Properties of glutamate receptors of Alzheimer's disease brain transplanted to frog oocytes

Abstract

It is known that Alzheimer's disease (AD) is a synaptic disease that involves various neurotransmitter systems, particularly those where synaptic transmission is mediated by acetylcholine or glutamate (Glu). Nevertheless, very little is known about the properties of neurotransmitter receptors of the AD human brain. We have shown previously that cell membranes, carrying neurotransmitter receptors from the human postmortem brain, can be transplanted to frog oocytes, and their receptors will still be functional. Taking advantage of this fact, we have now studied the properties of Glu receptors (GluRs) from the cerebral cortices of AD and non-AD brains and found that oocytes injected with AD membranes acquired GluRs that have essentially the same functional properties as those of oocytes injected with membranes from non-AD brains. However, the amplitudes of the currents elicited by Glu were always smaller in the oocytes injected with membranes from AD brains. Western blot analyses of the same membrane preparations used for the electrophysiological studies showed that AD membranes contained significantly fewer GluR2/3 subunit proteins. Furthermore, the corresponding mRNAs were also diminished in the AD brain. Therefore, the smaller amplitude of membrane currents elicited by Glu in oocytes injected with membranes from an AD brain is a consequence of a reduced number of GluRs in cell membranes transplanted from the AD brain. Thus, using the comparatively simple method of microtransplantation of receptors, it is now possible to determine the properties of neurotransmitter receptors of normal and diseased human brains. That knowledge may help to decipher the etiology of the diseases and also to develop new treatments.

Alzheimer's disease (AD) was first described in 1907 by Alois Alzheimer (1) in a 51-year-old woman with the following symptoms: progressive memory impairment; disordered cognitive function; altered behavior, including paranoia and delusion; and decline in language function. The brains of AD subjects are characterized by accumulation of neuritic plaques containing amyloid β protein and neurofibrillary tangles, accompanied by loss of certain types of receptors and neurons (e.g., refs. 2–8). Pathological changes begin in the transentorhinal allocortex with degeneration accentuated in the hippocampal formation and entorhinal, frontal, and temporal cortices (9, 10). All these are brain regions in which glutamate (Glu) is the principal excitatory neurotransmitter, and where it is involved in higher mental functions, such as learning and memory (for review, see ref. 11).

It is known that Glu receptors (GluRs) are involved in AD, and it is thought that GluR-mediated toxicity plays an important role in cell loss associated with the disease (4, 5, 8, 12, 13). Most of the evidence comes from animal models and binding experiments in different regions of the human brain (14–18), but so far, no direct analyses of the function of the receptors have been made.

It has been demonstrated recently that it is possible to induce the incorporation of functional human receptors into the plasma membrane of Xenopus oocytes by injecting them with membranes isolated from surgically resected, as well as from postmortem, brain tissues (refs. 19–21; see also review in ref. 22). This technique of microtransplantation allows a direct characterization of the original receptors, with their associated molecules, while they are still embedded in their natural lipidic environment. Using the transplantation of cell membranes from the human brain into oocytes, we have now investigated the functional characteristics of GluRs present in the brains of AD subjects. We also studied the expression of some GluR and GABA_A receptor subunits using Western blot (WB) and quantitative real-time PCR (qRT-PCR) analyses of the same tissues used for the electrophysiological studies.

Results

Transplantation of Functional Receptors from the AD Brain.

It has been shown (21) that, a few hours after injection, oocytes injected with cell membranes isolated from postmortem control or AD brains incorporate functional GABA receptors and GluRs. For example, Fig. 1 shows sample currents elicited by GABA, kainate (Kai), and Glu applied to oocytes injected with membranes isolated from the temporal cortices of a non-AD and AD subjects [oocytes from the same frog; subjects C3 and A2 of supporting information (SI) Table 1]. Glu applied alone generated a current of only a few nanoamperes that was greatly increased when Glu was coapplied with cyclothiazide (CTZ), a known positive allosteric modulator of AMPA-type GluRs (23). A CTZ concentration of 100 μM, added after Glu pretreatment of 30 sec, was sufficient to potentiate maximally membrane currents elicited by Glu (Glu currents) (data not shown). Aspartate alone did not generate appreciable currents, but the coapplication of aspartate plus glycine elicited currents of a few nanoamperes, caused by activation of NMDA receptors.

Fig. 1.

Sample records of currents elicited by GABA, Kai, and Glu plus CTZ in two oocytes from the same frog, one injected with non-AD membranes (C3, SI Table 1) and the other with AD membranes (A2). For all figures, the oocyte membrane potential was held at −80 mV.

In addition to the ionotropic GABA and GluRs mentioned above, some oocytes injected with brain membranes incorporated also human metabotropic receptors to Glu and acetylcholine. After activation by their respective neurotransmitters, these receptors generated oscillatory chloride currents caused by triggering of the phosphoinositide cascade (24). The NMDA and the metabotropic responses were small and inconsistent and, for now, they are not investigated further.

The GABA, Kai, and Glu plus CTZ currents were detected within 1 day after membrane injection, and the oocytes continued to incorporate receptors for a few more days. After this, the responses diminished in amplitude as the oocytes began to die ≥1 wk after injection (Fig. 2). The percentage of surviving oocytes and the amplitude of the currents were very variable among different membrane preparations and different batches of oocytes, but usually the oocytes injected with non-AD membranes survived longer and consistently incorporated more functional receptors than the oocytes injected with AD membranes. For instance, in oocytes injected with non-AD membranes (non-AD oocytes), the Kai current amplitude varied between 4.2 ± 0.3 nA (mean ± SEM; number of oocytes n = 3, C11, SI Table 1) and 109.3 ± 15.8 nA (n = 27, C2), whereas the Glu + CTZ current amplitude varied between 10.8 ± 0.8 nA (n = 3, C11) and 259 ± 25.2 nA (n = 30, C2). Oocytes from the same batches injected with AD membranes had Kai current amplitudes that ranged from 0 (A3) to 19.1 ± 3.9 nA (n = 18, A4), and Glu + CTZ current amplitudes ranged between 9.9 ± 1.7 (n = 8, A5) and 70 ± 13 nA (n = 18, A4).

Fig. 2.

Time course of incorporation of GABA receptors, GluR, and Kai receptors in oocytes from the same frog injected with non-AD membranes (squares, C2, SI Table 1; n = 30; approximately four oocytes per day) or AD membranes (circles, A4, n = 26).

Therefore, for each experiment, oocytes from the same batch were injected with membranes isolated from one non-AD and from AD brains (e.g., Fig. 3A). Despite the experimental variability, the GABA, Glu, and Kai currents were consistently larger in non-AD vs. AD oocytes. This significant difference is illustrated in Fig. 3B, which shows that the currents in AD oocytes were all reduced to somewhat less than 50%.

Fig. 3.

GABA, Glu + CTZ, and Kai currents were reduced in AD oocytes. (A) Mean current amplitudes in non-AD (C4, SI Table 1; n = 16) and AD (A2; n = 13) oocytes. All oocytes were from the same frog. (B) Overall current ratios (percentage of AD-non-AD) from oocytes (n = 134, 11 frogs) recorded 1–4 days after injection with non-AD membranes (five subjects) or AD membranes (four subjects). For each AD oocyte, the currents were normalized to the mean non-AD currents obtained for each frog.

Some Properties of Glu and Kai Currents of AD Brain Receptors.

The current–voltage (I–V) relationships of Kai and Glu currents generated in AD and non-AD oocytes were fairly linear (Fig. 4). The Kai current reversal potential was close to −8 mV in non-AD oocytes and 0 mV in AD-injected oocytes, whereas in both cases, Glu currents reversed at a membrane potential close to 0 mV (Fig. 4). To determine the neurotransmitter sensitivity of the transplanted receptors, we applied different concentrations of Kai or Glu plus CTZ to the oocytes and obtained dose–response curves very similar for AD and non-AD oocytes (Fig. 5 A and C). For example, the EC₅₀ and Hill coefficient (n_h) for Kai were, respectively, 70 μM and 1.4 for non-AD, whereas for AD oocytes, they were 60 μM and 1.6, and the differences were not significant (t test). The same was true for the dose–response relationships of the receptors activated by Glu in the presence of 100 μM CTZ. In this case, the EC₅₀ and n_h were 50 μM and 0.9 for non-AD and 60 μM and 0.8 for the AD oocytes.

Fig. 4.

Kai and Glu + CTZ I–V relationships in non-AD and AD injected oocytes. (A) I–V relationships of Kai (100 μM) currents, normalized to I_max at −100 mV (I_max = −63.9 ± 9.2 nA in non-AD oocytes and 23.4 ± 1.5 nA in AD oocytes. (B) I–V relationships of Glu (1 mM) plus CTZ (100 μM) currents from non-AD and AD oocytes normalized to I_max: −67.3 ± 15 nA (non-AD) and −25.3 ± 9.8 nA (AD) at −100 mV (one frog). Each relation is the mean of three oocytes and membranes from one subject.

Fig. 5.

Properties of Kai and Glu + CTZ currents in non-AD and AD injected oocytes.(A) Kai dose–current response relationships, obtained from 12 non-AD oocytes and 11 AD oocytes (three AD and three non-AD subjects, three frogs). (B) Inhibition of Kai currents by CNQX, obtained by coapplying Kai (100 μM) plus CNQX in non-AD oocytes (n = 6, two subjects) and AD oocytes (n = 6, two subjects). Data were from two frogs, and all values were normalized to the maximum. IC₅₀ and n_h in non-AD oocytes were 1.8 μM and 1, whereas in AD oocytes, they were 1.9 μM and 0.9. The differences were not significant. (C) Glu dose–current response relationships (n = 13, three non-AD subjects, three frogs; and n = 14, three AD subjects and three frogs). (D) Inhibition of Glu currents by CNQX. In non-AD oocytes, IC₅₀ and n_h were 0.98 μM and 0.8 (n = 8, two non-AD subjects), whereas in AD-injected oocytes, IC₅₀ and n_h were 0.7 μM and 0.7 (n = 8, two AD subjects and two frogs). The differences were not significant.

To begin to study the effects of GluR antagonists on receptors from the human brain, we examined the effect of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX), a potent AMPA/Kai receptor antagonist. For that purpose, different concentrations of CNQX were applied together with Kai (100 μM) or Glu (1 mM) plus CTZ (100 μM) to obtain dose–inhibition curves. As illustrated in Fig. 5 B and D, CNQX was equipotent in inhibiting the currents generated by AD and non-AD receptors, and the curves yielded very similar IC₅₀ and n_h values.

WB and qRT-PCR.

Because the amplitudes of the GABA, Kai, and Glu currents generated by AD oocytes were smaller than those elicited in non-AD oocytes, we used WB and qRT-PCR analyses of the same tissue samples to begin to determine whether the expression of some receptor subunits is specifically affected. WBs, obtained with an anti-GluR2/3 and with anti-GABA α1 subunit antibodies, showed that the quantity of those proteins was always smaller in the AD membranes (e.g., Fig. 6), and similar results were confirmed by qRT-PCR. Namely, the levels of mRNAs encoding the analyzed GABA and Glu subunits were significantly lower in AD compared with non-AD brains (Fig. 7). Interestingly, the GluR5 subunit was found to be low by qRT-PCR in both non-AD and AD. Accordingly, that subunit was not detected by WB in the non-AD and AD membranes tested (data not shown).

Fig. 6.

Glu and GABA subunit proteins in non-AD (C) and AD (A) membranes (three subjects each) detected with an anti-GluR2/3 and anti-GABA α1 antibodies.

Fig. 7.

GluR and GABA receptor subunit mRNAs were reduced in the AD brain. Columns represent the mean change in gene expression of three AD brains (A2, A6, and A7; SI Table 1) relative to gene expression of a non-AD brain (C7), shown as percentage.

Discussion

The properties of neurotransmitter receptors microtransplanted from fresh human brain to frog oocytes are similar to those of native receptors present in brain slices (19, 20). Remarkably, we find now that GABA and GluRs from postmortem brains kept frozen for >10 years are still functional and behave like the native receptors (see also ref. 21). Moreover, the same membrane preparation can be stored frozen for many months, thawed several times, and injected into oocytes (for review, see ref. 22). Therefore, using this comparatively simple methodology, it is now possible to study directly the electrophysiological characteristics and pharmacological properties of human receptors associated with many neurological diseases.

Although GluRs have been frequently implicated in AD (4, 5, 8, 12, 13), very little was known about the functional properties of GluRs in the normal and diseased human brains. Therefore, we focused our attention on GluRs, in particular nonNMDA receptors, microtransplanted from AD or non-AD brains to frog oocytes. So far, the main functional characteristics of the receptors, such as agonist dose–current response and I–V relationships, were essentially the same for AD and non-AD receptors. The same was true for the antagonistic effects of CNQX. Studies of more brains and more membrane preparations are needed to determine whether small differences observed, like that of the IC₅₀ for inhibition of Glu currents by CNQX (0.7 vs. 1 μM) are really significant.

One very significant difference between AD and non-AD oocytes was the smaller amplitude of Glu currents in AD oocytes, similar to the smaller GABA currents observed (21). Such decreases could result from many factors, e.g., (i) an impaired receptor–channel function, (ii) a decreased number of receptors in the membranes of the AD brain, or (iii) a lower ability of AD membranes to incorporate into the oocyte membrane. To begin to examine these possibilities, we compared the mRNA and protein expression of some GluR and GABA receptor subunits in the same membrane preparations used for the electrophysiological experiments. Because the responses to NMDA were rather small, we studied some nonNMDA receptors. So far, of the many subunits that can make up these receptors, we have quantified the mRNAs coding for the GluR2 and GluR3 AMPA receptor subunits and the GluR5 Kai subunit, together with six GABA_A receptor subunits. All of them occurred in smaller amounts in the AD brain (Fig. 7 and refs. 15–17, 25, and 26). Because the GluR5 protein or mRNA was not clearly detected, this suggested that the Kai currents observed in non-AD and AD oocytes are generated mainly by receptors made up of other Kai receptor subunits.

Interestingly, the GABA_A subunit mRNAs were all decreased in the temporal cortex AD tissues that we examined. This reduction applied also to the α1 subunit, which was previously found to be decreased in the AD hippocampus without a reduction of the protein level (27). In contrast, in our experiments with cerebral cortex membranes, the GABA α1 protein was also decreased. Because the GluR2/3 subunit protein content was also decreased, we conclude that the reduced Glu and GABA currents in AD oocytes result from a diminished number of corresponding receptors in the membranes of AD and not to a diminished incorporation of the AD membranes.

Thus, using the membrane transplantation method, it is now possible to study the functional and structural characteristics of receptors and ion channels of the human postmortem brain and also the effects of some of the substances used to alleviate brain diseases. The information thus obtained may help us to understand better the etiology of the diseases and to develop new medications for their treatment.

Materials and Methods

Microtransplantation of Brain Membranes into Xenopus Oocytes.

Portions of the temporal cerebral cortex from the postmortem brains of 8 AD and 11 age-matched control non-AD subjects (SI Table 1) were used to prepare cell membranes according to procedures previously described (19, 22). The membranes were suspended in water and kept frozen at −80°C until used for an experiment. Usually, one AD and one non-AD tissue sample were processed simultaneously, and the corresponding membranes were injected into oocytes of the same batch to obviate frog donor variations. Xenopus laevis oocytes were dissected from segments of ovary, defolliculated with collagenase (Type I; Sigma, St. Louis, MO), and maintained at 16°C in Barth's solution containing 100 units/ml penicillin/streptomycin. The next day, each oocyte was injected with membranes at a protein concentration of 1–2 mg/ml (cf. refs. 19 and 22 for more detail).

Electrophysiology.

One to 10 days after injection, Glu and GABA currents were recorded from voltage-clamped oocytes using two microelectrodes filled with 3 M KCl (28) and with the membrane potential held at −80 mV. The oocytes were superfused continuously with Ringer's solution (115 mM NaCl/2 mM KCl/1.8 mM CaCl₂/5 mM Hepes, adjusted to pH 7 with NaOH) at room temperature. All of the drugs (Sigma or Tocris Cookson, Bristol, U.K.) were applied by using a perfusion system (Warner Instruments, Hamden, CT). For agonist dose–current response curves, the substances were applied at 5-min intervals, and the EC₅₀ and IC₅₀ for the antagonist CNQX plus n_h were estimated by fitting the data to Hill equations (29). For recording Glu currents, Glu (1 mM) was applied for 30 s before coapplying Glu (1 mM) plus CTZ (100 μM). For the I–V relationships, the currents were normalized to the one obtained at −100 mV, and the drugs were applied at membrane potentials from −100 to 20 mV. All values are expressed as means ± SEM.

WB.

Identification of AMPA and Kai subunits was done by WB. As antibodies, we used an anti-GluR2/3 (Chemicon, Hampshire, U.K.) for AMPA and an anti-GluR5 (Sigma) for Kai receptors. For GABA receptors, we used an anti-GABA_A α1 subunit (Sigma). Membrane preparations (40 μg) were loaded onto an 8% SDS polyacrylamide gel (Pierce, Rockford, IL) and separated electrophoretically. Proteins were transferred to a nitrocellulose membrane, and unspecific binding was blocked by using Tris-buffered saline (TBS) and BSA 5% for 1 h at room temperature. The membranes were incubated for 90 min with the first antibody diluted 1:200, then washed for 45 min with 0.05% Tween 20 in TBS; the solution was changed every 15 min. Membranes were then incubated in TBS with the secondary antibody (anti-rabbit, conjugated with alkaline phosphatase; Sigma) diluted 1:400 for 1 h at room temperature. The blots were developed with SigmaFast BCIP/NBT (Sigma).

qRT-PCR.

The expression of GluR and GABA receptor genes in the brain samples was evaluated by qRT-PCR. The AMPA receptor subunits GluR2 and GluR3, the Kai receptor subunit GluR5, and the GABA subunits α1, α2, α5, γ2, β2, and β3 were chosen, and GAPDH was used as internal control. The primers were designed to meet the basic requirements for real-time PCR and subjected to BLAST (30) analysis to ensure that they did not prime to other human genes.

Total RNA was extracted from the same tissues used to prepare the membranes by using a method described elsewhere (31). mRNA was then isolated by using the Qiagen (Valencia, CA) oligotex kit and contaminating DNA eliminated by DNase digestion. Two micrograms of mRNA was used as template to obtain cDNA with Bio-Rad (Hercules, CA) iScript and then used to perform the qRT-PCR in triplicate with the SYBR green mix from Bio-Rad and an iCycler real-time PCR machine. For data analysis, threshold cycles obtained in the qRT-PCR were analyzed with the 2^−ΔΔCt method (32), which can be used to calculate the fold change in gene expression normalized to an endogenous reference gene (GAPDH) and relative to a control (non-AD brain).

Abbreviations

AD

Alzheimer's disease

AD oocytes

oocytes injected with membranes from an AD brain

Glu

glutamate

GluR

Glu receptor

Glu current

membrane current elicited by Glu

Kai

kainate

CTZ

cyclothiazide

WB

Western blot

qRT-PCR

quantitative real-time PCR

CNQX

6-cyano-7-nitroquinoxaline-2,3-dione

I–V

current–voltage.