et al. 1998), and reduced signaling through IL-1 receptors results in impaired learning (Goshen et al. 2007). However, overexpression, or elevated levels, of IL-1β also results in memory deficits (Moore et al. 2009) and decreased synaptic strength (Ross et al. 2003). In terms of its effects on postsynaptic receptors, IL-1β can increase the surface expression of hippocampal GABA_A receptors and potentiate GABA-evoked inhibitory currents (Wang et al. 2012), features that have been observed in mdx hippocampal neurons (Graciotti et al. 2008). A relationship between IL-1β and cholinergic signaling in the hippocampus has also been postulated, because IL-1β causes a reduction in extracellular acetylcholine in this brain region (Rada et al. 1991), which is an effect that is also correlated with memory defects (Taepavarapruk and Song 2010). This could perhaps be compared to the reduced sensitivity of the mdx hippocampus to cholinergic agonists (Coccurello et al. 2002) due to decreased nAChR binding sites (Ghedini et al. 2012).

IL-6.

Local increases in hippocampal IL-6 are observed following intense synaptic or neuronal activity such as that induced by LTP induction protocols (Balschun et al. 2004; Jankowsky et al. 2000). However, in contrast to the effects of IL-1β, Balschun et al. (2004) found that inhibition of IL-6-mediated signaling in “healthy” brains actually improved long-term memory, suggesting a possible negative regulatory role for IL-6 on LTP by limiting memory acquisition (Balschun et al. 2004; Li et al. 1997). These findings correlate with in vivo behavioral studies that have demonstrated that neutralization of IL-6 improved long-term spatial memory in rats (Balschun et al. 2004) and that IL-6 knockout mice displayed an enhanced capacity for learning and memory (Braida et al. 2004). Conversely however, another study found that hippocampus-dependent learning was impaired in IL-6 knockout mice (Baier et al. 2009). Overexpression of IL-6 results in alterations in both inhibitory and excitatory synapses and in abnormal dendritic spine formation. These structural alterations are associated with impairments in cognitive abilities and deficits in learning, similar to symptoms exhibited by individuals with autism (Wei et al. 2012). This is an interesting finding because a significant number of boys with DMD also display behaviors on the autistic spectrum (Banihani et al. 2015).

Prenatal exposure to IL-6 in rats leads to increased levels of IL-6 in the hippocampus that are associated with altered expression of NMDA and GABA_A receptors and with deficits in spatial learning (Samuelsson et al. 2006). This has led to the proposal that exposure to pathological levels of IL-6 during critical stages of neurodevelopment may contribute to deficits in cognitive function. Although it is currently unknown whether IL-6 is elevated locally in brain regions such as the hippocampus in mdx mice, given that elevated IL-6 levels are a hallmark of dystrophin-deficiency and are elevated in mdx plasma (Pelosi et al. 2015), it is possible that this may be a factor in dystrophinopathy-associated hippocampal dysfunction.

TNFα.

TNFα concentration is elevated in both patients with DMD and mdx mice, with glial cells representing likely sources of endogenous TNF (Stellwagen and Malenka 2006). Interestingly, in an mdx IL-6 mouse, where IL-6 levels are elevated, TNFα levels are also potentiated (Pelosi et al. 2015). Activation of hippocampal TNF type 1 receptors increases the frequency and amplitude of mEPSCs, which stimulates an increase in the expression of AMPA receptors and an overall increase in synaptic strength (Beattie et al. 2002; Dummer et al. 2002). Weakening of inhibitory synaptic input also contributes to the excitatory effects of TNFα on hippocampal neurons, which may be due to a persistent decrease of inhibitory synaptic strength, caused by downregulation of membrane-expressed GABA_A receptors (Pribiag and Stellwagen 2013). Although not consistent with the changes in electrophysiological properties of dystrophin-deficient hippocampal neurons, it remains to be elucidated whether this inflammatory molecule may have a compensatory role in cognitive dysfunction in dystrophinopathies.

Other Possible Contributory Factors in DMD-Related Cognitive Dysfunction

Nitric oxide.

In a manner similar to that proposed for elevated levels of peripherally released cytokines altering hippocampal function, others have discovered that peripherally released nitric oxide (NO), which is decreased in DMD, also exerts neuromodulatory effects. Patients with DMD and mdx mice display up to an 80% reduction of nitric oxide synthase (NOS) activity, because loss of dystrophin leads to a secondary loss of neuronal NOS from muscle (Brenman et al. 1995; Chang et al. 1996). Because muscle is the primary source for systemic NO, and this is reduced in patients with DMD, Deng et al. (2009) investigated whether this had an impact on hippocampal neurogenesis in the mdx mouse. Adult neurogenesis within the hippocampus is a proposed means by which alterations in synaptic plasticity and memory formation may occur. They found that neurogenesis was disrupted in mdx hippocampal tissue but that normal neurogenesis could be restored by increasing muscle, and subsequently serum, NO levels (Deng et al. 2009). This evidence further supports the potential role of peripheral molecules in hippocampal dysfunction in dystrophinopathies.

Mitochondrial dysfunction.

Loss of dystrophin in skeletal muscle results in structurally unstable fibers that are more porous to the extracellular environment, resulting in excessive influx of calcium. Poor calcium handling and subsequent activation of proteases and/or lipases leads to calcium overload within cellular mitochondria and subsequent muscle degeneration. Because mitochondrial ATP is crucial to a myriad of physiological functions, dysregulation of cellular energy homeostasis is likely to contribute to dystrophinopathy-associated pathophysiology, including impaired intracellular calcium signaling in skeletal muscle and in the brain (Timpani et al. 2015). Tracey et al. (1995) found evidence of raised inorganic phosphate ratio, relative to ATP, phosphocreatine and phosphomonomers, in DMD brains, and there is also evidence that the concentration of choline-containing compounds is increased in DMD brains (Kato et al. 1997; Rae et al. 1998), which is usually interpreted as being caused by increased membrane turnover and degradation or by decreased membrane stability, also seen in a number of other brain disorders (Anderson et al. 2002). However, in contrast, a more recent study showed a deficit in total choline in the brains of patients with DMD (Kreis et al. 2011).

The mdx mouse has raised levels of hippocampal choline-containing compound (Rae et al. 2002; Xu et al. 2015), inorganic phosphate, and pH, and a reduced level of total creatine (Tracey et al. 1996). The fact that these metabolic changes are also detected in muscle (Tracey et al. 1996) suggests bioenergetic similarities in other tissues that lack dystrophin. Uptake of glucose in mdx mice is increased relative to that in controls, reflecting an increase in metabolism, and this may indicate increased brain activity due to decreased GABA-evoked inhibition (Rae et al. 2002). The fact that the GABA agonist muscimol had a reduced dampening effect on glucose metabolism in mdx brains (Rae et al. 2002) is consistent with this theory, in addition to the reduction and abnormal clustering of synaptically located GABA_A receptors in the hippocampus, cerebellum, amygdala, and cerebral cortex (Knuesel et al. 1999, 2001; Kueh et al. 2008, 2011; Perronnet and Vaillend 2010; Sekiguchi et al. 2009). A recent study demonstrated a significant deficit in the amount of GABA within the hippocampal region of mdx mice and raised concentrations of the anti-oxidant molecule glutathione (Xu et al. 2015), which is likely to be a compensatory development against the increased quantity of reactive oxygen species produced by increased metabolism.

Such alterations in cellular metabolism, energy requirements, and redox status in mdx brain tissue might also be expected to influence neuronal responses to hypoxia in dystrophic brain tissue. Indeed, this has proved to the case, with CA1 hippocampal neurons displaying particular sensitivity to hypoxic insults, with greater and more rapidly developing decreases in synaptic transmission relative to control brain slices at Schaffer collateral-CA1 synapses (Godfraind et al. 2000; Mehler et al. 1992). This altered response to hypoxia of mdx mouse hippocampal neurons may be due to a combination of factors, including reduced GABA_A receptor clustering in dystrophic neurons, because GABA_A receptor activation exacerbates oxygen-glucose deprivation-induced neuronal injury (Muir et al. 1996). However, Godfraind et al. (2000) hypothesized that the more rapid failure of nerve conduction in the dystrophic tissue under hypoxic conditions was due to “excessive leakiness” of nerves and poor regulation of ionic homeostasis, possibly due to Na⁺-K⁺ and/or calcium ATPase activity. Disrupted cellular calcium ATPase activity, with a knock-on effect on cellular calcium homeostasis, has been well described in dystrophic muscle and also seems to be mirrored in certain neural tissues. Although studies have not shown that either elevated intracellular calcium levels or the number of calcium-positive neurons was actually associated with frank neuronal loss per se in the mdx mouse (Tuckett et al. 2014), modulation of synaptic plasticity (both LTP and LTD) is a calcium-dependent process that is absolutely contingent on tight spatiotemporal control of intracellular neuronal calcium concentration (for review see Baker et al. 2013; Malenka and Bear 2004). Therefore, any dysregulation of calcium homeostasis or calcium signaling pathways would undoubtedly disrupt both short- and long-term plasticity in affected neurons and might at least partially account for some of the cognitive deficits observed in the mdx mice.

Matrix metalloproteinases.

Matrix metalloproteinases (MMPs) are a large family of zinc-dependent extracellular and membrane-bound endopeptidases that can cause proteolysis of selected components of the extracellular matrix in skeletal muscle. MMP-9 is one member of the family that is expressed at low levels under basal conditions but which can be induced by numerous factors, including inflammatory cytokines (Hu et al. 2007). MMP-9 is also important in synaptic plasticity, thereby contributing to learning and memory. For example, MMP-9 knockout mice exhibited deficits in particular components of CA1 LTP and spatial learning (Nagy et al. 2006). Furthermore, pharmacological inhibition of MMP-9 resulted in destabilization of LTP (Wojtowicz and Mozrzymas 2010). MMP-9 is a key molecule in the regulation of dendritic spine morphology and has been implicated in several neurological disorders (Stawarski et al. 2014). In the context of the dystrophinopathies, MMP-9 is elevated in mdx skeletal muscle (Kherif et al. 1999; Li et al. 2009). Thus MMP-9 represents a molecular candidate, possibly secondary to raised cytokine levels, that could be an important contributory factor in hippocampal dysfunction in patients with DMD.

Stress.

Immune cells respond not only to tissue damage and infection but also to stress, both physical and psychological in nature, by secreting cytokines. Thus stress may also indirectly contribute to inflammation-related hippocampal dysfunction in patients with DMD. However, the hippocampus is also directly affected by stress factors and is, in fact, susceptible to damage by acute psychological stress, resulting in difficulties in memory retrieval and the acquisition and storage of new memories (Sapolsky 1996). In patients with DMD, the contributory effects of chronic corticosteroid exposure on behavior is controversial. Long-term treatment with corticosteroids, used in DMD to suppress chronic inflammation, is associated with changes in mood, memory, and attention (Brown et al. 2008), although not all studies have detected such changes (Banihani et al. 2015). Nonetheless, patients with DMD do exhibit an increased incidence of anxiety and depression (Banihani et al. 2015; Fitzpatrick et al. 1986; Pangalila et al. 2015). Mdx (Manning et al. 2014) and also mdx^3cv (Vaillend and Ungerer 1999) mice, which lack Dp427, Dp71, and Dp140, display stress-related behaviors comparable to symptoms of depression and anxiety, although not all behavioral assessments detected signs of anxiety in dystrophic mice (Sekiguchi et al. 2009). Enhanced freezing behavior in response to restraint stress, which is not painful but induces a psychological stress, was also observed in mdx mice, but it should be noted that this was not associated with any change in corticosteroid levels (Yamamoto et al. 2010). Nonetheless, we have observed the efficacy of the tricyclic antidepressant amitriptyline in alleviating depressive and anxiety-like behaviors displayed by mdx mice, changes that were linked to altered levels of hippocampal neurotransmitters (Manning et al. 2014). It remains to be seen if loss of dystrophin from the CNS contributes directly to a heightened stress responses or if these effects are secondary to the disease itself. Moreover, further research must be carried out to determine if cross talk between the stress axis and the immune system is important in cognitive dysfunction in DMD.

Conclusions

DMD is a devastating disease characterized by a progressive loss of physical function and premature death in affected boys. Despite the continuing absence of an effective cure, palliative treatments, including prednisone and respiratory ventilation, are extending the life expectancy and life quality of patients. As such, investigation into the cognitive deficits and comorbid CNS disorders associated with loss of dystrophin from neurons in the CNS is an emerging area of research. Boys with DMD exhibit nonprogressing cognitive dysfunction, with deficits in verbal, short-term, and working memory. In the genetically comparable dystrophin-deficient mdx mouse model of DMD, specific deficits in synaptogenesis and channel clustering at synapses is evident, and deficits in some, but not all, types of learning and memory have been identified. Key changes underlying hippocampal dysfunction in dystrophin-deficient models are alterations in the number and clustering of postsynaptic receptors. The absence of dystrophin results in malformed and dysfunctional synapses that are likely to disturb the precise and coordinated neural network activity crucial to the formation and consolidation of memories in the hippocampus. Indeed, loss of inhibitory input in the mdx hippocampus has been associated with abnormally enhanced NMDA-induced Schaffer collateral-CA1 LTP, which is sensitive to a GABA_A antagonist. It is still unclear how enhanced LTP in a dystrophin-deficient hippocampus results in deficits in learning and memory. However, other studies have revealed that loss of neuronal dystrophin is associated with alterations in cellular metabolism, energy requirements, and redox status, resulting in hippocampal neurons being particularly vulnerable to hypoxic insults, all of which may contribute to hippocampal dysfunction. Furthermore, psychological stress can have a negative impact on hippocampal function, resulting in difficulties in memory retrieval and the storage of new memories. We have also identified the potential importance of inflammatory mediators, which are chronically elevated in peripheral tissue and plasma of patients with DMD and mdx mice, in hippocampal function. Key proinflammatory cytokines such as IL-1β, TNFα, and IL-6 exert neuromodulatory effects on the hippocampus and have been linked to altered capacities for learning and the formation of memories. This novel research avenue may reveal the importance of neuroimmune interactions in the CNS of dystrophin-deficient patients. Although this is an area of research still in its infancy, our understanding of both normal cognitive function in healthy brains and specific cognitive deficits in the brains of those suffering from DMD will be improved by continuing research in this field.

GRANTS

We acknowledge funding support from Muscular Dystrophy Ireland.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.G.R. and D.O. drafted manuscript; M.G.R. and D.O. edited and revised manuscript; M.G.R. and D.O. approved final version of manuscript.