Rett syndrome is a neurodevelopmental disorder caused by mutations in the X-chromosomal MECP2 gene encoding for the transcriptional regulator methyl CpG binding protein 2 (MeCP2). Rett patients suffer from episodic respiratory irregularities and reduced arterial oxygen levels. To elucidate whether such intermittent hypoxic episodes induce adaptation/preconditioning of the hypoxia-vulnerable hippocampal network, we analyzed its responses to severe hypoxia in adult Rett mice. The occurrence of hypoxia-induced spreading depression (HSD)—an experimental model for ischemic stroke—was hastened in Mecp2−/y males. The extracellular K+ rise during HSD was attenuated in Mecp2−/y males and the input resistance of CA1 pyramidal neurons decreased less before HSD onset. CA1 pyramidal neurons were smaller and more densely packed, but the cell swelling during HSD was unaffected. The intrinsic optical signal and the propagation of HSD were similar among the different genotypes. Basal synaptic function was intact, but Mecp2−/y males showed reduced paired-pulse facilitation and higher field potential/fiber volley ratios, but no increased seizure susceptibility. Synaptic failure during hypoxia was complete in all genotypes and the final degree of posthypoxic synaptic recovery indistinguishable. Cellular ATP content was normal in Mecp2−/y males, but their hematocrit was increased as was HIF-1α expression throughout the brain. This is the first study showing that in Rett syndrome, the susceptibility of telencephalic neuronal networks to hypoxia is increased; the underlying molecular mechanisms apparently involve disturbed K+ channel function. Such an increase in hypoxia susceptibility may potentially contribute to the vulnerability of male Rett patients who are either not viable or severely disabled.
Rett syndrome is a genetic, X-chromosome–linked neurodevelopmental disorder caused by mutations in the MECP2 gene (Amir et al. 1999). It almost exclusively affects girls, starting within their first 6–18 mo. This apparent gender preference arises because Rett boys are usually not viable; if born, they are severely disabled and die prematurely. Rett syndrome primarily targets the brain. An initially normal development is followed by loss of motor capabilities and impaired cognitive function. Other characteristics are stereotyped hand movements, spasticity, breathing disturbances, epileptic seizures, and autistic behavior (Hagberg et al. 1983; Steffenburg et al. 2001). The MECP2 gene affected in Rett patients encodes for a transcriptional regulator, methyl CpG binding protein 2 (MeCP2), which is involved in long-term gene silencing (Bienvenu and Chelly 2006). Target genes being controlled include brain-derived neurotrophic factor (BDNF), the transcription factor DLX5 (Bienvenu and Chelly 2006), the modulator of Na+/K+ ATPase phospholemman (Deng et al. 2007), and a subunit of mitochondrial respiratory complex III (Kriaucionis et al. 2006). Since MeCP2 is required for the formation and maintenance of functional synapses and neuronal networks (Zoghbi 2003), Rett syndrome is considered a synaptopathy. At birth MeCP2 is highly expressed in those networks being vital at birth, i.e., respiratory, circulatory, and cardiac control. In the neocortex, hippocampus, and cerebellum it is expressed later in life when these networks mature (Shahbazian et al. 2002; Zoghbi 2003). Interestingly, MeCP2 is abundant in mature neurons, but not expressed in glia (Jung et al. 2003; Shahbazian et al. 2002).
Different Rett mouse models are available carrying MECP2 knockout (KO) mutations (Guy et al. 2001) or MECP2 truncations (Shahbazian et al. 2002). They provide accumulating evidence that MeCP2 deficiency affects synaptic function and plasticity. A shift of excitation/inhibition causes cortical hyperexcitability and epileptic seizures (Glaze 2005) and, among the affected molecular targets, is a modulator of the Na+/K+ ATPase (Deng et al. 2007) and the γ-aminobutyric acid type A (GABAA) receptor β3 subunit (Samaco et al. 2005). Also a reduced expression level of the GABAA receptor α2 and α4 subunits has been reported (Medrihan et al. 2008). In these genetic mouse models, however, the typical signs of Rett syndrome are milder than those in the patients.
Rett patients are threatened by sudden death resulting from cardiac dysregulation and/or breathing disturbances (Julu et al. 2001; Kerr et al. 1997). In the awake state phases of irregular, inadequate breathing occur, paralleled by drops in arterial O2 levels <60 mmHg (Julu et al. 2001). They apparently reflect overexcitability of postinspiratory neurons resulting in prolonged expiratory intervals and apneas lasting up to 1 min (Stettner et al. 2007, 2008). To which extent these apneas and intermittent hypoxic episodes affect neural function is largely unclear. Hippocampal, neocortical, and cerebellar circuits are among the most hypoxia/ischemia vulnerable parts of our brain (Pulsinelli et al. 1982; Schmidt-Kastner and Freund 1991). Accordingly, they may be potentially impaired by such repeated metabolic compromise.
We therefore screened for changes in synaptic function and, in a multiparametric approach, we analyzed the hippocampal responses to severe acute hypoxia, systematically comparing both transgenic males and females with wildtype mice. As hypoxic network response the hypoxia-induced spreading depression (HSD) was chosen, a concerted massive depolarization of neurons and glial cells that is critically dependent on intact neuronal networks and whose ignition is affected by either kind of excitatory/inhibitory imbalance (Müller 2000; Müller and Somjen 2000a,b; Somjen 2001). Multiple electrophysiological parameters of HSD were analyzed, complemented by an analysis of the associated intrinsic optical signals that report the propagation velocity of HSD and the severity of brain tissue invasion (Aitken et al. 1999; Andrew et al. 1999; Müller and Somjen 1999). Also, hypoxia-related synaptic failure and recovery were elucidated.
As a Rett model mice lacking exons 3 and 4 of the MECP2 gene [B6.129P2(C)-Mecp2tm-1-1Bird (Guy et al. 2001)] were used. Heterozygous female mice were obtained from Jackson Laboratories (Bar Harbor, ME) and bred with wildtype (WT) males (C57BL/6J) to generate heterozygous females (Mecp2+/−), hemizygous males (Mecp2−/y), and WT mice of either sex. Ether anesthetized mice were decapitated and the brain was rapidly removed from the skull and placed in chilled artificial cerebrospinal fluid (ACSF) for 1–2 min. At the same time blood samples were taken for hematocrit analysis. For later HIF-1α level analysis brain tissue samples from different regions were isolated and immediately frozen in liquid nitrogen. Acute neocortical/hippocampal tissue slices (400-μm-thick transverse slices) were cut from the forebrain using a vibroslicer (752M Vibroslice, Campden Instruments). The slices were then separated in the sagittal midline, transferred to an Oslo-style interface recording chamber, and left undisturbed for ≥90 min. The recording chamber was kept at a temperature of 35–36°C, continuously aerated with 95% O2-5% CO2 (400 ml/min), and perfused with oxygenated ACSF (3–4 ml/min). The ACSF contained (in mM): 130 NaCl, 3.5 KCl, 1.25 NaH2PO4, 24 NaHCO3, 1.2 CaCl2, 1.2 MgSO4, and 10 dextrose, aerated with 95% O2-5% CO2 to adjust pH to 7.4.
Morphometric cell analysis
Mice were deeply anesthetized with isoflurane (Abbott). Once respiration was severely depressed, final gasping was visible, the animal failed to respond to noxious pinch of tail or toe, and the heart stopped beating, the chest was opened ventrally and the animals were perfused transcardially, first with phosphate-buffered saline (PBS, 100 ml) to remove blood and subsequently with 150 ml PBS containing 4% paraformaldehyde (PFA, Riedel de Haen). The isolated brain was postfixed (>48 h) in 4% PFA containing 30% saccharose (NeoLab), before consecutive coronal sections (20 μm) were cut using a freezing microtome (Reichert–Jung). Sections were placed on microscope slides, air-dried, incubated for 4 h in propanol:ethanol (1:1), rinsed in a descending ethanol series (99.9, 96, 90, and 70%), and washed in demineralized water. Slides were then incubated in thionin solution (1.25 mg/ml) for 1–2 min; excessive thionin was washed out by water and an ascending ethanol series (70, 90, 96, and 99.9%). Sections were cleared in terpineol:xylol (1:1; Sigma–Aldrich) and xylol (Roth) and coverslipped with Depex (Serva). Morphometric analysis was performed with a digital microscope (Coolscope, Nikon) equipped with geometric cell analysis software (NIS-Elements AR 2.1, Nikon).
HIF-1α Western blot
Brain tissue samples were rapidly homogenized in a buffer containing 50 mM Tris-HCl, 1% SDS (sodium dodecyl sulfate), 1 mM sodium-orthovanadate, 2 mmol/l ethylenediaminetetraacetic acid (EDTA), 1 mM phenylmethylsulphonyl fluoride (PHSF), 1 μg/ml aprotinin, and 1 μg/ml leupeptin. Proteins were quantified using the DC-protein assay (BioRad). Protein (200 μg) in 1× Laemmli SDS sample buffer was boiled for 5 min and after centrifugation loaded onto a 10% SDS-PAGE gel. After electrophoresis, proteins were transferred to a nitrocellulose membrane. The filters were blocked with 5% nonfat milk and then incubated with antibodies directed against HIF-1α (Abcam) and β-actin (Sigma–Aldrich).
Hypoxia-protocol and electrical recordings
Severe hypoxia was induced by switching the recording chamber′s gas supply from carbogen (95% O2-5% CO2) to 95% N2-5% CO2 (carbogen aeration of the ACSF was continued) and it triggered HSD within a few minutes. O2 was resubmitted 20 s after the onset of HSD; within that time the extracellular DC-potential shift had reached its nadir. Extracellular recording electrodes were made from thin-walled borosilicate glass (GC150TF-10, Harvard Apparatus), filled with ACSF, and the tips were trimmed to a resistance of about 5 MΩ. Sharp microelectrodes for current-clamp recordings were made from thick-walled glass (GC 150F-10, Harvard Apparatus), filled with 2 M K-acetate + 5 mM KCl +10 mM HEPES [(N-(2-hydroxyethyl) piperazine-N′-(2-ethanesulfonic acid)] (pH 7.4). Their tips were beveled to a final resistance of about 80 MΩ.
Field excitatory postsynaptic potentials (fEPSPs) were elicited by 0.1-ms unipolar stimuli (Grass S88 stimulator with PSIU6 stimulus isolation units; Grass Instruments), delivered via steel microwire electrodes (50-μm diameter, A-M Systems). Orthodromic responses were elicited by stimulation of Schaffer collaterals and recorded in stratum radiatum of the CA1 region with a locally constructed extracellular DC-potential amplifier as previously described (Hepp et al. 2005). All electrophysiological data were sampled using an Axon Instruments Digitizer 1322A and pClamp 9.2 software (Molecular Devices). Fiber volleys were analyzed from representative traces only, which showed clear fiber volleys and high signal-to-noise ratios. Synaptic failure during hypoxia and posthypoxic recovery were analyzed in continuous DC-potential recordings, eliciting fEPSPs every 20 s (Hepp and Müller 2008).
Current-clamp recordings from CA1 neurons were performed with an intracellular recording amplifier (SEC-05L; npi Instruments, Tamm, Germany) (Hepp et al. 2005). Bridge balance and electrode-capacitance compensation were continuously controlled. Only CA1 neurons with a stable membrane potential of at least −55 mV were accepted. Their input resistance was probed every 10 s by a hyperpolarizing current (400–600 pA amplitude, 200-ms duration) and measured at the steady-state level of the voltage deflections, averaging five successive current pulses.
Changes in the extracellular K+ concentration ([K+]o) were recorded using double-barreled K+-selective microelectrodes of the twisted type and a differential electrometer amplifier (FD 223, World Precision Instruments) as previously described in detail (Hepp and Müller 2008). Two borosilicate glass capillaries—one with and the other without a filament (GC150-15 and GC100F-15, Harvard Apparatus)—were cut in halves, glued together with epoxy glue, and pulled on a vertical puller (Narishige PE-2). The designated ion-selective barrel was silanized by HMDS vapors (hexamethyldisilazane, 98%; Fluka), filled with the Potassium Ionophore I (Cocktail A, Fluka 60031), and backfilled with 150 mM KCl + 10 mM HEPES (pH 7.4). The reference barrel contained 150 mM NaCl + 10 mM HEPES (pH 7.4) (Hepp and Müller 2008; Müller and Somjen 2000a). Electrode resistances of the reference and the ion-selective barrel were 10–20 and 180–250 MΩ, respectively. K+-selective electrodes were calibrated by detecting their responses in standard solutions (0, 1, 2, 5, 10, 20, 50, and 100 mM K+). Their average slopes were 52.7 ± 4.3 mV/decade and their detection limits averaged 0.87 ± 0.62 mM K+ (n = 26). [K+]o was calculated directly from the electrode responses using the slope of the calibration.
Cell swelling during HSD was quantified using tetramethylammonium (TMA)-selective microelectrodes, determining changes in the extracellular background concentration (1.5 mM) of the cell-impermeant marker TMA ([TMA]o) according to the indicator-dilution technique (Hansen and Olsen 1980; Nicholson and Phillips 1981). The TMA-selective barrel was filled with the Corning 477317 K+ ion exchanger (IE190, World Precision Instruments) and backfilled with 150 mM TMA-Cl + 10 mM HEPES (pH 7.4). The reference barrel contained 150 mM NaCl + 10 mM HEPES (pH 7.4) (Müller 2000; Müller and Somjen 1999). TMA-selective electrodes were calibrated by detecting their responses in standard solutions (0, 0.1, 0.5, 1, 5, 10, 50, and 100 mM TMA solutions). Their average slopes were 59.5 ± 5.1 mV/decade and their detection limits averaged 0.35 ± 0.41 mM TMA (n = 31). Changes in relative interstitial volume (ISV) were normalized to control conditions according to the formula (Dietzel et al. 1980; Müller 2000): ΔISV(%) = 100 × [([TMA]baseline/[TMA]hypoxia) − 1].
The intrinsic optical signal (IOS) of HSD–an increase in light scattering—was monitored with a Polychrome II imaging system (Till Photonics, Gräfelfing, Germany) and a sensitive CCD camera (Imago QE, PCO Imaging, Kelheim, Germany) as previously described (Gerich et al. 2006; Hepp and Müller 2008). Interfaced hippocampal slices were illuminated (white light) at an angle of close to 45°, viewed with a ×5 objective (Epiplan, Zeiss), and images were taken every 2 s (15-ms exposure time). Hypoxia-induced changes in tissue reflectance were visualized by off-line image subtraction and normalized to prehypoxia baseline reflectance (Müller and Somjen 1999). They are displayed in a 256 gray-scale mode covering a range of ±20% brightness changes and were quantified in a small rectangular region of interest in CA1 st. radiatum close to the recording electrode. The propagation velocity of HSD was calculated from the wave-front progression of the reflectance increase. The invaded area was determined by counting those pixels whose brightness had increased by ≥5% (Gerich et al. 2006). Image analysis was performed with Tillvision 4.0 (Till Photonics) and MetaMorph Off-line 6.1 (Universal Imaging).
Determination of cellular ATP levels
Cellular adenosine triphosphate (ATP) content was determined from acute hippocampal tissue slices (hippocampal formation detached from neocortex) as previously reported (Foster et al. 2006; Gerich et al. 2006). Slices were transferred in perchloric acid (8%) and homogenized by sonication. ATP levels were quantified based on a coupled reaction of glucose-6P-dehydrogenase and hexokinase, by spectrophotometrically monitoring the formation of NADPH2 (Foster et al. 2006; Lamprecht and Trautschold 1974), and they were normalized to the protein content of the given sample determined by Bradford protein assay.
Since the experiments did not last longer than about 1 h, we used up to four slices from each brain. To ensure independence of observations, each experimental series was performed on at least four different animals of each genotype. All numerical values are represented as means ± SD; the number of experiments (n) refers to the number of slices investigated. Due to the different ages of male (38–60 days) and female mice (5–14 mo) used, cross-gender comparisons were not performed anywhere in the study. Significance of the observed changes (wildtype male vs. Mecp2−/y male or wildtype female vs. Mecp2+/− female) was tested using a two-tailed, unpaired Student's t-test and a significance level of P = 5%. In the case of significant changes, the P value is reported and in the diagrams significant changes are indicated by asterisks (*P < 0.05; **P < 0.01).
In acute hippocampal slices, severe hypoxia triggers within a few minutes hypoxia-induced spreading depression (HSD). It is characterized by a nearly complete depolarization of neurons and glial cells that is synchronized in neighboring cells, severely disturbed ionic distribution, and a negative shift in the extracellular DC potential (Hansen 1985; Müller and Somjen 2000a; Nicholson and Kraig 1981; Somjen 2001). Synaptic function and axonal conduction are blocked and, once triggered, HSD slowly spreads out from its ignition site (Müller and Somjen 1999; Somjen 2001). To screen for differences in the hypoxia susceptibility of wildtype (WT) and Rett mice, HSD was followed both electrically and optically (Fig. 1, Table 1). Hemizygous males (Mecp2−/y) already die around postnatal day 60, whereas heterozygous females (Mecp2+/−) develop symptoms after only several months. Therefore the age of the males (38–60 days) and females (5–14 mo) used differed.
In WT males HSD occurred within 170 ± 54 s of severe hypoxia; the sudden, negative DC shift (ΔVo) had an amplitude of −15.6 ± 2.5 mV and, measured at the half-amplitude level, it lasted 51 ± 14 s (n = 14). WT females generated similar HSD episodes (for details see Table 1, Fig. 1A). Comparing WT and Rett mice showed that the characteristic HSD parameters—amplitude, duration, and time to onset—determined for Mecp2+/− females did not differ significantly from those obtained in WT females. In Mecp2−/y males the onset of HSD, however, was significantly hastened, on average by 24%, compared with WT males (P = 0.0179). HSD amplitude and duration were unchanged and neither did the time course or profile of the DC-potential shifts clearly differ among the genotypes (Fig. 1A, Table 1).
In addition to the electrophysiological features of HSD the optical signs—an increase in light scattering referred to as intrinsic optical signal (IOS)—were monitored as well. The IOS coincides with the electrophysiological signs of HSD and is detectable as an increase in light reflectance of a given slice (Aitken et al. 1999; Andrew et al. 1999; Basarsky et al. 1998; Müller and Somjen 1999). In all Mecp2 genotypes the reflectance increase was most pronounced in the dendritic layers (st. radiatum, st. oriens) but relatively spared st. pyramidale. It preferentially invaded CA1 and dentate gyrus but did not spread into the CA3 subfield (Fig. 1B). As indicated by the reflectance increase, neocortical areas and parts of the diencephalon/basal ganglia also underwent HSD (Fig. 1B); nevertheless, we focused our detailed analysis on the hippocampus. Quantifying the dynamic changes in light reflectance in a small rectangular region close to the recording electrode (CA1, st. radiatum) yielded comparable increases in light reflectance for WT and Mecp2−/y males as well as WT and Mecp2+/− females. Also the time course of the IOS was similar (Fig. 1, B and C; Table 1). The relative area, i.e., those parts of the total hippocampal area invaded by the IOS (on average 54–56%), was indistinguishable among the different genotypes. The propagation velocity of HSD, determined from the spreading wave front of the IOS, was similar in WT and Mecp2−/y males; neither did it differ among WT and Mecp2+/− females (Fig. 1D).
Ionic disturbance and repetitive hypoxia
To identify possible mechanisms underlying the hastened onset of HSD in Mecp2−/y males, we quantified the changes in [K+]o because extracellular K+ accumulation is a key event in the generation and propagation of HSD (Grafstein 1956; Kager et al. 2002; Van Harreveld 1978). The [K+]o was monitored continuously with ion-selective microelectrodes and HSD was repeatedly induced to elucidate whether WT and Mecp2−/y males respond differently to repeated hypoxic treatment. During hypoxia the [K+]o shows a characteristic profile, consisting of an initial increase before HSD onset (threshold), followed by a massive rise as HSD is ignited (peak) and a transient undershoot of the prehypoxic baseline (undershoot) during posthypoxic recovery (Fig. 2A). In WT males, soon after O2 withdrawal, [K+]o started to increase at a rate of 5.8 ± 1.6 mM/min, reaching a level of 9.3 ± 4.9 mM before HSD was ignited. With the occurrence of the DC-potential deflection, [K+]o rapidly rose to its peak level of 73.7 ± 17.5 mM. On reoxygenation [K+]o recovered and transiently undershot its prehypoxic baseline, reaching a nadir of 2.1 ± 0.5 mM (n = 13, Fig. 2A). Repeated hypoxic treatment (20-min recovery between HSD episodes) did not reveal any significant changes in the characteristic electrophysiological signs of HSD, i.e., time to HSD onset, amplitude, and duration of the DC shift (Fig. 2, A and B). In Mecp2−/y males the initial rate of the K+ increase (5.6 ± 1.2 mM/min) and the K+ level reached before HSD onset (8.2 ± 3.6 mM) corresponded to the changes observed in WT males; the K+ peak level at the height of HSD, however, was significantly reduced, averaging 52.3 ± 29.5 mM (n = 12; P = 0.0358). The posthypoxic undershoot of the K+ baseline (2.1 ± 0.5 mM) was again indistinguishable from WT males (Fig. 2A). Inducing a second HSD again showed a significantly decreased K+ peak level at the height of HSD (P = 0.0431; Fig. 2C); during the third HSD episode no differences were observed.
The IOS responses remained unchanged when the hypoxic treatment was repeated. Comparing the reflectance increase in WT (n = 13) and Mecp2−/y (n = 12) males did not reveal any differences in its intensity or time course; neither did the propagation velocity of HSD or the degree of tissue invasion show any signs of adaptation or differences among the genotypes (Fig. 3).
Intracellular recordings during HSD
In search for the cause of the reduced K+ peak level during HSD in Mecp2−/y males, we performed intracellular recordings from CA1 pyramidal neurons in acute slices. Their resting membrane potential (point “a” in Fig. 4A) and input resistances in WT (−68.3 ± 10.3 mV; 25.3 ± 8.3 MΩ, n = 14) and Mecp2−/y males (−66.2 ± 11.6 mV; 27.7 ± 9.4 MΩ, n = 18) under control conditions were similar, but their responses to severe hypoxia showed some differences. Early during hypoxia, CA1 pyramidal neurons of WT males hyperpolarized by −3.2 ± 4.3 mV and their input resistance decreased by 43.0 ± 16.2% (point “b” in Fig. 4A). The initial hyperpolarization then turned into a slow depolarization. At a threshold potential of −44.8 ± 12.8 mV (point “c” in Fig. 4A) the sudden depolarization occurred driving the membrane potential to −18.4 ± 4.8 mV (point “d” in Fig. 4A) and then more slowly to a final peak of −10.7 ± 6.7 mV (point “e” in Fig. 4A). During reoxygenation cell impalement usually became unstable (arrow mark in Fig. 4A) and then was lost because of severe cell swelling. Accordingly, in most slices the posthypoxic recovery could not be followed completely and putative differences in that phase could not be elucidated. In Mecp2−/y males the initial hyperpolarization was comparable, but the decrease in input resistance (−27.9 ± 15.4%) was significantly less pronounced than that in WT males (P = 0.0290). The threshold potential, the final depolarization reached, and the massive decrease in input resistance corresponded to those in WT males (Fig. 4B). Since a similar massive depolarization was found in CA1 neurons of WT and Mecp2−/y males, incomplete neuronal depolarizations cannot be the cause for the reduced K+ peak level observed during HSD in Mecp2−y males.
Differences in ISV and cell swelling during HSD
Another reason for the reduced K+ peak level during HSD in Mecp2−/y males might be a wider interstitial space due to smaller neuronal size and/or a different packing density of neuronal elements. To elucidate this possibility the neuronal size and packing were analyzed and cell volume changes during HSD were monitored.
Rett mice were previously reported to have a reduced brain size—about 13% for the Rett mouse model used (Belichenko et al. 2008). Morphometric analysis of thionin-stained cortico-hippocampal sections (20–50 μm thickness, three slices for each genotype) of age-matched mice show an average reduction in the area of the hippocampal formation in Mecp2−/y males by 17.7% (Fig. 5A), extending earlier studies that have found that Mecp2 deficiency reduces neuronal sizes in the CA2 subfield by 15–25% (Chen et al. 2001). Analyzing the soma size of CA1 pyramidal neurons we obtained clear differences between WT (152 ± 29 μm2) and Mecp2−/y males (86 ± 19 μm2, n = 300 neurons each, P < 0.0001), amounting to an average reduction in neuronal size by 43% (Fig. 5B). Cell counting in an 80-μm-long section of st. pyramidale yielded an average of 24.3 ± 1.7 and 41.7 ± 6.2 neurons in WT and Mecp2−/y males, respectively, i.e., a 71.6% higher neuronal packing density in Mecp2−/y males (n = 10 slices each, P < 0.0001). The width of the cell layer did not differ among WT (50.1 ± 6.4 μm) and Mecp2−/y males (47.9 ± 6.4 μm).
We next asked whether the smaller neuronal size and higher packing density in Mecp2−/y males may affect the degree of cell swelling during HSD and used TMA-selective microelectrodes (Hansen and Olsen 1980; Nicholson and Phillips 1981) to monitor the restriction of the interstitial volume (ISV). Since such recordings require a constant TMA background level, 1.5 mM TMA-chloride was added to the ACSF (Müller 2000; Müller and Somjen 1999). Cell swelling already started before HSD onset, increasing [TMA]o to 1.75 mM, which corresponds to a restriction in ISV by 12.6 ± 10.7% (Fig. 6, A and B). The maximum degree of cell swelling occurred at the height of HSD, decreasing ISV by 50.3 ± 13.7% (n = 19) and 50.2 ± 12.3% (n = 12) for WT and Mecp2−/y males, respectively. On reoxygenation [TMA]o slowly recovered and transiently undershot its prehypoxic baseline, obviously due to overregulation of cell volume during posthypoxic recovery and a possible contribution of TMA loss from the interstitial space into the bathing medium (Fig. 6A). The undershoot indicates a transient widening of ISV by 20.8 ± 18.2% (n = 19) and 18.8 ± 19.6% (n = 12) for WT and Mecp2−/y males, respectively. Accordingly, despite differential neuronal size and packing density, cell swelling during HSD does not differ among WT and Mecp2−/y males.
Synaptic function, plasticity, and hypoxic failure
Another factor facilitating HSD ignition in Mecp2−/y males might be an increased excitability of the hippocampal network. Disturbed GABAergic inhibition was reported for this Rett mouse model in medullary neurons, based on decreased expression of α2 and α4 subunits of the GABAA receptor (Medrihan et al. 2008), and diminished basal inhibitory rhythmic activity was found in the hippocampal CA3 subfield (Zhang et al. 2008). Therefore synaptic function was assessed by monitoring orthodromically evoked field potentials (fEPSPs) in st. radiatum of the CA1 region under our experimental conditions (interface recording chamber, ∼35.5°C). Input–output curves recorded for stimulation intensities of 10–150 μA (unipolar stimuli) did not significantly differ among the genotypes (Fig. 7A). Half-maximum amplitude responses of WT and Mecp2−/y males were obtained with 40- to 50-μA stimuli; half-maximum amplitude responses of WT and Mecp2+/− females were obtained with 30- to 40-μA stimuli (Fig. 7A). Absolute fEPSP amplitudes were equal among the different genotypes (data not shown). Normalizing the fEPSP amplitudes to the fiber volley did not show any differences for WT and Mecp2+/− females, but yielded a significantly increased ratio for Mecp2−/y than WT males, indicating increased postsynaptic excitability in Mecp2−/y males (P = 0.0314; Fig. 7B). Yet obvious signs of pronounced hyperexcitability, i.e., multiple population spikes, were not observed on a regular basis.
As a paradigm for synaptic short-term plasticity we quantified paired-pulse facilitation at various interpulse intervals (25–200 ms). With stimulation intensities adjusted to yield nearly 50% response amplitudes, marked paired-pulse potentiation was observed. At an interpulse interval of 25 ms it averaged 209.8 ± 58.1% for WT males, but was significantly less pronounced in Mecp2−/y males: 165.9 ± 25.2% (n = 14 each; P ≤ 0.0362; Fig. 7, C and D). In WT and heterozygous females no differences were observed (WT 168.2 ± 33.8%, n = 24; Mecp2+/−177.8 ± 41.0%, n = 14; Fig. 7, C and D).
Furthermore, we monitored the loss of synaptic function during hypoxia, to decide whether an altered time course of synaptic failure may contribute to the increased HSD susceptibility of Mecp2−/y males. Severe hypoxia blocks synaptic function (fEPSPs) within about 2 min (Hansen et al. 1982), as indicated by the inability to elicit evoked responses. We observed a 50% reduction in fEPSP amplitudes within close to 80 s of severe hypoxia and it was followed by a complete block of synapses; the time course and extent of synaptic failure were indistinguishable among the different genotypes (Fig. 8A). Nonetheless some differences became evident during the posthypoxic recovery of synaptic function. Reoxygenation was started 20 s after HSD onset and it restored synaptic function within 7–8 min (Fig. 8B). Synaptic recovery tended to be slightly delayed in WT females, which showed 50% synaptic response amplitudes after 280 s of reoxygenation, whereas Mecp2+/− females regained 50% responses after 220 s of reoxygenation. Both WT and Mecp2−/y males regained 50% fEPSPs responses after 220 s of reoxygenation (Fig. 8B). Posthypoxic potentiation, a transient increase of fEPSP amplitudes immediately after synaptic recovery (Crepel et al. 1993; Frenguelli 1997; Gozlan et al. 1994), was observed in WT males only, averaging 121.7 ± 27.8% of control (n = 13, P ≤ 0.0390; Fig. 8B). The final degree of synaptic recovery reached after 20 min of reoxygenation, however, was indistinguishable among the different genotypes; fEPSPs stabilized at about 80% of their respective control amplitudes (Fig. 8B).
In Rett syndrome neuronal networks seem functionally disturbed by impaired inhibition and increased excitation, resulting in hyperexcitability (Medrihan et al. 2008; Zhang et al. 2008). The increased ratio of fEPSP/fiber volley observed in Mecp2−/y males (Fig. 7B) also suggests such increased excitability for the hippocampal network. We therefore asked whether seizure susceptibility is increased and applied stimuli of different severity: low Mg2+ solutions to potentiate N-methyl-d-aspartate (NMDA) receptors (Anderson et al. 1986; Mody et al. 1987), low Mg2+ solutions plus bicuculline to also diminish GABAergic inhibition (Borck and Jefferys 1999), and 4-aminopyridine (4-AP) to block K+ channel-mediated inhibition (Rutecki et al. 1987). DC potentials were recorded in st. radiatum of the CA1 subfield and the temperature was lowered to 33°C to prevent spontaneous spreading depression episodes (Mody et al. 1987).
Lowering [Mg2+]o to 0.25 mM induced spontaneous burst discharges with amplitudes of close to 0.2 mV, but evoked fEPSPs did not show any obvious signs of hyperexcitability (Fig. 9, Table 2). Combining low Mg2+ solution with 10 μM bicuculline induced pronounced interictal spikes of approximately 1.5-mV amplitudes and 0.5-s duration, and the evoked fEPSPs showed multiple population spikes, indicating hyperexcitability (Fig. 9). Application of 100 μM 4-AP induced massive seizure-like events that were characterized by an initial fast negative deflection by ≤2 mV followed by pronounced bursting activity, but were terminated early and typically lasted only about 1.5–2 s. In the presence of 4-AP evoked fEPSPs also showed multiple population spikes (Fig. 9). Quantitative analysis of the different types of spontaneous and seizure-like activity (for summary see Table 2) did not reveal any significant differences for slices obtained from WT and Mecp2−/y males. Accordingly an increased seizure susceptibility and/or severe hyperexcitability in Mecp2−/y males can also be excluded to underlie the increased hypoxia susceptibility.
Blood and tissue analyses
As a test for systemic adaptation of Mecp2−/y males to the repeated apnea-related systemic hypoxic episodes, we determined the hematocrit, tissue levels of hypoxia-inducible factor (HIF) 1α, and cellular ATP content. The hematocrit—determined from blood samples collected during dissection—was significantly higher in Mecp2−/y (49.8 ± 2.3, n = 8) than that in WT males (44.1 ± 1.4, n = 7, P = 0.0001; Fig. 10A), which suggests that during their life these Mecp2−/y males experienced repeated systemic hypoxic episodes. Analysis of brain tissue samples isolated from neocortex, cerebellum, lower brain stem, and hippocampus consistently showed a higher HIF-1α expression level in Mecp2−/y (n = 4) than that in WT males (n = 3; Fig. 10B). Since HIF-1α expression increases in a transient manner within minutes of systemic hypoxia (Jewell et al. 2001; Stroka et al. 2001), the underlying cause for the observed increase in HIF-1α expression is obviously the global ischemic condition inherent to the isolation of the brain tissue during dissection. Nevertheless this proves that in addition to hippocampus the other brain regions of Mecp2−/y males are also more sensitive to hypoxia/ischemia.
To decide whether tissue ATP content may be altered under control conditions or affected differently by metabolic compromise, ATP levels were determined in acute tissue slices (isolated hippocampal formation only). Significant differences, however, were not detected; slices from WT males contained 10.2 ± 3.8 nmol ATP/mg protein (n = 6) and those from Mecp2−/y males 13.8 ± 7.1 nmol ATP/mg protein (n = 8). Metabolic challenge by 100 μM CN− (25 min) did not cause significant alterations, neither in WT nor Mecp2−/y males (Fig. 10C). Therefore a lower ATP content or accelerated ATP consumption can be excluded to hasten the onset of HSD in Mecp2−/y males.
Rett syndrome impairs cortical function and, due to severe breathing disturbances and frequent apneas, it is associated with intermittent episodes of systemic hypoxia. In the present study we analyzed in a Rett mouse model whether the systemic hypoxic episodes experienced by the animals disturb the function of the highly anoxia/ischemia vulnerable hippocampal formation or modulate its responses to acute severe hypoxia. The increased hematocrit of Mecp2−/y males confirms a systemic adaptation to the recurrent hypoxic episodes, but evidence of preconditioning of the hippocampal formation was not found. Rather, we observed impaired synaptic short-term plasticity in Mecp2−/y males and an increased susceptibility of the hippocampus to acute severe hypoxia. Our study is the first report on an increased hypoxia susceptibility of telencephalic neuronal networks in Rett syndrome. In multiparametric analyses we obtained first evidence that disturbed K+ channel function is among the underlying molecular mechanisms.
HSD onset was hastened, as reported by the accelerated occurrence of the DC-potential shift in Mecp2−/y males. This clearly indicates a reduced hypoxia tolerance because in these mice the neurons and glial cells lost their membrane potentials within a shorter period of hypoxia compared with WT males. Once ignited though, the neuronal hypoxic depolarizations underlying the generation of HSD were complete in both WT and Mecp2−/y males. Neither did WT and Mecp2−/y males react differently when HSD was repeatedly induced. The invasion of the hippocampal formation by HSD, its propagation velocity, and the intensity of the reflectance increase did not differ among the WT and Rett mice. Cross-gender comparison was not performed because the age of males (38–60 days) and females (5–14 mo) used differed.
Despite the hastened onset of HSD in Mecp2−/y males, the time course of synaptic failure during hypoxia and the final degree of posthypoxic synaptic recovery did not differ. This suggests that the earlier HSD onset in Mecp2−/y males does not culminate in more pronounced neuronal damage, at least not with the short duration of hypoxia applied and the time span of posthypoxic recovery analyzed. Whether long-term neuronal damage differs cannot be investigated in acute slices, but rather would have required in vivo stroke models to be able to quantify neuronal damage several days after the insult.
So far, there are only single reports on increased hypoxic responses in Rett syndrome. Respiratory responses to hypoxia (5 min, 8% O2), i.e., the initial increase in minute ventilation as well as the respiratory depression following hypoxia, were more pronounced in Mecp2−/y and Mecp2+/− females than in WT mice (Bissonnette and Knopp 2006). In cultured cerebellar granule cells cell death on hypoxia (95% N2-5% CO2) and NMDA/AMPA-mediated excitotoxicity was more pronounced because short insults not threatening WT cells were already deleterious in Mecp2−/y cells (Russell et al. 2007). Activation of caspase-3 and mitochondrial release of apoptosis inducing factor were intensified (Russell et al. 2007), suggesting that mitochondria may be more vulnerable in Mecp2−/y males.
Other mechanisms that might increase the hypoxia vulnerability are a decreased Na+/K+ ATPase activity as found in neocortical neurons of MECP2-null mice and Rett patients (Deng et al. 1999) and reduced brain ATP levels reported for MECP2-null mice (Saywell et al. 2006). In hippocampal slices from Mecp2−/y males, however, we found normal cellular ATP levels that did not react differently to moderate chemical anoxia. Also the resting membrane potentials of CA1 pyramidal neurons in WT and Mecp2−/y males did not differ. Accordingly, such changes are unlikely to hasten the onset of HSD in Mecp2−/y males.
In general, conditions increasing neuronal excitability hasten HSD onset, whereas those reducing excitability postpone it (Aitken et al. 1991; Müller 2000; Müller and Somjen 2000b). Hyperexcitability arising from reduced inhibition and giant evoked potentials were reported for neocortical networks of Rett patients (Glaze 2005), which explains why Rett patients suffer from epileptic seizures (Steffenburg et al. 2001). GABAA receptor α2, α4, and β3 subunit expression is reduced in the brain of MECP2-deficient mice (Medrihan et al. 2008; Samaco et al. 2005), basal inhibitory rhythmic activity is diminished (Zhang et al. 2008), and increased mouse brain glutamine levels suggest changes in glutamate release and recycling (Viola et al. 2007). Such molecular disturbances potentially induce hyperexcitability and indeed we observed a slightly increased ratio of fEPSP/fiber volley amplitudes in Mecp2−/y males, which corresponds to observations on Mecp2308/y mice with a truncated MECP2 gene (Moretti et al. 2006). Signs of pronounced hyperexcitability such as multiple population spikes or markedly enhanced fEPSPs amplitudes, however, were not found. Also our epilepsy tests confirm that the seizure susceptibility in Mecp2−/y males is not noticeably increased.
Basal synaptic function, i.e., the synaptic response to single pulse stimulation, was intact in Mecp2−/y males and Mecp2+/− females, which extends earlier observations restricted to Mecp2−/y males only (Asaka et al. 2006). Paired-pulse facilitation was impaired in Mecp2−/y males but not in Mecp2+/− females. Accordingly, in female Rett mice the X-chromosomal inactivation pattern seems sufficient to guarantee synaptic function, at least on the level of field potential recordings. The degree to which paired-pulse facilitation was suppressed in Mecp2−/y males matches earlier studies that were performed on the same mouse model and/or mice carrying a truncated MEPC2 gene and that also found hippocampal long-term potentiation to be impaired (Asaka et al. 2006; Guy et al. 2001; Moretti et al. 2006). Paired-pulse facilitation is of presynaptic origin (Kuhnt and Voronin 1994), whereas early long-term potentiation is induced postsynaptically (Malenka and Bear 2004). Accordingly, in the hippocampal formation MeCP2 deficiency apparently affects both pre- and postsynaptic targets.
As a possible reason for the increased hypoxia susceptibility we suggest reduced K+ flux during hypoxia. Intracellular recordings from CA1 pyramidal neurons reported a less intense decrease in input resistance early during hypoxia in Mecp2−/y males. In this early phase of hypoxia K+ channels are activated (Hansen et al. 1982; Müller and Somjen 2000a), and both KATP and BK-type KCa channels are considered to mediate the initial hyperpolarization (Erdemli et al. 1998; Fujimura et al. 1997; Zawar and Neumcke 2000). We therefore propose that in Mecp2−/y males the function and/or expression level of these channel types may be impaired. The unequivocal identification of the very K+ channel affected and the type of modulation imposed will require in-depth single-channel characterization complemented by immunohistological studies. In line with these thoughts of altered K+ channel function we found the K+ peak level at the height of HSD to be attenuated in Mecp2−/y males, which cannot be explained by incomplete hypoxic depolarizations of the single neurons.
Other critical parameters are the tissue morphology and the interstitial volume. In Rett patient brain sections cortical and subcortical neurons are of smaller size and more densely packed (Bauman et al. 1995). Others observed a mild loss of cortical pyramidal neurons as well as reduced complexity of dendritic arborizations and spine densities (Belichenko et al. 1994), although the number of presynaptic terminals seems unchanged in MECP2-deficient neurons (Nelson et al. 2006). Furthermore, delayed neocortical maturation and growth were assumed to underlie the reduced cortical thickening (Fukuda et al. 2005). Each of these changes may have contributed to the smaller dimension of the hippocampal formation. Assuming that the ISV might be restricted as well could easily explain the hastened HSD onset (Chebabo et al. 1995) because extracellular K+ and glutamate could accumulate to critical levels within a shorter duration of hypoxia. A closer proximity of hippocampal neurons may also intensify ephaptic interactions (Francis et al. 2003), facilitating the generation of HSD and seizures. Our morphometric analyses revealed a smaller size and denser packing of CA1 pyramidal neurons in Mecp2−/y males and the TMA recordings verify that at least the relative degree of cell swelling during HSD does not differ in WT and Mecp2−/y males. This excludes less intense cell swelling and ISV restriction as the underlying causes for the reduced K+ peak levels observed during HSD. Whether the absolute interstitial volume fraction differs cannot be judged on the basis of our data. This would rather require pulsed iontophoretic application of TMA and detailed analysis of diffusion profiles (Nicholson 1991; Nicholson and Phillips 1981).
Oxidative stress, which is associated with hypoxia/ischemia, also seems augmented in Rett syndrome. Quantification of antioxidant enzyme activities in red blood cells of Rett patients shows a decreased activity of superoxide dismutase, whereas plasma malondialdehyde, a marker for lipid peroxidation, is increased (Sierra et al. 2001). Accordingly, reactive oxygen and nitrogen species could also contribute to the neuronal dysfunction of Rett syndrome. As we have previously reported for hippocampal slices, such changes in cellular redox status may modulate the susceptibility of neuronal networks to hypoxia (Gerich et al. 2006; Hepp et al. 2005) as well as the outcome (Hepp and Müller 2008). Whatever the molecular mechanisms of the accelerated HSD onset may be, they are likely to be of neuronal origin because only hippocampal neurons, but not glial cells, express MeCP2 (Jung et al. 2003).
In conclusion, preconditioning of Rett mice, as may be expected from the intermittent systemic hypoxic episodes they experience, does not seem to take place. Instead, in the case of acute metabolic compromise, the hypoxia tolerance of Mecp2−/y males is diminished. As evident from the hastened onset of HSD, their hippocampal network failed to tolerate the same duration of severe hypoxia as WT mice or Mecp2+/− females. The loss of neuronal function occurred earlier and the neurological outcome of metabolic compromise (e.g., stroke, reduced perfusion, brain concussion/edema) might be worse under the condition of MeCP2 deficiency. The increased HIF-1α expression levels found in Mecp2−/y males throughout the brain strongly suggest that this increased hypoxia susceptibility is not restricted to the hippocampus. In cortical/hippocampal networks it may well contribute to the cognitive dysfunction and mental disabilities of Rett patients. In medullary control circuits of cardiorespiratory control such an increased susceptibility to hypoxia could be life-threatening, especially in view of the respiratory disturbances and the concomitant decreases in arterial oxygen levels experienced by Rett patients. Furthermore, the increased hypoxia sensitivity could potentially contribute to the vulnerability of male Rett patients, who are either not viable or, if they survive until birth, show more severe disabilities than Rett girls.
This study was supported by the Deutsche Forschungsgemeinschaft Research Center for Molecular Physiology of the Brain and by Göttingen University (Ausstattungsmittel Juniorprofessur).
We thank Prof. Dr. George G. Somjen for a critical reading of the manuscript and Prof. Dr. Irmelin Probst for help with the determination of ATP levels.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society
- Aitken et al. 1999.↵
- Aitken et al. 1991.↵
- Amir et al. 1999.↵
- Anderson et al. 1986.↵
- Andrew et al. 1999.↵
- Asaka et al. 2006.↵
- Basarsky et al. 1998.↵
- Bauman et al. 1995.↵
- Belichenko et al. 2008.↵
- Belichenko et al. 1994.↵
- Bienvenu and Chelly 2006.↵
- Bissonnette and Knopp 2006.↵
- Borck and Jefferys 1999.↵
- Chebabo et al. 1995.↵
- Chen et al. 2001.↵
- Crepel et al. 1993.↵
- Deng et al. 1999.↵
- Deng et al. 2007.↵
- Dietzel et al. 1980.↵
- Erdemli et al. 1998.↵
- Foster et al. 2006.↵
- Francis et al. 2003.↵
- Frenguelli 1997.↵
- Fujimura et al. 1997.↵
- Fukuda et al. 2005.↵
- Gerich et al. 2006.↵
- Glaze 2005.↵
- Gozlan et al. 1994.↵
- Grafstein 1956.↵
- Guy et al. 2001.↵
- Hagberg et al. 1983.↵
- Hansen 1985.↵
- Hansen et al. 1982.↵
- Hansen and Olsen 1980.↵
- Hepp et al. 2005.↵
- Hepp and Müller 2008.↵
- Jewell et al. 2001.↵
- Julu et al. 2001.↵
- Jung et al. 2003.↵
- Kager et al. 2002.↵
- Kerr et al. 1997.↵
- Kriaucionis et al. 2006.↵
- Kuhnt and Voronin 1994.↵
- Lamprecht and Trautschold 1974.↵
- Malenka and Bear 2004.↵
- Medrihan et al. 2008.↵
- Mody et al. 1987.↵
- Moretti et al. 2006.↵
- Müller 2000.↵
- Müller and Somjen 1999.↵
- Müller and Somjen 2000a.
- Müller and Somjen 2000b.
- Nelson et al. 2006.↵
- Nicholson 1991.↵
- Nicholson and Kraig 1981.↵
- Nicholson and Phillips 1981.↵
- Pulsinelli et al. 1982.↵
- Russell et al. 2007.↵
- Rutecki et al. 1987.↵
- Samaco et al. 2005.↵
- Saywell et al. 2006.↵
- Schmidt-Kastner and Freund 1991.↵
- Shahbazian et al. 2002.↵
- Shahbazian et al. 2002.
- Sierra et al. 2001.↵
- Somjen 2001.↵
- Steffenburg et al. 2001.↵
- Stettner et al. 2007.↵
- Stettner et al. 2008.↵
- Stroka et al. 2001.↵
- Van Harreveld 1978.↵
- Viola et al. 2007.↵
- Zawar and Neumcke 2000.↵
- Zhang et al. 2008.↵
- Zoghbi 2003.↵