Neuronal hyperexcitability in dystrophin-deficient mdx hippocampal neurons: the importance of interleukin-6 and GABAergic regulation

Abstract

Duchenne Muscular Dystrophy (DMD) is a severe neuromuscular disorder arising from loss of the structural protein, dystrophin. It also often presents with cognitive deficits and susceptibility to epilepsy. Expressed in neurons of the hippocampus, dystrophin plays an important role in synapse formation, specifically the post-synaptic organisation of γ-aminobutyric acid A receptors (GABA_ARs). This study explored possible interactions between interleukin (IL)-6, which is elevated in DMD, and GABA_AR signalling in cultured hippocampal neurons of dystrophic mdx mice. Immunofluorescent imaging revealed altered development of network connectivity that displayed similar characteristics to dystrophin-expressing neurons cultured in elevated levels of IL-6. Mdx neurons dependably exhibited spontaneous oscillations. Calcium (Ca²⁺) signalling was further modulated by exposure to agonists and antagonists of GABA_A and GABA_BRs. IL-6-evoked Ca²⁺ responses were enhanced by muscimol, a GABA_AR agonist, in wildtype (WT) and mdx neurons, whilst bicuculline, a GABA_AR antagonist, only suppressed IL-6-evoked Ca²⁺activity in WT neurons. The GABA_BR agonist, baclofen, enhanced IL-6-evoked Ca²⁺ responses only in mdx neurons. Our findings support dysfunctional GABAergic signalling in hippocampal neurons that lack dystrophin, resulting in aberrant neuronal network excitability. The contribution of elevated levels of IL-6 further impact upon Ca²⁺ dyshomeostasis in dystrophic neurons and may underpin cognitive changes reported in dystrophinopathies.

Introduction

Loss of the sarcolemma-spanning structural protein, dystrophin, from skeletal muscle fibres results in increased susceptibility to contraction-induced damage, inflammation and myophagocytosis^1,2. These pathophysiological changes underpin progressive loss of mobility and early death³ in the most severe form of the X-linked dystrophinopathy, Duchenne Muscular Dystrophy (DMD). Individuals with DMD more commonly present with cognitive delays and neurodevelopmental disorders⁴, including lower intelligence quotients (IQs)^5–7 and poor verbal, short-term and working memory^8–10. Increased incidence of seizures and epilepsy are also reported in individuals with DMD^11,12.

In healthy individuals, dystrophin is expressed in several brain regions, including the hippocampus, which is essential for the acquisition of novel memories^13–16. The mouse model of DMD, dystrophin-deficient mdx mice¹⁷, have a poor capacity to learn new tasks and exhibit deficits in storing spatial memories^18–21. Furthermore, hippocampal long-term potentiation, the proposed molecular correlate of learning and memory formation, is suppressed in mdx mice²², and normalised when dystrophin is restored¹⁵, further implicating the protein in hippocampal function. Increased neuronal excitability in dystrophic hippocampal neurons has also been reported²¹.

Quantitatively, Dp71 is the main dystrophin gene product in the brain, being expressed in neurons of the hippocampal dentate gyrus and CA1 regions²³ and in glia^24,25. Dp71 is also found in the glial perivascular endfeet, where it plays an important role in blood brain barrier integrity, which is compromised in dystrophic mdx mice^26,27. Dp140, present in human hippocampal tissue²⁸, is also important in the brain microvasculature²⁹ and both of these isoforms have been linked to cognitive deficits³⁰. Full-length, Dp427 dystrophin, is expressed in CA1 hippocampal neurons^31,32, where it has an important role in both synaptogenesis³³ and receptor anchoring in the post-synaptic membrane²⁹. Dp427 is co-expressed with post-synaptic γ-aminobutyric acid (GABA)_A receptors (GABA_A Rs), where it facilitates clustering and stabilisation of inhibitory GABAergic synapses, important in signal transduction^34,35. Indeed, mutations in dystroglycan, a cell adhesion molecule present at inhibitory synapses, has been linked to severe cognitive deficits and epilepsy in muscular dystrophies through changes in GABAergic synaptic plasticity³⁶.

In dystrophic mdx hippocampal tissue, expression of GABARs is decreased³⁷. Evidence also indicates that Wnt signalling is a modulatory factor underpinning hippocampal GABAergic efficiency, with consequences for neuronal plasticity and cognitive function in dystrophic models³⁸. Interneurons are important in generating different oscillating rhythms in the hippocampus, acting as pacemakers, facilitating spontaneous feedback loop oscillation or amplifying extrinsic signals³⁹. Although GABAergic inhibitory interneurons account for a relatively small proportion of total neurons in the hippocampus⁴⁰, they are specialised regarding the temporal regulation of neuronal excitability and action potential generation. Moreover, loss of dystrophin impacts the hippocampal arrangement of molecular machinery which orchestrates the spatio-temporal configuration of GABAergic synaptic transmission⁴¹ and hippocampal interneurons are negatively impacted by epileptic activity⁴².

Interleukin-6 (IL-6) is a pleiotropic cytokine with neuromodulatory actions, exhibiting the capacity to modify post-synaptic expression of GABA_A receptors in cortical slices⁴³. Moreover, overexpression of IL-6 reduced immunostaining for inhibitory synapses⁴⁴ and synaptic network activity was altered in hippocampal neurons cultured in the presence of IL-6, an action that involved GABAergic interneurons⁴⁵. Given that the integrity of the blood brain barrier is impaired in the absence of dystrophin^26,27 and chronic inflammation in both human individuals with DMD^46,47 and mdx mice^48,49 is characterised by elevated circulating concentrations of pro-inflammatory cytokines, such as IL-6, the actions of this cytokine in the brain may be important in the reported dysfunctional activity of GABARs. To better understand this relationship, we have explored the role of IL-6 in neuronal connectivity in dystrophin-expressing neurons. We further examined intracellular Ca²⁺ activity in wildtype (WT) and dystrophic mdx neurons exposed to IL-6 in the presence or absence of GABA_AR and GABA_BR agonists and antagonists.

Results

Exposure to IL-6 increased synaptic connections in dystrophin-expressing cultured hippocampal neurons

To explore the cellular location of neuronal dystrophin, hippocampal neurons were cultured from dystrophin-expressing C57BL/6 mice mice. Both anti-dystrophin and anti-MANDRA-1 antibodies are directed towards the dystrophin c-terminus, and detect Dp71, Dp140 and Dp427, but not Dp40⁵⁰. Previous work from our group has demonstrated that in the absence of Dp427 (in mdx hippocampal tissue), overall staining to the above antibodies is reduced but not eliminated in CA1, CA3 and dentate gyrus regions²², likely reflecting the continued presence of DP71 and Dp140 isoforms in this region. In cultured neurons co-stained with anti-dystrophin and anti-MANDRA-1 antibodies, the pattern of staining overlapped. Staining was punctate and strongest in hippocampal neuronal cell bodies and dendrites (Fig. 1A). Consistent with post-synaptic localisation, anti-MANDRA-1 staining was co-expressed with anti-PSD-95, a marker of the post-synaptic density (Fig. 1B, n = 7 images, co-localisation indicated by arrows).

Fig. 1.

Elevated IL-6 increased dystrophin-expressing synaptic connections in cultured hippocampal neurons (A) The images show immunofluorescently-labelled hippocampal neurons cultured from C57BL/6 mice. Neurons have been labelled with an antibody (MANDRA-1) that identifies both Dp427 & Dp71 isoforms of dystrophin (green staining) and the Dp427 isoform alone (red staining) and (B) red MANDRA-1 staining co-expressed with the post-synaptic marker, PSD-95 (green staining). Arrows indicated co-expression of dystrophin with the post-synaptic marker (yellow labelling). (C) The representative immunofluorescent images and data plots illustrate the intensity (corrected total cell fluorescence, CTCF) of IL-6 staining in the neuronal cell bodies and processes under control culture conditions (black circles) and in culture medium supplemented with IL-6 (1 nM, > 48 h, grey open circles). Arrows indicate punctate expression of IL-6 in the cell bodies and processes. (D) The images and data plot illustrates the number of discrete clusters of dystrophin/10 µm in the processes of neurons cultured under control conditions or in culture medium supplemented with IL-6. Arrows highlight the increased expression of dystrophin clusters in treated neurons. (E) The images and data plots show clustering of IL-6 receptor (IL-6R) labelling (green staining) at synapses (indicated by arrows) labelled with the pre-synaptic marker, synaptophysin (red staining). The merged images illustrate regions where labelling is co-expressed (yellow staining). The number of IL-6R clusters and synaptophysin-labelled synapses under control and IL-6-exposed conditions are illustrated in the pooled data. * and ***p < 0.05 and p < 0.001, respectively. Scalebars: 35 µm.

As elevated levels of circulating IL-6 have been reported in mdx mice⁴⁸, we sought to determine if culturing medium supplemented with IL-6 had any effect on neuronal development and connectivity. Under standard culture conditions (i.e. no supplemental IL-6), the pro-inflammatory cytokine, IL-6, was detected in neuronal cell bodies as bright puncta of cytosolic IL-6 staining, consistent with compartmentalisation of IL-6 into vesicles (indicated by arrows, n = 18 images, Fig. 1C). Culturing media with supplemental IL-6 (1 nM, > 48 h) did not modify expression of IL-6 in the cell body, as measured by comparing relative fluorescence (p = 0.51, n = 18 images, Fig. 1C). However, in the neuronal processes, IL-6, which was expressed at low levels under control conditions, was more abundant in neurons cultured in media supplemented with IL-6 (p = 0.01, n = 18 images, Fig. 1C). Expression of IL-6 in the processes was punctate (indicated by arrows) but also strongly localised to structures morphologically similar to growth cones (indicated by arrowhead, Fig. 1C).

The density of dystrophin clusters, observed as distinct puncta on the processes of neurons, was also increased when culture medium was supplemented with IL-6 (0.56 ± 0.31 clusters/10 µm, n = the average of 3 processes from 28 images) compared to controls (0.39 ± 0.27 clusters/10 µm, n = the average of 3 processes from 26 images, p = 0.035, Fig. 1D). In both control and IL-6-exposed cultures, IL-6 receptor (IL-6R) – immunopositive puncta were clustered close to synaptophysin-expressing synapses (Fig. 1E, indicated by arrows). In control cultures, 0.425 ± 0.18 IL-6R immune-positive clusters/10 µm (n = the average of 3 processes each from 12 images) were detected, whereas in neurons grown in culture medium supplemented with IL-6, we detected 0.744 ± 0.17 IL-6R clusters/10 µm (n = the average of 3 processes each from 11 images, p = 0.0003, Fig. 1E). In parallel, the number of synaptophysin-expressing synaptic connections was also elevated in neurons cultured in IL-6-supplemented medium (0.75 ± 0.39 synapses/10 µm, n = the average of 3 processes from 10 images, Fig. 1E) compared to control cultures (0.45 ± 0.3 synapses/10 µm, n = the average of 3 processes from 13 images, p = 0.048, Fig. 1E). Thus, exposure to elevated IL-6 during network development in hippocampal neurons resulted in altered synaptic contacts. Furthermore, the density of clusters expressing dystrophin, IL-6 and IL-6Rs, which were located at synaptic contacts in neural processes, was increased.

Absence of dystrophin resulted in altered Ca²⁺ kinetics in cultured hippocampal neurons

Dysfunctional regulation of intracellular Ca²⁺ has previously been reported in dystrophic hippocampal neurons²¹, possibly due to reduced voltage-dependent block of NMDA receptors by magnesium⁵¹. Cultured hippocampal neurons from dystrophin-deficient mdx mice survived and developed in a comparable manner to dystrophin-expressing WT neurons and no obvious differences in immuno-labelled NMDA (NMDAε) receptor expression were apparent. Strong, punctate expression was present in the cytosolic compartment of WT and dystrophin-deficient mdx neurons and extended into primary neural processes (Fig. 2A). However, prominent differences were apparent in the cellular activity of dystrophic neurons, with the observation of increased numbers of Ca²⁺ spikes (defined as a 20% rise and fall relative to the baseline range), but more striking was the culmination of intracellular Ca²⁺ into larger amplitude, longer duration and sometimes cyclical oscillations, although manifestation of this increase in excitability was variable among mdx cultures.

Fig. 2.

Loss of dystrophin is associated with neuronal hyperexcitability (A) The brightfield and immuno-labelled images show NMDA receptor (NMDA R) expression in wildtype (WT) and mdx cultured hippocampal neurons. Scalebar: 35 µm. (B) The dot plots show pooled data of the (i) frequency, (ii) amplitude and (iii) duration (full width, half max: FWHM) of spontaneous Ca²⁺ peaks in WT (black circles) and mdx (red triangles) cultured neurons under basal conditions. (C) The representative Ca²⁺ traces illustrate increased spontaneous Ca²⁺ activity in mdx hippocampal neuronal cultures (red traces) in comparison to WT neurons (black/grey traces) and traces and (D) The data plot illustrates suppression of spontaneous neuronal activity in dystrophic mdx hippocampal neurons following incubation with the neurotoxin, tetrodotoxin (TTX). (E) The data plot illustrates enhanced Ca²⁺ responses induced by high (50 mM) K⁺ in mdx neurons. ** and ***p < 0.01 and p < 0.001, respectively.

Although baseline Ca²⁺ levels were similar in hippocampal neurons cultured from WT mice (37.4 ± 25.5 AUs, n = 150 neurons) and mdx mice (34.2 ± 20.5 AUs, p = 0.17, n = 274 neurons), the frequency of spontaneous Ca²⁺ spikes detected in a 10-min period where neurons were only exposed to HBSS, was significantly elevated in mdx neurons (mean: 10.5 ± 3.4 spikes per minute, range: 14.26, n = 169 neurons from 5 cultures) relative to WT neurons (mean: 8.3 ± 2.5 spikes per minute, range: 11.68, n = 103 neurons from 4 cultures, p < 0.0001, Fig. 2B i). Although the software did not differentiate between spikes and larger amplitude Ca²⁺ oscillations, (examples in Figs. 2Bii and 2C, red traces), manifestation of such Ca²⁺ events was recorded manually. 18% of WT neurons exhibited spontaneous Ca²⁺ oscillations (n = 24/133), in comparison to 46% of mdx neurons (n = 91 /198, p < 0.0001, Fisher’s exact test). Overall, the amplitude of each Ca²⁺ spike was elevated in mdx neurons (0.03 ± 0.02 AUs, n = 186 neurons from 5 cultures), compared to WT neurons (0.02 ± 0.02 AUs, n = 105 neurons from 4 cultures, p = 0.005, Fig. 2Bii), although the duration of responses (time in seconds between 1/2 amplitude on ascent and descent of the transient; full width, half max (FWHM)) was slightly decreased (WT: 0.07 ± 0.02, n = 103 neurons from 4 cultures vs mdx: 0.05 ± 0.01, n = 170 neurons from 5 cultures, p < 0.0001, Fig. 2Biii). The increase in Ca²⁺ activity in mdx hippocampal neurons appeared to be due to enhanced synaptic transmission, as the voltage-gated Na⁺ channel blocker, tetrodotoxin (TTX, 100 nM) significantly reduced baseline activity in mdx neurons (n = 60 neurons from 4 cultures, paired t-test, p = 0.002, Fig. 2D).

Neuronal viability was confirmed by exposure to high K⁺ (50 mM) at the end of each experiment. In neurons only perfused with HBSS, the peak amplitude of Ca²⁺ responses evoked by high K⁺ was notably larger in mdx (44.5 ± 32.7, n = 12 neurons from 3 cultures), relative to WT, neurons (15.8 ± 2.9, n = 17 neurons from 3 cultures, p = 0.003, Fig. 2E), providing further evidence for altered Ca²⁺ mobilisation in dystrophic hippocampal neurons.

Dystrophic hippocampal cultures displayed increased expression of synaptic IL-6Rs on neuronal processes

Abundant IL-6 staining was evident in the cytosol of cell bodies of both WT (n = 6 images) and dystrophin-deficient mdx (n = 15 images, Fig. 3A) cultured mouse hippocampal neurons, with lower levels in the processes. The relative expression levels of IL-6 in cell bodies (p = 0.76) and process (p = 0.21) were similar. However, the density of IL-6R immuno-positive clusters, located at synaptophysin-expressing synapses on the processes, was increased in mdx hippocampal neurons (1.37 ± 0.57 synapses/10 µm, n = 8 images) relative to WT neurons (0.90 ± 0.46 synapses/10 µm, n = 14 images, p = 0.046, Fig. 3B), mimicking the effect noted in dystrophin-expressing neurons cultured in the presence of elevated IL-6.

Fig. 3.

Dystrophic mdx neurons have increased numbers of interleukin (IL)-6 receptor-expressing synaptic contacts but similar Ca²⁺ responses (A) The immunofluorescent images and data plots illustrate the intensity of IL-6 staining (corrected total cell fluorescence, CTCF) in the cell bodies and processes of hippocampal neurons from wildtype (WT, black circles) and mdx (red triangles) neurons. Scalebar: 35 µm. (B) The immunofluorescent images and data plot show the number of IL-6 receptor (green staining) clusters located at synapses identified using anti-synaptophysin (red staining) /10 µm of neuronal processes in WT and mdx neurons. Yellow staining in the merged images indicates co-localisation. Scalebar: 35 µm. (C) (i) The representative traces from WT (black trace) and mdx (red trace) and the dot plots of pooled data illustrating (ii) the changes in amplitude and (iii) frequency of neuronal intracellular Ca²⁺ oscillations following application of IL-6. (D) (i) The representative traces and (ii) dot plot from WT and mdx neurons show suppression of the IL-6-evoked response in mdx neurons. *, ** and ***p < 0.05, p < 0.01 and p < 0.001, respectively.

Sensitivity to IL-6 was unaffected by loss of dystrophin

We explored the actions of IL-6 (1 nM, 3 min) on neuronal [Ca²⁺]_i, in the context of altered Ca²⁺ homeostasis in mdx cells. 61.5% of WT, and 67.5% of mdx, neurons responded with an increase in [Ca²⁺]_i to IL-6 with a mean amplitude of 3.15 ± 3.3 AU (n = 128 neurons from 14 cultures) and 3.18 ± 3.3 AU (n = 213 neurons from 20 cultures, p = 0.93, unpaired t-test), respectively, suggesting that both neuronal genotypes were equally sensitive to the neuromodulatory actions of this pro-inflammatory cytokine. Although the basal frequency of Ca²⁺ spikes was elevated in mdx relative to WT neurons (WT: 0.27 ± 0.19 vs mdx: 0.86 ± 1.01) confirming the strain difference (F(1,1027) = 88.5, p < 0.001 (two-way ANOVA), no change in the frequency of Ca²⁺ spikes (relative to baseline) was detected in mdx (0.75 ± 0.75, n = 347), when compared to WT neurons (0.42 ± 0.27, n = 164 neurons) following exposure to IL-6 (F(1,1027) = 0.25, p = 0.62). However, an interaction between treatment and strain was detected (F (1, 1027) = 6.7, p = 0.01).

The response to IL-6 was found to be reproducible in a subset of both WT (1^st response: 1.81 ± 1.76 vs 2^nd response: 1.17 ± 1.58, n = 26 neurons, p = 0.07) and mdx neurons (1^st response: 3.4 ± 2.9 vs 2^nd response: 2.95 ± 5.94, n = 21 neurons, p = 0.78). IL-6 binds to membrane-bound or soluble IL-6Rs, which activates the signal transducing protein, gp130, to stimulate intracellular signalling pathways⁵². By using a neutralising IL-6R antibody (xIL-6R, 10 µg/ml, 20 min incubation), to block IL-6-evoked signalling, we noted that the amplitude of spontaneous Ca²⁺ spikes was inhibited in mdx neurons (IL-6: 3.38 ± 5.4 vs IL-6 plus xIL-6R: -3.51 ± 4.6, n = 34 neurons, p < 0.0001), but not in WT neurons (IL-6: 1.43 ± 4.8 vs IL-6 plus xIL-6R: 1.5 ± 6.3, n = 18 neurons, p = 0.99, Fig. 3Di and ii). When treated with xIL-6R, an effect of the treatment was evident (F(1,100) = 10.0, p = 0.002), but strain differences were lost (F(1,100) = 58.8, p = 0.16), resulting in an interaction between factors (F(1,100) = 10.4, p = 0.002).

The GABA_AR agonist, muscimol, modulated Ca²⁺ kinetics in mdx hippocampal neurons.

Exposure to elevated levels of IL-6 is considered pathophysiological during early development of neural networks, and is believed to involve inhibitory interneurons⁴⁵. Furthermore, loss of dystrophin is known to interfere with synaptic GABA_AR stability^53,54. Nonetheless, punctate GABA_AR staining was similarly observed in the cell bodies and processes of cultured neurons in both WT (n = 14 images) and mdx (n = 6 images, Fig. 4A) mice, although that, obviously, does not preclude functional changes related to activation of these receptors⁵⁵.

Fig. 4.

IL-6-evoked Ca²⁺ responses are enhanced in cultured WT and mdx hippocampal neurons when exposed to the GABA_A receptor agonist, muscimol (A)The representative brightfield and immuno-labelled images of cultured hippocampal neurons from wildtype (WT) and dystrophin-deficient, mdx mice illustrate expression of GABA_A receptors in the cell bodies and processes of these neurons. Scalebar: 35 µm. (B) (i) The representative Ca²⁺ traces illustrate spontaneous activity in WT (black trace) and mdx (red trace) cultured neurons in the absence, presence, and following washout of the GABA_A receptor agonist, muscimol (1 µM, 20 min). The data plots of pooled data show the relative change (from baseline) of the (ii) frequency, (iii) amplitude and (iv) duration (time between 1/2 amplitude on ascent and descent of the transient; full width, half max (FWHM)) of Ca²⁺ oscillations in mdx hippocampal neurons (red triangles) as compared to WT (black circles) neurons. (C) (i) The representative traces from WT and mdx neurons and (ii) dot plot illustrate the enhanced response to IL-6 (1 nM, 3 min) in the presence of muscimol. *, ** and ***p < 0.05, p < 0.01 and p < 0.001, respectively.

As tonic GABA_AR-mediated inhibitory neurons are critical for regulating hippocampal neuronal hyperexcitability⁵⁶, we used pharmacological agonists and antagonists of GABA_A and GABA_B receptors to explore the spontaneous Ca²⁺ activity characteristics of mdx cultured hippocampal neurons (Fig. 2C). Given the differences in baseline neuronal activity, the relative changes in WT and mdx cells were compared by normalising each neuron to its own baseline (baseline = 1). Application of muscimol (1 µM, 20 min), a GABA_AR agonist that binds to ligand-gated, pentameric chloride ion-permeable channels⁵⁷, significantly suppressed the frequency of Ca²⁺ spikes in WT neurons (0.64 ± 0.52, n = 33 neurons from 4 cultures) but had little effect on mdx neurons (0.99 ± 0.91, n = 25 neurons from 3 cultures, p = 0.04) resulting in a difference between strains (Fig. 4B i & ii). The relative amplitude of Ca²⁺ spikes in WT (0.93 ± 0.33) and mdx (1.05 ± 0.45) neurons was generally unchanged in the presence of muscimol (p = 0.89, Fig. 4Biii), although some mdx neurons did exhibit large, high-amplitude Ca²⁺ oscillations (red trace, Fig. 4Bi). The relative change in duration (FWHM) of Ca²⁺ events did not differ between WT (1.04 ± 0.23) and mdx (1.08 ± 0.46) neurons (p = 0.97, Fig. 4B iv).

Although the frequency of spontaneous Ca²⁺ spikes returned to baseline levels in WT neurons (1.08 ± 0.55) following muscimol washout, the frequency of Ca²⁺ events in mdx neurons decreased (0.65 ± 0.53, p = 0.009, Fig. 4B ii). This likely reflects the development of large, longer duration and, sometimes, large amplitude, Ca²⁺ oscillations (noted in 62% (18 out of 29) of mdx neurons, p = 0.048, Fisher’s exact test, Fig. 4Bi). The relative change in the amplitude (2.1 ± 2.2, Fig. 4B iii) and duration (1.8 ± 0.89, Fig. 4Biv) of oscillations in mdx neurons was increased, with little change WT neurons (amplitude: 0.84 ± 0.27, p < 0.0001; duration: 1.04 ± 0.23, p < 0.0001), revealing differences in Ca²⁺ kinetics in neurons lacking dystrophin. Analysis by 2-way ANOVA detected an interaction between the effects of muscimol on oscillation frequency, which differed between strains (F (1, 111) = 13.9, p = 0.0003). The effect of muscimol (F (1, 108) = 5.96, p = 0.02) on the amplitude of Ca²⁺ events, differed between strains (F (1, 108) = 12.1, p = 0.0007) and resulted in an interaction between factors (F (1, 108) = 8.3, p = 0.005). Furthermore, the effect of muscimol (F (1, 108) = 13.4, p = 0.0004) on the duration of Ca²⁺ events, differed between strains (F (1, 108) = 15.98, p = 0.0001) and resulted in an interaction between factors (F (1, 108) = 13.38, p = 0.0004).

Distinct from spontaneous Ca²⁺ activity, exposure to IL-6 evoked a discrete increase in [Ca²⁺]_i. The peak amplitude of IL-6-evoked responses was compared in the presence or absence of muscimol. We found that the amplitude of Ca²⁺ responses to IL-6 was increased in the presence of muscimol in both WT (IL-6: 1.05 ± 3.67 vs IL-6 plus muscimol: 5.3 ± 5.82, n = 22 neurons, p = 0.002) and mdx neurons (IL-6: 1.05 ± 2.06 vs IL-6 plus muscimol: 4.61 ± 4.2, n = 25 neurons, p = 0.006, Fig. 4Ci and ii). Interestingly, the amplitude of Ca²⁺ responses evoked by high K⁺-induced neuronal depolarisation was no longer elevated in mdx neurons following exposure to muscimol (WT: 18.6 ± 16.5, n = 27 vs mdx: 20.7 ± 15.7, n = 28, p = 0.63).

The GABA_AR inhibitor, bicuculline, suppressed the amplitude of IL-6-evoked Ca²⁺ responses in WT but not mdx neurons

In WT neurons, incubation with bicuculline (100 µM, 20 min exposure), a GABA_AR antagonist, had the opposite effect to the GABA_AR agonist, muscimol, increasing the relative change in the frequency of spontaneous spikes (1.4 ± 0.53, n = 26 neurons from 3 cultures). Whereas muscimol had no effect on the frequency of Ca²⁺ events in mdx neurons, the GABA_AR antagonist increased the frequency in a similar manner to the WT neurons (1.3 ± 1.33, n = 40 neurons from 5 cultures, p = 0.86, Fig. 5A i and ii). Ca²⁺ spikes in the presence of bicuculline in WT neurons were smaller (0.69 ± 0.21, p = 0.026) and of shorter duration (0.75 ± 0.14, p = 0.001) compared to the amplitude (0.99 ± 0.35, Fig. 5A iii) and duration of Ca²⁺ events in mdx neurons (1.01 ± 0.39, Fig. 5A iv).

Fig. 5.

IL-6-evoked Ca²⁺ responses in the presence of the GABA_A receptor antagonist, bicuculline, is suppressed in WT hippocampal neurons (A) (i) The representative Ca²⁺ traces from WT (black trace) and mdx (red trace) hippocampal neurons and the data plots show the effects of the GABA_A receptor antagonist, bicuculline (100 µM, 20 min exposure) on the relative change in (ii) frequency, (iii) amplitude and (iv) duration (time between 1/2 amplitude on ascent and descent of the transient; full width, half max (FWHM)) of Ca²⁺ oscillations. (B) (i) The representative traces and (ii) dot plot illustrate suppression of the Ca²⁺ responses to IL-6 (1 nM, 3 min) in the presence of bicuculline, only in WT neurons. *, ** and ***p < 0.05, p < 0.01 and p < 0.001, respectively.

Compared to baseline, more WT (29%) and mdx (73%) neurons (p = 0.0002, Fisher’s exact test) exhibited cellular Ca²⁺ spikes following washout of bicuculline. The frequency of Ca²⁺ spikes in WT (1.16 ± 0.6) and mdx neurons remained elevated (1.31 ± 1.2) compared to baseline such that there was no difference between the strains (p = 0.8, Fig. 5A ii), although the amplitude of these Ca²⁺ events was reduced to a greater extent in WT (0.78 ± 0.25) than mdx (0.92 ± 0.6) neurons (p = 0.002, Fig. 5A iii). The duration of spikes was less in WT (0.78 ± 0.16) than mdx (0.97 ± 0.32, p = 0.02) neurons (Fig. 5A iv). No significant changes in the frequency of spikes in the presence or after washout of bicuculline were detected (2-way ANOVA). However, the relative change in the amplitude of Ca²⁺ events was different between strains (F (1, 135) = 17.1, p < 0.0001), as was the duration (F (1, 131) = 18.8, p < 0.0001).

The amplitude of the IL-6-evoked increase in [Ca²⁺]_i was suppressed in the presence of bicuculline only in WT neurons (IL-6: 4.2 ± 3.8 vs IL-6 plus bicuculline: 0.46 ± 2.27, n = 26, p = 0.0003), whereas the response to IL-6 in mdx neurons was insensitive to the actions of the GABA_AR antagonist (IL-6: 1.67 ± 2.88 vs IL-6 plus bicuculline: 1.42 ± 2.43, n = 43 neurons, p = 0.89, Fig. 5Bi and ii). Bicuculline did not modify the amplitude of the Ca²⁺ responses evoked by high K⁺, which remained higher in mdx (23.8 ± 19.9, n = 47 neurons) relative to WT neurons (12.8 ± 11.33, n = 28 neurons, p = 0.01).

A GABA_BR agonist suppressed Ca²⁺ responses evoked by high K⁺ in mdx neurons

GABA_B receptors are G-protein-coupled metabotropic receptors expressed pre- and post-synaptically⁵⁸, that are postulated to modulate hippocampal rhythmic network activity⁵⁹. Expression of GABA_BRs was less abundant than GABA_ARs, but had a similar punctate pattern in both cell bodies and processes of WT (n = 13 fields) and mdx (n = 9 fields, Fig. 6A) hippocampal cultured neurons.

Fig. 6.

The GABA_B receptor agonist, baclofen, potentiated IL-6-evoked Ca²⁺ responses in mdx hippocampal neurons (A) The representative brightfield and immuno-labelled images of cultured hippocampal neurons from wildtype (WT) and dystrophin-deficient mdx mice illustrate expression of GABA_B receptors in the cell bodies and processes of these neurons. Scalebar: 35 µm. (B) (i) The representative Ca²⁺ traces from wildtype (WT, black traces) and mdx (red traces) cultured hippocampal neurons illustrate neuronal activity in the absence, presence, and following washout, of the GABA_B receptor agonist, baclofen (100 µM, 20 min). The data plots of pooled data show the relative change (from baseline) of the (ii) frequency, (iii) amplitude and (iv) duration (time between 1/2 amplitude on ascent and descent of the transient; full width, half max (FWHM)) of Ca²⁺ oscillations in WT (black circles) and mdx hippocampal (red triangles) neurons. (C) (i) The representative traces and (ii) dot plot illustrate the enhanced response to IL-6 (1 nM, 3 min) in the presence of baclofen in mdx neurons. (D) The dot plot illustrates suppression of the amplitude of Ca²⁺ response evoked by high K⁺ in mdx neurons after exposure to baclofen. *, ** and ***p < 0.05, p < 0.01 and p < 0.001, respectively.

In the presence of baclofen (100 µM, 20 min), a GABA_BR agonist, the relative change in frequency of Ca²⁺ spikes from baseline was significantly increased in WT neurons to 1.4 ± 0.71 (n = 20 neurons from 3 cultures) but unchanged in mdx neurons (0.99 ± 0.17, n = 52 neurons from 4 cultures) resulting in a significant difference between strains (p = 0.0003, Fig. 6B i and ii). The relative change in amplitude of these Ca²⁺ events was similar in WT (0.85 ± 0.37) and mdx (0.92 ± 0.19) neurons (p = 0.51, Fig. 6B iii); as was the change in duration (0.91 ± 0.2 in WT neurons and 1 ± 0.15 in mdx neurons, p = 0.1, Fig. 6B iv).

Following washout of baclofen, the relative change in frequency of Ca²⁺ spikes remained significantly elevated in WT (1.3 ± 0.65), relative to mdx, neurons (0.79 ± 0.28, p < 0.0001, Fig. 6B ii). The change in amplitude of these spikes did not differ between strains with baclofen washout (WT: 0.85 ± 0.39 vs mdx: 0.96 ± 0.23, p = 0.2, Fig. 6B iii), but the change in duration of distinct Ca²⁺ spikes remained significantly altered (WT: 0.91 ± 0.2 vs mdx: 1.1 ± 0.17,p = 0.001, Fig. 6B iv). Furthermore, neuronal excitability remained elevated in mdx neurons following baclofen washout, as they continued to exhibit significantly more Ca²⁺ oscillations than WT neurons (63% vs 26% WT, p = 0.002, Fisher’s exact test). Overall, a small effect of baclofen was detected on the frequency of spikes (F (1, 143) = 4, p = 0.047, 2-way ANOVA) and a difference between strains was evident (F (1, 143) = 37.1, p < 0.0001). No effects of baclofen on the change in amplitude in either strain was detected. A difference between strains in the response to baclofen was detected for the duration of spikes (F (1, 147) = 15.2, p = 0.0001), but no effect of the agonist and no interaction between factors was detected.

The amplitude of acute Ca²⁺ increases evoked by IL-6 in WT hippocampal neurons (1.64 ± 3.6) was not significantly affected by baclofen (3.2 ± 3.6, n = 17, p = 0.11). However, IL-6-evoked Ca²⁺ responses in mdx neurons (1.3 ± 2.1) were significantly increased in the presence of baclofen (3.2 ± 1.8, n = 55, p = 0.002, Fig. 6C i and ii). Moreover, there was notable effect of baclofen exposure on Ca²⁺ responses induced by high K⁺ in mdx hippocampal neurons such that, following washout of baclofen, Ca²⁺ responses evoked by high K⁺ in mdx neurons (13.5 ± 12.35, n = 58) were lower than those evoked under control conditions (i.e. HBSS only, Fig. 2E) in mdx neurons (44.5 ± 32.7, n = 12 neurons, p < 0.0001) and were also reduced in comparison to WT high K⁺-evoked Ca²⁺ responses (21.9 ± 23.7, n = 26 neurons, p = 0.036, Fig. 6D).

Ca²⁺ responses evoked by IL-6 were unaffected by a GABA_BR antagonist

The GABA_B receptor antagonist, phaclofen (100 µM, 20 min), had no significant effect on the relative change in the frequency of spontaneous Ca²⁺ spikes detected in WT (0.94 ± 0.45, n = 26 neurons from 3 cultures) or mdx (0.99 ± 0.65, n = 51 from 3 cultures, p = 0.93) neurons (Fig. 7A i & ii). However, exposure to phaclofen did evoke large amplitude oscillations, which were detected in 66% of mdx neurons (example in red trace, Fig. 7A i), but only 36% of WT neurons (p = 0.004. Fisher’s exact test). However, the relative change in the amplitude of Ca²⁺ events decreased in mdx, (0.88 ± 0.26, n = 49) neurons compared to WT (1.04 ± 0.27, n = 31, p = 0.026, Fig. 7A iii). No significant differences in the relative duration of Ca²⁺ spikes between WT (0.98 ± 0.32) and mdx (0.91 ± 0.39, p = 0.62, Fig. 7A iv) were detected. Following washout of phaclofen, the frequency of Ca²⁺ spikes remained similar between strains (WT: 1.2 ± 0.77 vs. mdx: 0.91 ± 0.7, p = 0.13). The relative change in amplitude (WT: 0.98 ± 0.31 vs mdx: 0.91 ± 0.29, p = 0.46, Fig. 7A iii) and duration (WT: 1.04 ± 0.32 vs mdx: 0.86 ± 0.3, p = 0.08, Fig. 7A iv) of Ca²⁺ events was also similar between strains. However, following washout of phaclofen, increased prevalence of large amplitude, long duration Ca²⁺ oscillations (red trace, Fig. 7A i) was apparent in mdx neurons (mdx: 67% vs WT: 28%, p = 0.0002, Fisher’s exact test). No significant changes in the frequency of oscillations in the presence, or after washout, of phaclofen were detected (2-way ANOVA). However, the relative change in the amplitude of oscillations was different between strains (F (1, 160) = 6.67, p = 0.011), as was the duration of oscillations (F (1, 135) = 4.49, p = 0.04).

Fig. 7.

The GABA_B receptor antagonist, phaclofen, had no effect on IL-6-evoked Ca²⁺ responses in WT and mdx hippocampal neurons (A) (i) The Ca²⁺ traces and the data plots show the effects of the GABA_B receptor antagonist, phaclofen (100 µM, 20 min exposure), on the relative change in (ii) frequency, (iii) amplitude and (iv) duration (time between 1/2 amplitude on ascent and descent of the transient; full width, half max (FWHM)) of Ca²⁺ oscillations in WT and mdx hippocampal neurons. (B) (i) The representative traces and ii dot plot show that Ca²⁺ responses to IL-6 (1 nM, 3 min) were not altered in the presence of phaclofen in WT or mdx neurons. *p < 0.05.

In WT neurons, the relative change in amplitude of Ca²⁺ responses evoked by IL-6 (1.36 ± 2.9) was not significantly modulated by phaclofen (1.19 ± 1.4, n = 26 neurons, p = 0.14). Similarly, the amplitude of IL-6-evoked Ca²⁺ responses (1.39 ± 2.2) was not significantly affected by phaclofen in mdx neurons (1.46 ± 2.47, n = 41 neurons, p = 0.99, Fig. 7Bi and ii). Ca²⁺ responses to high K⁺ in mdx neurons (21.74 ± 16.14, n = 51) were similar in amplitude to those evoked in WT neurons (17.55 ± 9.87, n = 35, p = 0.18), following washout of phaclofen, showing a change from the baseline responses (Fig. 2E).

Freezing was a characteristic behaviour of mdx mice

Consistent with previous reports^16,60, mdx mice (8 – 12 weeks old) exhibited increased freezing (14.6 ± 5.8, n = 19) in the open field in comparison to WT mice (11.0 ± 4.8, n = 20, p = 0.04, Fig. 8A). Reluctance to explore exposed central regions of an arena can indicate elevated levels of anxiety in rodents⁶¹, however, the frequency of entries into the central area was not significantly reduced in mdx (22.17 ± 6.5, n = 18) compared to WT (25.6 ± 8.2, n = 20) mice (p = 0.17, Fig. 8B). Similarly, perception of stress, leading to increased faecal output is also linked to increased anxiety⁶². However, excretion rates were not different between WT (3.3 ± 2.5 boli, n = 20) and mdx (2.6 ± 1.7 boli, n = 19) mice (p = 0.34, Fig. 8C). We also examined exploratory and escape behaviours in the dystrophic mice but found that rearing (WT: 88.6 ± 18.6, n = 18 vs mdx: 96.7 ± 23.1, p = 0.26, n = 17, Fig. 8D) and jumping (WT: 2.9 ± 5.2, n = 20 vs mdx: 2.1 ± 2.8, n = 19, p = 0.58, Fig. 8E) frequency were similar in both strains. The frequency of repetitive self-grooming has been linked to autism spectrum disorder⁶³ but this behaviour did not differ between mdx mice (2.9 ± 1.2, n = 18) and WT (2.95 ± 1.2, n = 19, p = 0.91, Fig. 8F) mice. Cognitive processes, such as learning and memory are key functions facilitated by the hippocampus. It has been reported that mdx mice, 4 months and older, spend less time exploring the new object in a novel object recognition test – possibly suggesting that they have poorer recollection of the objects they explored 24 h earlier⁶⁴. However, our studies did not detect any difference between the discrimination index of the time spent exploring a novel object as compared to an object the mice had encountered 24 h earlier in WT (0.76 ± 0.04 n = 8) and mdx (0.79 ± 0.04, n = 5) mice (Fig. 8G).

Fig. 8.

Mdx mice exhibit increased freezing behaviours (A) The frequency of freezing (mouse did not move for more than 2 s) in the open field was increased in mdx mice as compared to wildtype (WT) controls. Other anxiety-related behaviours such as (B) the frequency of exploration of the exposed inner zone of the open field (number of times the mouse had four paws in the central area) and (C) stress-induced faecal excretion was similar in WT and mdx comparators. Exploration and escape behaviours such as the frequency of (D) rearing (number of times the mouse reared its front paws, against the wall or mid-air) and (E) jumping (number of times the mouse jumped with all four paws in the air) was similar in both strains. (F) Compulsive behaviours, such as repetitive grooming was similar in both strains. (G) The discrimination index, calculated from a novel object recognition trial, was similar in WT and mdx mice *p < 0.05.

Discussion

Even with an intact blood–brain barrier, IL-6 can gain access to the CNS⁶⁵. However, loss of dystrophin results in a compromised blood–brain barrier^26,27, making it more likely that elevated circulating IL-6 will have neuromodulatory consequences within the CNS. In the context of elevated levels of this pro-inflammatory cytokine in DMD^46,47, an observation that is mimicked in the dystrophin-deficient mdx mouse⁴⁸, we have examined the modulatory effects of IL-6 on neuronal network connectivity and activity in cultured hippocampal neurons from dystrophin-expressing mice. Dystrophin is expressed in the hippocampus, a structure critical for declarative and spatial learning^13–16. Cognitive deficits relating to learning are more common in individuals with DMD^5–10 and are also reported in dystrophic mdx mice^18–20. Our studies demonstrated an increased density of IL-6R-labelled synaptic contacts in dystrophic mdx mice. We also observed an increased prevalence of IL-6R-expressing synapses in dystrophin-expressing neurons that were cultured in the presence of elevated IL-6. Brief exposure to IL-6 had similar effects in both WT and mdx neurons, resulting in short-lived increases in [Ca²⁺]_i . However, the neurostimulatory effect of IL-6 was further altered in the presence of GABAergic pharmacological reagents.

The hippocampus is a region that is vulnerable to the deleterious effects of neuronal excitability induced by epilepsy, and individuals with DMD have an increased susceptibility to epilepsy¹². At a cellular level, aberrant Ca²⁺ homeostasis was prevalent in hippocampal neurons cultured from dystrophin-deficient mdx mice. This supports previous reports of modified neuronal Ca²⁺ in the absence of dystrophin²¹. Mdx neuronal cultures consistently exhibited spontaneous Ca²⁺ spikes, oscillations and sometimes, synchonised Ca²⁺ waves moving through the neuronal networks. However, manifestation of these Ca²⁺ events varied between cultures. This could be related to the heterogeneous nature of the cultures, which included variable proportions of primary cells from both Ammon’s horn and the DG, which, respectively, express Dp427 and Dp71, or the developmental switch of GABAergic neurons. It is also possible that protective effects of hippocampal astrocytes⁶⁶ have been impacted by loss of dystrophin, which fits with previous observations from our laboratory that dentate gyrus astrocyte numbers are decreased across the life-span of mdx mice⁶⁷.

Given that GABAergic inhibitory neurons are critical to the regulation of hippocampal neuronal hyperexcitability²¹, we used pharmacological modulators of GABA_A and GABA_B receptor function to manipulate inhibitory regulation of hippocampal network activity. In doing so, we revealed dynamic oscillatory changes in cellular Ca²⁺ in mdx neurons that were further influenced by the presence of the pleiotropic cytokine, IL-6. Behaviourally, mdx mice were more likely to freeze in an open arena, likely to be due to loss of dystrophin from amygdalar neurons^16,49,68, resulting in modification of local GABAergic neuronal circuit activity⁵⁵. However, we, and others have found little evidence of any additional behavioural consequences of neuronal dystrophin loss, including an intact recognition memory, at least in young adult mdx mice. The additional physiological challenges of aging may reveal behavioural consequences of the observed cellular changes⁶⁴.

Consistent with dystrophin playing a role in synapse structure and function⁶⁹, in dystrophin-expressing neurons, we detected co-localisation of dystrophin with the post-synaptic marker, PSD-95. Development of neural networks during the first 10–12 days of culture was modified when the culture medium was supplemented with IL-6, an environmental modification carried out to mimic elevated levels of this pro-inflammatory cytokine in dystrophic mdx mice. Both dystrophin and IL-6Rs were expressed at synapses on the neuronal processes, and when hippocampal neurons were cultured in the continued presence of IL-6, the number of IL-6R-positive puncta on neural processes increased. This likely reflects increased synaptogenesis in this developing network, which is a finding that is consistent with studies focussed on developing hippocampal networks in in utero embryos⁷⁰. These morphological changes may, in turn, impact upon function. Indeed, in humans, exposure to IL-6 in the early developmental period has been linked to deterioration in working memory in newborns⁷¹. Increases in excitatory, and reductions in inhibitory, synapse formation induced by overexpression of IL-6⁷² infer that similar changes occur in our neuronal preparation, underpinning interruptions to normal neural function. Indeed, exposure to pathophysiological concentrations of IL-6 during network development has been reported to reduce neuronal expression of Group-II metabotropic receptors (mGluR2⁄3) and L-type Ca²⁺ channels, which was associated with altered synaptic network activity^37,53,54. Consistent with such consequences, relative to dystrophin-expressing WT neurons, spontaneous neuronal excitability was increased in mdx cultured hippocampal neurons.

Cytosolic Ca²⁺ plays a critical role in neuronal plasticity, survival and cellular metabolism. It is tightly regulated to balance influx from the extracellular environment and release from intracellular stores⁷³. Dysfunctional regulation of intracellular Ca²⁺ has been observed in dystrophic skeletal^74,75 and cardiac muscle⁷⁶. It has also been reported in dystrophic neural tissue^21,77, which is consistent with the aberrant cellular Ca²⁺ activity observed in our study. Furthermore, depolarisation-induced Ca²⁺ responses in the presence of high K⁺ saline solution, was notably larger in mdx neurons. High extracellular K⁺ induces Ca²⁺ influx through L-type voltage-gated Ca²⁺ channels. The use of such a stimulus is useful for demonstrating global neuronal viability at the end of the experiment. However, it also resembles epilepsy- like activity, where elevated extracellular K⁺ induces neuronal depolarisation, network hyperactivity, further K⁺ accumulation and subsequent network depression⁷⁸. Prolonged and / or sustained depolarisation can result in impaired synaptic transmission⁷⁹. Alternatively, IL-6-mediated changes to expression of L-type Ca²⁺ channels^37,53,54 could contribute to the aberrant Ca²⁺ regulation noted in neurons lacking dystrophin. Similarly, changes to expression or function of voltage-gated K⁺ channels⁸⁰ or Na v channels which suppress GABAergic inhibition⁸¹, could contribute to changes in intrinsic neuronal excitability in mdx neurons. Further exploration is warranted to understand the physiological mechanisms contributing to this alteration in neuronal calcium dynamics.

Given that elevated Ca²⁺ influx can trigger a neuronal cytotoxic cascade that may contribute to neuronal dysfunction and neurodegeneration, and has also been linked to premature brain aging^82,83, altered Ca²⁺ homeostasis could contribute to deficits in hippocampal function in mdx mice. Indeed, reducing intracellular Ca²⁺ improves cognitive function in mdx mice²¹, indicating the fundamental role it has in cognitive dysfunction related to loss of dystrophin. Excitatory and inhibitory synaptic transmission in cultured hippocampal neuronal networks is manifested as spontaneous electrical events with associated Ca²⁺ transients⁸⁴. The capacity of TTX to reduce the frequency of such Ca²⁺ events indicates the importance of synaptic transmission for the generation of these Ca²⁺ transients in mdx neurons. Synaptic drive can be increased by removing the Mg²⁺ block from NMDARs, and, indeed, dystrophin deficiency may facilitate activation of hippocampal NMDA receptors through reduction of the voltage-dependent block of NMDA receptors by Mg²⁺⁵¹. While we saw no overt changes in expression patterns of NMDAR in mdx neurons, alternative experimental approaches are needed to assess functional changes in NMDAR-medicated activity in excitatory and inhibitory dystrophic hippocampal neurons. It would also be of value to explore the actions of IL-6, which can modulate NMDA receptor activity to evoke Ca²⁺ influx into hippocampal neurons⁸⁵ in mdx hippocampal neurons.

Given the importance of dystrophin for postsynaptic GABA_AR stabilisation⁴⁸ and reported reductions^{34,35,54,86,87}, or alterations, in subunit expression of GABARs in mdx hippocampal tissue⁸⁸, we chose to explore synaptic network activity by modulating GABAergic regulation. Based on the work of Nunez and McCarthy, who reported that GABAergic neurons have predominantly inhibitory actions by the second post-natal week⁸⁹, we used neurons that were at least eleven post-natal days in culture. However, it is recognised that this switch is a dynamic process⁹⁰, thus the GABA_AR agonist may have induced an excitatory response in some cultured neurons, and perhaps contributed to the oscillatory activity observed in mdx neurons. Indeed, the fact that GABA, itself, regulates the developmental switch in neuronal GABA responses⁹⁰, may imply that altered GABA signalling in dystrophic neurons could impact on the timing of this switch in mdx neurons.

When GABA receptor activity was modulated using pharmacological reagents, Ca²⁺ homeostasis, which already exhibited spontaneous Ca²⁺ oscillations in dystrophic hippocampal neurons, was further modified. We initially explored the importance of GABA_ARs, which are ligand-gated, heteropentameric Cl^- permeable ion channels, that induce post-synaptic inhibition in mature neurons. Predictably, increasing inhibitory regulation by applying the GABA_AR agonist muscimol, suppressed the frequency of Ca²⁺ spikes in dystrophin-expressing WT cultured neural networks. However, in some of the mdx neurons, muscimol had the striking effect of stimulating large amplitude, long-duration Ca²⁺ oscillations, perhaps, as discussed above, due to a delayed switch in GABA responses from excitatory to inhibitory, or perhaps, due to changes in expression of function other neuronal ion channels.

These Ca²⁺ oscillations continued long after washout of the drug, which may be due either to the time needed for the development of these Ca²⁺ events or a non-reversible effect of the agonist. Similar to previous studies, where picrotoxin, a GABA_AR antagonist, increased the frequency of Ca²⁺ oscillations in hippocampal neurons⁹¹, in our hands, incubation of WT neurons with bicuculline also increased the relative change in the frequency of spontaneous oscillations. The frequency of spontaneous oscillations was also increased in mdx neurons when GABA_ARs were inhibited, but again, the prevalence of the large amplitude and long-duration Ca²⁺ oscillations, either during exposure to, or following washout of, bicuculline was significantly increased in dystrophic neurons. Altered responses to modulation of GABA_ARs, which are dysfunctional throughout the lifespan of dystrophic mice^37,53,54, are likely to underpin the observed differences between strains. It is important to note however, that further functional characterisation is necessary to determine if the hyperexcitability in mdx cultured neurons is due to pre- or post-synaptic changes in the neural networks, although one might predict that if it is due to lost post-synaptic expression of dystrophin²⁹, that this is where changes would be apparent. Use of electrophysiological recordings would be beneficial to understand whether changes in excitatory or inhibitory post-synaptic potentials underpin the alterations in neural network calcium dynamics.

Several reports indicate the importance of NMDAR activity in the neurostimulatory actions of IL-6 in hippocampal neurons^45,85,92,93. However, GABAergic inhibitory neurons also play a prominent role in Ca²⁺ oscillations observed in IL-6-treated hippocampal cultures³⁷, suggesting an interplay between neurostimulatory and neuroregulatory mechanisms in neural networks. Thus, in the context of altered [Ca²⁺]_i homeostasis in mdx neurons, we explored the neuromodulatory actions of acute IL-6 exposure in the absence or presence of GABAergic agonists and antagonists. IL-6 stimulated small, short-lived [Ca²⁺]_i responses which were similar in amplitude and duration in both WT and mdx neurons. IL-6 can modify GABA_AR-mediated responses in cortical slices using postsynaptic mechanisms which likely involve interference with GABA_AR receptor expression, resulting in increased network excitability⁴³. Consistent with IL-6-induced increases in cellular excitability, the amplitude of Ca²⁺ responses evoked by IL-6 was larger in the presence of muscimol in both WT and mdx neurons. The GABA_AR antagonist, bicuculline, ameliorated IL-6-evoked Ca²⁺ responses, but only in WT neurons, further supporting the idea of that GABA_AR-mediated regulation neuronal function is aberrant in the absence of dystrophin.

Consistent with previous studies⁹⁴, we observed expression of GABA_BRs on the cell bodies and processes of both WT and mdx neurons, although less abundantly than GABA_ARs. Postulated to modulate hippocampal rhythmic network activity⁵⁹, we found that baclofen, a GABA_BR agonist, increased the frequency of spontaneous Ca²⁺ oscillations in WT neurons. There is little evidence linking GABA_BRs with post-synaptic dystrophin and, indeed, baclofen had no effect on the generation of spontaneous Ca²⁺ oscillations in mdx neuronal cultures. However, the capacity of GABA_BR activation to mediate the inactivation of voltage-gated Ca²⁺ channels and subsequently suppress both the pre-synaptic release of neurotransmitters, and activation of post-synaptic inwardly rectifying K⁺ channels⁹⁵, is likely to underlie the means by which baclofen inhibited the high K⁺-evoked Ca²⁺ responses in mdx neurons, to the extent that they were smaller than the same responses in WT neurons. This may indicate that GABA_BRs have a lower activation threshold in mdx hippocampal neurons, which would contrast with a reported enhancement of the sensitivity of extra-synaptic GABA_ARs in mdx mice^68,96. However, additional research is needed to explore this possible mechanism.

Activation of GABA_BRs in mdx mice potentiated the amplitude of IL-6-evoked Ca²⁺ responses. Many reports have implicated NMDAR-mediated Ca²⁺ influx in the neuromodulatory actions of IL-6^45,85,92,93 but GABA_BR activation attenuated NMDA-mediated synaptic activity⁹⁷. The GABA_BR antagonist, phaclofen, had little effect on the characteristic spontaneous oscillations observed in dystrophic neurons, with the innate strain difference persisting. However, more mdx neurons developed large, spontaneous oscillations both during, and following washout of this antagonist. High K⁺-evoked depolarisation remained elevated in mdx neurons, and IL-6-evoked Ca²⁺ responses were not modified in either WT or mdx neurons during exposure to phaclofen. Our observations of altered [Ca²⁺]_i homeostasis in mdx hippocampal neurons fits with the functional consequences of altered inhibitory regulation of neuronal network activity in the absence of dystrophin. Modulation of GABA_AR activity has the most obvious impact on Ca²⁺ kinetics in mdx neurons, exacerbating the prevalence of spontaneous Ca²⁺ oscillations. However, it was interesting to note that GABA_BRs also appeared to play a role in the intracellular regulation of Ca²⁺, although differential effects were noted depending upon the molecular mechanism underpinning the evoked Ca²⁺ response, be that through stimulation of IL-6 receptors or by stimulating cellular depolarisation with elevated extracellular K⁺.

Individuals with DMD have an increased incidence of behavioural issues, such as Attention-Deficit Hyperactivity Disorder (ADHD), anxiety, depression, and/or other mood disorders^12,98,99. Enhanced fearfulness, as interpreted by increased freezing as a maladaptive coping strategy, is the most consistently reported behavioural change reported in mdx mice^{16,49,68,100,101} and our observations in 8–12 weeks old mice supports these findings, although a potential limitation to this behavioural test may include cues from other mice, which can enhance or suppress these fear-related responses⁶⁰. GABAergic dysregulation is associated with mood-related illnesses, including anxiety spectrum disorders^102–104 and depression^105–107. Moreover, elevated levels of central IL-6 have been linked to imbalances in the neuronal circuitry that mediates autism-like behaviours⁷². Reductions in amygdalar local GABAergic neuronal circuit activity results in an enhancement of fear-motivated defensive behaviours, which can be normalised by restoring brain dystrophin⁵⁵. Exploration of an exposed open field and faecal output, both behaviours associated with anxiety¹⁰⁸, were similar to WT mice, as we previously reported⁴⁹. Exploratory and escape behaviours and compulsive-like behaviours were all similar between strains, as was recognition memory. That said, cognitive deficits may be masked by compensatory circuitry until they become apparent in older mice⁶⁴. Thus, the absence of dystrophin has clear consequences for the development and regulation of neuronal networks and leads to dysregulation of cytosolic [Ca²⁺]_i but, at least in these young mice, alternative pathways appear to compensate for these changes and, with the exception of freezing, preserve most cognitive behaviours.

Conclusions

Although obstacles persist in the translatability of pre-clinical findings in rodent models of DMD, dystrophin-deficient mdx mice do provide opportunities for exploring the contribution of dystrophin to cellular function in non-contracting cells, such as neurons. Mimicking human pathophysiological features, such as elevated levels of the pleiotropic cytokine, IL-6, in the context of dystrophic hippocampal neurons, increases our understanding of the multiple players contributing to the pathophysiology of this disorder. In this study, we have noted altered neuronal network formation in mdx-derived neurons that mimics dystrophin-expressing neurons cultured in elevated IL-6. Evidence of cytosolic Ca²⁺ dyshomeostasis was present in mdx neurons, which exhibited neuronal hyperexcitability, reminiscent of epilepsy-like activity, in addition to larger Ca²⁺ responses to the neurostimulatory effects of elevated extracellular K⁺.

Modulation of neurons with pharmacological reagents that targeted GABA_A and GABA_B receptors further disrupted Ca²⁺ kinetics, evoking large amplitude, long duration Ca²⁺ oscillations, which is consistent with the known role of dystrophin in post-synaptic clustering of GABA_A receptors⁵⁵, but also suggests a possible modulatory role for GABA_B receptor activation. Although IL-6-evoked Ca²⁺ responses were similar in WT and mdx neurons under control conditions, differences in the neurostimulatory effects of IL-6 were observed in the presence of agonists and antagonists of GABARs. Recognised limitations of this work include the reliance on calcium imaging cultured hippocampal neurons. Additional experimental approaches, such as electrophysiology with appropriate pharmacological tools in ex vivo tissues from older mdx mice will facilitate a more complete understanding of the molecular mechanisms involved in the observed neuronal hyper-excitability. Nonethless, our data support the concept that defective inhibitory regulation of hippocampal neurons caused by loss of dystrophin, results in Ca²⁺ dyshomeostasis and spontaneous neuronal hyperexcitability. Although cellular changes are apparent, with the exception of freezing in an exposed open field, cognitive behaviours appear to be intact, at least in the young mdx mice used in these experiments.

Methods and materials

Ethical approval

All animal experiments were approved and performed following guidelines set out by the HPRA (Health Products Regulatory Authority), Ireland and following project authorisation (AEI9130/P088), as well as individual authorisation (AEI9130/I303). Animals were euthanized in accordance with European Directive 2010/63/EU.

Animals and tissue collection

Dystrophin-expressing C57BL/6 mice were used in experiments examining the neuromodulatory actions of IL-6 on cultured neurons. Male hemizygotic dystrophin-deficient C57BL/10ScSn-Dmd^mdx/J (mdx, dystrophic) and C57BL/10ScSn wild type (WT) controls, purchased from Jackson Laboratories (Bar Harbor, Maine, US), were used for comparator studies. Breeding colonies of mice were established and maintained in the Specific Pathogen Free (SPF) facility, Biological Services Unit, University College Cork. Weaned mice were housed in individually ventilated cages with environmental enrichment and kept under an artificial light/dark cycle provided with light between 06:00 and 18:00 h, with free access to drinking water and standard chow.

Neuronal cell culture

Primary hippocampal neuronal cultures were prepared using neonatal mdx and WT mouse pups (3–5 days old), as previously described¹⁰⁹. In brief, mice underwent cervical dislocation and were then decapitated. Using aseptic technique, whole hippocampi (containing CA1 – CA3 regions and the dentate gyrus) were excised into cold HEPES-buffered saline solution (HBSS, in mM NaCl, 130; KCl 5.4; MgCl₂, 2; CaCl₂ 0.5; D-Glucose, 2 and HEPES, 10) before being chopped into smaller pieces. All solutions used in the preparation of cultured hippocampal neurons were syringe filtered (0.2 μm pore size; Corning Inc., New York, USA). Hippocampal tissue was digested using papain (2 mg/ml, 45 min, 37 °C Worthington, Ohio, USA) dissolved in HBSS supplemented with Glutamax supplement (0.5 mM), L-glutamic acid (0.025 mM), penicillin–streptomycin (P-S;1%) solution and triturated in the same solution that was further supplemented with Foetal Bovine Serum (FBS, 10%). The neuronal tissue was mechanically dissociated three times with a flame-polished glass Pasteur pipette and once more, gently, with a larger diameter plastic Pasteur pipette (Sarstedt, North Carolina, USA). Dissociated cells were then centrifuged at 100 g for two minutes prior to resuspension of the tissue pellet in Dulbecco’s modified eagle’s medium (DMEM) supplemented with FBS (10%), Glutamax supplement (0.5 mM), L-glutamic acid (0.025 mM) and P-S (1%). Dissociated cells were seeded onto glass coverslips pre-treated with poly-D-lysine hydrobromide (0.1 mg/ml, Sigma-Aldrich, Missouri, USA) in 35 mm plastic petri dishes (Corning Inc., New York, USA) and incubated at 37 °C, 95% O₂ / 5% CO₂ in culture medium consisting of low-glucose DMEM, serum replacement 2 (2%), Glutamax supplement (0.5 mM), L-glutamic acid (0.025 mM) and P-S (1%) for 1 h prior to addition of low-glucose DMEM (2 ml). Coverslips were then inverted after 1–2 h, and the cultures maintained at 37 °C, 95% O₂ / 5% CO₂.

Immunofluorescence and confocal microscopy

Hippocampal neuronal cultures were washed in HBSS, fixed with 4% paraformaldehyde (90 min, 4 °C) permeabilized with 0.1% Triton X-100 (5 min, room temperature) and blocked with 1% donkey serum (1 h, room temperature, all Sigma Aldrich, UK). Neurons were immunolabelled with primary antibodies against dystrophin (mouse anti-MANDRA-1, (1:100) Abcam, cat no.: ab7164 and rabbit anti-dystrophin (1:250), Abcam, cat no.: ab15277), IL-6 (mouse, 1:250), IL-6 receptors (IL-6Rα, rabbit, 1:250), PSD-95, (rabbit 1:250, ThermoFisher Scientific), synaptophysin (rabbit, 1:250, Abcam), NMDAε1 (mouse 1:250, Santa Cruz biotechnology, US), GABA_A receptors (mouse, 1:100, Abcam, Cambridge, UK) and GABA_B receptors (mouse, 1:250, AbCam) for 2 h at room temperature. Appropriate TRITC (red staining) and FITC (green staining) fluorophores (1:250, 2 h, 37 °C, Jackson Immunoresearch, PA, US) were used to visualise the primary antibody staining. Coverslips containing cultured neurons were mounted on glass slides using Dako-fluorescent mounting medium containing DAPI (blue staining, Agilent Pathology Solutions Santa Clara, California, USA).

Images were captured using a FVl0i-Olympus-confocal microscope with Fluoview software (FV10i-SW, Olympus Europe, Hamburg, Germany) or on an Olympus BX51 microscope with an Olympus DP71 camera. Cell Sens™(Olympus) was used to digitally capture the images. At least three different tissue preparations (each culture was prepared from three mice to ensure sufficient volume of cells) were used in each experiment. No non-specific fluorescent immuno-tagging was detected in negative control experiments, where tissues were incubated with primary antibodies in the absence of secondary antibodies or with secondary antibodies alone.

Calcium (Ca²⁺⁾ imaging

For Ca²⁺ imaging studies, cultured hippocampal neurons (up to 12 days in vitro) were loaded with Fluo-2AM (4 μM, 1 h, Thermo Fisher Scientific, Waltham, MA, USA) dissolved in HBSS in the dark, with residual dye washed off with HBSS prior to recording. Cytosolic changes in intracellular Ca²⁺ ([Ca²⁺]_i) were recorded from neuronal cell bodies using WinFluor fluorescence image capture and analysis program (John Dempster, University of Strathclyde, Scotland). Images were captured at 2 Hz using a Cairn optoscope 1200 (Cairn Research, Kent, UK), a charge-coupled device digital camera (Hamamatsu ORCA-ER, Hamamatsu Photonics, Hertfordshire, UK) and a 20 × water-immersion objective on a fixed stage upright microscope (BX51WI, Olympus, UK). Viability of neurons was confirmed by responsivity to a brief exposure to HBSS+50 mM KCl, which was added at the end of each experiment. A perfusion system continuously superfused the hippocampal neurons with HBSS and pharmacological reagents were added to the perfusate. Quantification of intracellular Ca²⁺ transients in the neuronal cultures was conducted using PeakCaller software (Hussman Institute for Autism, MatLab, R2022, 1984–2022, The MathWorks Inc. Natick, MA, US) to detect the kinetics and magnitude of the Ca²⁺ responses in WT and mdx cultured neurons. Following trace smoothing (set at 60), Ca²⁺ spikes with a 20% rise and fall relative to the baseline range were detected. These settings were maintained for all data analysis. At least three different tissue preparations (each culture was prepared from three mice to ensure sufficient volume of cells) were used in each experiment.

Behavioural assessments

Behaviours of WT and mdx mice (aged 8–12 weeks) placed in an open field arena for habituation were observed and analysed by two researchers, both blinded to the groupings. Freezing events (number of times the mouse did not move for more than 2 s) and anxiety-related behaviours, such as entries (the number of times the mouse had all four paws in central zone) into the exposed central area and faecal output (the number of boli excreted), which can indicate stress-induced defecation were recorded. Exploratory and escape behaviours such as rearing (the number of times the mouse reared its front paws against the wall or mid-air) and jumping (the number of times the mouse jumped against the wall, with all four paws in the air) were also recorded. Compulsive behaviours, such as repetitive grooming (the number of times the mouse groomed its face and fur with its forepaws) were also noted during the 10-min trial.

Novel Object Recognition comprised of a training phase, which took place 24 h after open field habituation. Mice were exposed to two identical objects (200 ml purple plastic bottles) for 10 min. For the trial, mice were returned to the texting arena after 24 h (all olfactory cues were removed with cleaning (70% ethanol). In the trial phase, one object was exchanged for a novel object (a 200 ml flat bottomed rectangular flask partially filled with coloured dye). Exploration of the objects by the mice was recorded on a camera (Akaso EK7000, 4 K camera) suspended 50 cm above the open field area. Behaviours were analysed by two researchers, blinded to the groups. Exploration of an object included directing the snout at a distance less than 1 cm to the object, sniffing or touching it with the snout, plus leaning on the object while sniffing or touching it with the snout. Turning or running around the object, climbing or sitting on the object was not considered as exploration.

Power calculations designed to detect significant differences in behaviours between strains were conducted using G*power3. Assuming α = 0.05 and power (1−β) = 0.8 and an effect size of 0.85, group sizes were determined as n = at least 17 mice per group.

Statistics

The data are presented as mean values ± the standard deviation (S.D.) of the mean. All data points are shown in graphs for immunofluorescence, however, statistical analyses were conducted on mean values with one animal taken as an n = 1. All responding neurons (n) per animal were included for analysis in Ca²⁺ imaging studies, with at least three different cultures. Fisher’s exact test, T-tests or two-way ANOVAs with Sidak’s multiple comparison tests were used (GraphPad Prism Version 8.4.2) where appropriate. p ≤ 0.05 was considered significant.

Author contributions

KAS, QX, MBV, PP, CD, SO’M, and CS were involved in the acquisition, analysis and interpretation of the data presented in this manuscript. MGR and DO’M conceived and designed the experimental work and contributed to the interpretation of the findings. All authors were involved in drafting the work and critically reviewing it for intellectual content. All authors approved the final version of the manuscript and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by funding from the Health Research Board, Ireland. ILP-POR-2017-040.

Data availability

Data are available upon request by contacting DO’M.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

All animal experiments were approved and performed following guidelines set out by the HPRA (Health Products Regulatory Authority), Ireland and following project authorisation (AEI9130/P088), as well as individual authorisation (AEI9130/I303). Animals were euthanized in accordance with European Directive 2010/63/EU.