Development of b-wave amplitude after a 4 hour preincubation period for Cav2.3-competent mice by superfusing nutrient solution (= before NiCl2), 30 µM NiCl2 for 30 min (= after NiCl2), and by washing out the NiCl2 (= washout). Note, such a reversible increase of the b-wave amplitude, as shown for 9 experiments, was not always observed. In nearly the same number of experiments, no significant increase was detected, which may be attributed to the 4 hour preincubation period, during which reciprocal signalling was partially reduced or completely lost for unknown reasons.
Vertebrate retina has routinely been isolated from bovine [1-4] or rat eyes . However, many animal models for disturbances in neuronal signalling rely on gene inactivation or transgenes in mice. To get access to the isolated retina from these model systems to investigate neuronal processing, the technique of retina isolation and recording in a closed chamber was transferred from the bovine to the murine eye.
The ERG recordings of the isolated and superfused retina from bovine eye were most successfully performed in a closed recording chamber, in which a defined and controlled oxygen concentration can be achieved. Under these conditions, repetitive ERG recordings (interval time 5 min) can be performed up to 10 hours . For rat and murine ERG recordings an open system was introduced recently, in which oxygen may escape easier, and stable recordings were shortened to about 90 min . Therefore, we tried to optimize murine ERG recordings in the same closed system as it was introduced for frog and bovine retina [7;8].
2. Optimization of murine ERG recordings and applications
In the first section the history to optimize murine ERG recordings will be described for the last six years, in which initially an incomplete ERG was recorded, which was mainly dominated by the a-wave. After changes for the retina isolation as well as for the recording conditions, a full ERG was obtained, which even could be subdivided for the b-wave response into at least two different fractions .
Finally, two important topics will be covered for successful ERG recordings.
Using still the bovine retina, it will be analysed that a trace metal free solution can be achieved by using a tricine-based buffer system, which chelates heavy metal divalent cations selectively. Such a buffering may be very helpful for ERG recordings in the presence of submicromolar copper ion concentrations, which may be as important as investigations with the artificial Ni2+-ions [3;10] or the native occurring Zn2+-ions .
Flashes of high light intensity will be used to trigger repetitive ERG responses and to test the stability of the optimized murine ERG recording conditions.
2.1. Development of methods to record full murine ERGs for basal and reciprocal signalling
The diameter of a murine eye (male, 10 – 15 weeks of age) is about 10-fold smaller than the diameter of the bovine eye (age of 9 – 12 month). Therefore, an optimal isolation procedure is critical for subsequent successful and stable ERG recordings from the isolated murine retina.
In the following three chapters, we introduce three different procedures, which were used in our laboratory to realize and to improve the reproducible ERG recording from the murine retina.
The first procedure, the retina isolation by a rotating razor blade, was used to mimic the way of working as done with the bovine retina. Instead of using a knife to open up the eye bulb, we added a razor blade to a rotator, which cut the murine eye bulb laying in the stator very harsh, as we noted later. Under these conditions, the ERG recordings were dominated by the a-wave.
The second and the third procedure used a more gentle way to open up the eye bulb. The access to the retina was achieved by opening up the eye bulb at the level of the cornea. Consecutively, the lens was removed, and the eye cup was gently cut apart. The second and the third procedure differ with respect to the waiting time, after which ERG recordings were started. Initially (second procedure), we realized that a full ERG was better achieved by incubating the retina in nutrient solution for four hours in darkness. One may assume that it recovered from the ischemic conditions during the isolation procedure. Although basic signalling recovered during these 4 hours, reciprocal signalling obviously vanished. Therefore, we omitted preincubation, and started immediately (third procedure).
The main focus of our research was directed to understand the mechanisms of reciprocal signalling via voltage-gated Ca2+ channels, sensitive to heavy metal cations (Ni2+, Zn2+, Cu2+). It was started first by using NiCl2, an often used non-selective Ca2+-channel blocker with some preference for Cav2.3/R-type and for Cav3.2/T-type Ca2+ channels, which have IC50 values at 10 µM. Therefore, an optimal recording of only basal retinal signal transduction was not sufficient, and we had to pay attention to more favourable conditions for ERG recording, preserving also signalling via feedback from amacrine cells to bipolar neurons.
2.1.1. Retina isolation by a rotating razor blade
Initial preparations of the murine retina followed the idea that the eye bulb should be opened up in a similar way as for the bovine retina by cutting the eye bulb at the ora serrata region. Therefore, a rotating razor blade was used, which led fast and reproducibly to a nearly full sized retina, isolated from the proximal cup of the remaining eye bulb. Under these conditions, it was possible to record photoreceptor responses reliably for at least 90 min . Interestingly, from the incomplete ERG, which was characterized by a large a-wave (Fig. 1A – 1C), the photoreceptor response was sensitive towards low concentrations of dihydropyridines (25 nM racemic isradipine). Under similar recording conditions, the bovine photoreceptor response was insensitive even towards 2 µM isradipine .
Detailed analysis of the mechanism for dihydropyridine sensitivity of the murine photoreceptor response (a-wave) led to the detection that preincubation of the murine retina in 1 – 10 mM D-aspartate, a well known antagonist of transsynaptic signalling , eliminated the sensitivity towards this classical L-type Ca2+ channel antagonist. Similarly, also the sensitivity of the a-wave towards the T-type Ca2+ channel antagonist mibefradil (1 - 10 µM) was reversibly occluded by D-aspartate preincubation leading to the conclusion that minor transsynaptic signal transduction was present before aspartate application . Such a very minor b-wave component could be deduced from our recordings by subtracting the mean values of murine ERGs in the presence of aspartate (1 mM) from those ERGs initially recorded before applying aspartate (Fig. 1D).
Overall, under the conditions of razor-blade isolated retina it was impossible to record a full ERG with a “healthy” b-wave component, as it was routinely observed for the isolated bovine retina in the same setup system. Therefore, the isolation protocol was changed completely.
2.1.2. Retina isolation under more gentle conditions followed by a preincubation period
The isolation of the murine retina was changed in a way that mechanical stress was reduced. In a well documented approach, the murine retina was isolated by opening up the eye bulb from the anterior side by dissecting the cornea, and by explanting the eye lens first . We assume that the rotating razor blade in the former procedure may have introduced too much mechanical tension on the retina leading to a trauma and a loss of full transretinal signalling capacity. The new isolation procedure can be followed up in a detailed animation, which is available in the internet as a supplement . In short, enucleated murine eyes (Fig. 2)
were placed into nutrient solution, fixed by a lancette and opened from the front side first by punching a tiny hole through the cornea to relieve the aqueous humour from the anterior and posterior chamber of the eye, and thus, to reduce the tension (Fig. 2A). Thereafter, the cornea was removed, and the iris disconnected, and a tiny triangle shaped incision was made into the sclera (Fig. 2B) for an easier separation later of the proximal eye cup, and for the immediate introduction of a small forceps. Finally, the lens was carefully removed (Fig. 2C). Consecutively, the opened eye-cup was held and separated apart by a forceps in order to detach the retina from the pigment epithelium by repetitive moving in the nutrient solution (Fig. 2D). The successive separation of the retina from the pigment epithelium has to be performed gently, and is supported by the cutting of the sclera layers from the outside. Finally, the isolated retina may be divided into 4 - 6 segments.
One of the posterior retina segments was finally mounted on a plastic mesh occupying the centre of the perfusing chamber by using a supporting tool (Fig. 2E). The ring, fixed by the submerged lever tool, is loaded with a retina segment, which was stored floating in the oxygenized nutrient solution. It is touched with the forceps at the region where formally the fibres of the N. opticus had left the eye cup. The retina segment adheres by itself and the supporting lever is turned upside down to move the retina loaded ring into the holder of the recording chamber (Fig. 2F). The view from the underlying nylon mesh further visualizes one of the two silver/silver-chloride electrodes, which are on both sides of the recording chamber. After closing the chamber by two gaskets and two glasses, the nutrient solution is perfused over the front and the rear side (Fig. 2F) of the recording chamber.
Based on the more gentle isolation procedure, full ERG recordings could be achieved, and optimal conditions for the recordings for mice were summarized and published . For example, light intensity for scotopic ERG recordings was as low as 6.3 mlx for the bovine retina but 10-fold higher for the murine retina.
ERG recordings even in the presence of low NiCl2 concentrations (up to 50 µM) could be easily followed up for several hours without loosing a full b-wave response (Fig. 3B).
Usually, the murine retina needed a while before a full ERG was seen. Early ERGs (Fig. 3A) were characterized a dominant a-wave, and the b-wave developed during the next 30 min . The superimposed traces in Fig. 3B reveal that over time (60 – 215 min) the amplitude did not decrease much, but the implicit time got shorter and the shape of the b-wave changed indicating that the transretinal signalling varied (compare to later results with the different mouse mutants, especially the ERGs from Cav2.3-deficient mice, which are lacking the later fraction of the b-wave ).
The major goal for murine ERG-recording was the demonstration of an increase of the b-wave amplitude after the superfusion of low NiCl2 (10 – 30 µM). Under these conditions (15 µM NiCl2) the ERG b-wave amplitude of the isolated bovine retina was increased up to 1.5-fold . For the bovine retina it was concluded that the pharmacoresistant Ca2+ channel Cav2.3 may trigger the GABA release necessary for reciprocal signalling backwards to the bipolar neurons . This conclusion should be validated or disproven in mouse models with selective Ca2+ channel inactivation, either lacking Cav2.3 or Cav3.2 or even both subunits of voltage-gated Ca2+ channels. Therefore, the murine retina had to be isolated in a way that not only the basal signal transduction but also the reciprocal signalling would be conserved over time. Under these conditions, retinas from Cav2.3-competent and Cav2.3-deficient mice were isolated and their sensitivity towards NiCl2 (15 and 30 µM) was analysed and compared between different genotypes . For the ERGs from Cav2.3-competent mice, the data for 30 µM NiCl2 are summarized (Tab. 1).
|Exp. #||Before NiCl2:|
under NiCl2 before NiCl2
under NiCl2 after Washout
As the full ERG-response with “healthy” b-waves was only achieved after a prolonged incubation period of 4 hours, we were afraid and realized that the capacity for maximal reciprocal signalling was partially or fully lost during such a preincubation. The consecutive large scatter for Ni2+-mediated stimulation of the b-wave amplitude led us to conclusion to modify the recording procedure for a more stable transretinal signalling.
2.1.3. Retina isolation under gentle conditions without a preincubation period
To overcome any variable loss of reciprocal signalling, the murine retina was isolated again by the more gentle procedure as outlined before . The adaptation period however was skipped to avoid any loss of reciprocal signalling. Early ERG recordings, which were still dominated by the photoreceptor response, were digitalized and stored, and mean values were calculated for the response before adding any NiCl2. Thereafter, NiCl2 superfusion (15 µM) was initiated within the first 15 – 20 min after introducing the retina segments into the recording chamber. The NiCl2 application routinely lasts for 30 min, a time, after which a new equilibrium was reached and stable full ERG responses were recorded. From those ERG traces a mean trace curve was calculated and the mean ERG trace without NiCl2 superfusion was subtracted from the mean trace under NiCl2. Such difference-ERG traces were normalized and compared between different genotypes .
To our surprise, even the photoreceptor response differed between the two mouse lines either lacking Cav3.2 (including the double knockout) and the Cav2.3-competent or -deficient mouse line. The larger hyperpolarization at the beginning of each experiment for Cav3.2-deficient mice or double knockout mice was the reason that the amplitude of the difference–ERG was much larger than in retinas from Cav2.3-competent or –deficient mice. Most interesting, however, was the result deduced from implicit times of the normalized difference-ERGs. Obviously, Cav2.3 / R-type and Cav3.2 / T-type Ca2+ channels may contribute to different fractions of the signal transduction in the retinal network. In Cav2.3-deficient mice, the implicit time for the apparent b-wave is significantly shorter (346 ± 12 ms; n = 8 retinas) than in Cav2.3-competent mice (474 ± 27 ms; n = 5 retinas; p < 0.001), which implies that the later b-wave fraction, which is missing in the Cav2.3-deficient mice, would be carried in Cav2.3-compentent mice by the pharmacoresistant R-type channel triggering synaptic neurotransmitter release. This interpretation is in line with the assumption  that Cav2.3 may act in the synapse of a third order neuron, e.g. after the bipolar – amacrine synapse, while Cav3.2 may rather be responsible for a T-type triggered neurotransmitter release in the bipolar – amacrine synapse. T-type Ca2+ currents were routinely identified in these bipolar neurons [15-17].
The inactivation of Ca2+ channel genes could also disturb the overall retinal organisation. However, at least the inactivation of Cav2.3 does not disturb macroscopic structures in the retina as evaluated by histochemical staining (Fig. 4). Visual inspection and quantification of the thickness for retinal layers confirmed that no massive loss of neurons has occurred, which may also be true for the other knockout models.
2.2. Tricine-containing nutrient solution to chelate traces of divalent heavy metals
The use of Ni2+ as Ca2+ channel antagonist for R- and T-type Ca2+-channels has a long tradition [18;19]. Only recently zinc and copper ions have been recognized as important endogenous modulators of neuronal excitability , and the most Ni2+-sensitive Ca2+ channels were also characterized for its capability to be blocked by these native heavy metal cations [21-24]. The disturbance of their homoeostasis may be responsible for several neurological diseases as for example Morbus Alzheimer [25;26]. Much more is known for the physiological role of Zn2+ than for Cu2+, which proves to be effective (submicromolar) even in lower concentrations than Zn2+.
The effect of ZnCl2 (10 – 15 µM) on transretinal signalling resembles the effect of NiCl2, which was investigated in detail for the isolated bovine retina .
The second most abundant trace element in the human body is represented by Zn2+. In total it reaches up to two grams in adult humans. Most Zn2+ in the body is located intracellularily. But as only a small labile portion in the liver may be available during Zn2+ deficiency, a regular dietary supply of Zn2+ during growth and aging is needed .
After Zn2+ deficiency, the multitude of physiological roles for Zn2+ are best seen, which include anorexia, retarded growth, weight loss, impaired immune function, delayed sexual maturation, testicular atrophy, epidermal hyperkeratinization, alopecia, hypogeuisa, and night blindness [27;28]. In ocular tissues, the concentration of Zn2+ is unusually high when compared to other tissues. Zn2+ was localized in retina, chorioid, ciliary body, iris, optic nerve, sclera, cornea, and lens [27;29].
As soon as the function of Cu2+ would also be investigated, contaminations of submicromolar heavy metal cations may count for the validity of experiments in the future. To test a well known buffer system, which can be used to calibrate the desired cation concentration, tricine was added in various amounts (15 mM, 10 mM, and 5 mM) to the normally used nutrient solution. The bovine retina was used in these experiments. While 15 and 10 mM inhibited the b-wave response, no substantial change was observed at 5 mM tricine. Sometimes we observed a slight decrease of the b-wave amplitude, which may be attributed to heavy metal contaminations in the nutrient solution. After a chelation of traces by tricine, the effect of NiCl2 (15 µM) was analysed leading to a similar and transient increase of the b-wave response (Fig. 5).
2.3. Stability testing for murine ERG responses under high intensity light conditions
So far, murine ERGs were recorded under low light intensity (63 mlx for 0.5 sec). But the testing of higher light intensities may be needed, because reciprocal signalling should be more pronounced at elevated light intensities, if the reciprocal signalling participates in light/dark adaptation. Therefore, we analysed the effect of increasing light intensities on the photoreceptor response of the isolated murine retina (Fig. 6).
During an experimental regime of early dose-response (Fig. 6A, 6B), light (Fig. 6C, 6D) and dark adaptation (Fig. 6E, 6F), and late dose-response (Fig. 6G, 6H), it gets obvious that the isolated murine retina yields rather stable photoreceptor responses, which are only slightly reduced by the slowly growing b-wave. A faster development of the b-wave response can be achieved by adding NiCl2. The superfusion of 15 µM NiCl2 leads within 30 min to a fast increase of the developing b-wave. The ERG traces before and after NiCl2 superfusion indicate that during this period b-wave gets maximally increased (Fig. 7), as reported recently .
To test the stability of the isolated murine retina for its transretinal signalling capability (Fig. 8), the b-wave amplitude was recorded and used as the readout during similar experiments as it was performed for photoreceptor responses. The light induced stress was even increased by recording the light intensity – response curve under continuous background light first at 630 mlx (not shown) and thereafter at 2000 mlx (Fig. 8E, 8F).
Increasing the light intensity beyond 63 to 630 mlx caused a reduction of the b-wave amplitude as reported in the literature, and reflecting the process of light adaptation. Consecutively, the apparent a-wave amplitude gets increased (Fig. 8B). For the b-wave, the implicit time gets shorter during increased light intensities. Interestingly, under these conditions of 15 min continuous background light (200 mlx), the stability of the b-wave response is remarkable. Even more astonishing is the observation that under 630 and 2000 mlx background light, the b-wave responses continue to be very pronounced tolerating the increased light intensity at the recording chamber.
The transfer of the ERG recording technique from the bovine retina model to the various murine knockout models is possible, and the higher sensitivity of the murine retina towards ischemia maybe overcome by a more gentle preparation technique as outlined in detail . Future investigations of the murine retina from different mouse models may benefit from such an isolation procedure.
Besides ERG recordings additional electrophysiological recording techniques (e.g. patch clamp registration) will be needed to characterize signalling through retinal circuits.
We thank Priv.-Doz. Dr. Matthias Lüke, Dr. Siarhei A. Siapich for their excellent and permanent assistance during the work with the bovine retina, which is an important prerequisite for the transfer to the murine retina. The authors would also like to thank Nadeen Albanna (animation), Renate Clemens (genotyping and breeding of mice), Mehran Maghsoodian and Petra Müller (retinal IHC) for their support, as well as the animal keepers of the central facility for their excellent and permanent assistance. This work was financially supported by the Köln Fortune Programme of the Faculty of Medicine, University of Cologne (to MA, WA and MB) and by the Center for Molecular Medicine Cologne (CMMC, to TS and JH, Faculty of Medicine, University of Cologne).