Ocular movements must be eliminated to ensure successful multifocal electroretinography recordings. In early investigations within our laboratory, an IM injectable combination of ketamine/xylazine without paralysis did not provide sufficient ocular stabilization for acceptable recordings, although such an anesthetic combination has been applied successfully by other investigators [1
]). In accompanying studies, the use of pentobarbital sodium anesthesia with mfERG and mfVEP recordings has resulted in a stable animal preparation and reproducible and robust functional results [11
]. We also have anesthetized with a volatile inhalant (isoflurane) successfully [13
] during electrophysiological recordings; however, with this anesthetic, experimental studies cannot be prolonged to avoid potential depression of the mfERG response [7
]. Additionally, the evoked response potential (ERP) and the visually-evoked response (VER) collected under isoflurane anesthesia have been shown to become depressed or virtually eliminated and the latency measures and the character of the evoked response markedly altered during general anesthesia with volatile anesthetics [15
]). Under propofol anesthesia alone, achievement of stable ocular position was accomplished in our earlier sweep-VEP study in rhesus macaques [19
In the present study, anesthesia with both pentobarbital sodium (PB) and propofol (PF) enabled stable ocular fixation and robust, reproducible mfERGs. Primarily, effects of anesthetic regimen were observed in the K1 mfERG measures of both the ring (retinal eccentricity) and quadrant analyses. Amplitude of the early- and late-wave components (N1, P1, and P2 for ring; N1, P1, N2, and P2 for quadrant) of the K1 mfERG and of the K1 LFC (RMS SNR) with PF anesthesia were significantly enhanced when compared to animals anesthetized with PB. K1 IT differed for the early-wave (N1 and P1) and late-wave (N2 and P2) components with PF anesthesia prolonging early-wave and PB prolonging late-wave components both for ring and quadrant analyses; however, except for the P2 wave component, the delay in IT (though statistically significant) was ≤ 0.8 ms, and so probably not biologically significant. K2.1 LFC and K1 HFC (RMS SNR) measures for ring and quadrant analyses did not differ between the two anesthetic regimens; however, disparities in amplitude (significantly larger for PB in p1 and n2) and/or IT (significantly more prolonged for PB in p1, n2, and p2) of the K2.1 mfERG were observed. If the notions that K1 and K2.1 mfERG responses reflect largely outer retinal and inner retinal activity, respectively, then the K2.1 in combination with the K1 results suggest that specific anesthetic influence on the outer and inner retina, may depend somewhat on the particular agent used.
Recent electrophysiologic studies [17
] have compared the effects of varied general anesthetics on mammalian focal ERG, PERG, oscillatory potentials (OPs), scotopic threshold response (STR), and scotopic and photopic ERGs. In those studies in which barbiturate anesthesia was administered [17
], electrophysiologic responses were depressed relative to other anesthetics or anesthetic combinations administered (ketamine/xylazine, alphaxalone/alphadolone, halothane, urethane or telazol), but PF was not administered in either study.
More commonly, investigations of PB or PF effects on visual function have found differences in neuronal responses subserving purported outer and inner retinal function. Kapousta-Bruneau [25
] reported suppression by sodium pentobarbital of the scotopic threshold response and oscillatory potentials in the isolated rat retina, but not on photoreceptor function. PF has been shown to enhance electrophysiological responsiveness in canines [26
]. A strong reversible increase of scotopic b-wave and of a-wave peak amplitude was noted after an increase (50%) in propofol infusion rate. Ver Hoeve, Danilov, Kim and Spear [19
] showed little effect (a slight decrease) on pattern ERG and VEP when PF rate was increased 100%. Relatedly, when anesthetic agents (including ketamine/xylazine, PF, PB, urethane, and isoflurane) were compared on the ERP of the rat whisker sensory system [21
] ERP was highest and the latency slowest under PF anesthesia. PB anesthesia resulted in a lower and less prolonged (~ 35%) ERP peak amplitude when compared to propofol anesthesia. Both ERP findings were generally consistent with our K1 mfERG results.
tudies in the retina and central nervous system have shown that signal transduction depends not only on the identity of the transmitter substance(s) but also upon the receptors at which it binds [27
]. GABA (the principal inhibitory neurotransmitter in the brain) and glycine are the primary fast inhibitory neurotransmitters in the central nervous system and retina [16
] acting on GABAA
and glycine receptors, respectively, which are Cys-loop pentameric ligand-gated ion channels that are chloride ion-selective [16
]. Although the effects of general anesthetics at the molecular level are not fully understood, the primary molecular target of intravenous anesthetics such as pentobarbital and propofol are the GABAA
receptors, which are found abundantly in outer and inner retinal neurons [28
]. Based upon the clinical features of glycine receptors, they appear to be the primary molecular targets for the volatile (halogenated ether) anesthetics [29
]; however, their localization in the outer and inner retina [27
] also indicates a role in visual processing and function.
Anesthetic potentiation of inhibitory synaptic receptors (mainly GABAA
) best matches the pharmacological profile of a wide variety of agents for producing general anesthesia in mammals [16
]. All general anesthetics interact with multiple molecular targets, and characteristically, each drug has a unique pattern [18
]. But though functional effects (anesthesia) of anesthetics may be equivalent, the processes at the receptor level may not be [40
]. The degree of anesthetic potentiation and the diverse actions of general anesthetics have been shown to depend upon the subunit compositions of the GABAA
receptor subtype [16
]. Olsen and Sieghart [36
] have alluded to more than 800 distinct GABAA
receptor subtypes that may exist within the brain introducing a multiplicity of possibilities for binding processes and optimization for potentiation and direct activation by general anesthetics, which may underlie the variable influences of pentobarbital sodium and propofol on retinal function. In fact, specific binding sites on the heteropentameric GABAA
receptor differ for propofol and pentobarbital sodium for their respective potentiating and directly activating effects [51
] and may likely contribute to their differential functional effects as we have shown in the K1 and K2.1 responses. Additionally, pentobarbitone/pentobarbital also exhibits a more widespread effect than does propofol at clinically relevant concentration levels on neuromodulatory receptors of the central nervous system [29
]. For example, pentobarbital modulates activity not only on GABAA
], but also effectively inhibits neuronal nicotinic acetylcholine, α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA), kainate, and glycine receptors [29
], all of which are found in retinal neurons [28
], and when inhibited would lead to response decrements. When full-field ERGs were recorded in children under disoprofol (or propofol), b-wave amplitude and cone b-wave implicit time were nearly indistinguishable from responses recorded with topical anesthesia [55
]. These findings indicate that propofol anesthesia does not result in hyperabnormal responses of (particularly outer) retinal function, but rather suggests further that pentobarbital sodium anesthesia may result in decrement in responsiveness. Although specific attribution of retinal cellular activity subserving the K2.1 waveform components has not as yet been definitively established, the lower K2.1 amplitudes (of p1 and n2) observed in this study may be accounted for by propofol effects in slowing desensitization and deactivation of the GABAA
], which would stabilize and enhance receptor channel opening, and therefore prolong the postsynaptic inhibitory effects.
In summary, K1 and K2.1 mfERG responses of rhesus macaques are robust and reproducible when collected under either pentobarbital sodium or propofol anesthesia. K1 response density amplitude is generally larger under propofol than pentobarbital sodium anesthesia. Late-wave components (N2 and P2) are more prolonged under pentobarbital sodium than under propofol anesthesia. K1 LFC RMS SNR (larger RMS SNR values for propofol than for pentobarbital) differences were also noted to be significant in response comparisons between pentobarbital sodium and propofol. These response differences likely arise from variable relative effects of the anesthetics on retinal γ-aminobutyric acid (GABAA), and perhaps in part on glycine and on glutamate receptors. Therefore, as this study demonstrates, one must be mindful of the differential effects on the mfERG in the usage of these anesthetics. Care must be taken in combining data collected under propofol and pentobarbital sodium anesthesia.
There are practical considerations in the use of propofol rather than pentobarbital sodium anesthesia for multifocal electrophysiological testing—propofol permits a continuous intravenous infusion versus periodic injections of pentobarbital sodium, and thus, provides a more consistent level of anesthesia during recordings. A briefer period for initial responsiveness or waking and full recovery is typically experienced [57
]. Propofol intravenous infiltration does not result in negative ancillary tissue effects or extravasation injury [57
], which results from inadvertent infiltration of the surrounding tissue layer with injections, and may elicit a local, limited inflammatory response or an extensive necrosis of the epidermis and underlying soft tissues [61
]. Emesis potential is much reduced with propofol when compared to pentobarbital sodium anesthesia [58
]. Therefore, propofol is an appealing alternative anesthetic in the collection of mfERG responses in rhesus macaques.