Background
Copper-, iron-, and zinc-ions are essential in human biochemical function. Their concentration in vivo is under tight control, and a dyshomeostasis may cause electrical imbalance and consecutively, region-selective neurodegeneration eventually facilitating cognitive deficits. In a systematic quantification of all three biometals in the human brain, age-associated changes in the elderly population were quantified in postmortem neocortical tissue. While Zn
2+ was unaffected in any disease pathologies of the brain, both Cu
2+ and Fe
3+ showed a gradual age-associated decline in healthy non-cognitively impaired individuals. Further, Cu
2+ was significantly reduced by 20%, and Fe
2+ significantly increased by 10–16% in severe Alzheimer disease (AD) compared with age-matched controls [
1].
Various ion channels were reported to be highly sensitive towards Zn
2+ and Cu
2+ at concentrations below 10 μM (summarized in [
2,
3]). Among them are glutamate-, glycine-, GABA-A-, acetylcholine- and P2X- receptors as well as voltage-gated Na
+ (Na
v1.5), K
+ (K2P, Kv1.3, mSlo1), and Ca
2+ (Ca
v3.2) channels [
4‐
6].
Polyvalent inorganic cations can interfere with the function of VGCCs through various mechanisms, which include electrostatic effects, pore block, and in some cases allosteric effects [
3]. Besides, endogenous Zn
2+ and Cu
2+ are increasingly recognized to be involved in central neurotransmission, although their exact role remains ambiguous. Both can be released spontaneously or in an activation-dependent manner from defined populations of neurons located primarily in limbic regions (hippocampus, amygdala) in the cerebral cortex and in the retina [
7] (for a summary see [
8]).
The isolated and superfused vertebrate retina represents an important model of a neuronal network for the functional investigation of postsynaptic excitation conduction [
9‐
11]. Using the bovine retina, the process of reciprocal inhibition during transretinal signaling was investigated in detail [
8,
12,
13]. Cu
2+ occupy an allosteric binding site on the domain I gating module of Ca
v2.3 channels and may interfere with voltage-dependent-gating. Several studies have shown that even very nominal Cu
2+ concentrations are sufficient for significant suppression of these channels, as electroretinographically demonstrated on the ex-vivo bovine retina [
14]. It was found that L-Glutamate, as excitatory amino acids, can stimulate Ca
v2.3-channels by being as trace metal chelators and inhibiting the suppressing effects of metal ions such as Zn
2+ and Cu
2+ [
2]. The naturally occurring excitatory amino acid Kainic acid has an embedded L-glutamic acid unit, and thus, a potent agonist at (non-NMDA) glutamate receptors. The systemic application produces epilepsy in rodent experiments [
15] and in this context, Ca
v2.3 channels can act to the pathogenesis of KA-induced seizures [
16,
17]. However, the structural similarity of Kainat to L-Glutamate led us to investigate the chelating effect in the presence of physiologically concentrations of Cu
2+.
To understand molecular mechanisms of the biometals, this technique of the isolated and superfused retina was successfully transferred to the isolated mouse retina from control and Ca
v2.3-deficient mice [
18,
19]. Those recordings of a full electroretinogram (ERG) from the isolated and superfused murine retina were optimized only recently [
11] by modifying the composition of the standard Ames solution [
20]. After gentle isolation of the retina, the development of the b-wave, indicative for transretinal signaling, was stabilized by adding barium chloride (0.1 mM) to the final perfusion solution, similar as it was reported originally for eyecup preparations of the tiger salamander [
21].
In the present report, recordings of the full ERG were possible and were compared between the Cav2.3-competent and the –deficient mouse retina to understand if Cav2.3 is involved in transretinal signaling as supposed from recordings in the bovine retina.
Methods
Materials
Glucose and the constituents of the nutrient solution used for retinal superfusion were purchased from Merck (p.a. grade). For the preparation of the murine retina [
11] a superfine scissor (WPI, Nr. 501,839), and ultrafine suturing forceps (WPI, Nr. 555063FT) were used. Further, a 27-gauge needle (Sterican, size 20: 0.4 mm × 20 mm Bl/LB) was used to punch a hole into the cornea of the extirpated eye bulb. The receptor antagonists UBP 301 and CNQX were purchased from Sigma Aldrich (Munic, Germany).
Animals
In order to compare ERG responses from mice deficient of the voltage-gated Ca2+-channel Cav2.3 (R-type), we used control mice with an identical genetic background. Both mouse lines were generated and bred in our animal facility. Meanwhile, they can also be ordered from MMRRC (ID 50523).
Ca
v2.3-deficient animals and control mice were kept as separate mouse lines derived from heterozygous parents (fourth backcrossing into C57Bl/6). Homozygous littermates were regularly interbred with each other and back-bred into C57Bl/6 (for further information on Ca
v2.3-deficient generation see [
22,
23]). In short, the cacna1e gene encoding Ca
v2.3 was disrupted in vivo by agouti-colored Ca
v2.3(fl|+) and deleter mice expressing Cre-recombinase constitutively [
24]. Thus, exon 2 was ablated by Cre-mediated recombination. Ca
v2.3-deficient mice were fertile, exhibited no obvious behavioral abnormalities and had the same lifespan as control mice. The Cav2.3-deficient mouse line, which was generated in the Cologne lab was transferred to the Mutant Mouse Resource & Research Centers (MMRRC) with the strain name B6J.129P2(Cg)-Cacna1e
tm1.1Tsch/Mmjax.
Adult male mice were used at the age of 12 to 18 month and kept at 20 to 22 °C in makrolon type II cages under a 12 h light-dark cycle (7:00 a.m./p.m.) with food and water provided ad libitum. All animal experiments were in line with the European Communities Council Directive 2010/63/EU for the care and use of laboratory animals as described in the UFAW handbook on the care and management of laboratory animals. All experiments were approved by the local institutional committee on animal care (UniKöln_Anzeige§4.17.007).
Methods
In order to reach maximum transretinal signaling realized by a full ERG, the bovine [
10] and the murine retina [
11] had to be superfused under different conditions (Table
1).
Table 1Comparison of nutrient solutions used for ERG recordings from bovine (Sickel-medium) or murine isolated retina (modified AMES-medium). AMES-medium was modified by increasing the pH slightly to alkaline (pH 7.7) and by adding 0.1 mM BaCl2
NaCl | 120 mM | 120 mM |
KCl | 2 mM | 3.1 mM |
CaCl2 | 0.15 mM | 1.15 mM |
MgCl2/MgSO4 * 7 H2O | 0.1 mM | – |
D-glucose | 5 mM | 6 mM |
MgSO4 | – | 1.24 mM |
KH2PO4 | – | 0.5 mM |
NaH2PO4 | 1.5 mM | – |
Na2HPO4 | 13.5 mM | – |
NaHCO3 | – | 22.6 mM (pH 7.4) |
45.3 mM (pH 7.7) |
L-glutamine, amino acids, vitamins and other components | – | Sigma-Aldrich Nr. A1420 |
Equilibration | Pure oxygen | Carbogen gasing containing 5% (v/v) of CO2 in O2 |
pH | 7.8 | 7.4–7.7 |
BaCl2 | – | 0.1 mM |
Bovine eyes were received from a local slaughterhouse and immediately stored in a nutrient solution (Sickel medium), which was aerated with pure oxygen, consisting of the following (in mM): NaCl (120), KCl (2.0), CaCl
2 (0.15), MgCl
2 (0.1), NaH
2PO
4 (1.5), Na
2HPO
4 (13.5) and glucose (5) with a final pH of 7.8. The bovine retina was isolated as described in detail by Luke et al. 2005 [
10]. After isolation, a plain retina segment was transferred into the recording chamber described below, which is placed in an electrically and optically isolated air thermostat. From the dark-adapted retina and in response to a single white flash, electroretinograms were recorded at intervals of 5 min at 30 °C, with a constant superfusion at 1 ml/min controlled by a roll pump. The flash intensity was set to 6.3 mlux, with the duration of light stimulation at 500 ms [
10].
After reaching a stable equilibrium of the light-evoked ERG responses, 5 mM tricine (tricine, Sigma # RES3077T-A701X) was added to the nutrient solution with 10 mM (HEPES) (Carl Roth, p.A., # H3375) and superfused for 30 min. Washout was started thereafter with Sickel medium.
Murine retinas were isolated from mice of our animal facility department, in which the light–dark regime was 12:12 h, and the light intensity between 5 and 10 lx at the surface of the animal cages.
DNA-containing tissue samples were collected from tail biopsies. DNA was extracted and used as template for genotyping. Transcripts of Ca
v2.3 were detected by RT-PCR (RT, reverse transcriptase) using primers, which flanked the deleted exon 2 and exon 3 region [
22]. In short, contaminating protein and RNA were enzymatically digested by protease and RNAse, respectively. For the PCR amplification of indicative Ca
v2.3 DNA-fragments, about 1 μg DNA was introduced and amplified with the forward primer (B45Hilx1) 5′- AAA AAC AGC CGG GGA AAG CTT AT-3′ and the reverse primer (a1eb45r) 5′-CTG CCC TTT CTT CTT GCC TGA C-3′. The sizes of DNA fragments expected are 1056 bp for the WT and 86 bp for the Ca
v2.3-KOs. PCRs for the genotyping were performed using a DNAEngine Peltier thermal cycler (BioRad, Germany) or a PTC-200 Peltier thermal cycler (MJ Research, Biozym Diagnostik, Germany) with the initial denaturation (94 °C for 10 min) followed by 34 cycles (denaturation at 94 °C for 60 s, annealing at 60 °C for 90 s, extension at 72 °C 4 min) and final extension at 72 °C for 10 min. The PCR products were separated by agarose gel electrophoresis and fluorescent bands were detected on a Herolab UVT-28 M transilluminator by UV irradiation (312 nm excitation wave length).
For Cav2.3, mouse lines were used as separate inbred strains for Cav2.3(+|+) and Cav2.3(−|-), each after the fourth backcrossing in C57Bl / 6 mice.
The mice used for the retina isolation were dark adapted overnight, sacrificed by cervical dislocation under dim red light, without the need for an anesthetic, and the eyes were extirpated immediately. Enucleated eyes were protected from light and transferred into carbogen (95% O
2 / 5% CO
2)-saturated modified Ames medium respectively [
11]. The isolation of the murine retina was started immediately post mortem and carried out under dim red light. The complete retina was transferred to the recording chamber [
18] and the electroretinogram was recorded via two silver/silver-chloride electrodes on either side of the isolated retina. The recording chamber containing the retina was placed in an electrically and optically isolated air thermostat. From the dark-adapted retina responses to a single white flash were recorded at intervals of 3 min at 27.5 °C and with a constant superfusion at 2 ml/min controlled by a roller pump. The duration of light stimulation was 500 ms, controlled by a timer operating a mechanical shutter system. The pre-stimulus delay was 380 ms. Unless noted otherwise, the flash intensity was set to 63 mlux at the retinal surface using calibrated neutral density filters.
As soon as the isolated retina was placed into the recording chamber, it was equilibrated for about 60 min in modified Ames-solution (
n = 20 ERGs) (Table
1). After reaching a stable equilibrium of the light-evoked ERG responses, 100 nM CuCl
2 was added to the modified Ames-medium and superfused for 30 min (
n = 10 ERGs). Washout was started thereafter with Cu
2+-free modified Ames-medium. Each ERG response contains 239 data points.
The ERG was amplified and bandpass-limited between 0.3 and 300 Hz (PowerLab 8/35; Animal Bio Amp FE136, ADIntruments, Oxford, UK). Light flash, heating unit, fan and roller pump were automatically controlled by National-Instruments (BNC-2120; DASY-Lab V8.0). For each experiment, a new retina was transferred to the recording chamber. The retina was superfused with nutrient solution and stimulated repetitively until the responses had reached a stable level (usually after 60 min of perfusion). Switching from one solution to another was performed with a three-way valve to prevent disturbance of the experimental conditions. Experimental protocols for the isolation, storage and incubation of the vertebrate retina can markedly alter phototransduction and transretinal signaling [
25]. To quantify such changes, we evaluated the a- and b-wave amplitudes and their implicit times.
Data analysis
The b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave. After reaching the equilibrium, the initial b-wave amplitude was set to 100% (= Pre) resulting preferably illustration of ERGs after treatment with drugs as well as after washout. Quantitative normally distributed data are presented as mean ± standard error of mean (SEM) and as percentage. Nonparametric tests are demonstrated as median [1. quartile – 3. quartile]. Two-sided, paired Student t-test was used for comparison of quantitative parameters in case of normal distribution. If not applicable, Wilcoxon test was used instead.
All analyses were performed with IBM® SPSS® Statistics V22.0 (IBM, Chicago, Illinois, USA).
Discussion
During transretinal signaling in the vertebrate retina, VGCCs are involved in basal synaptic signal transduction by triggering Ca
2+-mediated glutamate release. In the dark, the excitatory neurotransmitter glutamate is tonically released from ribbon synapses and the rate of release is modulated in response to graded changes in the membrane potential contrasting with action potential–driven bursts of release at conventional synapses [
27]. Mainly the two L-type VGCC containing the ion conducting subunits Ca
v1.3 (α1D) and Ca
v1.4 (α1F) appear to form the principle voltage-gated Ca
2+ influx pathways in rods and cones [
28], but also in the cochlear inner hair cells [
29]. The importance of both Ca
v1 L-type Ca
2+ channels during sensory signaling gets obvious after its ablation, either by recombinant technologies for Ca
v1.3, leading to deafness [
30] or after native gene loss for Ca
v1.4 leading to night blindness [
31].
Transsynaptic signaling is modulated by GABA and glycine, and by additional neurotransmitters and other signaling molecules like nitric oxide (NO), acetylcholine and dopamine. So far, these synaptic modulation and integration has not yet been fully understood. For the bovine retina, it was postulated that the R-type Ca
2+ channel may be involved in this reciprocal inhibition by triggering the release of inhibitory neurotransmitters [
13]. As previously demonstrated in our experiments on the murine retina, such an involvement was deduced by recording and calculating the amplitude changes caused by Ni
2+ application, a rather complicated indirect determination [
19]. The inhibitory reciprocal modulation by GABA was related to the expression of both, Ca
v2.3 / pharmacoresistant R-type and Ca
v3.2 / T-type voltage-gated Ca
2+ channels [
19], which are both highly sensitive towards toxic Ni
2+.
The ten different genes encoding the ion conducting subunits show different sensitivities towards bioavailable trace metal cations Zn
2+ and Cu
2+ [
3]. Both, Ca
v2.3- and Ca
v3.2-channels are most sensitive towards Zn
2+ and Cu
2+ [
2,
32]. Therefore, the sensitivity towards Cu
2+, the most effective divalent metal cation inhibitor, was investigated first in the isolated and superfused bovine retina [
14] and now in the retina of wild type and genetically modified mice lacking the expression of Ca
v2.3. The present report provides further support to the idea that the assumed R-type Ca
2+ channel containing Ca
v2.3 as ion conducting pore is involved in transretinal signaling and modulated by submicromolar Cu
2+ concentrations. This concentration mimics the in vivo situation and indicates that Cu
2+ and probably also other bioavailable trace metal cations could help to modulate the strength of signal propagated through the retinal network in vivo.
Major differences between the Cu
2+ effects on transretinal signaling, which will be discussed, are related to a genotype-related difference. The genotype-dependent stimulation in Ca
v2.3-competent mice shows that submicromolar Cu
2+ mediates its effect clearly via Ca
v2.3, the pharmacoresistant Ca
2+ channel, which, in the bovine retina, was thought to trigger GABA-release from amacrine cells onto ON-bipolar neurons. This release of inhibitory neurotransmitters was deduced from the sensitivity of the b-wave to stimulation by Ni
2+ [
12,
13], Zn
2+ [
8] and Cu
2+ [
14]. For the isolated murine retina expressing Ca
v2.3, the same concentration of CuCl
2 (100 nM) caused a similar effect, namely a significant 30% increase of the ERG b-wave amplitude, which was not observed in the retinae from Ca
v2.3-deficient mice.
It should be noted that the bovine retina [
10] was isolated and recorded in a different medium than the murine retina [
11], although in both cases, the solutions were optimized to achieve a maximal b-wave response, indicative of a “healthy” retina. The major difference between both nutrient solutions is related to the absence (bovine) or presence (murine) of an amino acid cocktail including glutamate, which can chelate submicromolar concentrations of divalent trace metal cations. Therefore, control recordings by chelating Cu
2+ after application of tricine failed for the murine retina (present report, Fig.
1) but not for the bovine retina [
14]. Attempts to record from the murine retina under conditions previously used for the bovine retina (i.e. with Sickel instead of Ames medium) failed, possibly reflecting the higher metabolic needs of the murine retina.
Another set of experiments was related to the effects of kainate, which is routinely used to induce experimental seizures in mice. As this drug is also able to chelate trace metal cations, we tested its effects on transretinal signal transduction in the isolated retina. Kainic acid (KA) probably affects more efficiently the high-affinity binding sites of the classical non-N-methyl-D-aspartate (non-NMDA) ionotropic glutamate receptors, which are expressed in the murine retina [
33,
34]. Under the conditions for recording murine ERGs in our setup, KA did not have a prominent and visible effect as a chelating agent, since no reduction of the ERG b-wave amplitude was observed. However, we failed to find any effect on the Cu
2+ mediated stimulation, because endogenous kainate receptors could not be blocked completely by KA-receptor antagonists or the antagonists themselves showed pronounced effects on transretinal signal transduction (Fig.
5).
In summary, regardless of the exact mechanisms of Cu
2+ effects, our findings show that genetic inactivation of Ca
v2.3 channels is sufficient to prevent Cu
2+-mediated stimulation on the b-wave amplitude. The results illustrate that Cu
2+-induced responses in Ca
v2.3 channel function can lead to changes in the normal neurotransmission. Based on our results using the isolated retina model and previous works [
35,
36], an integrative model of Cu
2+-depended cellular responses is presented. Therefore, the involvement of VGCCs could reflect a potential target for interventions of diseases as reported by several researchers [
4,
37,
38]. Further investigation will be necessary to establish the relevance of trace metal ions on VGCCs.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.