|Home | About | Journals | Submit | Contact Us | Français|
Disruption in the expression and function of synaptic proteins, and ion channels in particular, is critical in the pathophysiology of human neuro-psychiatric and neuro-degenerative diseases. However, very little is known regarding the functional and pharmacological properties of native synaptic human ion channels, and their potential changes in pathological conditions. Recently, an electrophysiological technique has been enabled for studying the functional and pharmacological properties of ion channels present in crude membrane preparation obtained from post-mortem frozen brains. We here extend these studies by showing that also human synaptic ion channels can be studied in this way. Synaptosomes purified from different regions of rodent and human brain (control and Alzheimer’s) were characterised biochemically for enrichment of synaptic proteins, and expression of ion channel subunits. The same synaptosomes were also reconstituted in Xenopus oocytes, in which the functional and pharmacological properties of the native synaptic ion channels were characterised using the voltage clamp technique. We show that we can detect GABA, AMPA and NMDA receptors and modulate them pharmacologically with selective agonists, antagonists and allosteric modulators. Furthermore, changes in ion channel expression and function were detected in synaptic membranes from Alzheimer’s brains. Our present results demonstrate the possibility to investigate synaptic ion channels from healthy and pathological brains. This method of synaptosomes preparation and injection into oocytes is a significant improvement over the earlier method. It opens the way to directly testing, on native ion channels, the effects of novel drugs aimed at modulating important classes of synaptic targets.
Ligand- and voltage-gated ion channels play essential roles in controlling cell excitability and, specifically in the nervous system, synaptic activity and plasticity (Voglis et al., 2006). Ion channels also represent important drug targets with several successfully marketed drugs targeting them as agonists, antagonists or allosteric modulators (Bagal et al., 2013). In the process of drug discovery we and others, have made extensive use of recombinant systems, where molecularly identified ion channel subunits are heterologously expressed in cell lines (Broad et al., 2006) or oocytes (Zwart et al., 2008; Zwart et al., 2014a) where their function and pharmacology can be evaluated. This has tremendous value for drug discovery in allowing high throughput screening of diverse compound libraries, as well as molecular dissection of drug action with the use, for example, of subunit chimeras or point mutations (Young et al., 2008).
One limitation of recombinant systems is that they are unlikely to replicate the structure and function of native ion channels with good fidelity since the subunit stoichiometry, accessory subunits, posttranslational modifications, and lipid environment are not completely known. This might affect, or even mislead, our understanding of drug effects in humans. Importantly, with the possible exception of human ion channel genetic mutations, recombinant ion channels do not reflect the pathological state occurring in tissues from human patients.
One way to circumvent some of the issues described above is to confirm the drug effects in native tissues, like primary neuronal cultures, brain slices or brain homogenates. While this is very useful, it is not uncommon that receptors and ion channels behave differently across species and, importantly, that the drugs themselves show species-specific pharmacology.
Human neuronal cells and tissues have been used, albeit in a limited way, to confirm drug effects on human native ion channels. This includes the use of human neuroblastoma cell lines (Carbone et al., 1990; Chini et al., 1992), the expanding area of human inducible pluripotent stem cell-derived neurons (Dage et al., 2014; Chatzidaki et al., 2015), human brain homogenates (Mitewa et al., 2013), and surgical brain specimens (Zwart et al., 2014b).
A novel approach, pioneered by Miledi and colleagues (Miledi et al., 2006), is based on the micro-transplantation into the cytosol of Xenopus oocytes of cellular membranes obtained from human frozen brains. These membranes undergo fusion with the plasma membranes of the oocytes, allowing the expression and the study of native human ion channels properties and pharmacology in normal and disease conditions (Palma et al., 2002; Bernareggi et al., 2007; Limon et al., 2012). We also used this approach recently to study the efficacy of novel antiepileptic agents at native AMPA receptors (Zwart et al., 2014b).
Most of the papers on human brain membranes micro-transplantation in oocytes utilized crude brain membrane preparations (Palma et al., 2002; Miledi et al., 2006; Bernareggi et al., 2007; Limon et al., 2012; Zwart et al., 2014b). It would be highly valuable if we could use this technique to study more specifically synaptic receptors and ion channels, since these are core to the emerging understanding of human synaptopathies (Selkoe, 2002; van Spronsen et al., 2010; Grant, 2012). There is at least one precedent where mammalian synaptosomes have been purified from rat cerebellum and successfully micro-transplanted into oocytes (Sanna et al., 1996).
Our aim was to expand this evaluation to synaptosomes from other rodent brain areas, as well as to explore its feasibility with human brain synaptosomes. Here we report on the successful recording of glutamatergic and GABAergic ion channel subtypes from rodent synaptosomes obtained from the prefrontal cortex, the hippocampus and the cerebellum, as well as, novel findings on human synaptosomes obtained from the cortex of control and Alzheimer’s brains.
Rat pre-frontal cortex, hippocampal and cerebellar brain samples were dissected from adult, female Wistar rats (Charles River, Margate, UK). Frozen fragments of human normal and AD cortex, hippocampus and cerebellum brain samples were obtained with ethical approval from the Oregon Alzheimer’s Disease Center, kept frozen at −80°C and utilised according to the Code of Ethics of the World Medical Association (Declaration of Helsinki), printed in the British Medical Journal (18 July 1964). Age, gender and post mortem interval (PMI) are listed in supplemental Table 1. All animal procedures were performed in accordance with the Animals (Scientific Procedures) Act 1986 and were approved by the Eli Lilly Animal Welfare Board.
Membrane samples were prepared according to the method developed and described by Miledi et al. (2006) and Eusebi et al. (2009), with minor modifications. Synaptosomes were purified with a sucrose gradient (w/wo osmotic shock) as described by Gray and Whittaker (1962) with minor modifications or with a Percoll gradient, as described by Dunkley et al. (2008) with minor changes. PSD was purified as described by Carlin et al. (1980) with minor modifications. In brief, pre-frontal cortex, hippocampal or cerebellar tissue was dissected from Wistar rats and homogenized ten times with a Teflon homogenizer in 9 mL/g of ice-cold 0.3 M sucrose containing 50 mM Tris-HCl (pH 7.4), 50mM EGTA, 50 mM EDTA and both protease (Roche) and phosphatase (Calbiochem) inhibitor cocktails. A scheme showing the protocol of membrane, synaptosome and PSD preparations is shown in supplemental figure S1. The homogenate was centrifuged at 1,500 g for 20 minutes at 4°C. The supernatant fraction (S1) was centrifuged at 16,000 g for 30 minutes at 4°C. The resulting pellet (P2) was split in two: one for total membranes and one for synaptosomes purification. For total membrane preparation, P2 was re-suspended in Tris 50 mM and centrifuged at 100000 g for 2 h at 4°C. For synaptosomes preparation, three different methods have been tested, described below.
Protocol A: Sucrose gradient with osmotic shock. P2 was re-suspended in 5 mM Tris-HCl (pH 8.0), homogenized three times with a Teflon homogenizer and left on ice for 45 minutes. The P2 fraction was homogenized again for ten times, re-suspended in sucrose to make a 34% (w/w) solution and layered onto a discontinuous sucrose density gradient consisting of, from bottom up, equal volume of sample, and buffer containing 0.85 M and 0.3 M sucrose, respectively. After centrifugation at 60,000 g for 2 hours with a Beckmann SW41Ti swing bucket rotor, the synaptosomal fraction (layer between 0.8 and 1.2 M sucrose) was collected and diluted with two volumes of ice-cold 50 mM TrisHCl (pH 7.4) and centrifuged at 48,000 g for 30 minutes. The resulting synaptosomal pellet was re-suspended in 50mM Tris. For PSD preparation, synaptosomes were extracted with 1.5% Triton for 30 min at 4°C, layered onto a buffer containing 0.85 M sucrose and 50 mM Tris pH 7.4 and centrifuged at 104,000 g for 1 h with a Beckmann SW41Ti swing bucket rotor. The final pellet (PSD) was re-suspended in Tris 50 mM.
Protocol B: Sucrose gradient without osmotic shock. P2 was re-suspended immediately in 50 mM Tris-HCl (pH 7.4) and sucrose to make a 34% (w/w) solution before been layered onto discontinuous sucrose density gradients, without the 45 minutes osmotic shock step.
Protocol C: Percoll gradient. 2 ml (4–5 mg/mL of the S1 fraction were layered gently onto each of 12 discontinuous Percoll gradients. These comprised 2 mL each of 23%, 15%, 10% and 3% Percoll (v/v) in 0.32 M sucrose buffer. The tubes were centrifuged for 5 min at 31000 g and the fraction number 4 was carefully collected with a Pasteur pipette and pooled with the corresponding fractions from other tubes.
Membranes, synaptosomes and PSD aliquots were stored at −80°C until use. Protein content was determined using the Pierce BCA protein assay kit (Thermo Scientific).
HEK293 cells stably expressing GluN1/GluN2B were grown in bulk and about 1 gram of scraped cells were homogenized and membranes for oocyte injection were prepared using the protocol as described in (Zwart et al., 2014b) for brain tissues.
2.5–5 micrograms of membranes, synaptosomes and PSD were added to an equal volume of 2× Laemmli sample buffer (Bio-Rad Laboratories) with 5% beta-mercaptoethanol and heated at 95°C for 3 min. Samples were then loaded into NuPageRTM Novex 4–12% Bis-Tris Midi Gel (Invitrogen) and run for approximately 1 hour at 200 V. Proteins were transferred to nitrocellulose membranes (Amersham Biosciences) for 2 h at 0.5 costant Ampere. Membranes were blocked with 5% non-fat dry milk in TBST (0.1% Tween 20) for 1 hour and then incubated overnight with the different primary antibodies (supplemental Table 2) diluted in TBST according to the antibody supplier data sheet. Membranes were then washed three times for 5 min with TBST and incubated with the appropriate anti-rabbit or anti-mouse secondary antibody diluted 1:40,000 for 1 hour (Sigma). After washing, Enhanced Chemiluminescence Detection Reagent (Amersham Biosciences) was applied for 5 min and images acquired with ImageQuant LAS 4000 (GE).
Xenopus oocytes (stages V–VI) were removed from sacrificed frogs and de-folliculated after treatment with collagenase type I (5 mg/mL calcium-free Barth’s solution) for 3 hours at room temperature. Unless otherwise stated, we injected 60 nL of membrane or synaptosome (1mg/mL) suspension per oocyte using a Drummond (Broomall, PA) variable volume micro-injector. In some experiments recombinant cDNAs encoding human NMDA receptor subunits GluN1 and GluN2B ligated in the pcDNA3.1 plasmid vector (InVitrogen, Carlsbad, CA) were injected in a 1:1 subunit ratio. Approximately 18.9 nL of the DNA mixture with a total concentration of 0.1 mg/mL was injected into the nuclei of the oocytes using a Nanoject Automatic Oocyte Injector (Drummond, Broomall, PA). After injection oocytes were incubated at 18°C in a modified Barth’s solution containing 88 mM NaCl, 1 mM KCl, 2.4 mM NaHCO3, 0.3 mM Ca(NO3)2, 0.41 mM CaCl2, 0.82 mM MgSO4, 15 mM HEPES, and 50 mg/L neomycin (pH 7.6 with NaOH; osmolarity 235 mOsm). Unless otherwise stated, experiments were performed on oocytes after 1–3 days of incubation for synaptosomes and 2–3 days for membranes and 3–5 days for cDNA injection.
Oocytes were placed in a recording chamber (internal diameter 3 mm), which was continuously perfused with a saline solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM BaCl2, 30 µM glycine, 10 mM HEPES (pH 7.4) at a rate of approximately 10 mL/min. Dilutions of drugs in external saline were prepared immediately before the experiments and applied by switching between control and drug-containing saline using a BPS-8 solution exchange system (ALA Scientific Inc., Westbury, NY). Between responses, the oocytes were washed for 2 minutes. Oocytes were impaled by two microelectrodes filled with 3 M KCl (0.5–2.5 MΩ) and voltage-clamped using a Geneclamp 500B amplifier (Axon Instruments, Union City, CA) according to the methods described by Stühmer (1998). The external saline was clamped at ground potential by means of a virtual ground circuit using an Ag/AgCl reference electrode and a Pt/Ir current-passing electrode. The membrane potential was held at −100 mV. The current needed to keep the oocyte’s membrane at the holding potential was measured. Membrane currents were low-pass filtered (four-pole low-pass Bessel filter, −3 dB at 10 Hz), digitized (50 Hz), and stored on disc for offline computer analysis. Data are expressed as mean ± S.E.M. All experiments were performed at room temperature.
BAPTA-AM, NMDA and (R,S)-AMPA hydrobromide, zolpidem, cyclothiazide, pregnenolone sulfate sodium salt, ifenprodil hemitartrate and (+)-MK801 maleate were obtained from Tocris (Bristol, U.K.); γ-aminobutyric acid (GABA) and (-)bicuculline methiodide were from Sigma-Aldrich (Poole, U.K.). Compound LY3130481 was synthesised internally at Eli Lilly and Company (Gardinier et al., 2016). Concentrated stock solutions of BAPTA-AM, zolpidem, cyclothiazide, pregnenolone sulfate, ifenprodil and MK801 were prepared in DMSO and stocks of NMDA, AMPA hydrobromide, GABA and bicuculline were prepared in distilled water. The stock solutions were stored in aliquots at −20 °C. Final dilutions in oocyte recording solution were made on the day of the experiments.
All analyses were performed using the Prism Graphpad software package. Concentration-inhibition curves were fitted according to the Hill equation. (GraphPad Software, San Diego, CA).
Three different protocols to purify synaptosomes (Protocol A, B and C, see methods) have been evaluated and optimised from the original description with the aim of identifying the one that allowed for the a) most efficient insertion of synaptosomes into the plasma membranes of Xenopus oocytes resulting in the largest ion channel currents, and b) the least toxicity for the oocytes. This optimisation includes the presence of high concentration of calcium chelating agents in the first steps of synaptosome purification and the re-suspension of synaptosomes in physiological buffer before injection into the oocytes. For each protocol, synaptosomes from rat PFC have been injected at the same concentration (1mg/mL), at the same volume (60 nL) the same day, in the same batch of cells, in order to avoid variability in oocytes batches. GABA, AMPA and NMDA currents were analysed by two microelectrode voltage clamp (TEVC). The best currents derived from synaptosomes purified using the sucrose gradients with osmotic shock (Fig 1). Therefore, for consistency, all subsequent experiments were performed on synaptosomes purified using Protocol A.
In order to confirm the enrichment of synaptic proteins, and decrease in contaminants, in the synaptosome fraction used for oocyte injection, we performed an extensive immunoblot analysis comparing three different subcellular fractions (total membranes, synaptosomes and PSD (Supplemental Fig S1) isolated from human cortex and rat prefrontal cortex (PFC), respectively. 5 µg of each fraction was subjected to SDS-PAGE, blotted onto nitrocellulose membranes and analysed for the presence of NMDA GluN2A and GluN2B subunits, synaptophysin (marker of the pre-synapse), PSD95 (marker of the post-synapse), and GFAP (marker of astroglia) (Fig. 2). A strong enrichment of GluN2A and 2B NMDA receptor subunits, as well as PSD95 was observed in both human and rat synaptosomes and PSD fractions compared to total membranes. In particular, the enrichment in synaptosomes versus total membranes of GluN2A is 3-fold stronger than for GluN2B, a result compatible with the reported extensive extra-synaptic distribution of the latter (Tovar et al., 1999; Traynelis et al., 2010). Synaptophysin was enriched in synaptosomes compared to total membranes, but absent in the PSD fraction, as expected for a pre-synaptic protein. The astroglia marker GFAP was also decreased in the synaptosome fraction, indicating a reduction in non-neuronal, non-synaptic, contaminants.
Immunohistochemical studies of the anatomical distribution of AMPA, NMDA and GABAA receptor subunits in rodent brain have been previously performed by several groups (Day et al., 1995; Wenzel et al., 1995; Pirker et al., 2000), but a head to head, semi-quantitative, comparison of the synaptic distribution in human and rat brain has not been previously reported, to the best of our knowledge. The proteome integrity of the human tissue used was assessed by evaluation of the HUSPIR ratio (Bayes et al., 2014) in purified synaptosomes. For each human brain region the HUSPIR ratio, which is the ratio between the full-length (Band 1) and degraded (Band 2) NR2B protein signal, was >1 (supplemental Fig S2). Fig 3 summarises the western blot results for human (A and B) and rat (C and D) brain regions. GluA1-4 AMPA subunits were found to be differentially distributed in the three different brain regions. GluA1 and GluA2 were particularly enriched in human cortex synaptosomes whereas in the rat they were expressed more in hippocampal synaptosomes. GluA3 was widely distributed in all the brain regions, both in human and rat brain. Interestingly, GluA4 was selectively enriched both in the human and rat cerebellar synaptosomes, although in the rat a lower level was found also in the hippocampus. The GluN1 NMDA receptor subunit was found in synaptosomes from all tested regions of the human and rat brain. However, the relative expression of GluN1 was higher in the cortex and in the cerebellum in human and rat, respectively. The GluN2A subunit was enriched in human cortex and rat PFC and hippocampus, with little but detectable presence in the cerebellum, whereas the GluN2B subunit was restricted to the forebrain in both species. Both GluN2C and 2D were expressed in the cortex of both human and rats, but they were expressed in the hippocampus only in rats. Surprisingly, we observed GluN2C only in synaptosomes from the human cortex and rat forebrain, while Wenzel et al. (1995) reported a selective enrichment of GluN2C in the rat cerebellum. Regarding the anatomical distribution of synaptic GABAA receptor subunit, we observed a widely and rather similar distribution of alpha 1 in both human and rat brain. The alpha 3 subunit was expressed in the forebrain of both species, but expressed in the cerebellum only in rats. The beta 1 subunit was expressed mainly in human cortex and rat hippocampus, while the beta 2 subunit was expressed broadly in both human and rat brain. Interestingly, the gamma 2 subunit was broadly expressed in human brain synaptosomes, but enriched in rat cerebellum.
The time course of incorporation of GABA and NMDA ion channels was evaluated by recording ion currents evoked by application of 1 mM GABA and 300 µM NMDA at different times after the injection of oocytes with rat PFC synaptosomes (Supplemental Fig S3 A–B). Consistent with previous findings with total membranes (Bernareggi et al., 2007) the responses were detectable within 5 hours after the injection and peaked between 16–42 hours. In order to assess if the fusion of synaptosomes with plasma membrane of the oocytes is Ca2+ dependent, oocytes were injected with rat PFC synaptosomes and incubated in Barth’s solution containing 100 µM BAPTA-AM, a cell permeable Ca2+ chelator, for 5 hours before recording (Supplemental Fig S3 C–D). In oocytes treated with BAPTA-AM significantly smaller GABA and NMDA-evoked responses were measured, suggesting that Ca2+ is important for the fusion of synaptosomes with the oocyte plasma membrane.
To determine the neurotransmitter sensitivity of the transplanted receptors, we applied different concentrations of GABA to the oocytes and we obtained concentration-response curves for rat prefrontal cortex, hippocampus and cerebellum (Supplemental Figure S4). Application of 100 µM of the GABAA receptor agonist GABA induced large inward currents that decayed to baseline upon washout of GABA. These currents were inhibited by 85% upon co-application of 80 µM of the GABAA receptor antagonist bicuculline (Straughan et al., 1971). Inhibition by bicuculline was reversed within 2 minutes of washout of the drug (Fig. 4A). Figure 4B shows that inward currents evoked by 100 µM GABA were potentiated by 144% by the GABAA receptor modulator zolpidem (Arbilla et al., 1985) and this potentiating effect of zolpidem was completely reversed within two minutes of washout of the drug.
In contrast to GABA, application of 100 µM agonist AMPA produced very weak ion current in oocytes injected with synaptosomes from rat prefrontal cortex (Fig. 4C). In order to detect AMPA receptor activity reliably, AMPA was routinely co-applied with 30 µM cyclothiazide, a positive allosteric modulator of AMPA receptors (Bertolino et al., 1993). Ion currents induced with the combination of 100 µM AMPA and 30 µM cyclothiazide were almost completely blocked by 30 µM of perampanel, an antagonist of AMPA receptors (Zwart et al., 2014b).
One batch of rat PFC synaptosomes gave small, but detectable inward currents upon application of 300 µM of NMDA. These currents were potentiated to 177% of control upon co-application of 300 µM of NMDA with 50 µM of pregnenolone sulfate, a positive allosteric potentiator of NMDA receptors (Malayev et al., 2002) (PS; Fig 4D). NMDA-induced inward currents were inhibited by 75% by 10 µM ifenprodil, a selective GluN2B antagonist (Williams, 1993) and by 100% by 1 µM MK801, a selective and non-competitive NMDA receptor antagonist (Huettner et al., 1987). The pharmacology of GABAA, AMPA and NMDA receptors were studied also in rat hippocampus (Supplemental Fig S5) and rat cerebellum (Supplemental Fig S6) synaptosomes.
A possible complication of this approach might be the loss of protein-protein complexes during the preparation procedure of synaptosomes. To address this point, we studied the effect of a novel TARP gamma-8-dependent AMPA receptor antagonist, LY3130481 (Gardinier et al., 2016; Kato et al., 2015; Maher et al., 2016). Transmembrane AMPA Receptor Regulatory Proteins (TARPs) are a family of scaffolding proteins that regulate AMPA receptor trafficking and function. TARP proteins have distinct regional expression in the brain. TARP gamma-8 is the predominant subtype expressed in the hippocampus, whereas its expression in the cerebellum is very low. To demonstrate that reconstituted synaptic ion channels in the oocyte membrane retain accessory proteins necessary for their function, we tested the effect of LY3130481 in oocytes injected with rat hippocampal and cerebellar synaptosomes (Fig. 5). Compound LY3130481 at 10 µM inhibited responses induced by 100 µM AMPA and 30 µM cyclothiazide by 55.0 ± 6.8%, whereas it left similar responses in oocytes injected with cerebellar synaptosomes unaffected.
As with the rodent synaptosomes, the injection of human synaptosomes into Xenopus oocytes also led to the reconstitution of native human synaptic GABAA and AMPA receptors in the plasma membranes of the oocytes. Figure 6A shows that 1 mM GABA evokes large inward currents that were reversibly blocked by the co-application of 80 µM bicuculline, with an inhibition of 89%. Zolpidem, a positive allosteric modulator of GABAA receptors, potentiated ion currents evoked by a low concentration of GABA (30 µM) to 212% (Fig. 6B). Upon co-application of the AMPA receptor agonist AMPA with the AMPA receptor allosteric modulator cyclothiazide large inward currents were observed that were nearly completely blocked by the AMPA receptor antagonist perampanel (Fig. 6C). Application of NMDA to oocytes injected with human synaptosomes did not evoke ion currents that were large enough to be characterized pharmacologically in more detail.
The poor expression of NMDA receptors after injection of either rat or human synaptosomes was puzzling. In order to ensure our protocols were suitable to record NMDA currents and that membrane microinjection could work also for NMDA receptors, we used two approaches. First we injected oocytes with plasmids containing recombinant human GluN1 and GluN2B subunits. Second, we purified and microinjected membranes from HEK293 cells overexpressing human GluN1/GluN2B NMDA receptors (Supplemental Fig S7). In both cases, robust NMDA currents with amplitudes ranging from a few hundreds of nA to over a µA were measurable in the oocytes. This suggests that NMDA receptors are either expressed at a lower overall density in native synaptosomes, or that they are particularly vulnerable to some form of “inactivation” during synaptosome preparation and microinjection. However, as shown before (Fig 3 and supplemental Figure 1) the actual NMDA subunit proteins were readily detectable in western blots from synaptosomes, and were not significantly degraded.
Alzheimer’s disease (AD) is characterised by loss of synaptic markers and degeneration of excitatory and inhibitory pathways (Selkoe et al., 2002). Previous studies revealed a significant decrease in the amplitude of GABA and glutamate currents in oocytes transplanted with total membranes from AD brains (Miledi et al., 2004; Bernareggi et al., 2007; Limon et al., 2012). Here, we decided to focus on synaptic ion channels, and therefore measured GABA, AMPA and NMDA protein levels and functional currents in oocytes injected with human control and AD synaptosomes (Figure 7). As expected, we observed a reduction in the protein levels of pre- (synaptophysin) and post-synaptic markers (PSD95) in synaptosomes purified from the cortex of two cases of AD (Braak stage 4), with a more pronounced loss for PSD95 (Fig 7B). NMDA GluN2A and 2B, AMPA GluA2 and GABA alpha3 receptor subunits were also significantly reduced (Fig 7B). We also found that the currents induced by GABA, AMPA and NMDA were significantly reduced, in parallel to the reduction in protein levels, in the AD samples (Fig 7A). While these data are from only a few individuals, and certainly need confirmation in a larger number of patients, the nice correlation between biochemistry and electrophysiology is promising.
We have used a combination of biochemical and electrophysiological techniques to show that synaptosomes isolated from both rodent and human brain can be successfully used to reconstitute native synaptic ion channels into the plasma membranes of Xenopus oocytes. Ion currents mediated by native, synaptic, GABA, AMPA and NMDA receptors have been recorded and their pharmacology evaluated. Our present results expand on the previously reported use of total human brain membranes in a series of papers by Miledi and colleagues (Palma et al., 2002; Miledi et al., 2006; Bernareggi et al., 2007; Limon et al., 2012) and also by ourselves (Zwart et al., 2014b) and will most likely lead to a better understanding of native human synaptic ion channels under normal and pathological conditions.
The more reliable ion currents measurable in the synaptosomal preparations were those activated by GABA. We could easily generate GABA dose-response curves in synaptosomes and compare their properties in different regions (Supplemental Figure S4) or total membranes (not shown). Overall our results with synaptosomes do not differ significantly from those reported previously with total human membranes (Miledi et al., 2004). The data were also similar to a previous report, utilising the same micro-transplantation technique, but with rat cerebellar synaptosomes (Sanna et al., 2006). Because we identified multiple GABA receptor subunits in the synaptosomal membranes biochemically, it is likely that GABA currents were mediated by a mixture of GABA receptor subtypes. One future goal is to try to modulate the currents from different synaptosomes with more selective GABA receptor modulators and try to better define their regional selectivity in human brain.
Glutamate receptors were more difficult to study in the synaptosomal preparations. While AMPA activated currents were measurable most of the time, they were relatively small and, therefore, were routinely investigated in the presence of the positive allosteric modulator cyclothiazide (Zwart et al., 2014b). This, nevertheless, allowed us to demonstrate that rat and human synaptic AMPA currents were blocked by perampanel, a novel anti-epileptic drug, in all regions tested. Furthermore, using the novel TARP gamma-8 selective AMPA receptor antagonist, LY3130481, we demonstrated that AMPA currents were selectively inhibited in the hippocampusbut not in cerebellum, consistent with the known regional expression of TARP gamma 8 in the brain. This finding also confirmed that reconstituted synaptic ion channels in oocyte membranes maintain the integrity of complexes with accessory proteins that are required for their function. While synaptosomal NMDA activated currents were even weaker than the AMPA-activated currents, we managed to get large enough currents to evaluate their pharmacology in the rat PFC. In this case, we showed that the currents were potentiated by pregnenolone sulfate, a GluN2A and 2B preferring potentiator (Malayev et al., 2002), and significantly blocked by the GluN2B selective antagonist ifenprodil (Williams et al., 1993). The two results in combination suggest a major contribution of the GluN2B subunit to these synaptic currents. However, in other regions, the NMDA currents were too weak to allow pharmacological studies. The presence of protease and phosphatase inhibitors in the buffers did not help. Likewise, increasing the holding potential, removing Mg2+ and/or adding glycine did not help. It is possible that a combination of low expression together with a specific susceptibility to metabolic degradation in post-mortem brain contribute to the lack of a robust current. On the other hand, we showed that our recording system is efficient for the study of both recombinant NMDA receptors, as well as of membranes purified from a cell line overexpressing recombinant NMDA receptors and micro-transplanted into the oocytes.
Despite these challenges, we decided to biochemically and electrophysiologically profile the expression of the various channel classes in synaptosomes purified from human control and Alzheimer’s brain. In the few individuals we studied, we identified a striking correlation between the level of ion channel protein expression and size of the functional currents. GABAA, AMPA and NMDA subunits were all expressed at much lower levels in Alzheimer’s versus control synaptosomes. In parallel, ion currents through these channels were significantly reduced, or absent, in the Alzheimer’s samples. Similar findings were reported by using total brain membranes (Miledi et al., 2004) but, to the best of our knowledge, this is the first direct functional evidence of dramatic ion channel loss in the surviving synapses from Alzheimer’s brains.
We would like to thank Drs Eleonora Palma, Marcia Roy and Seth Grant for early suggestions and for sharing protocols. RZ, ES, KMG, WP and JR are permanent employees of Eli Lilly and Company, FM is a post-doc funded by Eli Lilly and Co., GMS was an Erasmus student training in ES laboratory, and GM a visiting PhD student from the University of Florence. We would also like to thank Donna Farley (Eli Lilly, Indianapolis) for help with obtaining human brain tissues from the Oregon Alzheimer’s Disease Center (supported by NIH grant P30AG008017).
Conflicts of interest: The authors have no conflict of interest to declare.
Refer to JNC’s web site for supplementary material.