Ординатура / Офтальмология / Английские материалы / Ocular Neuroprotection_Levin, Polo _2003

ibility of the effect either spontaneously or as a result of therapy. Similarly, the ERG was also shown to be a significant adjunct to the clinician’s armamentarium as it not only permitted one to diagnose and characterize a pathology, even in instances where there were no other obvious clinical signs, but also the information gathered was instrumental in identifying the possible site of malfunction. In the following sections, we will describe the different techniques used to record and analyze the electroretinogram. The accent will be on the human ERG because the techniques used are now standardized. However, it should be noted that human techniques can be used with little modifications on all sorts of animal species, such as: dogs [1] rats [2], mice [3], pigs [4], birds [5], and rabbits [6], to name a few. Although our aim is not to conduct a review of the literature to support this claim, modified human ERG techniques were previously used to study the toxicity of inhaled trichloroethylene on the rabbit retina [7,8], determine the shortand long-term sequels of postnatal oxygen exposure in newborn rats (animal model of human retinopathy of prematurity) [9,10], evaluate the beneficial effect of an intravitreal injection of ciliary neurotrophic factor (CNTF) in rescuing the degenerating photoreceptors in RDS mice [3], and, more recently, study the potentially harming effect of Sildafil (Viagra) on the retinal function of mice affected with a retinal degeneration similar to a form of human retinitis pigmentosa (RP) [11]. Similarly, the electroretinogram is frequently used to quantify the functional consequences of the new genotypes that are generated through selective breeding or created through genetic manipulations [12].

II.GENERAL OVERVIEW OF ELECTRORETINOGRAPHY

The electroretinogram (ERG) represents the electrical response that is generated by the entire retina when stimulated by a flash of light of adequate energy. It is often compared to the electrocardiogram (ECG) in that it is similarly composed of a series of waves that are presumed to originate from different retinal cells. However, unlike the ECG, which represents the ongoing activity of the myocardium, in order to generate an ERG response one must stimulate the retina with light. The ERG is thus an evoked potential generated by the excited retina and recorded at a distance (usually the cornea) from the latter. It represents a weighted average of all the retinal cells excited by the flash of light.

It is fair to say that the ERG does not exist as such, unlike the ECG. It is a creation that enables us to synchronize the retinal activity and consequently greatly facilitate its study. This concept is of extreme importance if one wishes not only to understand the origin of the ERG but also to appreciate the need to normalize the ERG acquisition parameters in order to extract as much meaningful functional information as possible. Continuing with the ECG analogy, in the latter

case it is the electrode position that is critical for inter-laboratory comparisons, whereas with the ERG it is the stimulating parameters that are of the utmost importance.

Added to the above, one must also take into consideration the unique physiology of the retina, which permits it to work under a wide range of luminance levels. This is for the great part due to the two types of photoreceptors: rods (for night time or scotopic vision) and cones (for daytime or photopic vision), which are normally found in the retinas of most higher vertebrates. One of the aims of electroretinography, whether used as an investigative tool or as a diagnostic test, is to separate the cone ERG from the rod ERG in order to distinctly assess the function of each type of photoreceptor. As we will see later on, standardized ERG protocols were written with that specific aim in mind.

III.COMPONENTS OF THE ERG AND THEIR ORIGINS

The electroretinogram was the first biopotential ever recorded. As early as 1865, Holmgren published a study in which he reported the recording of the electrical response of the eye of frogs upon excitation by a light stimulus [13]. Some 12 years later, Dewar was first to report human ERG recordings [13]. It is, however, Granit who contributed most to our understanding of the ERG wave genesis. In his classic study, which made use of ether intoxication in cats, Granit showed that the ERG was basically composed of three different processes, each of which gives rise to a different wave [14]. The three processes, named PI, PII, and PIII to reflect their order of disappearance as the level of anesthesia deepened, were shown to correspond to the c-wave, b-wave, and a-wave, respectively. With the refinement of the stimulating and recording techniques, new waves were added to the three studied by Granit. However, it is customary in modern electroretinography to concentrate on the two major waves of the ERG: namely, the a- and b-waves as well as the oscillatory potentials (OPs), which are components of high frequency and low voltage normally seen on the rising limb of the b-wave (see Fig. 1 for ERG waves identification) [15]. In the following paragraphs we will briefly review the postulated site of origin of these waves.

A. The a-Wave

As shown at Figure 1, the a-wave, which is negative in polarity, is the first major component of the ERG. To date, the consensus situates the generator of the a- wave at the level of the photoreceptors [13]. Depending on the state of retinal adaptation, either the cones alone or the cones and the rods will contribute to the genesis of the a-wave [13]. It should be noted that there is no pure rod ERG a-

Figure 1 Representative photopic broadband (recording bandwidth: 1–500 Hz) electroretinogram recorded from a sedated Guinea pig. The vertical arrow points to the onset of the stimulus. Components of the ERG are identified as follows: a- wave (a), b-wave (b), and oscillatory potentials (OPs). The amplitude of the a-wave is measured from the prestimulus baseline to the trough of the a-wave, whereas that of the b-wave is measured from the trough of the a-wave to the peak of the b-wave. Peak times are measured from flash onset to peak. Vertical calibration: 30 V. Horizontal calibration: 20 ms.

wave as such because in scotopic condition, an a-wave is recordable only in response to flashes of light of intensities which are in the photopic range (see Sec. V). These responses are usually referred to as mixed cone and rod ERG because both photoreceptors are claimed to contribute to their genesis. Analysis of the leading edge of the a-wave, using a computational approach, is often used to further characterize the physiology of the photoreceptor response as measured with the ERG [16]. It is of interest to note that in some animal species (such as rats and mice, for example), the photopic (cone) ERG is devoid of a recordable

a-wave despite a large b-wave. The exact reason for this remains obscure to date. By convention, the amplitude of the a-wave is measured from baseline to the trough of the a-wave.

B. The b-Wave

This is the most prominent and consequently the most important component of the ERG response. As shown in Figure 1, the b-wave is positive in polarity and said to be generated as a second-order potential by the Mu¨ller cells, which are glial cells running through the entire thickness of the retina, from the outer to the inner limiting membrane [13]. Sieving et al. (1994) suggested that the photopic ERG would be generated as a result of the synchronized activation of the ON-depolarizing bipolar cells (ON-DBC) and OFF hyperpolarizing bipolar cells (OFF-HBC), each contributing in sequence to the shaping of the b-wave [17]. The ON-DBC provides the impetus to push the b-wave up (i.e., from the trough of the a-wave to the peak of the b-wave) while the OFF-HBC limits the amplitude of the b-wave by pulling the retinal potential from the b-wave peak to a baseline value. This concept of ERG b-wave genesis is known as the PUSH-PULL hypothesis. By convention, the amplitude of the b-wave is measured from the trough of the a-wave (when present) to the peak of the b-wave.

C. The Oscillatory Potentials

The oscillatory potentials (OPs) identify the high-frequency components of the ERG, which are often seen as ripples on the ascending limb of the b-waves (Fig. 1) [18]. Fourier analysis reveals that while the a- and b-waves of the ERG are of a frequency domain of about 60 Hz or less, the OPs are usually of a frequency domain greater than 100 Hz. As we will see later, this difference in frequency domain is put to a use when one wishes to selectively record the OPs with minimal contamination from the slower a- and b-waves of the ERG. Practically all the retinal components, with the exception of the photoreceptors and the Mu¨ller cells, have been suggested as possible candidate for the genesis of the OPs [18]. Although there is yet no consensus as to their exact origin, it is important to note that there is more and more evidence to suggest that the terminology “oscillatory potentials” is probably a misnomer, because there are studies published to date that clearly show that the OPs do not (collectively) represent an oscillation such as the vibration of a retinal element or membrane, as the name could suggest. Past studies have shown that the number of OPs will vary as a function of stimulus intensity [19] (see also Fig. 4), inter-stimulus interval [20], retinal adaptation [21,22], pharmacological manipulation [23,24], and pathology [25,26], to name a few. Results from the above-mentioned studies clearly suggest that each OP would be generated by a different, and most probably functionally independent

retinal element. Compared to the other ERG components, the OPs were also shown to be significantly more sensitive to alteration in the retinal environment, whether acquired or innate [18,25].

Finally, as it currently stands, flash-evoked ERG responses do not include component(s) specifically assigned to the activity of the ganglion cells or beyond. There have been reports suggesting that some long-latency post–b-wave components (photopic i-wave or long-latency OPs) might be attributable to the activation of the retinal ganglion cells and/or the optic nerve including the chiasm [27– 29]. However, these claims were not verified elsewhere.

IV. RECORDING PROCEDURE

A.Preparation of the Subject

As indicated at the onset, the electroretinogram is an evoked potential. Consequently, in order to record, one must synchronize the recording of the retinal potential with the stimulus that will generate this response. To achieve this goal, several criteria must be met, the most important of which being full cooperation from the subject. Taking again the ECG analogy, the heart will continue to beat and thus an ECG recording will be possible (maybe with some difficulties) whether the subject (human or animal) is compliant or not. This is somewhat different for the ERG because the subject must cooperate to a certain degree if one wishes to record good quality and reproducible responses. If the subject voluntarily keeps its eyes shut, an adequate ERG recording is simply impossible. It is for that reason that, in animal experimentation, the subjects are systematically anesthetized for the procedure and not for the pain that might be involved, because it is, depending on the type of recording electrodes, an almost harmless procedure. In this laboratory we have successfully used a mixture of ketamine and xylazine to record ERGs from a variety of animal species including rats, mice, Guinea pigs, rabbits, and birds. Human subjects, on the other hand, need not be anesthetized or even sedated—even newborns or young infants. At least this has been our experience of more than 25 years of recording ERGs in a pediatric center.

B.Electrode and Recording Parameters

1.Electrode Types and Position (active, reference, ground)

According to the ERG standard of the International Society for Clinical Electrophysiology of Vision (ISCEV) [15], the active electrode should be corneal or as close as possible to the cornea. It is for that reason that the electrode of choice is the contact lens electrode (Henkes, Burian-Allen, etc.; see Fig. 2, top). The

reason for this is simple. As mentioned above, the ERG is an evoked potential that is recorded on the surface of the eye but originates from the retina. Therefore, the recording site is at some distance from the origin of the potential. Consequently, the closer we are from the source of the potential, the larger this potential will be. This is true of all ERGs, human or animal, and for that reason contact lenses have been developed for several animal species, including the mouse. Another advantage to the contact lens electrodes is the fact that their construction also includes a blepharostat whose function is to keep the eyelids open during the procedure, thus optimizing the delivery of light to the retina (see Fig. 2, top). Contact lens electrodes must, however, be used with a topical anesthetic, as they are painful to wear, and also with a viscous interface (usually methylcellulose) in order to protect the cornea from a possible abrasion. These electrodes cannot be worn for more than 20–30 min because the risk of corneal abrasion increases with time. Also, several sizes must be purchased in order to accommodate the different sizes of eyes (especially when pediatric and adult patients are seen in the same clinic) as well as the waiting time due to sterilization (in a clinical setup). The latter point is of importance as these electrodes are quite expensive. As an alternative to the contact lens electrode, the fiber electrodes are gradually becoming more and more popular, the most popular one being the DTL fiber electrode, which is made of a nylon yarn coated with silver (Fig. 2, middle) [30,31]. Fiber electrodes offer several advantages, such as painless even after several hours of wear (in humans), no need for topical anesthesia and disposable. They must, however, be used with great care and their positioning controlled as even minor changes in position will have a significant impact on the amplitude of the resulting ERG. Regardless of the type of electrode used, with human subjects, reference and ground electrodes (surface electrodes) are usually positioned at the external canthi (or forehead) and earlobes (earclip electrodes) respectively; for animal subjects, they are respectively placed in the mouth and inserted in the tail (needle electrodes) or subcutaneously in the leg. Electrode impedance should be minimized as much as possible ( 5KΩ).

2. Delivering the Stimulus: The Ganzfeld

The electroretinogram being an evoked response, a good ERG examination, whether for clinical or investigative purpose, will try to record the electrical response from as much retinal tissue as possible. It is for that reason that it is recommended to use a light diffuser (or Ganzfeld: Fig. 2, bottom) whose purpose, by design, is to diffuse the stimulus (and light adapting background light when used) to retinal eccentricities as far as the ora serata [32]. This is even more important in experimental situations where the rod function must be accurately assessed because, as we know, rods are most numerous in the peripheral retina. The same applies to animal experimentation where human Ganzfeld or stimulator

Table 1 Normative Data Obtained in 40 Normal Subjects (22 men; 18 women, age: 5–82)

na: not applicable (e.g., wave not present in response).

specifically designed for animal experimentation are used [33]. It is important to stress here that an adequate stimulator with diffusing possibilities should always be used if one seeks reproducible responses of high diagnostic potentials. Placing the subject (human or animal) directly in front of the light source (flash or other) is clearly insufficient.

3. Amplification and Recording Bandwidth

As stated above, the electroretinogram is a far field potential in that it represents the electrical activity evoked from the retina but recorded at the cornea—that is, in humans, about 25 mm from the source. The amplitude of the resulting biopo-

Figure 2 In order to obtain signals of the highest possible quality, it is imperative that the recording electrode be in contact with the eye, whether the ERGs are recorded from human or animal subjects. This is best achieved with the corneal contact lens (top picture), which normally includes a built-in blepharostat in order to keep the eyelids open. Use of these electrodes requires topical anesthesia and thus limits the wearing time to some 30 min or so. Alternative ERG electrodes, such as the DTL fiber electrode (second picture: DTL fiber is seen running on margin of the lower lid: arrow head), placed deep in the conjunctival bag significantly extends the recording time because they are basically painless as well as harmless. Once the eye returns to its primary position (third picture), the electrode does not interfere with the optical axis. Finally, ERGs should always be evoked in full field (Ganzfeld: bottom picture) or full field–like condition in order to stimulate as much of the retina as possible.

Figure 3 Typical ERG responses representative of the ISCEV standards. Vertical line(s) indicate flash onset. Photopic responses (background set at 25 cd/m2: single flash (A), oscillatory potentials (B), flicker 30 Hz (C). Scotopic responses: pure rod (D), mixed rod-cone (E), and oscillatory potentials (F). Recordings were obtained using a UTAS-3000 system (LKC Technologies, Inc, Gaithersburg, MD) along with DTL fiber electrodes (DTL Plus ElectrodeTM, Retina Technologies, Scranton, PA) and dilated pupils. Bandwidths were between 1 and 500 Hz for the broadband signals (A,C,D,E) or 75 and 500 Hz for the OPs (B,F). SF ISCEV standard flash which was set at 3 cd/m2/s. DA dark adaptation. Vertical calibration: 25V (photopic tracings) and 50 V (scotopic tracings). Horizontal calibration: 25 ms.

tential is in the range of approximately 100 V. This value varies with the intensity of stimulation, the level of retinal adaptation (photopic usually yields smaller responses compared to scotopic: see Table 1), the type of recording electrodes (usually the largest responses are obtained with contact lens electrodes) [34] as well as species. Given the relatively low voltage generated, it is necessary to amplify this signal in order to obtain a measurable response. Similarly, it was

Figure 4 Photopic ERG luminance-response function. A gradual increase in the intensity of the stimulus (bottom to top) augments the amplitude of the a- and b- waves of the broadband ERG (left column). However, while the amplitude of the a-wave continues to grow regularly, that of the b-wave reaches a maximum at .64 log cd m 2 s, after which it decreases, a phenomenon known as the photopic hill. Similarly, the timing of the a-wave shortens with brighter flashes, whereas that the b-wave increases. The photopic waveforms also include another component identified as the i-wave, the origin of which is still debated [27]. A similar increase in flash energy will not only impact on the amplitude of the OPs but also on the number. At threshold, the OP response includes only one major OP—namely, OP2 (OP1 identifies the small notch seen prior to OP2 in some tracings). A gradual increase in intensity will add OP3 and OP4 to this initial response. Horizontal calibration: 25 ms. Vertical calibration: 50 (ERG) and 25 (OPs) V.

mentioned above that the ERG response includes several components, some of which are easily differentiated with their temporal frequency domain. Consequently, in order to record as complete an ERG response as possible, it is recommended to amplify the retinal signal at least 1000 times within a bandwidth of 1–500 Hz at least. However, if the selective recording of the OPs is sought, the low frequency cutoff of the recording bandwidth should be increased to 75–100 Hz while keeping the upper limit to 500 Hz. In the latter case, the amplification should be augmented to 10,000 times to compensate for the attenuation in signal that will result from restricting the recording bandwidth (see Figs. 3 and 4).