Momentary Brain Concentration of Trichloroethylene Predicts the Effects on Rat Visual Function
http://www.100md.com
《毒物学科学杂志》
Neurotoxicology Division, Experimental Toxicology Division
Human Studies Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711
ABSTRACT
The relationship between the concentration of trichloroethylene (TCE) in the brain and changes in brain function, indicated by the amplitude of steady-state pattern-elicited visual evoked potentials (VEP), was evaluated in Long-Evans rats. VEPs were recorded from visual cortex following stimulation of the eyes and, thus, reflect the function of the afferent visual pathway and, in broad terms, may be indicative of overall brain function. The concentration of TCE in the brain at the time of VEP testing (i.e., momentary brain concentration) was hypothesized to predict the amplitude of the VEP across a range of inhalation concentrations, both during and after exposure. Awake restrained rats were exposed to clean air or TCE in the following combinations of concentration and duration: 500 ppm (4 h), 1000 ppm (4 h), 2000 (2 h), 3000 ppm (1.3 h), 4000 ppm (1 h), and 5000 ppm (0.8 h). VEPs were recorded several times during the exposure session, and afterward for experimental sessions of less than 4 h total duration (i.e., concentrations from 2000 to 5000 ppm). The sample collection time for each VEP was about 1 min. Brain concentrations of TCE were predicted using a physiologically based pharmacokinetic (PBPK) model. VEP waveforms were submitted to spectral analysis, and the amplitude of the largest response component, occurring at twice the temporal stimulation rate (F2), was measured. Exposure to all air concentrations of TCE in the study reduced F2 amplitude. The reduction of F2 amplitude was proportional to momentary brain TCE concentration during and after exposure. A logistical function fit to the combined data from all exposure conditions described a statistically significant relationship with 95% confidence limits between brain TCE concentration and F2 amplitude. The results support the hypothesis that momentary brain concentration of TCE is an appropriate dose metric to describe the effects of acute TCE inhalation exposure on rat VEPs. The combination of the PBPK model predicting brain TCE concentration from the exposure conditions with the logistical function predicting F2 amplitude from the brain TCE concentration constitute a quantitative exposure-dose-response model describing an acute change in neurological function following exposure to an important hazardous air pollutant.
Key Words: trichloroethylene; visual evoked potentials.
INTRODUCTION
It is important to understand the relationships between exposure to toxic compounds, the concentrations of the compound in target tissues, and the production of adverse outcomes. Quantitative descriptions of such relationships can be referred to as "exposure-dose-response models." A key consideration in developing such models is determining the appropriate dose metric, in other words, the measure of dose to the target tissue which is best correlated with the adverse outcome of interest. The dose metric, once established, becomes a central link in the development and use of exposure-dose-response models. Exposure-dose-response models have numerous potential applications in risk assessment and risk management (e.g., Benignus et al., in press; Boyes et al., 2005a,b; Simmons, et al., 2005).
Trichloroethylene (TCE) is a volatile organic solvent that is used widely in a variety of industrial applications, such as metal cleaning and degreasing and microelectronic manufacture (Wu and Schaum, 2000). TCE is listed as a Hazardous Air Pollutant under the Clean Air Act, and the U.S. Environmental Protection Agency considers TCE to be one of 33 top priority urban air toxicants (U.S. EPA, 1999). TCE is found in many hazardous waste sites and may also contaminate water systems (Scott and Cogliano, 2000; Wu and Schaum, 2000). In 1998, the mean concentration of TCE in samples of outdoor ambient air from 115 monitors in the United States was 0.88 μg/m3 (range 0.01–3.9) (Wu and Schaum, 2000). The current workplace Permissible Exposure Limit established by the Occupational Safety and Health Administration is 100 ppm TWA (time weighted average), with a 300 ppm peak limit, intended to prevent acute narcosis. Acute exposure to high concentrations of TCE, such as might occur after spills or in other emergency situations, produces acute reversible deficits in the function of the central nervous system (Arlien-Sborg, 1992). These changes are expressed broadly as deficits of sensory, cognitive, and motor function in humans as well as rats.
When the concentrations of TCE in brain are predicted using a physiologically-based pharmacokinetic (PBPK) model, changes in the neurophysiology and behavior of rats can be predicted as a function of brain concentration of TCE at the time the outcomes are assessed (Boyes et al., 2000, 2003). The concentration of the compound in target tissue at the time of assessment can be referred to as the "momentary" concentration, as distinct from "peak" concentration, to emphasize the possibility that the magnitude of reversible effects tracks ascending and descending changes of target tissue concentration.
One set of techniques to evaluate visual function is visual evoked potentials (VEPs), involving presentation of visual stimuli to the eyes and recording responses elicited from electrodes over visual cortex (Herr and Boyes, 1995). VEPs have been used successfully to study the functional deficits produced by a number of neurotoxic substances (Dyer, 1985; Mattsson et al., 1992; Rebert, 1983). Pattern-elicited VEPs, in which the patterned visual stimulus is temporally modulated (e.g., on/off or contrast reversal) to elicit responses to the changing visual patterns, can be used in neurotoxicology studies of rats (Boyes and Dyer, 1983). When the stimulation rate is sufficiently high, the response takes on a steady-state characteristic of a continuous repetitive waveform that is amenable to analysis using Fourier techniques. The response elicited has primary spectral components at the temporal frequency of the stimulus modulation and its higher frequency harmonics (Boyes, 1994). In rats, the primary response component of steady-state pattern on-off VEPs occurs at twice the stimulation frequency, or "F2" (Boyes, 1994).
Acute TCE inhalation reduced the VEP F2 amplitude of rats in a manner that was related to the momentary TCE brain concentration, and was not related to the product of air TCE concentration (C) and duration (t) of exposure, or to the area under the curve of TCE concentration in the brain (Boyes et al., 2003). The nature of the relationship between momentary brain TCE concentration and decreased VEP amplitude was assessed previously with only a limited number of exposure conditions, making it difficult to develop quantitative relationships between target tissue dose and outcome. In addition, measurements were taken in previous experiments only during inhalation exposure. The testing of outcomes after termination of exposure, while TCE is being eliminated from brain tissue, enables a test of the generality of the hypothesis that momentary brain concentration predicts outcome after, as well as during, external exposure.
The current experiments evaluated visual function of rats, as measured by the amplitude of the F2 VEP component, during and after acute inhalation exposure to TCE. A range of external air concentrations and testing times was evaluated, and VEP measurements were made both during and after the termination of external exposure. It was hypothesized that momentary concentration of TCE in the brain during VEP testing would accurately predict the reduction of VEP amplitude irrespective of the atmospheric TCE concentration, exposure duration, or time of testing (during or after exposure). Brain concentrations were estimated using a PBPK model designed and evaluated to predict tissue concentrations in Long-Evans rats (Simmons et al., 2002). In addition, a logistic function was derived to model quantitatively the relationship between TCE concentration in the brain and VEP amplitude. The logistic curve was selected as the dose-effect curve model, because its shape has been found to fit empirical data frequently (Corso, 1967), and it allows a variety of outcomes to be compared on a common proportional scale ranging from 0 to 1 (e.g., Benignus et al., 1998, 2005).
MATERIALS AND METHODS
The experimental methods were similar in many respects to those described in detail elsewhere (Boyes et al., 2003) and, therefore, are described here more briefly.
Test compound.
Spectrophotometric grade 1,1,2-trichloroethylene (99.5% pure) (CAS # 79-01-6) was obtained from Aldrich Chemical Co., St. Louis, MO.
Test animals.
Male Long-Evans rats (LE) (Crl: (LE)BR) were obtained from Charles River Laboratories (Raleigh, NC) at approximately 60 days of age. The animal colony had an ambient temperature and relative humidity of 22 ± 2°C and 50 ± 10%, respectively, and a 12:12 h light:dark cycle (lights on at 6:00 A.M.). The illumination during the light cycle ranged from approximately 190 lux (bottom shelf of the animal rack) to 390 lux (top shelf of the animal rack). The rats were housed individually in polycarbonate cages with kiln-dried pine chip bedding (Northeastern Products, Warrensburg, NY), with ad libitum access to tap water and rat chow (LabDiet PMI #5001, Richmond, IN). All aspects of the care and treatment of laboratory animals were approved by the Institutional Laboratory Animal Care and Use Committee of the National Health and Environmental Effects Research Laboratory and were in compliance with applicable federal guidelines for laboratory animal experimentation.
Surgery.
After approximately 1 week to acclimate to the animal colony, rats were anesthetized (sodium pentobarbital; 50 mg/kg ip), placed in a standard stereotaxic device, and prepared for cranial surgery. Recording electrodes, constructed from stainless steel screws (00-90 x 1/16 in.) soldered to Nichrome wires, were implanted into the skull, epidurally, in the following locations: 1 mm anterior to lambda and 4 mm left of the midline overlying the left primary visual cortex; and 2 mm anterior to bregma and 2 mm lateral left and right to midline for ground and reference electrodes, respectively. The electrode assembly was encased in dental acrylic and the wound sutured. Approximately 1 week was allowed for surgical recovery before electrophysiological testing. Details of this preparation have been described previously (Boyes et al., 2003; Herr et al., 1992).
Inhalation exposure.
A J-tube inhalation system was used to generate vapors of TCE and deliver them into a head-only inhalation chamber (Boyes et al., 2003; McGee et al., 1994). The concentration of TCE in the exposure chamber was monitored online using an infrared (IR) spectrophotometer (Miran, 1A, Foxboro Co, East Bridgewater, MA). The head-only exposure chamber (10 x 10 x 17 cm) was constructed of stainless steel, with the exception of a glass front plate to allow the rat to view the video monitor displaying visual stimulus outside of the chamber and a glass side plate allowing observation of the experimental subject. Awake (nonanaesthetized) rats were restrained in a plastic cone with the eyes, nose and ears exposed. When placed in the exposure chamber, a gas-tight latex seal was established around the upper torso of the rat to isolate the atmosphere of the head-only exposure chamber. Body temperature was monitored during experimental sessions. A flexible cable entering the head-only exposure chamber through a gas-tight port was connected to the rat's headset to allow recording of electrophysiological signals.
Visual stimuli.
A video monitor for presentation of visual stimuli was located approximately 15 cm from the rat's eyes outside the glass face of the exposure chamber. Visual stimuli were generated with a computer-based system described elsewhere (Boyes et al., 2003; Hamm et al., 2000). Stimulus brightness and contrast were calibrated to achieve a linear relationship between input signal voltage and screen luminance. The entire assembly was covered in a black cloth so that the primary source of light for the rat was the video screen.
The visual stimulus pattern was a sinusoidal vertical grating of 0.16 cycles per degree visual angle, selected because this value of spatial frequency is approximately at the peak of the contrast sensitivity function of pigmented rats (Birch and Jacobs, 1979), and peak contrast of 60%, selected because it provides a strong visual response under baseline conditions and yet lies well within the linear input/output range of the stimulus monitor. Mean stimulus illuminance was approximately 10 cd/m2, measured at the position of the rat's eyes. The pattern was modulated in an on/off manner with a sinusoidal temporal profile at approximately 5 Hz. Further details are presented in Boyes et al. (2003).
Evoked potential recording.
Electrophysiological potentials were amplified, band-pass filtered (1–100 Hz; roll off = 12 dB/octave), and sampled in one sec epochs (512 data points per epoch) using a Pathfinder II signal averager (Nicolet Biomedical, Madison, WI), so that each sample epoch contained approximately five cycles of the stimulus temporal modulation. Each VEP was constructed from the average of 50 1-s epochs, a process that took approximately 1 min for the collection of data for each waveform. To provide an estimate of recording noise level, "plus-minus" averages in which alternate single sweeps were inverted in polarity before adding to the signal averager were collected simultaneously. In this way, any reliable evoked potential signal would average toward zero volts, and the result would provide an estimate of the inherent system-wide noise which, being random, should approximate that of the random noise in the signal channel. The resulting waveforms were submitted to spectral analysis (SPECTM Program, Nicolet Biomedical, Madison, WI). Spectral amplitude of both signal and plus-minus waveforms was measured at twice the stimulus rate (F2). The F2 amplitude of the signal was the primary dependent variable of the initial data analysis, since F2 is the largest response component in rats presented this stimulus paradigm (Boyes, 1994), and previous experiments indicated that exposure to TCE reduced F2 amplitude (Boyes et al., 2003). For subsequent quantitative modeling of the relationship between F2 amplitude and brain TCE concentration, F2 amplitudes were normalized using both baseline recording levels and the plus-minus noise data as described below.
PBPK model.
Exposure concentrations and time points for VEP sampling were selected using an initial version of the PBPK model (Boyes et al., 2000) with the goal of yielding different combinations of exposure concentrations and durations that resulted in equivalent blood TCE concentrations. After the experiments were completed, an improved version of the PBPK model became available (Simmons et al., 2002). The tissue concentrations predicted from the revised model were compared to TCE tissue concentrations measured in brain, blood, liver, and fat of exposed LE rats, and the model was shown to have good predictive accuracy (Simmons et al., 2002). In addition, a linear relationship was observed between blood and brain TCE concentrations over a range between 0.1 and 100 mg/1 (Boyes et al., 2003). In the current report, we evaluate the VEP results as a function of predicted brain TCE concentration using the improved version of the PBPK model (Simmons et al., 2002).
Experimental design.
Rats (n = 8–10/group) were exposed to TCE in concentrations of 500, 1000, 2000, 3000, 4000, or 5000 ppm. Total exposure durations differed for each concentration to provide total concentration–duration products of 4000 ppm-h, with the exception of 500 ppm, for which the total duration of exposure was 4 h. The times of VEP sampling and the associated estimates of brain TCE concentrations are depicted in Figure 1. VEP sample times are also presented in Table 1. Sample times greater than 4 h were excluded due to decrements in the VEP amplitude, presumably from fatigue or adaptation, observed in recordings from control animals. VEPs were recorded repeatedly from each rat during and after exposure to TCE concentrations between 2000 and 5000 ppm. For rats exposed to 500 or 1000 ppm, VEPs are presented only during the exposure phase of the experiment which lasted for 4 h. In addition, VEPs were sampled from rats exposed to only clean air at times matching those of the TCE-exposed rats. Two or three such control rats were run at matching sample times for each of the TCE concentrations (overall n = 13), with the exception of 500 ppm, for which an additional 8 matched controls were run.
Statistical Analysis and Curve Fitting
Data.
Data from one rat exposed to 5000 ppm were eliminated due to high levels of electrical noise during the recording session. To fit a logistic dose-effect function, data from all the experiments were pooled into a single data base. This data set consisted of 489 observations on 79 subjects (58 exposed to TCE and 21 clean air controls), including data from baseline (preexposure), several during-exposure recordings and, for 2000–5000 ppm TCE, several postexposure recordings as well. Measurements of the plus-minus average, as described above, were considered as measures of background noise. The F2 amplitude values were corrected for background noise by subtracting the mean F2 amplitudes from plus-minus records. Next, each F2 amplitude measure was converted to a proportion-of-baseline score by dividing by the mean preexposure baseline amplitude. These transformed F2 data were called normalized baseline-adjusted data, and have the properties that the baseline (no-TCE value) value of F2 is 1.0, and the maximum effect value (F2 entirely suppressed) is 0.0.
The baseline-adjusted F2 data were screened for outliers using a box-and-whisker plot (Tukey, 1977; Systat Inc. 1990). Thirteen observations qualified as outliers (>3 times the interquartile range), seven of which came from a single subject. Logistic functions were fit, as described below, to the entire data set, and also to the data set after removal of outliers, including all the data from the one subject with seven outliers. The logistic function showed a similar fit with or without the inclusion of outliers, with derived parameters differing by less than 0.15%. The outlier-trimmed data showed a greater resemblance to the original nontransformed data, and therefore the logistical function based on that analysis is presented here.
Curve fitting.
The adjusted F2 amplitude data were used as the dependent variable in fitting the dose-effect curve. The data contained a variable number of repeated measures for each rat. To account for the correlations between the variable numbers of repeated measures, a mixed-model curve fitting procedure was used to fit the dose-effect curve (SAS Proc NLMIXED). The independent variable for these data was the concentration of TCE in the brain as estimated for each exposure scenario using the PBPK model.
A logistic model, which has been used previously for similar types of data (Benignus et al., 1998, 2005; Corso, 1967), was fit empirically to the data from the TCE-exposed rats. The logistic curve has a variable ogive-like shape but unlike the ogive, the upper and lower asymptotes are independently variable, thus permitting a better fit to some data. The lower and upper asymptotes are zero and one, respectively. There are two parameters, 1 and 2, which determine the location and slope, respectively, of the curve. The asymptotes make it necessary to transform raw outcome data so that the resultant data points range in value from zero to one.
Other nonlinear curves might have been selected, but do not have some of the desirable characteristics of the logistic. Quadratic functions can be made to fit well within the range of the data, but behave poorly outside that range, thus making quadratic functions a poor choice for extrapolation. The ogive does not permit asymmetry about mid value of dose. Logarithmic and power functions do not have convenient (or any) asymptotes. Thus, the logistic was selected as the preferred model.
The logistic model has the form
(1)
in which E is the measure of the effect of TCE (proportion of baseline VEP amplitude), e is the base of natural logarithms, D is the estimated momentary concentration of TCE in the brain, ln(D) is the natural logarithm of D, and 1 and 2 are empirical coefficients which determine the location and slope, respectively, of the fitted curve. The fitted curve and its 95% confidence limits were calculated as a function of predicted brain TCE concentration.
Residual analysis.
To test the possibility that either the air TCE concentration, the duration of exposure, or their product (C x t) might influence F2 amplitudes beyond the effect of the internal dose, an analysis was performed of the residuals from fitting the dose-effect curve. The residuals are the absolute differences between the observed data and the line of best fit. In the ANOVA (SAS Proc MIXED) the residuals from the fitted curve were treated as the dependent variable, with duration and concentration and their interaction as effects. Step-down tests employed t-tests. A significant effect would indicate that concentration or duration, or the C x t product, had affected F2 amplitudes in addition to the influence of momentary brain concentration.
RESULTS
The target concentrations of TCE in the headspace air of the exposure chamber were 500, 1000, 2000, 3000, 4000, and 5000 ppm. The TCE concentration measured in the head-only exposure chambers for each experimental condition was 511 ± 5.8, 938.5 ± 13.6, 1966.2 ± 60.2, 2988.4 ± 86.9, 3970 ± 90.9, and 4870.5 ± 63.9 (mean ± standard deviation), respectively.
Visual evoked potential waveforms averaged across the group exposed to 3000 ppm TCE are presented in Figure 2, along with the spectral analysis of each group average waveform. The group average waveforms from the other exposure concentrations resembled those from the 3000-ppm group and are not presented. Before the onset of exposure to TCE, the VEP waveforms showed 10 clear cycles of the response and manifested a steady-state evoked potential profile (Boyes, 1994). The spectral analysis of the baseline waveforms showed predominant response amplitude at F2. After the onset of exposure to TCE, VEP waveforms decreased in amplitude, and the strong steady-state response characteristics diminished. This was accompanied by a decrease in amplitude of the F2 spectral component. Spectral amplitude at the stimulus modulation rate (F1) was more variable, and not altered systematically by exposure to TCE. After termination of exposure, the VEP waveform gradually began to recover a resemblance to the original steady-state characteristics, and the F2 amplitude of the spectra increased.
The nature of the change in VEPs caused by TCE exposure is illustrated in Figure 3, where two of the group average waveforms from Figure 2 were superimposed and then replotted. The two waveforms selected were from baseline (time = 0) and from the time of maximum F2 suppression immediately after termination of exposure (time = 1.33 h). Inspection of these two waveforms reveals that every other cycle of the evoked response was suppressed by TCE exposure, with relatively little change in the intervening cycles.
The amplitude of the F2 component was examined as a function of estimated brain concentration during and/or after exposure. Figure 4 presents F2 amplitude during exposure for all exposure conditions, and after exposure for groups exposed to 2000, 3000, 4000, or 5000 ppm TCE. The amplitude of F2 was reduced by TCE exposure. There was some variation among the groups in the baseline F2 amplitudes prior to exposure, which is occasionally observed for different sets of rats tested at different times. The amount of F2 amplitude reduction observed during or after exposure to any given concentration was equivalent when expressed as a function of the brain TCE concentration expected at that sampling time.
Previously it was shown that for a C x t product of 4000 ppm-h, that the C x t product did not adequately describe the effect of TCE exposure on F2 amplitude, whereas peak brain TCE concentration did (Boyes et al., 2003). Data from the current experiment are presented as a function of the momentary brain TCE concentration predicted from the PBPK model, and also as a function of the C x t product for the exposure occurring up to the point of VEP recording (Fig. 5). The results indicate that the data are better described by momentary brain concentration than by C x t product.
In order to describe quantitatively the relationship between brain TCE concentration and F2 amplitude, the data from all exposure conditions were combined, normalized, and plotted as a fitted logistical dose-effect curve (Fig. 6). Any apparent increase in the variability of the data in Figure 6 over that of nontransformed values of Figure 5 may be attributed to variability of the baseline or noise level values which were used in calculating the F2 amplitude transformations. The fit of the dose-effect curve to the data was statistically significant, giving 1 = 4.48, t = 7.86, p < 0.0001 and 2 = –1.50, t = –9.16, df = 76, p < 0.0001. Table 2 gives the covariance matrix for the parameter estimates. The mean estimate derived from the logistic curve for the brain concentration causing a 10% amplitude deficit (ED10) was 4.5 mg TCE/l, and causing a 50% deficit (ED50) was 20.0 mg TCE/l.
Analysis of the residuals from the fitted curve (ANOVA) indicated that the duration of exposure, the concentration, or the C x t combination were not statistically significant (p-values were 0.602, 0.367, and 0.337, respectively). Thus, chamber air concentration, exposure duration, or the C x t product had no significant explanatory power beyond that conveyed by momentary brain TCE concentration, which was significantly linked to changes in F2 amplitude.
DISCUSSION
Inhalation exposure to TCE reduced the amplitude of the F2 component of pattern VEPs of Long-Evans rats in a manner that was associated with the momentary concentration of TCE in the brain. The limited set of data which was available previously suggested a linear relationship between brain TCE concentration and change in F2 amplitude (Boyes et al., 2003). The current data set is more extensive and clearly shows a curvilinear relationship that was described by a logistic function. The parameters of the logistic curve may be useful in predicting the level of F2 amplitude impairment caused by a variety of TCE exposure conditions.
Momentary brain concentration of TCE appeared to determine the change in F2 amplitude. This outcome extends previous results that momentary brain TCE concentration was a better predictor of VEP F2 amplitude than was the product of air concentration and duration, or the area under the curve (AUC) reflecting the total cumulative dose (Boyes et al., 2003). In the current study, the residuals of the logistic function associated with TCE inhaled-air concentration, the exposure duration, or the C x t product were not statistically significant, indicating that these factors had no explanatory power beyond that of the momentary concentration of TCE in brain. The amplitude of F2 is a reflection of massed firing of neurons in visual cortex, responding to afferent visual input originating in the retina. To the extent that F2 amplitude reduction reflects general neurological impairment, these results suggest that momentary brain TCE concentration is an appropriate measure of internal dose (i.e., a dose metric) for evaluating acute neurological effects of exposure to TCE.
Acute exposure to volatile organic compounds such as TCE does not necessarily have selective actions restricted to the visual system, but rather suppress neuronal activity broadly throughout the nervous system. Human volunteer subjects exposed to TCE vapors while performing a visual/motor performance task showed no observable adverse effects (NOAEL) at 300 ppm for 2 h (Vernon and Ferguson, 1969). In the same study, the lowest dose producing significant impairment of human visual/motor performance (LOAEL) was 1000 ppm for 2 h (Vernon and Ferguson, 1969). Through PBPK modeling, it was estimated that the human blood TCE concentration associated with this NOAEL was approximately 4.8 mg/l, and with the LOAEL was about 18 mg/l (Boyes et al., 2005a). Considering that in rats (and presumably humans) there is approximately a 1:1 correspondence between TCE concentration in blood and brain (Boyes et al., 2003), there is an interesting similarity between the tissue dose of the human neurobehavioral NOAEL (4.8 mg/l) and the rat VEP ED10 measured here (4.5 mg/l), and also between the human neurobehavioral LOAEL (18 mg/l) and the rat VEP ED50 (20 mg/l). These results suggest that the rat VEP is sensitive to acute TCE exposure at concentrations similar to those of human neurobehavioral performance.
The relationships between the target tissue concentration of organic solvents and neurotoxic outcomes have been of interest. Bruckner and Peterson (1981) showed an association between brain toluene concentration and CNS depression in mice both during and after exposure to either 4000 ppm for 3 h or 10,600 ppm for 10 min. Single inhalation exposures to TCE (Kishi et al., 1993) or toluene (Kishi et al., 1988) altered shock avoidance behavior of rats in a manner consistent with solvent concentration in the blood (TCE or toluene) or brain (toluene). In a meta-analysis of previously published data, Benignus et al. (1998) showed that rat avoidance behavior and human choice reaction time could be well described by blood toluene level at the time of behavioral assessment. The momentary concentration in blood was also predictive of TCE-induced deficits in performance of a signal detection task (Boyes et al., 2000). In contrast, Warren et al. (1996) could find no consistent relationship between response rate on a fixed-ratio operant task and the time course or concentration of perchloroethylene in brains of repeatedly treated rats. Van Asperen et al. (2003) compared the visual discrimination performance of rats following inhalation exposure to a constant concentration or to repeated spiked concentrations of toluene. Performance of the task, as indicated by the percentage of intertrial interval responding, was not related to the concentration of toluene in brain at the end of the exposure session. Similarly, it is unlikely that brain solvent concentration accounts for development of behavioral tolerance following repeated exposure to TCE (Bushnell and Oshiro, 2000; Oshiro et al., 2001). Thus, it appears that momentary brain solvent concentration is generally a good predictor of simultaneous behavioral outcomes, in particular for single exposure sessions, but additional factors may be necessary to explain behavioral consequences of repeated exposures or complex exposure protocols.
There is often a need in risk assessment to estimate the exposure concentrations associated with a particular outcome for exposure durations which were not assessed experimentally. A common practice in doing so is to apply Haber's rule, that C x t = k, or the ten Berge equation Cn x t = k (ten Berge, et al., 1986), to adjust the critical air concentration for a desired exposure duration. We have shown previously that using a PBPK model to predict the target tissue concentration is preferable to Haber's or ten Berge's duration adjustments (Boyes et al., 2003, 2005a: Simmons et al., 2005). The analysis presented here extends this principle to include prediction of effects lingering after the termination of exposure. The Haber or ten Berge approaches cannot account for effects after termination of exposure, when the external air concentration is theoretically zero and the C x t (or Cn x t) product would also be zero, leading to a prediction of no effect. Alternatively, the Haber or ten Berge products could be interpreted as the total cumulative exposure during the inhalation phase of the experiment, in which case the values never decrease after termination of exposure and cannot account for a gradual return to baseline. When using target tissue dose as the metric, however, the level of impairment can be predicted over the course of elimination of the compound according to the amount of residual toxic compound in the target tissue.
One issue of interest is whether the neurological impairments from exposure to TCE are caused by the parent compound by one or more of the potentially active metabolites. There are many identified metabolites of TCE (Lash et al., 2000), at least two of which, trichloroethanol (TCOH) and chloral hydrate, likely have neuroactive properties (Krasowski et al., 1998; Peoples and Weight, 1998). We argued previously that the acute effects of TCE on F2 amplitude were probably caused by TCE and not a metabolite because clear dose-response relationships were observed above the dose of metabolic saturation, about 400 ppm, where no additional metabolites would be formed with increasingly higher doses (Boyes et al., 2003). The current data provide additional evidence, albeit indirect, that the parent compound is the active species. Very similar effects were seen during and after exposure when plotted as a function of brain TCE concentration (Figure 4). The time course of TCE in rat tissues is shorter, by various amounts, than that of its principal metabolites including trichloroacetic acid, TCOH, and chloral hydrate (Clewell, et al., 2000; Fisher et al., 1991; Larson and Bull, 1992; Prout et al., 1985; Templin et al., 1995). If a metabolite were the active species, then the F2 amplitude curves during and after exposure would have been displaced in proportion to the time course of the active metabolite formation and clearance. This was not observed, supporting the case that the effects observed were attributable to the TCE parent compound.
The visual system is thought to be comprised of multiple parallel systems, one aspect of which concerns whether the responses of the system show linear and nonlinear relationships to the physical and temporal modulation of the visual stimulus (Regan, 1989). The response patterns of visual cells characterized as "linear" include firing rates that are the linear sum of the responses to independent stimulation of the excitatory and inhibitory portions of their receptive fields (Enroth-Cugell and Robson, 1966). Linear responses also include sustained firing to continuous stimulation, and responding to modulated visual stimuli at the temporal frequency (F1) of stimulus modulation (Enroth-Cugell and Robson, 1966; Hochstein and Shapley, 1976; Kaplan and Shapley, 1982). Visual responses characterized as "nonlinear," on the other hand, include firing rates that are not the linear sum of stimulating the excitatory and inhibitory components of the receptive fields. In addition, nonlinear responses are characterized by firing in a transient or phasic fashion to continuous visual stimuli, and responding to modulated visual stimuli primarily at twice the rate (F2) of temporal stimulation (Enroth-Cugell and Robson, 1966; Hochstein and Shapley, 1976; Kaplan and Shapley 1982; Lennie and Perry, 1981). Depending on the stimulus conditions, individual cells may respond at both F1 and F2 (e.g., Lennie and Perry, 1981). The linear and nonlinear response profiles originate in retinal bipolar cells, are transmitted to retinal ganglion cells, and are passed on through the visual pathway (Awatramani and Slaughter, 2000). The ratio of the response amplitude at second (F2) to that at the first (F1) harmonic frequency has been used as an index of nonlinearity (Hochstein and Shapley, 1976). Linear and nonlinear properties also may be observed in the integrated responses of the visual system (e.g., Regan and Regan, 1989; Zemon and Radcliff, 1984). The current results suggest that TCE exposure suppresses nonlinear activity patterns in the visual system.
The visual system also contains On and Off visual pathways, which respond differentially to ionotropic and metabotropic glutamate receptor agonists and antagonists (Awatramani and Slaughter, 2000; Schiller, 1984; Slaughter and Miller, 1981). At a systemic level, selective blocking of either the On or the Off response of the transient (nonlinear) visual system might be expected to reduce VEP F2 amplitude and produce a response profile resembling that of Figure 3, in which every other cycle of the response to an On-Off stimulus was diminished. Unfortunately, it is not clear whether the On or the Off cycle was suppressed, since the phase lag (temporal delay) between the stimulus modulation and the response is ambiguous in steady-state recordings (Regan, 1989). Toluene, a volatile organic compound with acute sedative actions similar to TCE (Arlien-Sborg, 1992), blocks the actions of glutamate at ionotropic NMDA receptors (Cruz et al., 1998, 2000), but to our knowledge, the actions of volatile organic compounds on metabotropic glutamate receptors have not been evaluated to date (Bushnell et al., 2005). Other potential target sites include GABAA (Beckstead et al., 2000), nicotinic acetylcholine receptors (Bale et al., 2002), and voltage sensitive calcium channels (Tillar et al., 2002), which could all conceivably alter the temporal response profiles of visual neurons.
In summary, the results of these experiments establish a quantitative relationship between the concentration of TCE in brain tissue and changes in the amplitude of VEPs recorded from awake rats. Similarly, the PBPK model can be used to predict brain TCE concentrations for exposure conditions of interest. Together these components enable the quantitative prediction of graded decrement in neurological function for acute exposure scenarios involving inhalation of TCE vapors.
NOTES
This manuscript has been reviewed by the National Health and Environmental Effects Research Laboratory, U.S. EPA and approved for publication. Mention of trade names and commercial products does not constitute endorsement or recommendation for use.
ACKNOWLEDGMENTS
The authors thank Philip Bushnell, David Herr, Ambuja Bale, Linda Birnbaum, Elaina Kenyon, and Jane Caldwell for comments on an earlier version of the manuscript. Conflict of interest: none declared.
REFERENCES
Arlien-Sborg, P. (1992). Solvent Neurotoxicity. CRC Press, Boca Raton, FL.
Awatramani, G. B., and Slaughter, M. M. (2000). Origin of transient and sustained responses in ganglion cells of the retina. J. Neurosci. 20, 7078–7095.
Bale, A. S., Smothers, C. T., and Woodward, J. J. (2002). Inhibition of neuronal nicotinic acetylcholine receptors by the abused solvent, toluene. Br. J. Pharmacol. 137, 375–383.
Beckstead, M. J., Weiner, J. L., Eger, E. I., Gong, D. H., and Mihic, S. J. (2000). Glycine and gamma-aminobutyric acid(A) receptor function is enhanced by inhaled drugs of abuse. Mol. Pharmacol. 57, 1199–1205.
Benignus, V. A., Boyes, W. K., and Bushnell, P. (1998). A dosimetric analysis of acute toluene exposure in rats and humans. Toxicol. Sci. 43, 186–195.
Benignus, V. A., Bushnell, P. J., and Boyes, W. K. (2005). Toward cost-benefit analysis of acute behavioral effects of toluene in humans. Risk Anal. 25, 447–456.
Birch, D., and Jacobs, G. H. (1979). Spatial contrast sensitivity in albino and pigmented rats. Vision Res. 19, 933–937.
Boyes, W. K. (1994). Rat and human evoked potentials and the predictability of human neurotoxicity from rat data. Neurotoxicology 15, 569–578.
Boyes, W. K., Bercegeay, M., Ali, J., Krantz, T., McGee, J., Evans, M., Raymer, J. H., Bushnell P. J., and Simmons, J. E. (2003). Dose-based duration adjustments for the effects of inhaled trichloroethylene on rat visual function. Toxicol. Sci. 76(1), 121–130.
Boyes, W. K., Bushnell, P. J., Crofton, K. M., Evans, M., and Simmons, J. E. (2000). Neurotoxic and pharmacokinetic responses to trichloroethylene as a function of exposure scenario. Environ. Health. Perspect. 108 (Suppl. 2), 317–322.
Boyes, W. K., and Dyer, R. S. (1983). Pattern reversal visual evoked potentials in awake rats. Brain Res. Bull. 10, 817–823.
Boyes, W. K., Evans, M. V., Eklund, C., Janssen P., and Simmons, J. E. (2005a). Duration adjustment of acute exposure guideline level values for trichloroethylene using a physiologically-based pharmacokinetic model. Risk Anal. (in press).
Boyes, W. K., Simmons, J. E., Eklund C., Benignus, V. A., Janssen, P., Bushnell, P. J. (2005b). Applications of dosimetry modeling to assessment of neurotoxic risk. Environ. Toxicol. Pharmacol. 19, 599–605.
Bruckner J. V., and Peterson, R. G. (1981). Evaluation of toluene and acetone inhalant abuse. I. Pharmacology and pharmacodynamics. Toxicol. Appl. Pharmacol. 61, 27–38.
Bushnell, P. J., and Oshiro, W. M. (2000). Behavioral components of tolerance to repeated inhalation of trichloroethylene (TCE) in rats. Neurotoxicol. Teratol. 22, 221–229.
Bushnell, P. J., Shafer, T. J., Bale, A. S., Boyes, W. K., Simmons, J. E., Eklund, C., and Jackson, T. L. (2005). Developing an exposure-dose-response model for the acute neurotoxicity of organic solvents: Overview and progress on in vitro models and dosimetry. Environ. Toxicol. Pharmacol. 19, 607–614.
Corso, J. W. (1967). The Experimental Psychology of Sensory Behavior. Holt-Rinehart and Winston, New York.
Clewell, H. J., III., Gentry, P. R., Covington, T. R., and Gearhart, J. R. (2000). Development of a physiologically based pharmacokinetic model of trichloroethylene and its metabolites for use in risk assessment. Environ. Health Perspect. 108, 283–305.
Cruz, S. L., Balster, R. L., and Woodward, J. J. (2000). Effects of volatile solvents on recombinant N-methyl-d-aspartate receptors expressed in Xenopus oocytes. Br. J. Pharmacol. 131, 1303–1308.
Cruz, S. L., Mirshahi, T., Thomas, B., Balster, R. L., and Woodward, J. J. (1998). Effects of the abused solvent toluene on recombinant N-methyl-d-aspartate and non-N-methyl-d-aspartate receptors expressed in Xenopus oocytes. J. Pharmacol. Exp. Ther. 286, 334–340.
Dyer, R. S. (1985). The use of sensory evoked potentials in toxicology. Fundam. Appl. Toxicol. 5, 24–40.
Enroth-Cugell, C., and Robson, J. G. (1966). The contrast sensitivity of retinal ganglion cells of the cat. J. Physiol. 187, 517–552.
Fisher, J. W., Gargas, M. L., Allen, B. C., and Andersen, M. E. (1991). Physiologically based pharmacokinetic modeling with trichloroethylene and its metabolite, trichloroacetic acid, in the rat and mouse. Toxicol. Appl. Pharmacol. 109, 183–195.
Hamm, C. W., Ali, J. S., and Herr, D. W. (2000). A system for simultaneous multiple subject, multiple stimulus modality, and multiple channel collection and analysis of sensory evoked potentials. J. Neurosci. Meth. 102, 95–108.
Herr, D. W., Boyes, W. K., and Dyer, R. S. (1992). Alterations in flash and pattern reversal evoked potentials after acute or repeated administration of carbon disulfide (CS2). Fundam. Appl. Toxicol. 18, 328–342.
Herr, D. W., and Boyes, W. K. (1995). Electrophysiological analysis of complex brain systems: Sensory evoked potentials and their generators. In Neurotoxicology: Approaches and Methods (L. W. Chang and W. Slikker, Eds.), pp. 205–221. Academic Press, New York.
Hochstein, S., and Shapley, R. M. (1976). Quantitative analysis of retinal ganglion cell classifications. J. Physiol. 262, 237–264.
Kaplan, E., and Shapley, R. M. (1982). X and Y cells in the lateral geniculate nucleus of Macaque monkeys. J. Physiol. 330, 125–143.
Kishi, R., Harabuchi, I., Ikeda, T., Yokota, H., and Miyake, H. (1988). Neurobehavioral effects and pharmacokinetics of toluene in rats and their relevance to man. Brit. J. Indust. Med. 45, 396–408.
Kishi, R., Harabuchi, I., Ikeda, T., Katakura, Y., and Miyake, H. (1993). Acute effects of trichloroethylene on blood concentrations and performance decrements in rats and their relevance to humans. Brit. J. Indust. Med. 50, 470–480.
Krasowski, M. D., Fin, S. E., Ye, Q., and Harrison, N. L. (1998). Trichloroethanol modulation of recombinant GABAA, glycine and GABA 1 receptors. J. Pharmacol. Exp. Therap. 284, 934–942.
Larson, J. L., and Bull, R. J. (1992). Species differences in metabolism of trichloroethylene to the carcinogenic metabolites trichloroacetate and dichloroeacetate. Toxicol. Appl. Pharmacol. 115, 278–285.
Lash, L. H., Fisher, J. W., Lipscomb, J. C., and Parker, J. C. (2000). Metabolism of trichloroethylene. Environ. Health Perspect. 108 (Suppl. 2), 177–200.
Lennie. P., and Perry, V. H. (1981). Spatial contrast sensitivity of cells in the lateral geniculate nucleus of the rat. J. Physiol. 315, 69–79.
Mattsson, J. L., Boyes, W. K., and Ross, J. F. (1992). Incorporating evoked potentials into neurotoxicity test schemes. In Target Organ Toxicology Series: Neurotoxicity (H. A. Wilson and C. L. Mitchell, Eds.), pp.125–145. Raven Press.
McGee, J. K., Evansky, P. A., Terrell, D., Walsh, L. C., Davies, D. W., and Costa, D. L. (1994). A new vapor and gas test atmosphere generator with broad concentration and flow output capabilities. In Measurement of Toxic and Related Air Pollutants: Proceedings of the 1994 U.S. EPA / AWMA International Symposium, pp. 1011–1016. Air and Waste Management Association Publication No. VIP-39.
Oshiro, W. M., Krantz, Q. T., and Bushnell, P. J. (2001). Characterizing tolerance to trichloroethylene (TCE): Effects of repeated inhalation of TCE on performance of a signal detection task in rats. Neurotoxicol. Teratol. 23, 617–628.
Peoples, R. W., and Weight, F. F. (1998). Inhibition of excitatory amino acid-activated currents by trichloroethanol and trifluoroethanol in mouse hippocampal neurones. Br. J. Pharmacol. 124, 1159–1164.
Prout, M. S., Proven, M. W., and Green, T. (1985). Species differences in response to trichloroethylene. Toxicol. Appl. Pharmacol. 79, 389–400.
Rebert, C. S. (1983). Multisensory evoked potentials in experimental and applied neurotoxicology. Neurobehav. Toxicol. Teratol. 5, 659–671.
Regan, D. (1989). Human Brain Electrophysiology: Evoked Potentials and Evoked Magnetic Fields in Science and Medicine. Elsevier, New York.
Regan, M. P., and Regan, D. (1989). Objective investigation of visual function using a nondestructive zoom-FFT technique for evoked potential analysis. Can. J. Neurol. Sci. 16, 168–179.
Scott, C. S., and Cogliano, V. J. (2000). Trichloroethylene health risks—state of the science. Environ. Health Perspect. 108 (Suppl. 2), 159–160.
Schiller, P. H. (1984). The connections of the retina On and Off pathways to the lateral geniculate nucleus of the monkey. Vision Res. 24, 923–932.
Simmons, J. E., Boyes, W. K., Bushnell, P. J., Raymer, J. H., Limsakun, T., McDonald, A., Sey, Y. M., and Evans, M. V. (2002). A physiologically-based pharmacokinetic model for trichloroethylene in the male Long-Evans rat. Toxicol. Sci. 69, 3–15.
Simmons, J. E., Evans, M. V., and Boyes, W. K. (2005). Moving from external exposure concentration to internal dose: Duration extrapolation based on physiologically-based pharmacokinetic derived estimates of internal dose. J. Toxicol. Environ. Health A 68, 927–950.
Slaughter, M. M., and Miller, R. F. (1981). 2-Amino-4-phosphonobutyric acid: A new pharmacological tool for retina research. Science 211, 182–185.
Templin, M. V., Stevens, D. K., Stenner, R. D., Bonate P. L., Tuman D., and Bull, R. J. (1995). Factors affecting species differences in the kinetics of metabolism of trichloroethylene. J. Toxicol. Env. Health 44, 435–447.
ten Berge, W. F., Zwart, A., and Appelman L. M. (1986). Concentration-time mortality response relationship of irritant and systemically acting vapours and gases. J. Hazard. Mater. 13, 301–309.
Tillar, R., Shafer, T. J., and Woodward, J. J. (2002). Toluene inhibits voltage-sensitive calcium channels expressed in pheochromocytoma cells. Neurochem. Int. 41, 391–397.
Tukey, J. W. (1977). Exploratory Data Analysis. Addison-Wesley, Reading, MA.
U.S. EPA (1999). National Air Toxics Program: The Integrated Urban Strategy. U.S. Environmental Protection Agency, Federal Register 64 (137), Monday, July 19, 38706–38740.
van Asperen, J., Rijcken, W. R. P., and Lammers, J. H. C. M. (2003). Application of physiologically based toxicokinetic modeling to study the impact of the exposure scenario on the toxicokinetics and the behavioural effects of toluene in rats. Toxicol Lett. 138, 51–62.
Vernon, R. J., and Ferguson, R. K. (1969). Effects of trichloroethylene on visual-motor performance. Arch. Environ. Health 18, 894–900.
Warren, D. A., Reigle T. G., Muralidhara, S., and Dallas, C. E. (1996). Schedule-controlled operant behavior of rats following oral administration of perchloroethylene: Time course and relationship to blood and brain solvent levels. J. Toxicol. Environ. Health 47, 345–362.
Wu, C., and Schaum, J. (2000). Exposure assessment of trichloroethylene. Environ. Health. Perspect. 108 (Suppl. 2), 359–363.
Zemon, V., and Ratliff, F. (1984). Intermodulation components of the visual evoked potential: Responses to lateral and superimposed stimuli. Biol. Cybern. 50, 401–408.(William K. Boyes, Mark Be)
Human Studies Division, National Health and Environmental Effects Research Laboratory, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, North Carolina 27711
ABSTRACT
The relationship between the concentration of trichloroethylene (TCE) in the brain and changes in brain function, indicated by the amplitude of steady-state pattern-elicited visual evoked potentials (VEP), was evaluated in Long-Evans rats. VEPs were recorded from visual cortex following stimulation of the eyes and, thus, reflect the function of the afferent visual pathway and, in broad terms, may be indicative of overall brain function. The concentration of TCE in the brain at the time of VEP testing (i.e., momentary brain concentration) was hypothesized to predict the amplitude of the VEP across a range of inhalation concentrations, both during and after exposure. Awake restrained rats were exposed to clean air or TCE in the following combinations of concentration and duration: 500 ppm (4 h), 1000 ppm (4 h), 2000 (2 h), 3000 ppm (1.3 h), 4000 ppm (1 h), and 5000 ppm (0.8 h). VEPs were recorded several times during the exposure session, and afterward for experimental sessions of less than 4 h total duration (i.e., concentrations from 2000 to 5000 ppm). The sample collection time for each VEP was about 1 min. Brain concentrations of TCE were predicted using a physiologically based pharmacokinetic (PBPK) model. VEP waveforms were submitted to spectral analysis, and the amplitude of the largest response component, occurring at twice the temporal stimulation rate (F2), was measured. Exposure to all air concentrations of TCE in the study reduced F2 amplitude. The reduction of F2 amplitude was proportional to momentary brain TCE concentration during and after exposure. A logistical function fit to the combined data from all exposure conditions described a statistically significant relationship with 95% confidence limits between brain TCE concentration and F2 amplitude. The results support the hypothesis that momentary brain concentration of TCE is an appropriate dose metric to describe the effects of acute TCE inhalation exposure on rat VEPs. The combination of the PBPK model predicting brain TCE concentration from the exposure conditions with the logistical function predicting F2 amplitude from the brain TCE concentration constitute a quantitative exposure-dose-response model describing an acute change in neurological function following exposure to an important hazardous air pollutant.
Key Words: trichloroethylene; visual evoked potentials.
INTRODUCTION
It is important to understand the relationships between exposure to toxic compounds, the concentrations of the compound in target tissues, and the production of adverse outcomes. Quantitative descriptions of such relationships can be referred to as "exposure-dose-response models." A key consideration in developing such models is determining the appropriate dose metric, in other words, the measure of dose to the target tissue which is best correlated with the adverse outcome of interest. The dose metric, once established, becomes a central link in the development and use of exposure-dose-response models. Exposure-dose-response models have numerous potential applications in risk assessment and risk management (e.g., Benignus et al., in press; Boyes et al., 2005a,b; Simmons, et al., 2005).
Trichloroethylene (TCE) is a volatile organic solvent that is used widely in a variety of industrial applications, such as metal cleaning and degreasing and microelectronic manufacture (Wu and Schaum, 2000). TCE is listed as a Hazardous Air Pollutant under the Clean Air Act, and the U.S. Environmental Protection Agency considers TCE to be one of 33 top priority urban air toxicants (U.S. EPA, 1999). TCE is found in many hazardous waste sites and may also contaminate water systems (Scott and Cogliano, 2000; Wu and Schaum, 2000). In 1998, the mean concentration of TCE in samples of outdoor ambient air from 115 monitors in the United States was 0.88 μg/m3 (range 0.01–3.9) (Wu and Schaum, 2000). The current workplace Permissible Exposure Limit established by the Occupational Safety and Health Administration is 100 ppm TWA (time weighted average), with a 300 ppm peak limit, intended to prevent acute narcosis. Acute exposure to high concentrations of TCE, such as might occur after spills or in other emergency situations, produces acute reversible deficits in the function of the central nervous system (Arlien-Sborg, 1992). These changes are expressed broadly as deficits of sensory, cognitive, and motor function in humans as well as rats.
When the concentrations of TCE in brain are predicted using a physiologically-based pharmacokinetic (PBPK) model, changes in the neurophysiology and behavior of rats can be predicted as a function of brain concentration of TCE at the time the outcomes are assessed (Boyes et al., 2000, 2003). The concentration of the compound in target tissue at the time of assessment can be referred to as the "momentary" concentration, as distinct from "peak" concentration, to emphasize the possibility that the magnitude of reversible effects tracks ascending and descending changes of target tissue concentration.
One set of techniques to evaluate visual function is visual evoked potentials (VEPs), involving presentation of visual stimuli to the eyes and recording responses elicited from electrodes over visual cortex (Herr and Boyes, 1995). VEPs have been used successfully to study the functional deficits produced by a number of neurotoxic substances (Dyer, 1985; Mattsson et al., 1992; Rebert, 1983). Pattern-elicited VEPs, in which the patterned visual stimulus is temporally modulated (e.g., on/off or contrast reversal) to elicit responses to the changing visual patterns, can be used in neurotoxicology studies of rats (Boyes and Dyer, 1983). When the stimulation rate is sufficiently high, the response takes on a steady-state characteristic of a continuous repetitive waveform that is amenable to analysis using Fourier techniques. The response elicited has primary spectral components at the temporal frequency of the stimulus modulation and its higher frequency harmonics (Boyes, 1994). In rats, the primary response component of steady-state pattern on-off VEPs occurs at twice the stimulation frequency, or "F2" (Boyes, 1994).
Acute TCE inhalation reduced the VEP F2 amplitude of rats in a manner that was related to the momentary TCE brain concentration, and was not related to the product of air TCE concentration (C) and duration (t) of exposure, or to the area under the curve of TCE concentration in the brain (Boyes et al., 2003). The nature of the relationship between momentary brain TCE concentration and decreased VEP amplitude was assessed previously with only a limited number of exposure conditions, making it difficult to develop quantitative relationships between target tissue dose and outcome. In addition, measurements were taken in previous experiments only during inhalation exposure. The testing of outcomes after termination of exposure, while TCE is being eliminated from brain tissue, enables a test of the generality of the hypothesis that momentary brain concentration predicts outcome after, as well as during, external exposure.
The current experiments evaluated visual function of rats, as measured by the amplitude of the F2 VEP component, during and after acute inhalation exposure to TCE. A range of external air concentrations and testing times was evaluated, and VEP measurements were made both during and after the termination of external exposure. It was hypothesized that momentary concentration of TCE in the brain during VEP testing would accurately predict the reduction of VEP amplitude irrespective of the atmospheric TCE concentration, exposure duration, or time of testing (during or after exposure). Brain concentrations were estimated using a PBPK model designed and evaluated to predict tissue concentrations in Long-Evans rats (Simmons et al., 2002). In addition, a logistic function was derived to model quantitatively the relationship between TCE concentration in the brain and VEP amplitude. The logistic curve was selected as the dose-effect curve model, because its shape has been found to fit empirical data frequently (Corso, 1967), and it allows a variety of outcomes to be compared on a common proportional scale ranging from 0 to 1 (e.g., Benignus et al., 1998, 2005).
MATERIALS AND METHODS
The experimental methods were similar in many respects to those described in detail elsewhere (Boyes et al., 2003) and, therefore, are described here more briefly.
Test compound.
Spectrophotometric grade 1,1,2-trichloroethylene (99.5% pure) (CAS # 79-01-6) was obtained from Aldrich Chemical Co., St. Louis, MO.
Test animals.
Male Long-Evans rats (LE) (Crl: (LE)BR) were obtained from Charles River Laboratories (Raleigh, NC) at approximately 60 days of age. The animal colony had an ambient temperature and relative humidity of 22 ± 2°C and 50 ± 10%, respectively, and a 12:12 h light:dark cycle (lights on at 6:00 A.M.). The illumination during the light cycle ranged from approximately 190 lux (bottom shelf of the animal rack) to 390 lux (top shelf of the animal rack). The rats were housed individually in polycarbonate cages with kiln-dried pine chip bedding (Northeastern Products, Warrensburg, NY), with ad libitum access to tap water and rat chow (LabDiet PMI #5001, Richmond, IN). All aspects of the care and treatment of laboratory animals were approved by the Institutional Laboratory Animal Care and Use Committee of the National Health and Environmental Effects Research Laboratory and were in compliance with applicable federal guidelines for laboratory animal experimentation.
Surgery.
After approximately 1 week to acclimate to the animal colony, rats were anesthetized (sodium pentobarbital; 50 mg/kg ip), placed in a standard stereotaxic device, and prepared for cranial surgery. Recording electrodes, constructed from stainless steel screws (00-90 x 1/16 in.) soldered to Nichrome wires, were implanted into the skull, epidurally, in the following locations: 1 mm anterior to lambda and 4 mm left of the midline overlying the left primary visual cortex; and 2 mm anterior to bregma and 2 mm lateral left and right to midline for ground and reference electrodes, respectively. The electrode assembly was encased in dental acrylic and the wound sutured. Approximately 1 week was allowed for surgical recovery before electrophysiological testing. Details of this preparation have been described previously (Boyes et al., 2003; Herr et al., 1992).
Inhalation exposure.
A J-tube inhalation system was used to generate vapors of TCE and deliver them into a head-only inhalation chamber (Boyes et al., 2003; McGee et al., 1994). The concentration of TCE in the exposure chamber was monitored online using an infrared (IR) spectrophotometer (Miran, 1A, Foxboro Co, East Bridgewater, MA). The head-only exposure chamber (10 x 10 x 17 cm) was constructed of stainless steel, with the exception of a glass front plate to allow the rat to view the video monitor displaying visual stimulus outside of the chamber and a glass side plate allowing observation of the experimental subject. Awake (nonanaesthetized) rats were restrained in a plastic cone with the eyes, nose and ears exposed. When placed in the exposure chamber, a gas-tight latex seal was established around the upper torso of the rat to isolate the atmosphere of the head-only exposure chamber. Body temperature was monitored during experimental sessions. A flexible cable entering the head-only exposure chamber through a gas-tight port was connected to the rat's headset to allow recording of electrophysiological signals.
Visual stimuli.
A video monitor for presentation of visual stimuli was located approximately 15 cm from the rat's eyes outside the glass face of the exposure chamber. Visual stimuli were generated with a computer-based system described elsewhere (Boyes et al., 2003; Hamm et al., 2000). Stimulus brightness and contrast were calibrated to achieve a linear relationship between input signal voltage and screen luminance. The entire assembly was covered in a black cloth so that the primary source of light for the rat was the video screen.
The visual stimulus pattern was a sinusoidal vertical grating of 0.16 cycles per degree visual angle, selected because this value of spatial frequency is approximately at the peak of the contrast sensitivity function of pigmented rats (Birch and Jacobs, 1979), and peak contrast of 60%, selected because it provides a strong visual response under baseline conditions and yet lies well within the linear input/output range of the stimulus monitor. Mean stimulus illuminance was approximately 10 cd/m2, measured at the position of the rat's eyes. The pattern was modulated in an on/off manner with a sinusoidal temporal profile at approximately 5 Hz. Further details are presented in Boyes et al. (2003).
Evoked potential recording.
Electrophysiological potentials were amplified, band-pass filtered (1–100 Hz; roll off = 12 dB/octave), and sampled in one sec epochs (512 data points per epoch) using a Pathfinder II signal averager (Nicolet Biomedical, Madison, WI), so that each sample epoch contained approximately five cycles of the stimulus temporal modulation. Each VEP was constructed from the average of 50 1-s epochs, a process that took approximately 1 min for the collection of data for each waveform. To provide an estimate of recording noise level, "plus-minus" averages in which alternate single sweeps were inverted in polarity before adding to the signal averager were collected simultaneously. In this way, any reliable evoked potential signal would average toward zero volts, and the result would provide an estimate of the inherent system-wide noise which, being random, should approximate that of the random noise in the signal channel. The resulting waveforms were submitted to spectral analysis (SPECTM Program, Nicolet Biomedical, Madison, WI). Spectral amplitude of both signal and plus-minus waveforms was measured at twice the stimulus rate (F2). The F2 amplitude of the signal was the primary dependent variable of the initial data analysis, since F2 is the largest response component in rats presented this stimulus paradigm (Boyes, 1994), and previous experiments indicated that exposure to TCE reduced F2 amplitude (Boyes et al., 2003). For subsequent quantitative modeling of the relationship between F2 amplitude and brain TCE concentration, F2 amplitudes were normalized using both baseline recording levels and the plus-minus noise data as described below.
PBPK model.
Exposure concentrations and time points for VEP sampling were selected using an initial version of the PBPK model (Boyes et al., 2000) with the goal of yielding different combinations of exposure concentrations and durations that resulted in equivalent blood TCE concentrations. After the experiments were completed, an improved version of the PBPK model became available (Simmons et al., 2002). The tissue concentrations predicted from the revised model were compared to TCE tissue concentrations measured in brain, blood, liver, and fat of exposed LE rats, and the model was shown to have good predictive accuracy (Simmons et al., 2002). In addition, a linear relationship was observed between blood and brain TCE concentrations over a range between 0.1 and 100 mg/1 (Boyes et al., 2003). In the current report, we evaluate the VEP results as a function of predicted brain TCE concentration using the improved version of the PBPK model (Simmons et al., 2002).
Experimental design.
Rats (n = 8–10/group) were exposed to TCE in concentrations of 500, 1000, 2000, 3000, 4000, or 5000 ppm. Total exposure durations differed for each concentration to provide total concentration–duration products of 4000 ppm-h, with the exception of 500 ppm, for which the total duration of exposure was 4 h. The times of VEP sampling and the associated estimates of brain TCE concentrations are depicted in Figure 1. VEP sample times are also presented in Table 1. Sample times greater than 4 h were excluded due to decrements in the VEP amplitude, presumably from fatigue or adaptation, observed in recordings from control animals. VEPs were recorded repeatedly from each rat during and after exposure to TCE concentrations between 2000 and 5000 ppm. For rats exposed to 500 or 1000 ppm, VEPs are presented only during the exposure phase of the experiment which lasted for 4 h. In addition, VEPs were sampled from rats exposed to only clean air at times matching those of the TCE-exposed rats. Two or three such control rats were run at matching sample times for each of the TCE concentrations (overall n = 13), with the exception of 500 ppm, for which an additional 8 matched controls were run.
Statistical Analysis and Curve Fitting
Data.
Data from one rat exposed to 5000 ppm were eliminated due to high levels of electrical noise during the recording session. To fit a logistic dose-effect function, data from all the experiments were pooled into a single data base. This data set consisted of 489 observations on 79 subjects (58 exposed to TCE and 21 clean air controls), including data from baseline (preexposure), several during-exposure recordings and, for 2000–5000 ppm TCE, several postexposure recordings as well. Measurements of the plus-minus average, as described above, were considered as measures of background noise. The F2 amplitude values were corrected for background noise by subtracting the mean F2 amplitudes from plus-minus records. Next, each F2 amplitude measure was converted to a proportion-of-baseline score by dividing by the mean preexposure baseline amplitude. These transformed F2 data were called normalized baseline-adjusted data, and have the properties that the baseline (no-TCE value) value of F2 is 1.0, and the maximum effect value (F2 entirely suppressed) is 0.0.
The baseline-adjusted F2 data were screened for outliers using a box-and-whisker plot (Tukey, 1977; Systat Inc. 1990). Thirteen observations qualified as outliers (>3 times the interquartile range), seven of which came from a single subject. Logistic functions were fit, as described below, to the entire data set, and also to the data set after removal of outliers, including all the data from the one subject with seven outliers. The logistic function showed a similar fit with or without the inclusion of outliers, with derived parameters differing by less than 0.15%. The outlier-trimmed data showed a greater resemblance to the original nontransformed data, and therefore the logistical function based on that analysis is presented here.
Curve fitting.
The adjusted F2 amplitude data were used as the dependent variable in fitting the dose-effect curve. The data contained a variable number of repeated measures for each rat. To account for the correlations between the variable numbers of repeated measures, a mixed-model curve fitting procedure was used to fit the dose-effect curve (SAS Proc NLMIXED). The independent variable for these data was the concentration of TCE in the brain as estimated for each exposure scenario using the PBPK model.
A logistic model, which has been used previously for similar types of data (Benignus et al., 1998, 2005; Corso, 1967), was fit empirically to the data from the TCE-exposed rats. The logistic curve has a variable ogive-like shape but unlike the ogive, the upper and lower asymptotes are independently variable, thus permitting a better fit to some data. The lower and upper asymptotes are zero and one, respectively. There are two parameters, 1 and 2, which determine the location and slope, respectively, of the curve. The asymptotes make it necessary to transform raw outcome data so that the resultant data points range in value from zero to one.
Other nonlinear curves might have been selected, but do not have some of the desirable characteristics of the logistic. Quadratic functions can be made to fit well within the range of the data, but behave poorly outside that range, thus making quadratic functions a poor choice for extrapolation. The ogive does not permit asymmetry about mid value of dose. Logarithmic and power functions do not have convenient (or any) asymptotes. Thus, the logistic was selected as the preferred model.
The logistic model has the form
(1)
in which E is the measure of the effect of TCE (proportion of baseline VEP amplitude), e is the base of natural logarithms, D is the estimated momentary concentration of TCE in the brain, ln(D) is the natural logarithm of D, and 1 and 2 are empirical coefficients which determine the location and slope, respectively, of the fitted curve. The fitted curve and its 95% confidence limits were calculated as a function of predicted brain TCE concentration.
Residual analysis.
To test the possibility that either the air TCE concentration, the duration of exposure, or their product (C x t) might influence F2 amplitudes beyond the effect of the internal dose, an analysis was performed of the residuals from fitting the dose-effect curve. The residuals are the absolute differences between the observed data and the line of best fit. In the ANOVA (SAS Proc MIXED) the residuals from the fitted curve were treated as the dependent variable, with duration and concentration and their interaction as effects. Step-down tests employed t-tests. A significant effect would indicate that concentration or duration, or the C x t product, had affected F2 amplitudes in addition to the influence of momentary brain concentration.
RESULTS
The target concentrations of TCE in the headspace air of the exposure chamber were 500, 1000, 2000, 3000, 4000, and 5000 ppm. The TCE concentration measured in the head-only exposure chambers for each experimental condition was 511 ± 5.8, 938.5 ± 13.6, 1966.2 ± 60.2, 2988.4 ± 86.9, 3970 ± 90.9, and 4870.5 ± 63.9 (mean ± standard deviation), respectively.
Visual evoked potential waveforms averaged across the group exposed to 3000 ppm TCE are presented in Figure 2, along with the spectral analysis of each group average waveform. The group average waveforms from the other exposure concentrations resembled those from the 3000-ppm group and are not presented. Before the onset of exposure to TCE, the VEP waveforms showed 10 clear cycles of the response and manifested a steady-state evoked potential profile (Boyes, 1994). The spectral analysis of the baseline waveforms showed predominant response amplitude at F2. After the onset of exposure to TCE, VEP waveforms decreased in amplitude, and the strong steady-state response characteristics diminished. This was accompanied by a decrease in amplitude of the F2 spectral component. Spectral amplitude at the stimulus modulation rate (F1) was more variable, and not altered systematically by exposure to TCE. After termination of exposure, the VEP waveform gradually began to recover a resemblance to the original steady-state characteristics, and the F2 amplitude of the spectra increased.
The nature of the change in VEPs caused by TCE exposure is illustrated in Figure 3, where two of the group average waveforms from Figure 2 were superimposed and then replotted. The two waveforms selected were from baseline (time = 0) and from the time of maximum F2 suppression immediately after termination of exposure (time = 1.33 h). Inspection of these two waveforms reveals that every other cycle of the evoked response was suppressed by TCE exposure, with relatively little change in the intervening cycles.
The amplitude of the F2 component was examined as a function of estimated brain concentration during and/or after exposure. Figure 4 presents F2 amplitude during exposure for all exposure conditions, and after exposure for groups exposed to 2000, 3000, 4000, or 5000 ppm TCE. The amplitude of F2 was reduced by TCE exposure. There was some variation among the groups in the baseline F2 amplitudes prior to exposure, which is occasionally observed for different sets of rats tested at different times. The amount of F2 amplitude reduction observed during or after exposure to any given concentration was equivalent when expressed as a function of the brain TCE concentration expected at that sampling time.
Previously it was shown that for a C x t product of 4000 ppm-h, that the C x t product did not adequately describe the effect of TCE exposure on F2 amplitude, whereas peak brain TCE concentration did (Boyes et al., 2003). Data from the current experiment are presented as a function of the momentary brain TCE concentration predicted from the PBPK model, and also as a function of the C x t product for the exposure occurring up to the point of VEP recording (Fig. 5). The results indicate that the data are better described by momentary brain concentration than by C x t product.
In order to describe quantitatively the relationship between brain TCE concentration and F2 amplitude, the data from all exposure conditions were combined, normalized, and plotted as a fitted logistical dose-effect curve (Fig. 6). Any apparent increase in the variability of the data in Figure 6 over that of nontransformed values of Figure 5 may be attributed to variability of the baseline or noise level values which were used in calculating the F2 amplitude transformations. The fit of the dose-effect curve to the data was statistically significant, giving 1 = 4.48, t = 7.86, p < 0.0001 and 2 = –1.50, t = –9.16, df = 76, p < 0.0001. Table 2 gives the covariance matrix for the parameter estimates. The mean estimate derived from the logistic curve for the brain concentration causing a 10% amplitude deficit (ED10) was 4.5 mg TCE/l, and causing a 50% deficit (ED50) was 20.0 mg TCE/l.
Analysis of the residuals from the fitted curve (ANOVA) indicated that the duration of exposure, the concentration, or the C x t combination were not statistically significant (p-values were 0.602, 0.367, and 0.337, respectively). Thus, chamber air concentration, exposure duration, or the C x t product had no significant explanatory power beyond that conveyed by momentary brain TCE concentration, which was significantly linked to changes in F2 amplitude.
DISCUSSION
Inhalation exposure to TCE reduced the amplitude of the F2 component of pattern VEPs of Long-Evans rats in a manner that was associated with the momentary concentration of TCE in the brain. The limited set of data which was available previously suggested a linear relationship between brain TCE concentration and change in F2 amplitude (Boyes et al., 2003). The current data set is more extensive and clearly shows a curvilinear relationship that was described by a logistic function. The parameters of the logistic curve may be useful in predicting the level of F2 amplitude impairment caused by a variety of TCE exposure conditions.
Momentary brain concentration of TCE appeared to determine the change in F2 amplitude. This outcome extends previous results that momentary brain TCE concentration was a better predictor of VEP F2 amplitude than was the product of air concentration and duration, or the area under the curve (AUC) reflecting the total cumulative dose (Boyes et al., 2003). In the current study, the residuals of the logistic function associated with TCE inhaled-air concentration, the exposure duration, or the C x t product were not statistically significant, indicating that these factors had no explanatory power beyond that of the momentary concentration of TCE in brain. The amplitude of F2 is a reflection of massed firing of neurons in visual cortex, responding to afferent visual input originating in the retina. To the extent that F2 amplitude reduction reflects general neurological impairment, these results suggest that momentary brain TCE concentration is an appropriate measure of internal dose (i.e., a dose metric) for evaluating acute neurological effects of exposure to TCE.
Acute exposure to volatile organic compounds such as TCE does not necessarily have selective actions restricted to the visual system, but rather suppress neuronal activity broadly throughout the nervous system. Human volunteer subjects exposed to TCE vapors while performing a visual/motor performance task showed no observable adverse effects (NOAEL) at 300 ppm for 2 h (Vernon and Ferguson, 1969). In the same study, the lowest dose producing significant impairment of human visual/motor performance (LOAEL) was 1000 ppm for 2 h (Vernon and Ferguson, 1969). Through PBPK modeling, it was estimated that the human blood TCE concentration associated with this NOAEL was approximately 4.8 mg/l, and with the LOAEL was about 18 mg/l (Boyes et al., 2005a). Considering that in rats (and presumably humans) there is approximately a 1:1 correspondence between TCE concentration in blood and brain (Boyes et al., 2003), there is an interesting similarity between the tissue dose of the human neurobehavioral NOAEL (4.8 mg/l) and the rat VEP ED10 measured here (4.5 mg/l), and also between the human neurobehavioral LOAEL (18 mg/l) and the rat VEP ED50 (20 mg/l). These results suggest that the rat VEP is sensitive to acute TCE exposure at concentrations similar to those of human neurobehavioral performance.
The relationships between the target tissue concentration of organic solvents and neurotoxic outcomes have been of interest. Bruckner and Peterson (1981) showed an association between brain toluene concentration and CNS depression in mice both during and after exposure to either 4000 ppm for 3 h or 10,600 ppm for 10 min. Single inhalation exposures to TCE (Kishi et al., 1993) or toluene (Kishi et al., 1988) altered shock avoidance behavior of rats in a manner consistent with solvent concentration in the blood (TCE or toluene) or brain (toluene). In a meta-analysis of previously published data, Benignus et al. (1998) showed that rat avoidance behavior and human choice reaction time could be well described by blood toluene level at the time of behavioral assessment. The momentary concentration in blood was also predictive of TCE-induced deficits in performance of a signal detection task (Boyes et al., 2000). In contrast, Warren et al. (1996) could find no consistent relationship between response rate on a fixed-ratio operant task and the time course or concentration of perchloroethylene in brains of repeatedly treated rats. Van Asperen et al. (2003) compared the visual discrimination performance of rats following inhalation exposure to a constant concentration or to repeated spiked concentrations of toluene. Performance of the task, as indicated by the percentage of intertrial interval responding, was not related to the concentration of toluene in brain at the end of the exposure session. Similarly, it is unlikely that brain solvent concentration accounts for development of behavioral tolerance following repeated exposure to TCE (Bushnell and Oshiro, 2000; Oshiro et al., 2001). Thus, it appears that momentary brain solvent concentration is generally a good predictor of simultaneous behavioral outcomes, in particular for single exposure sessions, but additional factors may be necessary to explain behavioral consequences of repeated exposures or complex exposure protocols.
There is often a need in risk assessment to estimate the exposure concentrations associated with a particular outcome for exposure durations which were not assessed experimentally. A common practice in doing so is to apply Haber's rule, that C x t = k, or the ten Berge equation Cn x t = k (ten Berge, et al., 1986), to adjust the critical air concentration for a desired exposure duration. We have shown previously that using a PBPK model to predict the target tissue concentration is preferable to Haber's or ten Berge's duration adjustments (Boyes et al., 2003, 2005a: Simmons et al., 2005). The analysis presented here extends this principle to include prediction of effects lingering after the termination of exposure. The Haber or ten Berge approaches cannot account for effects after termination of exposure, when the external air concentration is theoretically zero and the C x t (or Cn x t) product would also be zero, leading to a prediction of no effect. Alternatively, the Haber or ten Berge products could be interpreted as the total cumulative exposure during the inhalation phase of the experiment, in which case the values never decrease after termination of exposure and cannot account for a gradual return to baseline. When using target tissue dose as the metric, however, the level of impairment can be predicted over the course of elimination of the compound according to the amount of residual toxic compound in the target tissue.
One issue of interest is whether the neurological impairments from exposure to TCE are caused by the parent compound by one or more of the potentially active metabolites. There are many identified metabolites of TCE (Lash et al., 2000), at least two of which, trichloroethanol (TCOH) and chloral hydrate, likely have neuroactive properties (Krasowski et al., 1998; Peoples and Weight, 1998). We argued previously that the acute effects of TCE on F2 amplitude were probably caused by TCE and not a metabolite because clear dose-response relationships were observed above the dose of metabolic saturation, about 400 ppm, where no additional metabolites would be formed with increasingly higher doses (Boyes et al., 2003). The current data provide additional evidence, albeit indirect, that the parent compound is the active species. Very similar effects were seen during and after exposure when plotted as a function of brain TCE concentration (Figure 4). The time course of TCE in rat tissues is shorter, by various amounts, than that of its principal metabolites including trichloroacetic acid, TCOH, and chloral hydrate (Clewell, et al., 2000; Fisher et al., 1991; Larson and Bull, 1992; Prout et al., 1985; Templin et al., 1995). If a metabolite were the active species, then the F2 amplitude curves during and after exposure would have been displaced in proportion to the time course of the active metabolite formation and clearance. This was not observed, supporting the case that the effects observed were attributable to the TCE parent compound.
The visual system is thought to be comprised of multiple parallel systems, one aspect of which concerns whether the responses of the system show linear and nonlinear relationships to the physical and temporal modulation of the visual stimulus (Regan, 1989). The response patterns of visual cells characterized as "linear" include firing rates that are the linear sum of the responses to independent stimulation of the excitatory and inhibitory portions of their receptive fields (Enroth-Cugell and Robson, 1966). Linear responses also include sustained firing to continuous stimulation, and responding to modulated visual stimuli at the temporal frequency (F1) of stimulus modulation (Enroth-Cugell and Robson, 1966; Hochstein and Shapley, 1976; Kaplan and Shapley, 1982). Visual responses characterized as "nonlinear," on the other hand, include firing rates that are not the linear sum of stimulating the excitatory and inhibitory components of the receptive fields. In addition, nonlinear responses are characterized by firing in a transient or phasic fashion to continuous visual stimuli, and responding to modulated visual stimuli primarily at twice the rate (F2) of temporal stimulation (Enroth-Cugell and Robson, 1966; Hochstein and Shapley, 1976; Kaplan and Shapley 1982; Lennie and Perry, 1981). Depending on the stimulus conditions, individual cells may respond at both F1 and F2 (e.g., Lennie and Perry, 1981). The linear and nonlinear response profiles originate in retinal bipolar cells, are transmitted to retinal ganglion cells, and are passed on through the visual pathway (Awatramani and Slaughter, 2000). The ratio of the response amplitude at second (F2) to that at the first (F1) harmonic frequency has been used as an index of nonlinearity (Hochstein and Shapley, 1976). Linear and nonlinear properties also may be observed in the integrated responses of the visual system (e.g., Regan and Regan, 1989; Zemon and Radcliff, 1984). The current results suggest that TCE exposure suppresses nonlinear activity patterns in the visual system.
The visual system also contains On and Off visual pathways, which respond differentially to ionotropic and metabotropic glutamate receptor agonists and antagonists (Awatramani and Slaughter, 2000; Schiller, 1984; Slaughter and Miller, 1981). At a systemic level, selective blocking of either the On or the Off response of the transient (nonlinear) visual system might be expected to reduce VEP F2 amplitude and produce a response profile resembling that of Figure 3, in which every other cycle of the response to an On-Off stimulus was diminished. Unfortunately, it is not clear whether the On or the Off cycle was suppressed, since the phase lag (temporal delay) between the stimulus modulation and the response is ambiguous in steady-state recordings (Regan, 1989). Toluene, a volatile organic compound with acute sedative actions similar to TCE (Arlien-Sborg, 1992), blocks the actions of glutamate at ionotropic NMDA receptors (Cruz et al., 1998, 2000), but to our knowledge, the actions of volatile organic compounds on metabotropic glutamate receptors have not been evaluated to date (Bushnell et al., 2005). Other potential target sites include GABAA (Beckstead et al., 2000), nicotinic acetylcholine receptors (Bale et al., 2002), and voltage sensitive calcium channels (Tillar et al., 2002), which could all conceivably alter the temporal response profiles of visual neurons.
In summary, the results of these experiments establish a quantitative relationship between the concentration of TCE in brain tissue and changes in the amplitude of VEPs recorded from awake rats. Similarly, the PBPK model can be used to predict brain TCE concentrations for exposure conditions of interest. Together these components enable the quantitative prediction of graded decrement in neurological function for acute exposure scenarios involving inhalation of TCE vapors.
NOTES
This manuscript has been reviewed by the National Health and Environmental Effects Research Laboratory, U.S. EPA and approved for publication. Mention of trade names and commercial products does not constitute endorsement or recommendation for use.
ACKNOWLEDGMENTS
The authors thank Philip Bushnell, David Herr, Ambuja Bale, Linda Birnbaum, Elaina Kenyon, and Jane Caldwell for comments on an earlier version of the manuscript. Conflict of interest: none declared.
REFERENCES
Arlien-Sborg, P. (1992). Solvent Neurotoxicity. CRC Press, Boca Raton, FL.
Awatramani, G. B., and Slaughter, M. M. (2000). Origin of transient and sustained responses in ganglion cells of the retina. J. Neurosci. 20, 7078–7095.
Bale, A. S., Smothers, C. T., and Woodward, J. J. (2002). Inhibition of neuronal nicotinic acetylcholine receptors by the abused solvent, toluene. Br. J. Pharmacol. 137, 375–383.
Beckstead, M. J., Weiner, J. L., Eger, E. I., Gong, D. H., and Mihic, S. J. (2000). Glycine and gamma-aminobutyric acid(A) receptor function is enhanced by inhaled drugs of abuse. Mol. Pharmacol. 57, 1199–1205.
Benignus, V. A., Boyes, W. K., and Bushnell, P. (1998). A dosimetric analysis of acute toluene exposure in rats and humans. Toxicol. Sci. 43, 186–195.
Benignus, V. A., Bushnell, P. J., and Boyes, W. K. (2005). Toward cost-benefit analysis of acute behavioral effects of toluene in humans. Risk Anal. 25, 447–456.
Birch, D., and Jacobs, G. H. (1979). Spatial contrast sensitivity in albino and pigmented rats. Vision Res. 19, 933–937.
Boyes, W. K. (1994). Rat and human evoked potentials and the predictability of human neurotoxicity from rat data. Neurotoxicology 15, 569–578.
Boyes, W. K., Bercegeay, M., Ali, J., Krantz, T., McGee, J., Evans, M., Raymer, J. H., Bushnell P. J., and Simmons, J. E. (2003). Dose-based duration adjustments for the effects of inhaled trichloroethylene on rat visual function. Toxicol. Sci. 76(1), 121–130.
Boyes, W. K., Bushnell, P. J., Crofton, K. M., Evans, M., and Simmons, J. E. (2000). Neurotoxic and pharmacokinetic responses to trichloroethylene as a function of exposure scenario. Environ. Health. Perspect. 108 (Suppl. 2), 317–322.
Boyes, W. K., and Dyer, R. S. (1983). Pattern reversal visual evoked potentials in awake rats. Brain Res. Bull. 10, 817–823.
Boyes, W. K., Evans, M. V., Eklund, C., Janssen P., and Simmons, J. E. (2005a). Duration adjustment of acute exposure guideline level values for trichloroethylene using a physiologically-based pharmacokinetic model. Risk Anal. (in press).
Boyes, W. K., Simmons, J. E., Eklund C., Benignus, V. A., Janssen, P., Bushnell, P. J. (2005b). Applications of dosimetry modeling to assessment of neurotoxic risk. Environ. Toxicol. Pharmacol. 19, 599–605.
Bruckner J. V., and Peterson, R. G. (1981). Evaluation of toluene and acetone inhalant abuse. I. Pharmacology and pharmacodynamics. Toxicol. Appl. Pharmacol. 61, 27–38.
Bushnell, P. J., and Oshiro, W. M. (2000). Behavioral components of tolerance to repeated inhalation of trichloroethylene (TCE) in rats. Neurotoxicol. Teratol. 22, 221–229.
Bushnell, P. J., Shafer, T. J., Bale, A. S., Boyes, W. K., Simmons, J. E., Eklund, C., and Jackson, T. L. (2005). Developing an exposure-dose-response model for the acute neurotoxicity of organic solvents: Overview and progress on in vitro models and dosimetry. Environ. Toxicol. Pharmacol. 19, 607–614.
Corso, J. W. (1967). The Experimental Psychology of Sensory Behavior. Holt-Rinehart and Winston, New York.
Clewell, H. J., III., Gentry, P. R., Covington, T. R., and Gearhart, J. R. (2000). Development of a physiologically based pharmacokinetic model of trichloroethylene and its metabolites for use in risk assessment. Environ. Health Perspect. 108, 283–305.
Cruz, S. L., Balster, R. L., and Woodward, J. J. (2000). Effects of volatile solvents on recombinant N-methyl-d-aspartate receptors expressed in Xenopus oocytes. Br. J. Pharmacol. 131, 1303–1308.
Cruz, S. L., Mirshahi, T., Thomas, B., Balster, R. L., and Woodward, J. J. (1998). Effects of the abused solvent toluene on recombinant N-methyl-d-aspartate and non-N-methyl-d-aspartate receptors expressed in Xenopus oocytes. J. Pharmacol. Exp. Ther. 286, 334–340.
Dyer, R. S. (1985). The use of sensory evoked potentials in toxicology. Fundam. Appl. Toxicol. 5, 24–40.
Enroth-Cugell, C., and Robson, J. G. (1966). The contrast sensitivity of retinal ganglion cells of the cat. J. Physiol. 187, 517–552.
Fisher, J. W., Gargas, M. L., Allen, B. C., and Andersen, M. E. (1991). Physiologically based pharmacokinetic modeling with trichloroethylene and its metabolite, trichloroacetic acid, in the rat and mouse. Toxicol. Appl. Pharmacol. 109, 183–195.
Hamm, C. W., Ali, J. S., and Herr, D. W. (2000). A system for simultaneous multiple subject, multiple stimulus modality, and multiple channel collection and analysis of sensory evoked potentials. J. Neurosci. Meth. 102, 95–108.
Herr, D. W., Boyes, W. K., and Dyer, R. S. (1992). Alterations in flash and pattern reversal evoked potentials after acute or repeated administration of carbon disulfide (CS2). Fundam. Appl. Toxicol. 18, 328–342.
Herr, D. W., and Boyes, W. K. (1995). Electrophysiological analysis of complex brain systems: Sensory evoked potentials and their generators. In Neurotoxicology: Approaches and Methods (L. W. Chang and W. Slikker, Eds.), pp. 205–221. Academic Press, New York.
Hochstein, S., and Shapley, R. M. (1976). Quantitative analysis of retinal ganglion cell classifications. J. Physiol. 262, 237–264.
Kaplan, E., and Shapley, R. M. (1982). X and Y cells in the lateral geniculate nucleus of Macaque monkeys. J. Physiol. 330, 125–143.
Kishi, R., Harabuchi, I., Ikeda, T., Yokota, H., and Miyake, H. (1988). Neurobehavioral effects and pharmacokinetics of toluene in rats and their relevance to man. Brit. J. Indust. Med. 45, 396–408.
Kishi, R., Harabuchi, I., Ikeda, T., Katakura, Y., and Miyake, H. (1993). Acute effects of trichloroethylene on blood concentrations and performance decrements in rats and their relevance to humans. Brit. J. Indust. Med. 50, 470–480.
Krasowski, M. D., Fin, S. E., Ye, Q., and Harrison, N. L. (1998). Trichloroethanol modulation of recombinant GABAA, glycine and GABA 1 receptors. J. Pharmacol. Exp. Therap. 284, 934–942.
Larson, J. L., and Bull, R. J. (1992). Species differences in metabolism of trichloroethylene to the carcinogenic metabolites trichloroacetate and dichloroeacetate. Toxicol. Appl. Pharmacol. 115, 278–285.
Lash, L. H., Fisher, J. W., Lipscomb, J. C., and Parker, J. C. (2000). Metabolism of trichloroethylene. Environ. Health Perspect. 108 (Suppl. 2), 177–200.
Lennie. P., and Perry, V. H. (1981). Spatial contrast sensitivity of cells in the lateral geniculate nucleus of the rat. J. Physiol. 315, 69–79.
Mattsson, J. L., Boyes, W. K., and Ross, J. F. (1992). Incorporating evoked potentials into neurotoxicity test schemes. In Target Organ Toxicology Series: Neurotoxicity (H. A. Wilson and C. L. Mitchell, Eds.), pp.125–145. Raven Press.
McGee, J. K., Evansky, P. A., Terrell, D., Walsh, L. C., Davies, D. W., and Costa, D. L. (1994). A new vapor and gas test atmosphere generator with broad concentration and flow output capabilities. In Measurement of Toxic and Related Air Pollutants: Proceedings of the 1994 U.S. EPA / AWMA International Symposium, pp. 1011–1016. Air and Waste Management Association Publication No. VIP-39.
Oshiro, W. M., Krantz, Q. T., and Bushnell, P. J. (2001). Characterizing tolerance to trichloroethylene (TCE): Effects of repeated inhalation of TCE on performance of a signal detection task in rats. Neurotoxicol. Teratol. 23, 617–628.
Peoples, R. W., and Weight, F. F. (1998). Inhibition of excitatory amino acid-activated currents by trichloroethanol and trifluoroethanol in mouse hippocampal neurones. Br. J. Pharmacol. 124, 1159–1164.
Prout, M. S., Proven, M. W., and Green, T. (1985). Species differences in response to trichloroethylene. Toxicol. Appl. Pharmacol. 79, 389–400.
Rebert, C. S. (1983). Multisensory evoked potentials in experimental and applied neurotoxicology. Neurobehav. Toxicol. Teratol. 5, 659–671.
Regan, D. (1989). Human Brain Electrophysiology: Evoked Potentials and Evoked Magnetic Fields in Science and Medicine. Elsevier, New York.
Regan, M. P., and Regan, D. (1989). Objective investigation of visual function using a nondestructive zoom-FFT technique for evoked potential analysis. Can. J. Neurol. Sci. 16, 168–179.
Scott, C. S., and Cogliano, V. J. (2000). Trichloroethylene health risks—state of the science. Environ. Health Perspect. 108 (Suppl. 2), 159–160.
Schiller, P. H. (1984). The connections of the retina On and Off pathways to the lateral geniculate nucleus of the monkey. Vision Res. 24, 923–932.
Simmons, J. E., Boyes, W. K., Bushnell, P. J., Raymer, J. H., Limsakun, T., McDonald, A., Sey, Y. M., and Evans, M. V. (2002). A physiologically-based pharmacokinetic model for trichloroethylene in the male Long-Evans rat. Toxicol. Sci. 69, 3–15.
Simmons, J. E., Evans, M. V., and Boyes, W. K. (2005). Moving from external exposure concentration to internal dose: Duration extrapolation based on physiologically-based pharmacokinetic derived estimates of internal dose. J. Toxicol. Environ. Health A 68, 927–950.
Slaughter, M. M., and Miller, R. F. (1981). 2-Amino-4-phosphonobutyric acid: A new pharmacological tool for retina research. Science 211, 182–185.
Templin, M. V., Stevens, D. K., Stenner, R. D., Bonate P. L., Tuman D., and Bull, R. J. (1995). Factors affecting species differences in the kinetics of metabolism of trichloroethylene. J. Toxicol. Env. Health 44, 435–447.
ten Berge, W. F., Zwart, A., and Appelman L. M. (1986). Concentration-time mortality response relationship of irritant and systemically acting vapours and gases. J. Hazard. Mater. 13, 301–309.
Tillar, R., Shafer, T. J., and Woodward, J. J. (2002). Toluene inhibits voltage-sensitive calcium channels expressed in pheochromocytoma cells. Neurochem. Int. 41, 391–397.
Tukey, J. W. (1977). Exploratory Data Analysis. Addison-Wesley, Reading, MA.
U.S. EPA (1999). National Air Toxics Program: The Integrated Urban Strategy. U.S. Environmental Protection Agency, Federal Register 64 (137), Monday, July 19, 38706–38740.
van Asperen, J., Rijcken, W. R. P., and Lammers, J. H. C. M. (2003). Application of physiologically based toxicokinetic modeling to study the impact of the exposure scenario on the toxicokinetics and the behavioural effects of toluene in rats. Toxicol Lett. 138, 51–62.
Vernon, R. J., and Ferguson, R. K. (1969). Effects of trichloroethylene on visual-motor performance. Arch. Environ. Health 18, 894–900.
Warren, D. A., Reigle T. G., Muralidhara, S., and Dallas, C. E. (1996). Schedule-controlled operant behavior of rats following oral administration of perchloroethylene: Time course and relationship to blood and brain solvent levels. J. Toxicol. Environ. Health 47, 345–362.
Wu, C., and Schaum, J. (2000). Exposure assessment of trichloroethylene. Environ. Health. Perspect. 108 (Suppl. 2), 359–363.
Zemon, V., and Ratliff, F. (1984). Intermodulation components of the visual evoked potential: Responses to lateral and superimposed stimuli. Biol. Cybern. 50, 401–408.(William K. Boyes, Mark Be)